Every January, San Francisco parents sit down with a list of 70-odd elementary schools and rank them, knowing the house they live in only partly decides where their kid ends up. The question I hear most this fall is simple. If the district finally ends the lottery, does my address start to count?

Quite possibly. On October 13 the SFUSD school board votes on a direction for replacing the SFUSD lottery, and two of the three options still on the table guarantee a seat near home. If that is where it lands, the San Francisco neighborhoods with the strongest school scores, most of them on the west side, start to matter to homebuyers in a way they have not since 2011. Nothing is final until April 2027, no zone lines exist yet, and the first students placed under new rules would start in August 2028.

In brief

• The October 13 vote sets a policy direction. Adoption comes April 13, 2027, and the new system launches for students starting in August 2028, the 2028-29 school year.
• Three designs remain: small zones of roughly two to five schools, a neighborhood guarantee with the right to apply anywhere, or a neighborhood guarantee plus citywide programs only.
• As of October 2026, 18 of San Francisco's 92 ranked neighborhoods score 4.2 or higher out of 5 on our school score, and 14 of them sit on the west side.
• The cheapest of those 18 is Anza Vista at about $1.25 million in September 2026. The Outer Sunset scores 4.5 at about $1.69 million, roughly $510,000 under the citywide median.
• Our school scores reflect state test results at nearby schools. They are not attendance boundaries, and the district has not drawn any.

What the SFUSD board is actually voting on October 13

The board is choosing a direction, not a map. SFUSD's own governance calendar lists the October 13 item as the board voting to "confirm enrollment policy direction," and the district's school reorganization page, updated September 21, describes the goal as a system that would promise a seat closer to home while keeping access to programs across the city.

Staff presented four models at a September 1 study session, according to the San Francisco Parent Coalition's meeting summary. Large zones were dropped because the travel distances did not work. That leaves three:

• Small zones. Families choose among roughly two to five schools near home, plus designated citywide programs.
• Neighborhood plus all schools. A guaranteed local school, with the option to apply to any school in the district.
• Neighborhood plus citywide programs. A guaranteed local school, with choice limited mostly to language immersion, magnet and other programs designated as citywide.

Pushing from outside is KTAP, short for Kids, Teachers and Parents, a group launched this week by tech executive John Jersin, former SF Parents staffer Sarah Montoya and two Lowell High teachers. It wants the neighborhood guarantee. The San Francisco Standard reported, and Hoodline followed on October 1, that an internal KTAP poll of 600 voters found 35 percent backing neighborhood zones, 35 percent expanded magnets, 17 percent a simplified lottery and just 2 percent the current system. It is the group's own poll, so treat it as advocacy, but that 2 percent is the clearest read yet on how little love the lottery has left.

One date needs straightening out. Some coverage says the new system will be ready for 2027-28. What the district and the Standard actually describe is a system built during 2027-28, with families applying under the new rules for the 2028-29 school year, which starts in August 2028. That matches our August look at SFUSD school closures, which reported the overhaul had slipped to 2028-29. The date has not moved again. It has just been described loosely.

Skepticism is earned here. The board approved a zone-based plan in December 2020 for a 2023 launch. In May, SFUSD governance head Hong Mei Pang told the Standard that simulations showed that 2020 policy "cannot guarantee that students living within a zone would be assigned to a school there," which is why the district is starting over. School closures are sequenced behind the new system, which is why they now sit at no earlier than 2029-30.

How much your address matters under the lottery today

Some, but less than most buyers assume. For an elementary attendance-area school, SFUSD's tiebreakers run in this order: siblings, then pre-K or TK at the same school, then families living in the parts of the city with the lowest average test scores (the CTIP1 tiebreaker), and only then families who live in the attendance area. At citywide schools the attendance area does not count at all.

The current choice system has run since 2011, and KQED reported in August that it has left long waitlists at the most sought-after campuses while others sit underenrolled.

The lottery mostly works, and mostly is the problem. In the 2025-26 cycle, 92 percent of students got one of their ranked choices, but only 69 percent of kindergarten, sixth and eighth grade assignments were a first choice. At Presidio Middle, 87 percent of applicants with the feeder tiebreaker got a seat in recent years, against 3 percent with no tiebreaker at all. So geography already leaks in through feeder patterns. A neighborhood guarantee would make it the main event.

Where San Francisco's strongest public school scores sit right now

They cluster on the west side. Of the 92 San Francisco neighborhoods on our best-schools ranking, 18 score 4.2 or higher out of 5 as of October 2026, against a citywide school score of 3.3, and 14 of those 18 are in the Richmond, the Sunset, the Presidio or the neighborhoods west of Twin Peaks.

At the top, tied at 4.5, are Jordan Park / Laurel Heights, the pocket around Laurel Village just north of Geary, and the Outer Sunset, the fog-belt grid running out Judah and Noriega to Ocean Beach. The Outer Sunset page lists Sunset Elementary at 4.8 and Francis Scott Key and A.P. Giannini Middle at 4.3 each. The four exceptions are close in: North Panhandle, Anza Vista, Clarendon Heights and Cole Valley, all on the slopes and flats just east of the line.

Zoom out and the split gets starker. The typical west-side neighborhood scores 4.0 for schools. The typical neighborhood everywhere else scores 2.8. Twenty of the 24 neighborhoods at 4.0 or higher are west-side, and only one of the 32 neighborhoods below 3.0 is (Stonestown, at 2.4).

I would not read that as a verdict on any school, or on the kids in it. These scores track test results, and test results reflect a lot more than teaching, household income included. What they do tell you is where today's highest-scoring campuses are, which is exactly what a neighborhood guarantee would hand to the people who live closest.

What homes cost in the strongest school-score neighborhoods

Less than you would guess, at least in part of the map. San Francisco's citywide median was about $2.20 million in August 2026, per our San Francisco city page, and 8 of the 15 top-scoring neighborhoods with their own price history sit below it.

Here are those eight, cheapest first, using the latest month on each neighborhood page.

The cheapest 4.2-plus neighborhood is Anza Vista, the small hill east of Masonic around the old City Center site, at about $1.25 million in September 2026. Its honest trade-off is the safety score, 3.3, the lowest in the table. The Outer Parkside, off Taraval near Ocean Beach, and Midtown Terrace, the curving streets off Clarendon Avenue below Sutro Tower, pair 4.2-plus school scores with safety scores of 4.5 and 4.7.

Want to compare how all 92 neighborhoods line up on schools, safety and amenities at once? See the full San Francisco neighborhood ranking, then sort by the factor you care about most.

The comparison that sticks with me is Noe Valley. Its median was about $2.53 million in August 2026 with a school score of 2.4. The Outer Sunset costs about $840,000 less and scores 2.1 points higher. Divide price by school score and Noe works out to roughly $1.05 million per point, against about $376,000 in the Outer Sunset and $284,000 in Anza Vista. People pay for Noe's 24th Street, its sun and its commute down 280, and under a lottery the school score barely entered the math. A neighborhood guarantee would put it back in.

The top end tells the same story from the other side. Jordan Park / Laurel Heights, at about $3.52 million in July 2026, was about $2.59 million in August 2025. Presidio Heights, at 4.3, ran about $8.03 million in August, on thin sales that swing month to month. Those blocks were never priced on schools first. As we found in our analysis of how school scores move Bay Area home values, the relationship is real but uneven, and it is weakest exactly where a lottery cuts the cord between house and classroom.

Why the school-score map may not carry over into zones

The biggest reason is that no lines exist. Under a small-zones model, an Outer Sunset address might share a zone with schools in the Central Sunset or Parkside, which would blur any one neighborhood's edge. Under either neighborhood-guarantee model, the guaranteed school depends on attendance boundaries the district has not proposed, and citywide programs, such as Mandarin and Spanish immersion, would still draw from everywhere.

Then there is what the score measures. Our school score is built from California's CAASPP state test results at the public schools serving each area, which is a snapshot of where students test today. It is not a promise about a building in 2028, and a guarantee could shift who enrolls where. Some of the strongest-scoring schools have long waitlists precisely because families from across the city chose them. If those seats go to neighbors first, demand and results could rearrange in ways nobody can model yet.

And the district can still change course. The board can revise the direction in the spring. KTAP's poll shows magnets as popular as zones. And students already enrolled will not be moved, per the district's reorganization page, so the change reaches families entering TK, kindergarten or SFUSD for the first time.

So I would not buy a house on the October 13 vote. I would ask, of any west-side listing, which elementary school is closest, how full it is, and whether it feeds Giannini, Presidio or Roosevelt Middle, because those feeder patterns are already where geography counts.

The housing math underneath the school map

If school seats follow addresses, the number of addresses on the west side becomes a schools question. San Francisco Planning says only 10 percent of new housing over the past 15 years was built in the western and northern neighborhoods, even though they cover more than half the city's land. Those are the same neighborhoods holding 14 of the 18 strongest school scores.

The Family Zoning Plan, signed December 12, 2025 and in effect since January 12, 2026, allows mid-rise buildings on corridors like Geary, Judah, Taraval and Noriega. That is the quiet link between school board politics and a fair outcome. A neighborhood guarantee in a part of the city that barely builds turns a good school score into a scarce, expensive ticket. More homes near those campuses spreads the ticket around, and it keeps the enrollment that keeps a school open.

How we read the school data

Every score here comes from Houseberry's San Francisco best-schools ranking, which rates 92 neighborhoods on a 0 to 5 scale from state test results at the schools serving each one, as displayed on October 1, 2026. A few neighborhood pages show a slightly refreshed figure (Merced Heights reads 3.9 on its page versus 3.6 on the ranking), so treat tenths as approximate. Prices are the latest month on each neighborhood's 12-month history, July to September 2026. The west side means the Richmond, Sunset, Presidio and west of Twin Peaks, 37 neighborhoods in all. Neighborhoods without their own price history were left out of price comparisons.

The useful move this fall is to stop treating the school score as background and put it next to price, safety and commute for every block you are considering. Open the San Francisco school-score ranking, pick the five neighborhoods you would actually live in, and see what the October 13 vote could mean for each of them.

Sources

The San Francisco Standard, As SFUSD weighs replacing its lottery, a new group is pushing the boldest overhaul, September 30, 2026

Hoodline, KTAP pushes SFUSD to overhaul school lottery system, October 1, 2026

SFUSD, School Reorganization, updated September 21, 2026

SFUSD, Board meeting calendar and agendas

SFUSD, Elementary tiebreakers

San Francisco Parent Coalition, SFUSD Board Meeting Summary, September 1, 2026

The San Francisco Standard, SFUSD's enrollment overhaul is years behind schedule, May 18, 2026

The San Francisco Standard, SFUSD's school lottery is broken, August 18, 2026

KQED, As SFUSD welcomes back students, officials plan to revamp enrollment, August 17, 2026

San Francisco Planning, Family Zoning Plan

Houseberry, best neighborhoods for schools in San Francisco

Houseberry, San Francisco city guide and price history

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

On June 10, 2026, Santa Clara sent the state the biggest housing number it has ever attached to its Caltrain station: 8,300 homes. The filing is a Notice of Preparation for the Santa Clara Station Area Specific Plan, the first formal step in an environmental impact report, and it covers 244 acres on both sides of the tracks next to the Santa Clara Transit Center.

Alongside the homes, the plan would allow 2.1 million square feet of office, 1.7 million square feet of light industrial space and 19.5 acres of open space. Nothing here is approved. No Draft EIR has been published, and the City Council has not voted on a plan. The comment window on the notice ran June 12 to July 13 and is now closed.

My read, up front: this is the right place in Santa Clara for this much housing, and the number worth watching is whether 8,300 survives the airport, the EIR and the council.

Where the 244 acres are

The plan area runs from El Camino Real and Franklin Street on the southwest to De La Cruz Boulevard, Reed Street and Martin Avenue on the northeast, according to the notice. That puts it between the Old Quad, the historic grid of bungalows around Santa Clara University and the 1863 depot, and the warehouse and office blocks along Brokaw Road and Coleman Avenue that run up toward San Jose Mineta International Airport.

The state listing names 192 parcels, all in ZIP code 95050. Their current General Plan designations run the whole range, from Low Density Residential to Regional Commercial to Light Industrial. In plain terms, the plan area includes some existing homes, a lot of low-slung industrial land and the park-and-ride lot at El Camino and Benton Street where Caltrain riders leave their cars.

Key takeaways

If you only have a minute, this is what the June filing says and does not say.

• Santa Clara's Station Area Specific Plan would allow up to 8,300 homes on 244 acres around the Santa Clara Caltrain station, per the June 2026 Notice of Preparation.
• That ceiling is at least 50 percent higher than the 5,000 to 5,500 homes in the preferred concept shown to the Planning Commission in October 2025.
• The plan is unapproved. No Draft EIR has been released, and no adoption hearing has been scheduled as of October 1, 2026.
• The two Houseberry neighborhoods closest to the plan area, East Central and South 101, score 3.6 and 2.9 out of 5 and both sell below the Santa Clara citywide median.
• The whole footprint sits inside the county's Airport Influence Area, which is likely to set the real height and density limits.

Santa Clara Station Area Specific Plan by the numbers

The notice is short on detail and long on square footage. Here is the full development program as filed, with the math that makes it readable.

Thirty-four homes per acre sounds modest for a rail station, and it is misleading in a useful way. The 244 acres include streets, existing houses, the rail corridor and the open space, so the parcels that actually redevelop would run much denser than the average.

The retail figure is the one I would push on. About 16 square feet of shops per new home is thin for a neighborhood meant to be walked. A June letter in the Silicon Valley Voice made a similar case, arguing for restaurants and gathering places alongside the units.

The jobs math cuts the other way. Roughly 10,500 employees against 8,300 homes is close to balanced, which is unusual in Silicon Valley, where job growth has long outpaced housing.

8,300 is a ceiling, and it keeps going up

Every version of this plan has been bigger than the last. Santa Clara, San Jose and VTA finished an earlier Santa Clara Station Area Plan that was folded into the city's General Plan in November 2010. It covered about 440 acres in both cities and planned for 2,200 or more new homes.

When the current effort came to the Planning Commission on October 8, 2025, the preferred concept called for roughly 5,000 to 5,500 homes and about 2 million square feet of commercial space, according to a summary of the study session. Eight months later the environmental notice set the ceiling at 8,300.

Treat that as an envelope, not a promise. Cities usually study the largest version of a plan in the EIR so they can adopt something smaller without redoing the analysis. The adopted plan could land at 8,300 or well below it.

But the direction matters. Santa Clara's state housing target for 2023 to 2031 is 11,632 homes. This one plan, at its ceiling, would cover about 71 percent of that number on land that is already next to a train.

How tall it could get, and why the airport decides

The airport, not the neighbors, will probably set the height limit. In its comment letter on the notice, Caltrans pointed out that the entire plan footprint falls inside the Santa Clara County Airport Influence Area, including safety zones and traffic pattern zones that limit height, density and noise exposure.

The heights the consultants floated are already modest. At a July 2025 Station Area Task Force meeting, the draft framework showed a base of about 85 feet, or seven floors, up to 13 floors near the station and the Brokaw corridor, about 40 feet where the plan meets existing houses, and 64 feet on light industrial parcels. Anything above the base would need FAA sign-off.

Here is the honest downside. Living here means jet noise on approach, freight trains on the Union Pacific line and a lot of construction for a long time. Planning commissioners in October also pressed staff on traffic along Coleman Avenue and on who pays for a proposed Benton or Brokaw crossing under or over the tracks. Fair questions, and the plan does not answer them yet.

What it means for the Old Quad and South 101

The plan area sits between two of the neighborhoods we score in Santa Clara, and both are mid-table or lower. Santa Clara East Central, the college-town grid around Santa Clara University and the Old Quad, scores 3.6 out of 5 overall and 4.0 for amenities, the best amenities score in the city. South 101, the mix of homes, tech campuses and big-box retail toward Reed Street and De La Cruz, scores 2.9, last of the seven.

Prices tell the same story. Both sell below the citywide median, and North 101, across the freeway near Levi's Stadium, sits lower still.

The median sale price in Santa Clara East Central was about $1.67 million in September 2026, roughly 10 percent below the Santa Clara citywide median of about $1.86 million in August 2026. South 101 was about $1.65 million in August, and its 2.5 safety score is the lowest in the city, which is worth researching block by block rather than treating as a verdict.

For a buyer, the practical question is what 8,300 mostly multifamily homes do to neighborhoods like these. The way we look at it, a plan this size adds choices more than it threatens values. Today the entry point near the station is a $1.6 million bungalow. A finished plan would add condos, townhomes and apartments at the platform, and it would add the shops and parks that move an amenities score. The full Santa Clara neighborhood ranking is the place to compare the rest of the city.

Caltrain now, BART later

The transit case is real today. The Santa Clara Transit Center already runs Caltrain, ACE and VTA buses, and VTA lists BART Silicon Valley Phase II construction running from 2024 to 2036, ending at a ground-level Santa Clara station next to Caltrain. Anyone buying near here this decade should plan around Caltrain and treat BART as the bonus.

Pieces of the plan area are already moving. Our coverage of University Station, the 408 homes proposed across from the depot, was the first chapter. The city also approved 142 for-sale townhomes at 1400 Coleman Avenue) in November 2025, 29 of them affordable, and the staff report notes the site sits inside the station focus area. VTA still owns a 2.4-acre park-and-ride site at the station that it wants to redevelop with 25 percent affordable housing, though its last developer agreement expired.

This is also not the only big number in town. Santa Clara is separately moving to put 6,000 homes on 45 acres at Tasman East near Levi's Stadium.

Quick answers on the station plan

Is the Santa Clara Station Area Specific Plan approved?

No. As of October 1, 2026, the city has issued only a Notice of Preparation. A Draft EIR, a public draft of the plan, a Planning Commission recommendation and a City Council vote all still have to happen.

Will it rezone homes in the Old Quad?

Some existing residential parcels are inside the boundary, since the state listing includes Low and Medium Density Residential designations. The southwest edge is Franklin Street and El Camino Real. If you own near that edge, check whether your parcel number is among the 192 listed in the notice.

When will the Draft EIR come out?

The city has not posted a date. A May 2025 staff report targeted a public Draft EIR in spring 2026 and adoption hearings in summer 2026, and the notice itself arrived about a year behind that schedule. The regional grant funding the plan expires December 31, 2026.

The dates that decide it

The next moves come in a fixed order, and none of them is on a calendar yet.

• A public draft of the Specific Plan, with the height and street standards the task force has been working through.
• The Draft EIR, which opens a public comment period of at least 45 days.
• A Planning Commission hearing and recommendation.
• A City Council vote to adopt the plan and certify the EIR.

The scoping meeting was June 29 on Zoom, and senior planner Rebecca Bustos is the city contact listed on the filing. The city keeps a project page for the Station Area Specific Plan where the draft documents should appear.

Santa Clara planned for 2,200 homes here in 2010. This time the target is nearly four times larger, the state is pushing harder and BART is under construction. If the council keeps the number close to 8,300, the Old Quad gets a real neighborhood on the other side of the tracks. We will be reading the Draft EIR the day it posts, and Houseberry will update this piece when it does.

Sources

• CEQAnet, Santa Clara Station Area Specific Plan, SCH 2026060462
• City of Santa Clara, Notice of Preparation, June 2026
• Caltrans District 4, comment letter on the Notice of Preparation
• City of Santa Clara, Station Area Specific Plan project page
• City of Santa Clara, File 25-410, VTA cooperative agreement amendment, May 13, 2025
• City of Santa Clara, 1400 Coleman Avenue staff report, November 2025)
• CitizenPortal, Planning Commission study session summary, October 8, 2025
• CitizenPortal, Station Area Task Force summary, July 17, 2025
• Dyett and Bhatia, Santa Clara Station Area Plan (2010)
• VTA, BART Silicon Valley Phase II
• VTA, Santa Clara Transit Center transit-oriented development
• Silicon Valley Voice, letter on the Station Area Specific Plan, June 2026
• Silicon Valley Voice, Planning Commission recommends Housing Element update, January 2023
• Houseberry, Santa Clara neighborhood rankings and price history, October 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

File 2026-6313 sits on Tuesday's Alameda City Council consent calendar, the part of the agenda that passes without discussion. It also decides who keeps running an Airbnb on the island.

The Alameda short-term rental ordinance limits new rentals to a host's own primary home, keeps most ADUs out of the nightly market and gives existing hosts 90 days to get a permit. If the council gives it final passage on October 6, 2026, the new Section 30-19 takes effect 30 days later, in early November.

So, can you Airbnb your ADU in Alameda? Only if you live in it, or if you have already been renting out a pre-2017 unit for at least a year.

Key takeaways

• New short-term rentals in Alameda must be on the property where the host lives, under the ordinance up for final passage October 6, 2026.
• ADUs and junior ADUs cannot be rented for 30 days or less unless the host lives in that unit as a primary residence.
• An owner who lives on a property with four or fewer units can rent one other unit there nightly, one at a time.
• Hosts with at least a year of operation before adoption keep one off-site unit or a pre-2017 ADU, but that status does not pass to a buyer.
• Existing hosts get a 90-day amnesty to apply, and city staff report back to the council in January 2028.

What the Alameda short-term rental ordinance allows

You can rent nightly where you live, and almost nowhere else. The ordinance recognizes two kinds of stays, both 30 consecutive days or less.

"Home sharing" puts guests in your unit. A "vacation rental" puts them in a different unit on the property you live on. There is no cap on nights, since tying rentals to the host's address does that job, as the Alameda Post explained in its council preview.

Tenants can host only in their own unit, with the landlord's written permission, and a rent-program unit can only be home shared, capped at the allowable rent. Every host needs a permit, a business license, $1 million in liability insurance and a local contact who can show up within 60 minutes of a complaint. Stays still carry the city's 14 percent transient occupancy tax.

The September 15 introduction passed 3-0, with Councilmember Jensen abstaining and Councilmember Boller recused, per the Alameda Post.

If you own an ADU or a duplex

This is the split most owners will miss. A backyard ADU is treated far more strictly than the second unit of a duplex.

An ADU or junior ADU cannot be a short-term rental unless the host lives in it as a primary residence. So the classic setup, owner in the Victorian and guests in the cottage out back, is closed to anyone starting fresh. Flip it, with you in the cottage and guests in the main house, and it fits the vacation-rental rule.

A duplex, triplex or fourplex works the other way. Live in one unit and you can rent one other regular unit nightly. Five or more units, and it is home sharing only. Also excluded: SB 9 units, deed-restricted affordable units, and any property with an Ellis Act eviction in the past five years, which knocks out every unit on the lot.

I think the ADU line is right. California spent a decade clearing the way for backyard units so they would house people, which is why the Planning Board asked staff to separate modern ADUs from older cottages. The cost lands on owners who built a cottage counting on nightly income. That is a real loss. Some will now lease long term, which is the point.

The legacy clause ends at the sale

Existing hosts are protected. Their buyers are not. A host with at least a year of operation before adoption can keep going without living on site, including in an ADU created before 2017.

They still need a permit, and they have to apply during the 90-day amnesty after the ordinance takes effect, which lands around early February 2027 if Tuesday goes as planned. Miss it and the rental stops. A fourth violation means permanent revocation.

And the text is blunt about transfers. The legacy exception "shall not apply to subsequent purchasers."

How to underwrite an Alameda income property now

If you will not live there, underwrite long-term rent only. A listing touting Airbnb history is showing income the next owner cannot legally earn.

Island prices leave little room for hopeful math. Here is each area's latest median next to the scores on Houseberry's Alameda neighborhood rankings.

The East End, the older half of the main island between Park Street and the Fruitvale and High Street bridges, carries a 4.4 amenities score and the walkable Park Street strip visitors book for. Its pre-war two-flats and backyard cottages are where this ordinance rewrites the most spreadsheets. The Alameda citywide median was about $1.42 million in August 2026, on Houseberry's numbers.

Owner-occupant buyers come out fine. Live in one side of a duplex and the other side can still help with the mortgage. The trade-off is becoming an innkeeper on your own lot, on a 60-minute response clock.

Two Alameda votes on Tuesday, one direction

The same October 6 consent calendar carries final passage of the ADU condo ordinance, which we broke down in our guide to selling an Alameda ADU separately. Together they say backyard units are for people who live here, as renters or owners.

Watch whether anyone pulls the item off consent Tuesday. After that, the next checkpoint is the January 2028 staff report on whether units moved back into long-term housing.

Sources

• City of Alameda, File 2026-6313, final passage of Section 30-19 (Short Term Rentals), Legistar
• Alameda Post, Council Advances Housing Rules, AC Transit Warns of Route Cuts, September 18, 2026
• Alameda Post, September 15 City Council Preview, September 8, 2026
• Alameda Post, Planning Board Backs Downtown Zoning Overhaul, Advances Short-Term Rental Rules, July 17, 2026
• Alameda Post, Planning Board Considers Regulating Short-Term Rentals, February 14, 2025
• Houseberry, Alameda city data and 12-month price history

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

$264.9 million. That is what Los Angeles investor Preylock paid in late September for Assembly at North First, a six-building research campus in North San Jose that sold for $192 million in 2021, according to Hoodline's report on the East Bay Times story. Same buildings, five years later, roughly 38 percent more money.

The short answer for anyone shopping for a home nearby: the deal is a real jobs signal for North San Jose, but it has not shown up in home prices yet. Every neighborhood we track around the campus still sells below the San Jose median, and most are flat or slightly down over the past year.

The numbers that matter

• Assembly at North First, a 490,000 square foot campus on 27 acres along North First Street, sold for $264.9 million in September 2026, or about $541 per square foot.
• The same campus sold for $192 million (about $392 per square foot) in 2021 and for $82 million in 2016, so its price has more than tripled in ten years.
• It is 100 percent leased. Humanoid robot maker Figure AI occupies about 98,700 square feet, alongside AI startup Hark and Teradyne's LitePoint.
• Golden Triangle, the North San Jose business district around the campus, had a typical home price of about $1.06 million in September 2026, roughly 35 percent below the $1.63 million San Jose citywide figure for August.
• Five of the six nearby neighborhoods in our comparison were flat or down between October 2025 and September 2026. Berryessa was the exception, up about 8 percent.

What actually sold on North First Street

The campus sits on North First Street, bounded by Headquarters Drive and Rose Orchard Way, in the stretch of North San Jose most locals still call the Golden Triangle. The seller was EQT Exeter, and Newmark ran the sale. Hoodline's write-up lists addresses running from 3930 to 4000 North First Street plus 90 Headquarters Drive, and reports a $172.2 million Barclays loan behind the purchase.

One discrepancy worth knowing about. The Registry reported the price as $268.3 million, and quoted Newmark calling it the largest sale of an advanced manufacturing or R&D campus in Bay Area history. The East Bay Times and Hoodline figure is $264.9 million. We use the East Bay Times figure. The gap is about 1 percent and does not change the story.

For context, Hoodline also reports the January 2026 assessed value at $242.3 million and an initial asking price near $285 million. So Preylock paid above the assessor and below the ask. Normal haggling, in other words.

Vacant in 2021, full in 2026

The price jump is about tenants, not about the buildings getting nicer. Lam Research left the campus around 2016, and it sat empty for roughly five years, through a renovation, until EQT Exeter bought it vacant for $192 million in 2021. That was a bet that somebody would want this much flexible R&D space in one place.

Somebody did. Figure AI moved its headquarters and robot production there from a 27,900 square foot space in Sunnyvale, taking about 98,700 square feet for manufacturing, fleet operations and engineering. Hark and LitePoint filled in the rest. A campus that was a vacancy problem five years ago is now a fully leased AI and robotics hub.

This is the second North San Jose AI hardware story we have covered in a week. MiTAC's $76.8 million AI server lab permit on Automation Parkway is a short drive away. And it fits the pattern we wrote about in why AI robotics firms are leasing Sunnyvale R&D space instead of San Francisco towers. Robots need loading docks and heavy power, and 1980s Silicon Valley campuses have both.

North San Jose home prices near the campus, September 2026

Here is the part the commercial coverage skips. Commercial investors just paid a premium to own space next to these neighborhoods. Homebuyers have not done the same.

The table below pulls our scores from the San Jose neighborhood rankings (and Santa Clara's, for North 101) and the latest typical price from each neighborhood's 12-month price history. Scores are out of 5.

Two things jump out. First, the cheapest address in the group is the one closest to the campus. Golden Triangle is mostly office parks, light rail and big-box retail between US-101, I-880 and Highway 237, with sparse housing, so its $1.06 million typical price is drawn from a small pool of homes.

Second, the best scores in the group belong to Berryessa North, the residential blocks around the I-680 and Berryessa Road interchange near the BART station. It scores 4.0 overall and 4.3 on safety, and it sat at about $1.29 million in September 2026, about $340,000 under the citywide figure.

Berryessa Station is the cautionary row. Despite the name, the BART station is a good distance south of most of it, and its typical price slid from about $1.21 million in October 2025 to $1.13 million in September 2026. A transit name on a neighborhood is not the same as a walk to the train.

Does a big campus sale push up nearby home prices?

Not directly, and not quickly. A commercial sale price is a bet on rent from tenants. It tells you investors expect AI and robotics companies to keep paying to be in North San Jose. It does not tell you those employees will buy houses next door.

What it does do is lower the odds that this corridor goes dark again. A fully leased campus means steady jobs on the light rail line, steady lunch traffic for the strip centers on North First, and a stronger case for the housing San Jose already planned here.

That last point matters most to me. North San Jose has had room on paper for about 32,000 new homes since 2005, and a long-running traffic lawsuit kept most of it stalled until a settlement in late 2022. Jobs on North First Street are exactly why those homes should get built near the light rail stops instead of an hour away in the Central Valley.

What this does and does not mean for buyers

It does mean North San Jose is a working employment district with new, well-funded tenants. If you work at Figure, Cisco or one of the Tasman Drive campuses, the neighborhoods in that table are your shortest commutes in the city.

It does not mean prices nearby are about to jump. Five of the six neighborhoods were flat or down over the past year, while the campus price rose 38 percent over five. Commercial and residential markets run on different math, and right now the residential side is cooler.

The honest downside: this is a place built for office parks first. Amenities scores near the campus are low, Rincon de los Esteros backs onto the flood-prone baylands toward Alviso, and Golden Triangle has very little conventional housing. You trade neighborhood feel for a short commute and a lower price.

So the practical move is to compare the blocks, not the headline. Start with our San Jose rankings, check the safety and school scores for the specific neighborhood, and treat the campus sale as one more reason North San Jose's job base looks stable. That is the neighborhood-first habit we built Houseberry around, and a $264.9 million deal is a good reminder of why it pays.

Sources

• Hoodline, San Jose AI and robotics hub sells for $264.9M (October 2026, citing the East Bay Times)
• The Registry, Preylock buys EQT Exeter's The Assembly campus in North San Jose (September 30, 2026)
• The Real Deal, EQT Exeter buys Assembly at North First for $192M (October 2021)
• The Real Deal, Figure AI leases 99K square feet for new San Jose headquarters (March 2025)
• Hoodline, A plan for 32,000 new homes in North San Jose is finally moving forward (December 2022)
• Houseberry San Jose neighborhood rankings and neighborhood price histories (September 2026)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

11.7 years. That is how long it took a 20-home building at 3945 Judah Street, on the corner of 45th Avenue in San Francisco's Outer Sunset, to go from a filed building permit to a finished one. The city's Department of Building Inspection marked permit 201412314770 complete on September 28, 2026, almost twelve years after it was filed on December 31, 2014.

Construction itself took about 3.4 years. The rest was paperwork, a redesign, an appeal, and a cleanup of old gas station tanks. And in a neighborhood where new apartment buildings of this size barely happen, the finish is rarer than the delay.

The numbers that matter

The short read on 3945 Judah, all from public records pulled September 30, 2026:

• 11.7 years from permit filing (Dec. 31, 2014) to DBI completion (Sept. 28, 2026), with 6.3 of those years spent before DBI approved the plans.
• 20 homes, five of them below market rate, on a 0.11-acre lot that used to be an auto repair shop.
• $3.4 million is the revised permit valuation, up from a $1.5 million estimate at filing.
• 5 buildings of 10 or more homes have been finished in the Sunset/Parkside planning area since 2015, by our count of DataSF permits. This is the slowest of them.
• $1.69 million was the Outer Sunset median in August 2026 on Houseberry, up about 15 percent in eleven months.

Where the 11.7 years went, phase by phase

Most of the wait happened before anyone could legally build. DBI's own record for the permit shows four dates that split the timeline cleanly: filed December 31, 2014, approved April 13, 2021, issued October 27, 2021, and first construction document May 11, 2023.

Put another way, it took longer to get the plans approved than it took to demolish the old building, pour the podium, frame five stories, and pass final inspection combined.

The gap between issuance and construction (1.5 years) is the part people outside the business tend to miss. A separate demolition permit for the auto shop was filed in January 2021 and not issued until December 2021, and SF YIMBY reported the single-story shop came down in 2022. Shoring permits followed. Only then could the concrete podium go in.

The building filed in 2014 was not the one that got built

The permit number is from 2014, but the 20-home design is from 2018. According to SF YIMBY's reporting, owner Brendan Quinlan bought the parcel in 2013 for $900,000 and first proposed a four-story building with ground-floor retail and six apartments. The application was resubmitted in 2015 and later closed.

What changed was HOME-SF, the density bonus program former District 4 Supervisor Katy Tang championed and the city adopted in 2017. It let the project add a floor and more than triple its home count in exchange for making 25 percent of the units affordable. Hoodline reported it was only the third HOME-SF project approved anywhere in the city when the Planning Commission voted 6-0 for it on November 7, 2019.

That is the honest downside of this story. Some of the delay was the price of a better project. Six apartments became 20. Two affordable units became five.

The appeal, and the two old tanks under the lot

Neighbors appealed within days. Hoodline covered the filing in November 2019, with opponents citing groundwater, parking, height, and density, and floating a 16-unit alternative.

The San Francisco Board of Appeals denied the appeal 4-0 on January 29, 2020, according to a meeting summary. The substance was real, though. Two closed underground storage tanks sat under the former gas station, and the board made clear that the Department of Public Health, through its Maher soil and groundwater program, had to sign off on a site mitigation plan before building permits could issue. DBI's plan approval came about 15 months later.

Cleaning up a gas station site before stacking homes on it is a good rule. It is also, on a lot this size, a year-plus of consultants and lab work that a 20-unit budget has to carry.

How rare is a 20-unit finish in the Sunset?

Very. We pulled every new-construction permit marked complete in the Sunset/Parkside planning area from DataSF and found five buildings of 10 or more homes finished since 2015, totaling 233 homes. Everything else finished in that stretch was small, mostly one- and two-home buildings.

The fastest two tell you what speeds things up. 1360 43rd Avenue, the 135-home educator building two blocks east on the same N-Judah line, was filed as a mayoral priority project and finished in 5.2 years. 3701 Noriega, 12 condos, took 4.7. The two slowest besides Judah, 2898 Sloat and 2255 Taraval, both ran past nine years.

Five buildings in roughly eleven years, across a planning area that takes in most of the Sunset and Parkside. That is the supply picture behind every Sunset price chart. It is also why our earlier look at the 199-home proposal at 1234 Great Highway and the 45 homes planned for the church at 4114 Judah matter more than their unit counts suggest.

What 20 homes do and do not do to Outer Sunset prices

Twenty homes will not move the Outer Sunset median. What they add is a kind of housing the neighborhood almost never gets: new rentals with an elevator, a block from an N-Judah stop and about four blocks from Ocean Beach. The building is leasing as Seadrift, and a Corcoran listing advertised a one-bedroom at $3,600 a month.

Meanwhile, ownership prices kept climbing without much help from new supply. On Houseberry's Outer Sunset page, the neighborhood median rose from $1.47 million in September 2025 to $1.69 million in August 2026.

So the Outer Sunset is tracking the city almost exactly, at about $510,000 below the citywide median. It also ranks second of 92 neighborhoods on our San Francisco overall ranking, with a 4.2 out of 5, which helps explain why demand keeps pace with the rest of the city even out by the beach.

What this does not mean: a finished building is not a sign the pipeline is speeding up. Our piece on San Francisco housing filings hitting their lowest September since 2012 found just 30 net new homes proposed citywide that month. And the deal this building took is already gone. After a July 2026 Board of Supervisors vote, SF Planning's HOME-SF page now lists affordability requirements of 7 to 10 percent, a fraction of the 25 percent 3945 Judah committed to in 2019.

Questions people are asking about 3945 Judah

When did 3945 Judah Street finish?

DBI marked the building permit complete on September 28, 2026. Follow-up permits for emergency radio coverage, evacuation signage, and a final inspection closed out in August and September 2026. Ground-floor retail build-outs are still working through permits.

Does the building have affordable units?

Yes. Five of the 20 homes are below market rate. SF YIMBY listed two at about 50 percent of area median income, two at 80 percent, and one at 110 percent. San Francisco lists these units through its DAHLIA housing portal under the name Seadrift.

What is the unit mix?

Ten one-bedrooms, nine two-bedrooms, and one three-bedroom, over about 2,400 square feet of ground-floor retail, with seven car spaces and 20 bike spaces.

What was on the site before?

A single-story auto repair shop on the site of a former gas station, which is why the underground tank cleanup had to happen before construction permits could issue.

If you are weighing the Outer Sunset against the rest of the west side, the new-construction count is the number to keep in mind. We keep score on schools, safety, and prices for every San Francisco neighborhood, and the Sunset's scores are strong. The part no score captures is how few new homes get built here, and how long the ones that do take.

Sources

• DataSF, DBI Building Permits, permit 201412314770
• DataSF, DBI Building Permits dataset
• SF YIMBY, Construction Underway for 3945 Judah Street, January 2024
• SF YIMBY, Construction Tops Out for 3945 Judah Street, August 2024
• Hoodline, 20 apartments to rise on the site of former gas station in Outer Sunset, November 2019
• Hoodline, Outer Sunset residents file appeal over 20-unit apartment building, November 2019
• Citizen Portal, Board of Appeals upholds HOME-SF authorization for 3945 Judah
• SF Planning, HOME-SF program page
• Corcoran, 3945 Judah Street rental listing
• Houseberry, Outer Sunset neighborhood page

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The last empty block of Swenson's Ohlone project finally has its building permit. San Jose issued permit 2026-132634-MF on September 4, 2026, for 272 apartments at 860 West San Carlos Street, and every one of them will be income-restricted affordable housing.

The city's own permit record lists the job as "Block A Family Affordable Apts," with a declared construction value of $80.8 million and 512,400 square feet of floor area. That is the largest valuation on any new apartment permit San Jose has issued this year, and it lands on a corridor where the headline projects around Diridon Station have mostly stalled.

The short version

Here is what changed, and what it means if you live or shop for a home nearby.

• San Jose issued the building permit for 272 affordable homes at 860 West San Carlos Street on September 4, 2026, after the Planning Director approved the design on May 20, 2026.
• The state tax-credit filing sets rents for households earning 30 to 70 percent of area median income, which in 2026 means roughly $61,650 to $143,850 a year for a family of four.
• The seven-story building completes Swenson's three-block Ohlone plan, about 809 homes in all, with completion projected for mid-to-late 2028.
• Buena Vista and Diridon, the Houseberry neighborhoods closest to the site, saw typical prices fall 7.4 and 2.9 percent in the 12 months to August 2026, while San Jose citywide rose about 1 percent.

What the 860 West San Carlos permit actually says

The permit is what turns an approval into a construction site. Here is what the city recorded on September 4:

• Permit number: 2026-132634-MF, new construction, multiple attached dwellings
• Owner: GR Block A LLC. Applicant and contractor: Swenson and Swenson Builders
• Homes: 272. Declared valuation: $80,819,610. Floor area: 512,400 square feet
• Plan review signoffs: building, electrical, plumbing and mechanical all marked complete

The design behind it went through a Planning Director hearing on May 20 as file H25-045, a Site Development Permit for a seven-story building on about 2.17 gross acres at the southwest corner of West San Carlos and Sunol streets. The city used State Density Bonus Law to grant three concessions and two waivers, including a waiver on floor area ratio, and cleared the environmental review as an addendum to the original Ohlone EIR.

That works out to about 125 homes per acre, dense by San Jose standards and ordinary for a site this close to a regional rail hub.

Divide by the number of homes and Block A still leads, at about $297,000 of construction value per home. The next highest in the chart, the Berryessa Family Apartments on Bernard Drive, works out to about $201,500. Big family-size units and a two-level concrete garage are not cheap to build.

From a 12-story market-rate tower to a seven-story affordable building

This site has been promised a building for a long time. The Ohlone Mixed-Use Project started environmental review in 2009 as up to 800 homes and 30,000 square feet of retail on about 8.25 acres of mostly paved former industrial land at West San Carlos and Sunol.

Swenson and its partner Republic Urban Properties built it in phases. Silver at Ohlone, with 268 apartments, broke ground in 2016. Patina@Midtown added 269 more in an eight-story building. The deal also set aside four acres to expand Del Monte Park.

Block A was always the tall one. San Jose approved a 12-story, 263-unit market-rate design in 2022, and Swenson later told reporters that version proved financially infeasible. A 13-story all-affordable version followed in 2024. The March 2026 filing cut it to seven stories, 79 feet, with five wood floors over a two-level concrete podium. The mix is 107 one-bedrooms, 96 two-bedrooms and 69 three-bedrooms, over a 242-space garage.

Shorter, same unit count, much cheaper to build. I will take a seven-story building that gets built over a 13-story rendering that does not, every time.

Who can live at Block A, and what the rent looks like

Block A will house working households earning roughly 30 to 70 percent of area median income, with most homes aimed at the 60 and 70 percent tiers.

There are two versions of that mix floating around, and they are both right. The city's hearing notice uses broad legal ceilings: 55 very-low-income homes at or below 50 percent of AMI and 217 low-income homes at or below 80 percent. The state tax-credit application reviewed in December 2025 sets the actual rent tiers, which are deeper. Some coverage described the 217 as moderate-income. The state filing does not.

Here is the tax-credit mix, with 2026 income ceilings. We derived those from HUD's FY2026 limits for Santa Clara County, where the four-person median family income is $205,500 and the 50 percent limit for a family of four is $102,750.

The state filing's 2025 rent schedule, utilities included, runs from $1,130 for a one-bedroom at the deepest tier to $2,638 at the top tier. Three-bedrooms run $1,567 to $3,656. So the 70 percent homes are not cheap. They are priced for a household with two steady paychecks that still cannot touch a $1.08 million condo near the station.

The 69 three-bedrooms matter more than they look. Large affordable units are the hardest kind to finance here, and many downtown projects skip them.

Leasing will run through the city's affordable housing portal once the building nears completion. Nothing is open for applications yet.

$80.8 million on the permit, $149.8 million all in

The permit valuation covers construction only. The real cost of the building is almost twice that.

The state tax-credit staff report puts the total development cost at $149.8 million, or about $550,900 per home. The permit's $80.8 million works out to roughly $297,000 per home. The gap is land, fees, financing and design. The financing stack includes about $5.7 million a year in federal tax credits and a $38 million tax-exempt bond allocation, according to the CTCAC report.

Here is the honest downside. Half a million dollars per affordable apartment is a hard number to defend to taxpayers, and it is still roughly what subsidized housing costs in the Bay Area right now. The building also trades the old plan's ground-floor retail for ground-floor apartments facing the street, one of the three density bonus concessions. Fine for residents, but a quieter frontage than West San Carlos deserves.

What 272 affordable homes do to the West San Carlos corridor

The corridor between Diridon Station and Bascom Avenue is where San Jose's housing is actually going up, and Block A is the biggest affordable piece of it.

A few blocks east, Urban Catalyst's Aquino at 486 West San Carlos is nearly topped out with 278 apartments, 14 of them affordable. Meanwhile Google's 80-acre Downtown West, approved in 2021, is still flattened lots and parking, which we dug into in our look at how much approved Diridon housing is actually getting built.

So the pattern on this side of downtown is clear. The mega-plan waits. The mid-rise, tax-credit-financed blocks keep coming, like the moderate-income plan for the Arena Hotel lot on The Alameda and the 18 townhomes proposed on a long-vacant Alameda lot. Together they are most of what this side of downtown will actually get this decade.

On prices, the best research I know is a Stanford study by Rebecca Diamond and Tim McQuade. It found that tax-credit developments in lower-income neighborhoods pushed nearby home prices up about 6.5 percent, while ones in higher-income areas led to declines of about 2.5 percent. The Buena Vista and Diridon medians sit well below San Jose's, which puts this site closer to the first case.

The neighborhoods around 860 West San Carlos, by the numbers

The site sits between two of our San Jose neighborhoods: Buena Vista, the older residential grid that stretches west from here toward Bascom, and Diridon, the station district to the northeast. This stretch of West San Carlos is often called Midtown. Both score in the bottom quarter of the city on our San Jose neighborhood rankings.

Buena Vista's typical home price was about $941,500 in August 2026, down from roughly $1.02 million in September 2025. Diridon's slipped from about $1.11 million to $1.08 million. Over the same 12 months, the San Jose citywide figure rose about 1 percent to roughly $1.63 million.

That softness predates the building and tracks the stalled Google campus and a housing stock heavy on older, smaller homes. For a buyer, the practical read is that this is one of the few parts of central San Jose where the entry price is still under $1 million, and the 0.9-point safety gap between Diridon and Buena Vista is the difference to research street by street.

When Block A could open, and what to watch next

The state filing assumed construction from May 2026 to July 2028. The building permit came four months after that start date, so treat mid-to-late 2028 as the realistic window for move-ins.

Three things are worth watching from here:

• Construction activity on the fenced lot at Sunol and West San Carlos, which tells you the financing actually closed.
• A listing on the city's affordable housing portal, usually months before first move-ins.
• Whether market-rate projects on the corridor, like the Gifford at 470 West San Carlos, reach a building permit.

Seventeen years after this land first went into environmental review, the last block of Ohlone is going to be 272 homes for households earning roughly $43,000 to $144,000 a year. If you are weighing Buena Vista or the streets around Diridon, look at the block first, and let the Houseberry scores tell you what the listing photos will not.

Sources

• City of San José, Active Building Permits dataset, permit 2026-132634-MF, accessed Sept. 30, 2026
• City of San José, Planning Director Hearing agenda, May 20, 2026, file H25-045 and ER26-016
• California Tax Credit Allocation Committee, staff report CA-25-715, Block A Family Apartments, December 2025
• HUD, FY2026 Income Limits, San Jose-Sunnyvale-Santa Clara HUD Metro FMR Area
• CEQAnet, Ohlone Mixed-Use Project, SCH 2009042075
• San Francisco YIMBY, Updated affordable housing plans filed for 860 West San Carlos Street, March 2026
• San Francisco YIMBY, Plans shift to affordable housing at 860 West San Carlos Street, September 2024
• The Real Deal, Swenson switches to affordable apartments at downtown San Jose site, May 20, 2026
• First American, The Ohlone revived
• San Francisco YIMBY, Construction nearly topped out for 486 West San Carlos Street, September 2026
• Planetizen, Google razed 80 acres in downtown San Jose, March 2026
• Stanford GSB, Diamond and McQuade, Who Wants Affordable Housing in Their Backyard?
• Houseberry, San Jose neighborhood rankings and neighborhood price history, August 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Every October the same question lands in my group chats, usually from someone who just moved to a condo or a street with three porches on it: where do people actually go? The honest answer is that Bay Area Halloween runs on a few dozen specific blocks, and most of them are in neighborhoods you can name.

The best trick-or-treat neighborhoods in the Bay Area for Halloween 2026 are Cole Valley/Parnassus Heights, Noe Valley, Jordan Park / Laurel Heights and Sea Cliff in San Francisco, Crocker Highlands and Rockridge in Oakland, Elmwood in Berkeley, Easton Addition in Burlingame, Willow Glen in San Jose, and Downtown Los Gatos. What they share is not a high score. It is houses packed close to the sidewalk, flat or gentle blocks, and neighbors organized enough to close a street.

Halloween falls on Saturday, October 31, 2026, so there is no school-night curfew this year. Expect the famous blocks to fill early.

The short version

If you only read one part of this guide, read this.

• San Francisco closed more than 40 street segments to cars for Halloween 2025, according to Mission Local's list, and most of the busy ones opened between 3 and 4 p.m.
• Sunset on October 31, 2026 is about 6:13 p.m. in San Francisco, and daylight saving time does not end until the next morning, so the first hour of trick-or-treating happens in daylight.
• The biggest single-block crowds in the region are on Alameda's Gold Coast, where one corner estimated about 2,500 visitors in 2025, per The San Francisco Standard.
• 16 of the 27 Houseberry neighborhoods in this guide score 4.0 or higher on safety, and Crocker Highlands in Oakland tops the list at 4.8.
• Two of San Francisco's busiest Halloween neighborhoods, Noe Valley and Bernal Heights, rank 61st and 73rd of 92 on our overall score. Halloween rewards density and organizing, not rankings.

What makes a block good for trick-or-treating

A good trick-or-treat block is one where a small person can hit 20 doors in 10 minutes without crossing a busy street. That comes down to lot size, stairs, terrain and whether someone on the block filed for a closure.

That is why this list looks different from our usual rankings. Sea Cliff and Jordan Park / Laurel Heights sit in the top five of our San Francisco neighborhood rankings, and they are great Halloween spots. But so is Bernal Heights at 73rd, because Cortland Avenue and three side streets go car-free every year. When we look at a neighborhood for daily life, schools and safety carry weight. On October 31, the physical layout of the blocks matters most.

Every name below matches the neighborhood's page on Houseberry, and where a famous street sits inside a neighborhood with a different name, I name both. Scores come from our neighborhood rankings as of September 2026.

San Francisco: the city of closed streets

San Francisco does Halloween by permit. In 2025 the city had car-free Halloween closures on more than 40 street segments, from Terrace Drive in Saint Francis Wood to 45th Avenue in the Outer Sunset, Mission Local reported on October 30. The 2026 list usually surfaces in the last week of October, so treat last year's hours below as a guide, not a promise.

Cole Valley/Parnassus Heights: Belvedere Street

Belvedere Street between Parnassus Avenue and 17th Street is the city's best-known Halloween block, a steep run of Victorians that climbs from Cole Valley/Parnassus Heights toward Ashbury Heights. In 2025 it was closed to cars from 4 to 10 p.m., and Shrader Street between Frederick and Beulah closed from 3 to 10 p.m. The SF Standard's guide puts it bluntly: go early and expect crowds. Candy runs out on the popular houses, per Mommy Poppins.

Take the N Judah to Carl and Cole and walk up. The neighborhood scores 4.1 overall and 4.4 on safety. The trade-off is the hill. Belvedere is a climb, and the stoops are tall.

Noe Valley: 24th Street and the flat blocks

Noe Valley is the flat-blocks pick. SFGate describes the parade of costumed toddlers along 24th Street between Church and Castro starting around 3 p.m., and the side streets are flatter than most of the city. Fair Oaks Street, just across Dolores on the Mission side, closed from 22nd to 26th streets from 5 to 8:30 p.m. in 2025.

Noe Valley scores only 3.2 overall, 61st of 92 in San Francisco, mostly because of school scores. Its safety score is 4.2. On Halloween, that overall number is irrelevant. The J Church stops at 24th Street.

Jordan Park / Laurel Heights and Presidio Heights

Jordan Park / Laurel Heights quietly runs one of the biggest closure clusters in the city. Commonwealth Avenue and Jordan Avenue were car-free from California to Geary from 4 to 10 p.m. in 2025, and Iris Avenue from 3 to 10 p.m. The streets are wide and flat. It is also our top-ranked San Francisco neighborhood, at 4.2 overall and 4.6 on safety.

Next door, Presidio Heights (4.2 overall, 4.3 safety) works as a loop: Sacramento Street merchants, then Washington Street and the Presidio Terrace cul-de-sac off Arguello, the route the SF Standard recommends.

Sea Cliff and Lake Street

Sea Cliff has the full-size-candy-bar reputation, and several hundred trick-or-treaters show up each year, according to Mommy Poppins. Bay City News found the heaviest decorating between 28th and 32nd avenues in 2025, per Local News Matters. If the core gets too packed, the Standard suggests staying near California Street and 32nd Avenue.

A few blocks east, Lake Street is both a Houseberry neighborhood and a Slow Street. The 5th-to-9th Avenue and 23rd-to-25th Avenue stretches draw the crowds, and the block between 6th and 7th avenues closed from 3:30 to 9 p.m. in 2025. Both neighborhoods score 4.5 on safety.

North Panhandle, Glen Park and Bernal Heights

These three are the block-party neighborhoods. In North Panhandle, Grove Street between Baker and Central and Lyon Street between Fulton and Hayes closed from 3 to 10 p.m. in 2025, and the 1500 and 1600 blocks of Grove host a kids' costume contest and a haunted house.

Glen Park closes Chenery Street between Elk and Diamond, 4 to 9:30 p.m. last year, and Glen Park BART is a block away. In Bernal Heights, Cortland Avenue from Bennington to Gates plus Moultrie, Anderson and Ellsworth streets went car-free from 4:30 to 8:30 p.m. Glen Park scores 4.2 on safety. Bernal Heights scores 3.5.

West Portal, Saint Francis Wood and the Outer Sunset

West Portal has had a Wawona Street block party since the 1980s, per Mommy Poppins, a short walk from West Portal Station. Wawona did not appear on Mission Local's 2025 closure list, so check before you count on it. Saint Francis Wood closed Terrace Drive from Portola to Santa Clara for the longest window on the list, 2 to 10 p.m.

The Outer Sunset had more separate closures than any other area in 2025, including 36th, 42nd, 43rd and 45th avenues. It scores 4.2 overall and 4.3 on safety, second in the city overall.

Oakland, Berkeley and Alameda: the porch-dense East Bay

The inner East Bay's Halloween map is short and well known. These blocks pull people from all over the region, so arrive before dusk if you want a parking spot.

Crocker Highlands and Trestle Glen

Trestle Glen Road east of Lakeshore Avenue is Oakland's densest trick-or-treat corridor. Red Oak Realty notes the houses sit close together and the street gets crowded. The road runs through two Houseberry neighborhoods: Crocker Highlands, which is first of 130 Oakland neighborhoods at 4.3 overall and 4.8 on safety, and Trestle Glen, at 3.5 overall and 3.4 on safety.

Haddon Hill, the cluster of big houses with decorated stoops and front-yard graveyards near Lake Merritt, is also on every list. We do not map it as its own neighborhood, so start from our Oakland neighborhood rankings if you are looking at homes there.

Rockridge

Rockridge does two Halloweens. The merchant parade and trick-or-treat on College Avenue ran from Broadway to Alcatraz on the Sunday before Halloween in 2025, starting at the Rockridge BART plaza, per 510 Families. On the night itself, parents on the Berkeley Parents Network point to Boyd Avenue between Hudson and Clifton. Rockridge scores 3.8 overall, 3.4 on safety and 4.6 on amenities.

Elmwood and Claremont: Russell Street

Russell Street from College Avenue up to the Claremont Hotel is Berkeley's big one, with elaborate yard graveyards and haunted houses, per SFGate. It starts in Elmwood and climbs into Claremont, both 3.9 overall and 3.7 on safety. The Claremont Elmwood Neighborhood Association sponsors it, and the City Council approved up to $1,000 per council office toward the 2025 event, calling it one of the best trick-or-treat spots in Berkeley in the council item. Take younger kids early. Crowds peak after sunset.

Alameda: the Gold Coast and the East End

Alameda's Gold Coast is the region's biggest single draw. One resident near Grand and Paru streets estimated about 2,500 visitors at his corner and gave out 5,700 pieces of candy, the Standard reported on October 31, 2025. Families come from Oakland and San Francisco for the flat, lit streets. We map Alameda's main island as two large neighborhoods, West End and East End, rather than using the Gold Coast name, so check the Alameda rankings for the blocks you care about.

In the East End (4.1 overall, 3.9 on safety), Berkeley Parents Network regulars send people to Thompson Avenue between High Street and Fernside Boulevard.

The Peninsula: fewer blocks, bigger candy budgets

The Peninsula's famous Halloween streets are narrower and more residential, and the numbers can be startling.

Easton Addition, Burlingame: Cabrillo Avenue

Five blocks of Cabrillo Avenue between Adeline Drive and Carmelito Avenue in Easton Addition are the Peninsula's headline block. One family told the Burlingame B it buys 20,000 pieces of candy a year. SFGate points out the flat street and close-together houses with no stairs. Easton Addition scores 4.0 overall, with a 4.8 school score and 3.2 on safety.

Old Palo Alto, College Terrace and Evergreen Park

Old Palo Alto is big enough on Halloween that Palo Alto Online has run traffic advisories about streets closing there. It scores 4.6 overall and 4.7 on safety. In Evergreen Park, residents told Palo Alto Online their count went from zero to hundreds in a few years, and a College Terrace resident described hundreds of kids coming through. Both neighborhoods lean on the Nextdoor treat map.

The Willows, Menlo Park

Pope Street in The Willows shows up in both Mommy Poppins and Silicon Valley agent roundups, with the same advice: bring little kids early. The Willows scores 4.3 overall and 4.3 on safety.

South Bay, Marin and the Tri-Valley

Willow Glen, San Jose

Willow Glen is San Jose's Halloween center. The Willow Glen Business Association's daytime trick-or-treat along Lincoln Avenue ran from Minnesota Avenue to Willow Street in 2025, with a toddler session from 11 a.m. and a school-age session from 3 to 4:30 p.m., per a local agent's recap. The flat grid of bungalows around Lincoln carries the evening. Willow Glen scores 3.7 overall and 4.0 on safety. Naglee Park near downtown also appears in Silicon Valley agent roundups, though our scores for it are still in progress.

Downtown Los Gatos: Tait Avenue in Almond Grove

Tait Avenue, from West Main Street to Los Gatos-Saratoga Road in the historic Almond Grove area of Downtown Los Gatos, closes to cars for the night. Residents have historically handed out 1,000 to 2,000 pieces of candy each, the Los Gatan reported in 2025. The neighborhood scores 4.1 overall and 3.6 on safety.

Sycamore Park, Mill Valley, and Vintage Hills, Pleasanton

Mill Valley closed Sycamore Avenue from Park Avenue to Valley Circle and Walnut Avenue from Park to Locust for Halloween 2025, the city's newsletter said. Those streets sit in the flat, sidewalked part of town Houseberry maps as Sycamore Park, 4.5 overall and 4.4 on safety.

In the Tri-Valley, Vintage Hills east of downtown Pleasanton made Bay Area Telegraph's 2025 list for walkable streets and high participation. It scores 4.3 overall and 4.3 on safety.

Every neighborhood in this guide, side by side

Here is the full list with the street that draws the crowd and the two Houseberry scores most people ask about.

Seven of the 12 San Francisco neighborhoods rank in the city's top 20 overall. Three rank below 50th. And two of the East Bay's busiest Halloween destinations, Russell Street and the Gold Coast, sit in neighborhoods scoring under 4.0 on safety. On Halloween night, closed streets and a lot of people outside do much of the work.

Timing, parking and getting there

Go before dark if your kids are small. With sunset around 6:13 p.m. on October 31, 2026, the 4:30 to 6 p.m. window gives you daylight, shorter lines and houses that still have candy. The famous blocks get loud and packed after dusk, especially on a Saturday.

Take transit where you can. The N Judah drops you at Cole Valley, the J Church at 24th Street in Noe Valley, BART at Glen Park and Rockridge, and Muni Metro at West Portal Station. Parking near Belvedere, Russell Street, the Gold Coast and Cabrillo Avenue fills early.

Check the 2026 closure list before you go. San Francisco's permits change year to year, and a block that was closed in 2025 may not be this year. If you are hosting, Nextdoor's treat map is how many neighborhoods, including those Palo Alto blocks, tell visitors which porches are open.

And bring more candy than you think if you live on one of these streets. The Cabrillo Avenue family buying 20,000 pieces is not overdoing it.

If you are house hunting and wondering what Halloween looks like on a given block, it is a surprisingly good read on a neighborhood: how close the houses are, how many people walk, and whether neighbors organize. We built Houseberry to answer the rest of that question. Start with our San Francisco neighborhood map and rankings and click into any neighborhood above.

Sources

• Mission Local, Here's every car-free street in S.F. for Halloween this year (Oct. 30, 2025)
• The San Francisco Standard, On Halloween, Alameda's Gold Coast is a living nightmare (Oct. 31, 2025)
• The San Francisco Standard, San Francisco trick-or-treating guide (Sept. 11, 2023)
• Local News Matters / Bay City News, Where to find SF's best Halloween-decorated homes (Oct. 27, 2025)
• SFGate, Best trick-or-treating neighborhoods in the Bay Area
• Mommy Poppins, Best neighborhoods to trick-or-treat in San Francisco
• Red Oak Realty, Where to trick or treat in the East Bay
• Berkeley Parents Network, Where to go trick or treating
• City of Berkeley, Russell Street Halloween Festivities council item (Oct. 28, 2025)
• 510 Families, Rockridge Halloween Parade (2025)
• The Burlingame B, Halloween on Cabrillo Avenue (Oct. 31, 2023)
• Palo Alto Online, Traffic advisory: Halloween festivities close some Old Palo Alto streets (Oct. 31, 2018)
• Palo Alto Online, Destination: trick or treat (Oct. 25, 2015)
• Michelle Elliott, Halloween in Willow Glen (2025)
• Nancy Dinshaw, Best trick-or-treat spots in Silicon Valley (Oct. 11, 2025)
• Los Gatan, Witches, skeletons and goblins flock to Los Gatos streets (Oct. 30, 2025)
• City of Mill Valley, MVConnect newsletter (Oct. 23, 2025)
• Bay Area Telegraph, 5 neighborhoods in the 925 where people love to trick or treat (Oct. 30, 2025)
• NOAA solar position formula, San Francisco sunset for Oct. 31, 2026 (Houseberry calculation)
• Houseberry neighborhood rankings, San Francisco

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

On January 2, business owners along Gate 5 Road in Sausalito told NBC Bay Area the water was the worst they had seen in more than 20 years. That was a normal king tide season with a storm on top. This fall the Bay starts a step higher.

A Kelvin wave, the slow bulge of warm ocean that El Niño pushes up the coast, could raise Bay Area sea levels by 6 to 12 inches starting in October, according to NC State marine scientist Dillon Amaya in ABC7's September 18 report. October's own tides are modest, so the real test is when that extra water lines up with the Bay Area king tides on November 24 to 26, December 23 to 25 and January 21 to 22. The blocks that feel it first are the ones that already get wet on an ordinary king tide.

The takeaways

• A Kelvin wave could add 6 to 12 inches to Bay Area water levels from October 2026, with three or four more waves possible over the winter and levels expected to settle by spring 2027.
• NOAA predicts the season's highest tide at the San Francisco gauge at 7.23 feet on December 24, 2026, at 10:37 a.m., before any Kelvin wave lift is added.
• At that gauge, predicted tides reach 7 feet on 7 days between October and January. Add a foot and the count is 54.
• Standard homeowners policies typically exclude flooding, and new flood insurance generally takes 30 days to start, per California's insurance department.

What a Kelvin wave actually does to the Bay

A Kelvin wave is not something you would see from the beach. Amaya describes it to ABC7 as "a bulging of the nearest 20 to 30 kilometers of the ocean" that travels up the coastline, which means it raises the baseline for weeks and every high tide rides on top of it.

KION and CBS put the first wave here in the first or second week of October. They also note that September and October tides are among the year's lowest, so the October bump alone is unlikely to set records. The worry is later waves.

And the water is already running high. A National Weather Service coastal flood advisory covered San Francisco, San Pablo and Monterey Bays from September 28 to October 1, with tides forecast to about 7.03 feet and no storm in sight, per Hoodline. On the Peninsula, forecasters said tides were running about a foot above normal and could spill onto trails at Redwood Shores, Bair Island and the Palo Alto Baylands, The Almanac reported.

This is the next chapter of our August piece on checking flood risk before a super El Niño winter, which covered rain and buyer homework. This one is about tides, dates and specific shorelines.

The king tide dates, with NOAA's predicted heights

The California Coastal Commission lists this season's statewide king tides as November 24 to 26, December 23 to 25, and January 21 to 22, with some Northern California spots also peaking December 22 and January 20 or 23 (Coastal Commission). Here is what NOAA's tide tables predict for two Bay stations, San Francisco at the Golden Gate and Alameda.

Heights are feet above each station's mean lower low water, so compare down a column, not across. Alameda reads higher because the tide range grows toward the south Bay. And every peak lands in daylight, which makes them easy to go look at.

The late October row is the sleeper. If the first Kelvin wave delivers its full foot during the October 26 to 29 high tides, simple addition puts San Francisco near 7.67 feet, higher than a normal Christmas king tide. That is an if, not a forecast.

We ran that count ourselves from NOAA's predictions. The lift will not be a flat foot all winter, so read the band as a what-if range. But it is why the whole season matters, not just three marked mornings.

Which shoreline neighborhoods feel it first

The short answer is the places that already flood on a regular king tide. Nothing below is a new exposure rating. It is where agencies and local reporting already show tidal water reaching streets, lots and shorelines.

Southern Marin

Tam Junction, the Mill Valley crossroads where Shoreline Highway meets Almonte Boulevard, sits just up the road from the Manzanita Park and Ride, the lot Caltrans partly closes when high tides are projected. Hoodline put the closure threshold at about 6.2 feet. Caltrans' US-101 and Highway 1 sea-level rise project covers the corridor from Marin City near Donahue Street to the Richardson Bay Bridge, but its draft environmental document is not due until summer 2027. This winter, the fix is still sandbags and tide gates.

In Sausalito, the Marinship waterfront holds the floating homes at Waldo Point Harbor and the Gate 5 Road businesses that flooded in January. The City Council has authorized $250,000 for El Niño work, including raising the road near Varda Landing before the king tides, ABC7 reported. Marinship's median sale price was about $1.01 million in August 2026 in our data, with a 4.1 overall score. That is the honest picture: a well-liked neighborhood with a known low spot and a city spending on it.

Alameda

Bay Farm Island is Alameda's top-ranked neighborhood in our data, with a 4.6 overall score, 4.9 for safety, and a median of about $1.38 million in July 2026. It is also where the City Council voted on April 7 to move shoreline adaptation to 60 percent design with a $2.64 million state grant, including an interim raise of Island Drive, the Alameda Post reported. Construction is tentatively aimed at around 2030.

On the main island, January's king tides pushed waves over the dunes onto Shore Line Drive, and USGS researcher Patrick Barnard told KQED that about 60 percent of Alameda's groundwater already sits close to the surface.

San Francisco's east waterfront

Mission Bay is the newer neighborhood wrapped around Mission Creek and the UCSF Mission Bay campus, just south of Oracle Park. The Port of San Francisco says stormwater already floods the Mission Creek area, and its Mission Creek and Mission Bay resilience page points to the Waterfront Flood Study plan finalized in August 2026. Tidal water reached the Embarcadero as recently as June (ABC7). Mission Bay's median was about $1.5 million in August 2026, in our numbers.

The Peninsula bayfront

Redwood Shores was named in this week's advisory. It is also the top-ranked neighborhood in Redwood City in our data, with a 5.0 school score and a median near $2.38 million in July 2026. Next door, Foster City finished a $90 million, 6.5-mile levee raise in February 2024 (Foster City), which is exactly the kind of fact worth checking when you compare Foster City's neighborhoods. East Palo Alto's shoreline story is its own, and we covered it in our piece on Palo Alto's living levee and the SAFER Bay project.

Those prices are why levee money matters. A rising Bay is also a good argument for adding homes on higher ground near transit, while funding the levees that protect the homes already here.

What the insurance warning means for owners and buyers

A standard homeowners policy typically will not pay for rising bay water. In a September 22 release, the California Department of Insurance said homeowners, renters and commercial policies typically exclude flooding, and that flood insurance through the National Flood Insurance Program generally takes effect 30 days after purchase, except when it is required as part of a home purchase (CDI).

That 30-day clock turns the tide table into a calendar. Coverage bought after about October 25 will not be active for the November king tides. Coverage bought after about November 23 misses the Christmas peak. If you are closing on a shoreline home, the lender-required policy is the exception, but the premium still belongs in your monthly math before you write the offer.

How to read your own block this season

Go look. Pick one of the king tide mornings in the table, stand at the lowest corner of the street at the listed time, and watch where the water goes. The California King Tides Project collects public photos from these mornings, and they are worth scrolling for your shoreline before you visit.

Notice the small tells while you are there: fresh sandbags by garages, flap gates on storm drains, a city notice on a pole. Ask a neighbor what January 2 looked like. Then ask the city who maintains the nearest levee or seawall, and what is funded.

The water this winter is a preview of what ordinary tides look like a few decades from now. The blocks that handle it well will show it in daylight on a weekday in December. If you are weighing a move to the water, we would start with the ranked neighborhoods on Houseberry's Marin page and then spend a king tide morning on the one street you actually care about.

Sources

• ABC7: El Niño-driven Kelvin wave could raise Bay Area sea levels by up to 1 foot in October (Sept 18, 2026)
• The Guardian: California rushes to prepare for massive Kelvin wave and predicted sea level rise (Sept 30, 2026)
• iHeart: Kelvin wave could lift California sea levels by a foot (Sept 28, 2026)
• KION: Bay Area sea levels expected to rise nearly 1 foot due to planetary wave (Sept 6, 2026)
• Hoodline: SF Bay coastal flood advisory (Sept 28, 2026)
• The Almanac: National Weather Service issues coastal flood advisory (Sept 28, 2026)
• California Coastal Commission: California King Tides Project dates
• NOAA tide predictions, San Francisco station 9414290
• NOAA tide predictions, Alameda station 9414750
• California Department of Insurance: Commissioner Lara urges Californians to review coverage (Sept 22, 2026)
• NBC Bay Area: King tides cause major flooding in Marin County (Jan 2, 2026)
• Caltrans: US-101, State Route 1 Sea-Level Rise Project
• Alameda Post: Council advances Bay Farm Island shoreline adaptation
• KQED: For this Bay Area island city, water is coming from all sides
• Port of San Francisco: Mission Creek / Mission Bay resilience
• Foster City: Levee Improvements Project completed
• Houseberry neighborhood data for Marinship, Bay Farm Island, Mission Bay and Redwood Shores, July and August 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

A 750-square-foot backyard unit in Alameda would likely sell for somewhere around $580,000 as a condo. The house in front of it sits in a city where Houseberry's typical home price hit $1.42 million in August 2026.

That gap is the whole story. Alameda is one council vote away from letting homeowners sell an ADU separately from the main house, and for the first time a backyard cottage on the island could become its own deed, its own mortgage and its own entry-level home.

Yes, you will be able to sell an ADU separately in Alameda, once the new rules take effect. The City Council gave the ordinance a 4-0 first vote on September 15 and is set to adopt it on the October 6, 2026 consent calendar. It takes effect 30 days later, which puts the first applications in early November.

The numbers that matter

• Alameda's new Section 30-5.19 lets an ADU be conveyed as a condominium through a ministerial map with no public hearing, per the ordinance text up for final passage on October 6, 2026.
• ADUs can run up to 1,200 square feet under the updated rules, and ADUs under 750 square feet pay no development impact fees.
• Alameda builds roughly 50 ADUs a year, according to city planning staff as reported by the Alameda Post.
• Houseberry's Alameda citywide median was about $1.42 million in August 2026, versus $974,000 in Oakland and $1.73 million in Berkeley.
• California's first ADU condo sale closed in San Jose in June 2026 at $530,000 for 749 square feet.

What Alameda actually approved

The ordinance does two jobs. It rewrites Alameda's ADU code (Section 30-5.18) and adds a new condo section (30-5.19) that uses the state law written for exactly this, AB 1033, which let cities opt in to separate ADU sales starting in 2024.

The building rules first, because they decide what can be sold later. Per the council file (2026-6314):

• ADUs top out at 1,200 square feet. Junior ADUs, which sit inside the main house, top out at 500.
• A single-family lot can carry one attached ADU, one JADU and one detached new-construction ADU of up to 800 square feet.
• A lot with an existing multifamily building can add up to eight detached ADUs, capped at the number of existing units. A proposed multifamily project gets up to two.
• Detached ADUs can rise 18 feet, or 20 if the roof pitch matches the main house, with 4-foot side and rear setbacks. No off-street parking is required.
• ADU permits get ministerial review within 60 days of a complete application.

The condo section is written for ADUs. A junior unit carved out of the main house is not a candidate in the text we read, which makes sense. You cannot really sell a bedroom suite with a shared wall and a shared furnace.

The Alameda Post's account of the September 15 meeting had Mayor Marilyn Ezzy Ashcraft framing it as a first-time-buyer foothold, while staff planner Tristan Suire called it "one tool in the toolbox." That is the honest description.

The six steps, and the two you do not control

Selling an ADU as a condo in Alameda takes six steps, and two of them depend on someone else saying yes. Here is the order the ordinance sets up.

The lender step is the one that will quietly kill the most deals. Every lienholder has to sign a consent to the condo plan, and the ordinance language gives them "sole and absolute discretion." If you have a big first mortgage on the property, your bank is being asked to let part of its collateral walk out the door. San Jose's planning staff flagged the same problem to KQED in 2025. Owners with little or no debt have a much easier path.

The tenant step is slower but predictable. If someone already rents the ADU, they get the first chance to buy it at the same price you would offer the public, for at least 60 days after they receive the state Department of Real Estate subdivision report.

Before the map is approved, the ADU also needs a safety inspection (a certificate of occupancy or a HUD housing quality standards report), and the two owners need a homeowners association with recorded CC&Rs. Yes, a two-member HOA. It will handle shared landscaping, exterior upkeep and utilities, and it will need rules both sides can live with for decades.

What the paperwork is likely to cost

Expect roughly $10,000 to $50,000 in soft costs to turn an existing, permitted ADU into a sellable condo, before any construction or utility work. Alameda has not published a dollar figure, so that range comes from other places doing the same thing.

In Sebastopol, a councilmember estimated $10,000 to $20,000 for a typical conversion, including $2,000 to $5,000 for a surveyor, $3,000 to $8,000 for an attorney to draft CC&Rs, and $1,000 to $3,000 in title and recording fees. Industry guides put the wider range at $15,000 to $50,000 once utility separation enters the picture.

Utilities are the wild card. A 1920s Alameda lot with one gas meter and one sewer lateral serving both buildings is a very different project than a new ADU built with its own meters. The ordinance only requires that you notify the utilities and prove it to the city, but a buyer's lender will want to know who pays which bill.

Against a sale in the high $400,000s to $600,000s, even $50,000 is not fatal. It just means this is a project for owners who are sure they want to sell, not a casual option to keep in the drawer.

What would an ADU condo sell for in Alameda?

A 500 to 750 square foot ADU condo in Alameda would plausibly land around $390,000 to $580,000, or roughly a quarter to 40 percent of what a typical Alameda house costs. That estimate multiplies Alameda's Redfin median of $773 per square foot in August 2026 by typical ADU sizes, and it lines up with San Jose's real sale.

On our own numbers, the West End median was about $1.45 million in August 2026, the East End about $1.40 million, and Bay Farm Island about $1.38 million in July 2026. Very few houses on the island trade anywhere near the ADU condo price band.

Two cautions on that math. Per-square-foot prices are an average across big houses and small condos, and small units sometimes command more per foot, sometimes less. And a condo with a two-member HOA and a shared driveway is a new product with no local sales history yet. Appraisers will be guessing for a while, which is why we label these as estimates and not comps.

Where the lots make this pencil

The East End and the older blocks of the West End are where Alameda ADU condos are most likely to show up first, because that is where the long, narrow lots are. Bay Farm Island is a harder case.

The East End, the part of the island between Park Street and the Fruitvale and High Street bridges, scores 4.1 overall and 4.4 for amenities on Houseberry. Its pre-war housing stock sits on deep lots with space behind the house and, often, a side path. That side path matters more than people think. A condo buyer needs a way in that does not run through someone else's kitchen door. Alameda's historic building study list counted about 4,000 buildings, a little under half of them Victorians, per SFGate.

The West End, the western half of the main island out toward Alameda Point, scores 3.4 overall and 3.0 on safety, with a 4.5 on public spaces. Its mix of older cottages and postwar homes on standard lots gives it plenty of candidates, and it is also where Alameda's bigger housing pipeline sits, as we covered in our piece on the 2,300 homes stalled at Alameda Point.

Bay Farm Island, the largely postwar planned community across San Leandro Bay next to Oakland Airport, ranks first in Alameda at 4.6 overall. But many of its homes sit inside the Community of Harbor Bay Isle association structure, on smaller, regular lots. Adding a second HOA layer on top of an existing one is the kind of thing lawyers enjoy and buyers do not.

How Alameda compares with Berkeley, San Jose and Oakland

Alameda joins San Jose and Berkeley among Bay Area cities with a working ADU condo route. Oakland, as far as we could find in late September 2026, has not adopted one.

Berkeley's version, per city staff's presentation and The Real Deal, is built around its rent-control system, which is why the tenant window is longer. Alameda's 60-day floor is simpler. San Jose went first, and our look at that first sale covers how a developer ran it as a proof of concept.

Oakland is the notable holdout. Its median of about $974,000 in August 2026 is the lowest of the four, so the price gap between an ADU condo and a house is narrower there. Maybe that makes the policy less urgent in Oakland. Or maybe it means Oakland ADU condos would be the only ones a lot of buyers could actually reach.

What this does and does not mean

It does mean an Alameda owner with a permitted ADU and a cooperative lender can, starting in November, turn that unit into cash without selling the house. It does not mean every backyard unit becomes a condo. At about 50 ADUs a year, this will be a trickle. The Planning Board discussion in May noted that converting every existing ADU would touch less than 3 percent of Alameda's rental stock.

Can my tenant stop the sale? No. They get the first offer at the public price for at least 60 days, not a veto.

Do I have to live on the property? The condo section does not add an owner-occupancy rule, and the broader update drops the owner-occupancy requirement for JADUs, per the Alameda Post's June report.

What if I have not built the ADU yet? Then the smarter move is to design it as a future condo from day one: its own meters, its own entrance path, under 750 square feet if you want to skip impact fees. Our rundown of California's 2026 ADU laws covers the state-level rules that apply alongside Alameda's.

I am for this. Alameda has spent decades being an island where the cheapest way into a house was a $1 million-plus mortgage. A $500,000 cottage with its own deed, a few blocks from Park Street, is exactly the kind of starter home the Bay Area stopped building. It will not fix affordability. It adds a rung to a ladder that has been missing a few.

If you are weighing whether a particular block has the lot depth and the neighborhood numbers to make it work, start with how Alameda's neighborhoods rank and work backward to the parcel. That is the order we would do it in.

Sources

• City of Alameda, Ordinance text amending Sections 30-5.18 and adding 30-5.19 (Legistar)
• City of Alameda, Council File 2026-6314, October 6, 2026 consent calendar (Legistar)
• Alameda Post, Council Advances Housing Rules, AC Transit Warns of Route Cuts
• Alameda Post, Planning Board Supports ADU-to-Condo Conversions
• Alameda Post, Planning Board Endorses New ADU-to-Condo Conversions
• California Legislature, AB 1033 bill status
• City of Berkeley, ADU Separate Sale Implementation presentation, January 2026
• The Real Deal, Berkeley approves ADU sales, January 22, 2026
• KQED, San Jose developers pioneer new California law selling ADUs as condos
• Sebastopol Times, Sebastopol City Council approves ADU condo conversions
• Redfin, Alameda housing market, August 2026
• SFGate, Step back in history with Alameda's Victorians
• Houseberry, Alameda city data and 12-month price history

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

One bill in Tuesday's housing package is already law. SB 328, Senator Tim Grayson's bill for the former Concord Naval Weapons Station, carries an urgency clause, so it took effect the moment Governor Newsom signed it on September 29, 2026. Almost everything else he signed that day waits for January 1, 2027.

That split is the most useful way to read the package. The Governor's office listed 34 housing and homelessness bills signed in San Francisco on Tuesday. We already took two of them apart, AB 956 and its two detached ADUs and SB 1116's small-lot splits, so this piece covers what is new since our September 2 list, sorted by who actually feels it.

What you need to know

Four groups of Bay Area owners and builders get something new, on three different clocks.

• SB 328 is in effect now. The Concord base redevelopment, planned at 12,272 homes with 25 percent affordable, can lean on environmental reviews the city already finished instead of starting new ones.
• AB 1751 makes qualifying townhome subdivisions a ministerial approval starting January 1, 2027. No hearing, no CEQA review. San Francisco is exempt.
• SB 1117 limits ADU impact fees to the square footage above 750, starting January 1, 2027.
• AB 1815 and AB 2058 stop cities from stacking their own building standards and inspection fees on factory-built homes.
• Nine of the 11 bills on our September 2 list are confirmed signed, and we found no vetoes. AB 1684 and SB 677 had not appeared in a published signing list as of this morning.

The September 2 scorecard, settled

Nine of the eleven bills we flagged on September 2 are now law. AB 956, SB 1117, SB 1116, SB 1014 and AB 2074 came through on Tuesday. The insurance pair, SB 1301 on non-renewal notice and AB 1680 on FAIR Plan oversight, were signed September 27 along with AB 739 on HOA managing agent fees, according to the Governor's September 27 update.

One correction to our September 21 piece. We wrote that none of the eleven had moved. AB 2576, the historic-site carve-out from SB 79, was actually on the Governor's September 14 list. We missed it.

That leaves AB 1684, the right to install air conditioning in a common interest development, and SB 677, the SB 79 cleanup bill. Today is the deadline, and under Article IV, Section 10 of the state constitution a bill he neither signs nor vetoes by September 30 becomes law anyway. So a veto is the only outcome that matters for those two, and we have not seen one.

Here is everything new since then that reaches a Bay Area lot, in one place.

Read the third column before anything else. Only one of these changes anything this year, and it is the one attached to a 2030 groundbreaking.

If you live near the Concord base, the legal risk just got smaller

SB 328 does not approve a single house. What it does is take the biggest legal risk off the table for the largest housing plan in the Bay Area.

The Naval Weapons Station's inland area covers 5,028 acres inside Concord city limits, next to North Concord/Martinez BART. Concord's plan with master developer Brookfield calls for 12,272 homes, 3,068 of them deed-restricted affordable, plus about 6 million square feet of commercial space, according to The Real Deal's account of the May approval. In May the City Council unanimously approved the money side of the deal with the Navy, roughly $628 million over about 30 years, with the first transfer of about 1,000 acres scheduled for 2030.

Under SB 328, projects consistent with the city's area plan can rely on environmental documents Concord already completed. The land deal has to restrict at least 25 percent of the homes to lower-income households, for 55 years on rentals and 45 years on owner-occupied homes. Grayson's office put it plainly: the bill will "leverage already completed environmental reviews to help get shovels in the ground sooner." Concord Mayor Laura Nakamura called it certainty to move "from vision to reality."

Why that matters: a fresh CEQA lawsuit against the Specific Plan could have added years to a project that has already spent almost two decades in planning. The trade-off is real too. The reviews it leans on are old, starting with a notice of preparation filed in 2007, and Highway 4 traffic and the remediation still ahead have not stood still since. The park acreage is settled. The traffic math is not.

What does that mean for existing homes? Concord's citywide median was about $804,600 in August 2026, on our Concord city page. Sun Terrace, the older single-family pocket at the city's far northern edge near Olivera Road and Highway 4, ran about $729,100 in August and scores 2.6 of 5 overall, with amenities at 2.6. North Willow Pass, between downtown and the base, was about $745,100 and scores 3.0. At the other end of our 42-neighborhood Concord ranking, hillside Crystyl Ranch scores 4.3 at about $1.4 million.

The north Concord neighborhoods score lowest on everyday amenities, and they sit closest to a plan with 6 million square feet of commercial space and 55 acres set aside for schools. That is the upside, on a 2030s timeline rather than a 2027 one. (If you want something moving sooner, the city is also working on a roughly 900-home plan for the old Coast Guard housing off Haleakala Street.)

If you want to build or buy a townhome, AB 1751 is the big one

AB 1751, the Missing Middle Townhome Ownership Act, turns a qualifying townhome subdivision into a checklist approval starting January 1, 2027. No hearing, no planning commission vote, and no CEQA review of the map.

The details, from the enrolled text:

• A townhome is a single-family dwelling of no more than three occupiable stories that shares a wall or sits a fire separation distance from its neighbor. New parcels can be as small as 600 square feet.
• Eligible land is zoned for multifamily housing, or zoned only for single-family and underutilized, meaning no permanent home on it unless that home is abandoned and uninhabitable.
• Very high fire hazard severity zones, wetlands, prime farmland, flood zones and conservation land are out. So is demolishing rent-controlled homes or units taken off the rental market in the past 15 years.
• Construction workers must earn at least $28 an hour, adjusted each year for inflation.
• A city can deny only with written findings of a specific health or safety impact it cannot mitigate. San Francisco is exempt entirely.

The line that decides where this bites is the density test. A project must reach at least 75 percent of the default density in state housing element law. By our reading, for a city of more than 25,000 people in a metro area our size that default is 30 homes per acre, so the floor lands at 22.5. Smaller towns such as Atherton or Moraga use a 20-per-acre default, which puts their floor at 15.

We ran that test against five townhome projects we have reported on this year. Two clear it.

Sunnyvale's 137 townhomes on Kifer Road and Santa Clara's 16 at 860 Civic Center would qualify on density. Cupertino's 27 on Bandley Drive at 17.4 per acre, San Jose's 18 at The Alameda and Emory and Saratoga's 132 on Marshall Lane would not. The bill rewards the tight three-story rowhouse over the roomier suburban townhome most cities still ask for.

So watch the cities with plenty of multifamily-zoned land that still send townhome maps through hearings: Sunnyvale and Santa Clara, San Jose's urban villages, the Concord corridors. The fire exclusion does the opposite in the Oakland and Berkeley hills and much of Orinda, which sit largely in very high fire hazard zones. And the honest limit: AB 1751 removes the hearing and the lawsuit. It does not lower the price of land, which is what kills most Bay Area townhome deals.

If you have a backyard, SB 1117 is the fee fix

SB 1117 means that starting January 1, 2027, a city can charge impact fees only on the part of an ADU above 750 square feet.

Units of 750 square feet or less were already exempt. Above that line, current law lets a city charge fees proportional to the main house on the whole unit, so crossing the line by one square foot could trigger fees on all of it. Take a 1,000-square-foot ADU behind a 2,000-square-foot house. By our reading, today's fee is half of what a new house would pay. Under SB 1117 it is one-eighth.

Pair that with AB 956's second detached unit and the 2027 backyard gets noticeably cheaper to fill. We covered what AB 956 trades away in its own piece, and the rules already in force this year in July, so we will not repeat either here.

If you are betting on factory-built homes, the fight moved to City Hall

Eleven days after vetoing a state backstop for housing factories' surety bonds, Newsom signed the two bills aimed at the other bottleneck, local rules. AB 1815, by Oakland's Buffy Wicks, bars cities from holding factory-built housing to building standards that exceed or differ from the state code, except where local climate, geology or topography requires it. It also creates a presumption in favor of treating a project as factory-built.

AB 2058 caps what local agencies can charge to inspect and permit factory-built housing, bars inspectors from taking units apart to check them, and lets state-approved quality assurance agencies handle installation inspections once rules are written, due by January 1, 2029. For the Bay Area's big prefab ADU market, the practical gain is one set of rules instead of a different one in every city.

Permits get two hard deadlines

AB 1621 limits cities to two rounds of plan-check review and forces permit appeals to be decided within 30 business days for projects of 25 units or fewer and 45 business days above that. Missing those deadlines becomes a Housing Accountability Act violation, which lets a builder go to court to compel approval.

SB 1014, another Grayson bill, requires an itemized list of required onsite and offsite improvements within 30 business days. The sewer upgrade that shows up after the budget is set is how small infill projects die. This makes the city name it early.

Questions people are asking

When do the new California housing laws take effect?

Most take effect January 1, 2027. SB 328 took effect immediately on September 29, 2026. AB 2074 gives major transit cities until July 1, 2027 to designate hub districts, and most of SB 1301's insurance protections begin January 1, 2028.

Does AB 1751 apply in San Francisco?

No. The bill says it does not apply to the City and County of San Francisco. It applies to every other Bay Area city, subject to the site exclusions and the density floor.

Can I build townhomes on my single-family lot under AB 1751?

Not if your house is on it. Single-family-zoned land qualifies only when it has no permanent home, or the home is abandoned and uninhabitable. For an occupied lot, SB 9 and the SB 1116 changes are the paths to look at.

Does SB 328 mean construction starts soon at the Concord base?

No. It removes a legal hurdle. The Specific Plan still has to be finished, and the first land transfer from the Navy is scheduled for 2030.

Faster lanes, same traffic

Line the biggest three up and a pattern shows. The state keeps building faster lanes, and much of what Bay Area cities approve is designed for the slow one. Three of our five townhome filings fall short of AB 1751's floor. The Concord base has its fast lane and a 2030 start.

Whether any of this reaches a street near you depends less on Sacramento than on how your city zoned the land, which is the layer we keep coming back to when we compare neighborhoods on Houseberry. Concord's ranking is a fair place to see what that looks like block by block.

Sources

Office of the Governor, "Governor Newsom signs housing legislation," September 29, 2026

Office of the Governor, legislative update, September 27, 2026

Office of the Governor, bills signed September 14, 2026

CalMatters Digital Democracy, SB 328 bill record

CalMatters Digital Democracy, AB 1751 bill record

AB 1751 enrolled text, September 3, 2026, via LegiScan

CalMatters Digital Democracy, SB 1117 bill record

CalMatters Digital Democracy, AB 1815 bill record

CalMatters Digital Democracy, AB 2058 bill record

CalMatters Digital Democracy, AB 1621 bill record

Office of Sen. Tim Grayson, SB 328 signing release, via EIN Presswire

Contra Costa News, Governor Newsom signs SB 328, September 30, 2026

Contra Costa News, Concord council to review Navy financial agreement, May 13, 2026

The Real Deal, Concord approves Brookfield deal for Navy station homes, May 27, 2026

CEQAnet, 10268 Bandley Drive Residential Development Project, June 2026

Houseberry, Concord city page, neighborhood pages and rankings, retrieved September 30, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The ballot question for Walnut Creek Measure U is one sentence long, and the key phrase sits in the middle of it: redesignate a 2.4-acre site "from office uses to mixed use residential uses." That is the whole decision. On November 3, Walnut Creek voters choose whether the northeast corner of North California Boulevard and Ygnacio Valley Road, directly across from Walnut Creek BART, stays zoned for offices or opens up for 200-plus homes aimed at residents 55 and older.

When we wrote about this corner in July, the city council still had a choice to make. It made it. On July 21 the council voted unanimously to send the initiative to voters instead of adopting it outright, and it is now on the ballot as Measure U. This is the next chapter.

In brief

Here is where things stand one week before ballots go out.

• A yes vote rezones the 2.4-acre corner from office to a new senior-focused mixed-use district. A no vote leaves the office zoning in place.
• It passes with a simple majority of Walnut Creek voters. No argument against it was filed for the county voter guide.
• City analysts studied two build-outs: about 225 senior homes, or about 280 homes plus roughly 60,000 square feet of commercial space.
• Mail ballots go out October 5, and October 19 is the last day to register.

What Measure U actually says

Measure U amends three documents at once: the Walnut Creek General Plan, the North Downtown Specific Plan, and the Municipal Code. According to the [county's impartial analysis](https://www.contracostavote.gov/m9013-en/), it creates a new Mixed Use Residential Senior and Commercial district and applies it to one parcel.

The specs, from the ballot materials and the city's July staff analysis:

• Site: 2.4 acres at the northeast corner of North California Boulevard and Ygnacio Valley Road, across from the BART station
• Use: housing "serving primarily persons 55 years or older," plus compatible commercial space
• Base density: one unit per 425 square feet of net lot area
• Height: up to 89 feet, per the [city's §9212 analysis as reported from the July 21 meeting](https://citizenportal.ai/articles/8579254/California/Contra-Costa-County/Walnut-Creek-City/Walnut-Creek-places-Transit-Village-senior-housing-initiative-on-November-ballot-after-receiving-9212-report)
• Residential floor area ratio: 3.5 base, 4.5 maximum
• Proponent: Friends of Walnut Creek Senior Housing, backed by Hall Equities Group and property owner Mark Hall

Run the density math and one unit per 425 square feet works out to roughly 246 homes if the full 2.4 acres counted as net lot area. It will not, quite, once streets and setbacks come out. That is why the city's two scenarios bracket the likely range at 225 to 280.

Why a rezoning ended up on your ballot

Walnut Creek does not require a public vote to rezone a parcel. Measure U is on the ballot because it started as a citizen initiative, and state election law gives a council only two moves once an initiative qualifies: adopt it word for word, or put it before voters.

The signature drive was fast even by Bay Area standards. Proponents turned in 7,248 signatures on May 4 after about 25 days of gathering, [Contra Costa News reported](https://contracosta.news/2026/05/07/walnut-creek-senior-housing-initiative-moves-forward-following-record-signature-effort/). It qualified in June with 5,000 valid signatures against a 4,987 threshold.

The council could have adopted it in July and skipped the campaign. It chose the ballot unanimously and declined to write an argument against it. I read that as a council comfortable with the housing and uncomfortable being the ones who hand a developer a rezoning without a public vote. Fair enough. The opposite critique, which one resident put bluntly back in April as a "backdoor approach," is also on the record.

Downtown's scores explain the site better than the campaign does

The corner sits at the north edge of Downtown Walnut Creek, and our data on that neighborhood makes the case for housing here more plainly than any mailer. Downtown scores 4.6 out of 5 for nearby merchants and 4.8 for public spaces, the kind of numbers you only get where a pharmacy, a grocery store, Broadway Plaza, and a BART platform are all a short walk apart.

It also carries a 2.3 safety score and ranks 24th of 30 Walnut Creek neighborhoods overall at 3.3, according to our [Walnut Creek neighborhood rankings](https://www.houseberry.com/Nhranking/best-overall/Walnut-Creek-CA/). That trade-off is real. Downtowns are busier, and busy places log more incidents.

Here is how the site's surroundings compare with the other places people weigh when they want a walkable downtown and a BART station in central Contra Costa.

Two rows matter most for Measure U. [Downtown Walnut Creek](https://www.houseberry.com/Neighborhood/Downtown-Walnut-Creek-Walnut-Creek/) leads the group on amenities at 4.2. And Rossmoor, the 1,800-acre gated 55-plus community on the southwest side of town, scores just 0.9 out of 5 for nearby merchants. That gap is the argument for putting age-restricted homes downtown instead of at the edge of town, where most errands start with car keys.

Downtown prices have been softening, not climbing

Downtown Walnut Creek's median home price was about $800,360 in August 2026, down roughly 4 percent from about $832,900 in September 2025, based on the neighborhood's 12-month price history on Houseberry. Over the same stretch, Downtown Lafayette's median rose about 8 percent to roughly $1.56 million.

The steepest drop is next door. [North Civic](https://www.houseberry.com/Neighborhood/North-Civic-Walnut-Creek/), the small residential pocket around Civic Park a few blocks north of downtown, slid about 14 percent to roughly $470,300 over those 12 months. It mixes single-family homes and condos, and one small neighborhood's 12-month swing can reflect the mix of what sold, so treat it as a signal rather than a verdict.

So the honest reading is that 200-some senior homes are not arriving into a market that is running hot. If anything, a busier north end of downtown, with ground-floor shops where there is now an office block, is one of the few near-term changes that could give that corner more pull.

Lafayette and Pleasant Hill already ran versions of this experiment

Lafayette broke ground on June 25 on a seven-story, 87-foot, roughly 90-unit building at 1001 Oak Hill Road, steps from its BART station. We covered [Lafayette's seven-story BART project](https://www.houseberry.com/blog/lafayette-bart-seven-story-groundbreaking/) when it happened, in a town that measures its downtown in one- and two-story storefronts. Measure U's 89-foot limit is in the same range.

Pleasant Hill BART is a different model. The station sits in Contra Costa Centre, an unincorporated area run by the county, where apartments were built right up to the platform years ago through county planning, not a city ballot. Downtown Pleasant Hill proper, a couple of miles away, scores 3.7 on amenities and saw its median dip about 3 percent to roughly $697,400 in August 2026.

The lesson from both is plain. The 680 corridor is putting its density next to its BART stations, and the only real question in Walnut Creek is which process gets it there.

What Measure U does not settle

A yes vote rezones land. It does not approve a building. The proponents' own land use counsel told the council in April that any project "needs to be consistent and follow the city's set processes," which means design review and a formal application still come later.

There are real downsides to weigh, too.

• The city estimated $6.4 million to $7.7 million in affordable housing in-lieu fees. That suggests the affordable obligation could be paid as a fee rather than built on the site.
• Zoning passed by voters through an initiative generally cannot be changed by the council alone later, unless the measure itself allows it.
• The office space goes away. In an East Bay office market this soft, I do not lose sleep over that, but it is a trade.

The city's analysis also put ongoing net revenue at $257,000 to $552,000 a year, against $66,000 to $83,000 in added service costs. That is a modest fiscal win, not a windfall.

Measure U, answered quickly

What does a yes vote on Measure U do?

It changes the 2.4-acre corner at North California Boulevard and Ygnacio Valley Road from office zoning to a senior-focused mixed-use district that allows roughly 225 to 280 homes plus shops. A developer still has to bring a project through the city.

What happens if Measure U fails?

The site keeps its office designation. The owner could still apply for housing through the normal planning process, but the initiative route would be closed.

Who can live in the homes?

The district is written for housing serving primarily people 55 and older. Final eligibility rules would come with an actual project.

Who supports it?

[Greenbelt Alliance backs Measure U](https://www.greenbelt.org/blog/vote-yes-measure-walnut-creek/), alongside East Bay YIMBY and the Non-Profit Housing Association of Northern California. No formal ballot argument against it was filed.

Between now and November 3

The [City of Walnut Creek's election page](https://www.walnutcreekca.gov/government/departments/city-clerk/elections-voter-registration/november-3-2026-election-4243) lists October 5 for mail ballots and October 19 as the registration deadline. The same ballot carries the Connect Bay Area transit tax, a half-cent sales tax in Contra Costa that [helps fund BART operations](https://contracosta.news/2026/07/01/connect-bay-area-ballot-measure-to-save-public-transit-officially-qualifies-for-november-2026-election/amp/), which makes it a strange week to be voting on a BART-adjacent building without also voting on the trains.

My view has not moved since July. Two hundred-plus homes on a hard corner across from a BART station is exactly where central Contra Costa should be adding housing, and Measure U gets there with a clear density limit and no surprise height. If you are weighing homes near downtown, the neighborhood around that corner is about to get more interesting either way, and our [earlier piece on the senior housing initiative](https://www.houseberry.com/blog/walnut-creek-senior-housing-transit-village/) has the backstory. We will keep scoring the blocks around it the same way we score everything else, by what is within walking distance of the front door.

Sources

• [Contra Costa County Elections: Measure U ballot question and impartial analysis](https://www.contracostavote.gov/m9013-en/)
• [Mercury News: Walnut Creek voters to consider zoning change for more senior housing](https://www.mercurynews.com/2026/09/30/walnut-creek-voters-to-consider-zoning-change-for-more-senior-housing/)
• [Greenbelt Alliance: Vote yes on Measure U](https://www.greenbelt.org/blog/vote-yes-measure-walnut-creek/)
• [Citizen Portal: Walnut Creek places initiative on November ballot after §9212 report](https://citizenportal.ai/articles/8579254/California/Contra-Costa-County/Walnut-Creek-City/Walnut-Creek-places-Transit-Village-senior-housing-initiative-on-November-ballot-after-receiving-9212-report)
• [Citizen Portal: Walnut Creek council orders report on the initiative, April 21, 2026](https://citizenportal.ai/articles/7856390/california/contra-costa-county/walnut-creek-city/walnut-creek-council-orders-report-on-citizen-led-senior-housing-transit-village-initiative)
• [Contra Costa News: Senior housing initiative moves forward after record signature effort](https://contracosta.news/2026/05/07/walnut-creek-senior-housing-initiative-moves-forward-following-record-signature-effort/)
• [City of Walnut Creek: November 3, 2026 election](https://www.walnutcreekca.gov/government/departments/city-clerk/elections-voter-registration/november-3-2026-election-4243)
• [Contra Costa News: Connect Bay Area measure qualifies for November 2026](https://contracosta.news/2026/07/01/connect-bay-area-ballot-measure-to-save-public-transit-officially-qualifies-for-november-2026-election/amp/)
• [Houseberry: Walnut Creek neighborhood rankings](https://www.houseberry.com/Nhranking/best-overall/Walnut-Creek-CA/)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

$1,917 a square foot. That is the construction value San Jose recorded on September 22, 2026, for a 40,042-square-foot research lab MiTAC is building inside an existing industrial building at 1750 Automation Parkway in North San Jose. The total is $76,756,558, the second-largest building permit the city issued in the past 30 days, behind only a 512,400-square-foot affordable apartment complex.

MiTAC is the Taiwanese group whose computing arm builds servers for AI and cloud data centers. The permit does not say what the lab will design or test. But the company, the address and the price per foot all point the same way. MiTAC San Jose is shaping up as a serious AI server engineering site, and for anyone house hunting on the Berryessa side of town it is a jobs signal, not a housing story on its own.

Quick take

• Permit 2026-139041-CI, issued September 22, 2026, covers a $76.76 million R&D lab build-out of 40,042 square feet at 1750 Automation Parkway, ZIP 95131.
• That works out to about $1,917 per square foot, roughly double the next-priciest build-out in San Jose's 30-day permit feed.
• It is filed as a lab, not a data center. Nothing in the permit describes a server farm.
• No headcount, product line or opening date has been announced.
• The nearby neighborhoods we compare ran from about $1.13 million to $1.46 million in September 2026 on our pages.

What the MiTAC permit actually says

The permit is specific about scale and cost and silent about purpose. We pulled it straight from the city's Last 30 Days Building Permits dataset on September 29. It is listed as a tenant improvement, subtype "R & D Lab," with building, electrical, plumbing and mechanical reviews all marked complete. The contractor is Megawatt Construction. The owner of record is MLC V SC Automation LLC.

It is the second MiTAC permit at the address this month. A $500,000 demolition-only permit was issued on September 3, which is the usual order of things: clear the old interior, then build the new one.

The building has a recent history. 1750 Automation Parkway is one of four single-story buildings, about 340,900 square feet in all, in a campus zoned for industrial park use. Battery maker QuantumScape leased the 80,641-square-foot building in 2021 on a term running to September 2032, and in December 2025 it subleased the neighboring 1762 building to space company Momentus as it trimmed its footprint. The permit does not say whether MiTAC's space is a sublease or a direct deal. Either way, MiTAC's lab covers roughly half of the 1750 building.

Why $1,917 a foot reads as a serious lab

The per-foot number is the tell. Ordinary office build-outs in the same feed ran $329 to $396 a square foot, and even the two jobs that are explicitly about servers came in under $1,000.

Equinix's data center build-out on Great Oaks Boulevard in south San Jose works out to $994 a foot. Supermicro's server-room job on Rock Avenue, a short drive from Automation Parkway, is $952. MiTAC's lab is about twice either one.

A few caveats keep this honest. Permit valuation is what the applicant declares for construction. It is not a sale price, and it typically does not cover the servers and test gear that roll in later. So the $76.8 million is mostly walls, power, cooling and plumbing. That is exactly why it is interesting. You do not spend that per foot on desks.

MiTAC's newly expanded Newark office offers a clue about what this kind of space does. The company said in August that it houses a new product introduction unit for advanced cooling and an engineering lab for validating AI and cloud workloads. A San Jose lab doing similar work would need heavy electrical service and liquid-cooling plumbing. That is a reasonable read of the number, not something the permit states.

Who MiTAC is, and why it keeps adding Bay Area space

MiTAC has spent the last three years turning itself into a serious AI server maker. In April 2023 Intel sold its Data Center Solutions Group server business to MiTAC, handing it the right to build and sell Intel-designed servers. MiTAC Computing now sells AI and GPU servers built on AMD and Intel platforms, and NVIDIA named it in August 2025 among the system makers bringing its RTX PRO Blackwell servers to market.

The Bay Area footprint has grown fast. MiTAC Information Systems took about 348,000 square feet at Fremont Technology Center in mid-2025, then committed in August 2026 to essentially all of the roughly 470,000-square-foot Campus at Bayside in Fremont. We looked at what that means for the neighborhoods across I-880 in our piece on the Fremont Bayside campus. San Jose is the new piece.

The backdrop is a push to build AI hardware in the United States. In April 2025 NVIDIA said it would work with partners to produce up to $500 billion of AI infrastructure in the U.S. over four years, with supercomputer assembly in Texas. And North San Jose already has the region's biggest homegrown example. Supermicro in April 2026 added a 32.8-acre, 714,000-square-foot campus on Qume Drive and Commerce Drive, which it calls its largest in the country. MiTAC's lab sits in the same North San Jose industrial belt.

A lab is not a data center

This is a lab permit, and the difference matters to neighbors. San Jose's own permit feed files the Equinix job as "Data Center" and MiTAC's as "R & D Lab." A lab is people at benches testing machines during work hours. A data center is racks running around the clock with chillers and backup generators out back, which is where the noise fights in San Jose come from.

Nothing in this permit points to that kind of equipment yard, and the site sits inside a business park zoned for industrial use. For Berryessa residents, a lab like this is not the kind of building that tends to produce noise complaints.

Where the people working here are likely to live

The practical commute shed is northeast San Jose and Milpitas. The site is about 0.9 mile from the Hostetter light rail station on Capitol Avenue, roughly 1.8 miles from Berryessa BART, and about 2.4 miles from Milpitas BART. Here is how the nearby neighborhoods compare on our scores and September 2026 typical prices.

Two things jump out. Berryessa Station posts the same 4.0 overall and 4.3 safety scores as Berryessa North for about $167,000 less. The catch is the name. Berryessa BART actually sits in Lundy, not in Berryessa Station, so do not buy there for a walk to the train.

The other is the school gap across the city line. Northwestern Milpitas scores 4.6 on schools for about $1.37 million, against 3.7 in the Berryessa pair. For a household that plans to stay a decade, that 0.9-point school gap may matter more than a shorter drive to Automation Parkway. Morrill is the San Jose answer, with a 4.2 school score and the highest overall score in the table, at the top of the price range.

Midtown Milpitas is the cheapest line, and its 2.2 safety score is the lowest. That is a trade to research street by street, not a verdict on the neighborhood.

What this does and does not mean for nearby home prices

It does add jobs that cannot go remote. You cannot validate a liquid-cooled GPU server from a laptop in another state. Engineers and technicians have to be in the room, and those rooms are clustering in North San Jose, around Supermicro's campuses and now MiTAC's lab.

It does not, on its own, move a neighborhood median. A 40,042-square-foot lab is a modest headcount compared with a manufacturing floor, and MiTAC has not said how many people will work there. Plenty of the people who fill these jobs already live somewhere in the South Bay and will simply change their commute. One permit is a data point, not a trend line.

The cumulative story is the one to watch. North San Jose keeps adding hardware jobs, and the housing math only works if the city keeps adding homes near Berryessa BART and along the Capitol Avenue light rail line to match. More jobs with no new homes is how a commute shed turns into a bidding war.

Next, watch for MiTAC to say what the lab is for and how many people it will hold, and for the permit's final inspection date, which is still blank in the city's data. When we compare neighborhoods for a buyer weighing a job in this corridor, we start with the commute shed and then let the school and safety scores do the sorting, and the full San Jose neighborhood rankings on Houseberry are the fastest way to run that same screen yourself.

Sources

• City of San Jose Open Data, Last 30 Days Building Permits (permits 2026-139041-CI and 2026-135876-CI)
• QuantumScape lease for 1750 Automation Parkway, SEC exhibit 10.1 (November 2021)
• QuantumScape Form 8-K on the 1762 Automation Parkway sublease (December 2025)
• Peninsula Land and Capital, 1750-1762 Automation Parkway property page
• DatacenterDynamics, Intel to sell dedicated server business to MiTAC (April 2023)
• NVIDIA, RTX PRO Servers with Blackwell partner announcement (August 2025)
• MiTAC Computing Technology product site
• MiTAC Computing Technology USA, Newark office grand opening (August 2026)
• GlobeSt, MiTAC Information takes 348K SF at Fremont industrial center (July 2025)
• ConnectCRE, MiTAC leases entirety of 470K-SF Fremont campus (August 2026)
• Nextgov, NVIDIA plans for U.S.-manufactured AI supercomputers (April 2025)
• Supermicro, adds largest Silicon Valley campus (April 2026)
• Hoodline, Supermicro North San Jose campus (April 2026)
• Hostetter station, VTA light rail
• Houseberry, San Jose and Milpitas neighborhood rankings (September 2026)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

San Francisco finished 405 new homes in the first half of 2026, and the pipeline behind them is drying up. Project applications filed September 1 to 28 proposed just 30 net new homes, the fewest for that stretch of September since 2012. On September 29, the Board of Supervisors' own analysts released a report showing the city is sitting on 147 publicly owned properties that could help change that number, from Muni parking lots in the Richmond to a seismically unsafe office tower at 170 Otis Street. We went through the report's appendices property by property and sorted all 147 by supervisorial district, so you can see which San Francisco neighborhoods hold the public land.

The short answer: the waterfront districts lead on paper, but most of those sites are Port piers that state law makes hard to build homes on. Take the Port out and the district with the most public land in play is District 7, west of Twin Peaks, with 19 sites.

Key takeaways

• The Budget and Legislative Analyst flagged 147 properties across 15 departments with potential for housing, joint use, consolidation or sale.
• District 3 (26 sites) and District 10 (25) top the raw count, but 29 of those 51 are Port waterfront properties.
• Excluding the Port, District 7 leads with 19 sites, and the three west side districts together hold 33.
• The report puts unit estimates on nine specific sites. Together they could hold about 1,812 homes, more than four times what the whole city finished in the first half of 2026.
• Land is not the main obstacle. Funding is. More than 12,600 affordable homes are already waiting for money.

Who did this work, and why it matters

This analysis belongs to the Board of Supervisors' Budget and Legislative Analyst's Office. Its policy analysis report, Opportunities for Using Vacant, Underutilized, Surplus, and other Public Properties for Housing and other Public Purposes, is dated September 29, 2026, was requested by Supervisor Danny Sauter, and lists Fred Brousseau and Daniela Estrada as project staff. Everything in this post that counts or describes a site comes from their appendices. The district tally and the neighborhood lens are ours.

It is a bigger piece of work than the headline suggests. The analysts reviewed land held by 28 city agencies plus the school district and City College. They screened for four things: seismically unsafe or poor-condition buildings, parking lots and garages, vacant or soon-to-be-vacant property, and parcels where the current building uses only part of the lot.

They had to go looking because nobody else was. A 2002 ordinance asks departments to report their own surplus land every year. The report found that no department submitted a single new surplus property for eight straight years, from fiscal 2017-18 through 2025-26. And the agencies that control the most land, including the Port, Muni, Rec and Park and the school district, are exempt from that law entirely.

The San Francisco Standard's Maggie Angst and the Chronicle's Alyce McFadden covered the report the morning it came out, and both are worth reading for Sauter's side of it.

The 147 sites, district by district

Here is the count by supervisorial district, straight from the report's Appendix A (44 vacant or soon-to-be-vacant properties) and Appendix B (103 properties with development or joint-use potential).

Two things jump out. First, the Port's 32 properties sit almost entirely in District 3 (17, the northeast waterfront from Fisherman's Wharf to South Beach) and District 10 (12, the southern piers around Pier 80 and Pier 96). The report is candid that direct waterfront land falls under the state's Public Trust Doctrine, which bars most non-maritime uses. Seawall lots back from the water are a different story, but a vacant pier shed is not a housing site in any near-term sense.

Second, the west side shows up far more than its reputation would suggest. District 7 alone has 19 properties, and District 1 (Richmond) and District 4 (Sunset) add 14 more. That matters because the west side is where most of the city's recent rezoning for taller buildings is aimed, and its public land is mostly the easiest kind to reuse: small surface parking lots.

The west side's secret inventory is Muni parking lots

Ten of the fourteen neighborhood Muni surface lots in the report sit in Districts 1, 4 and 7. These are the lots you have parked in without thinking about who owns them: 8th and Clement and 9th and Clement in the Inner Richmond, 18th and Geary and 21st and Geary further out, 8th and Irving and 20th and Irving in the Sunset, West Portal at 14th Avenue, and Claremont and Ulloa up the hill.

None of them is huge. The biggest, Claremont and Ulloa, is about 12,900 square feet. But they sit a block or two from the L, N and K lines and the Geary buses, in neighborhoods like the Inner Richmond and West Portal that score well on schools and safety and where a typical home runs about $2.45 million and $2.62 million respectively, per Houseberry's September 2026 figures. Small lot, great location, expensive neighborhood. That is the kind of site where housing built over retained parking has its best shot at penciling out.

The report also notes that SFMTA already reviewed about 90 of its own properties under a joint development program adopted last year and is ranking them with a consultant. So the lots are not news to Muni. What is new is seeing them listed next to everything else the city owns.

Nine sites, about 1,812 homes

The analysts only put unit estimates on a handful of properties, and they call those estimates illustrative. Still, they are the most concrete numbers in the report.

Add those up and you get 1,812 homes on nine parcels. For scale, that is more than four times the 405 homes San Francisco actually finished in the first half of this year, a figure we dug into in our look at why the city's housing pipeline built just 400 homes. The pipeline holds tens of thousands of proposed homes that are not getting built, and new proposals have slowed to a trickle. Public land cannot fix that by itself, but it removes one cost from the equation.

The two biggest numbers belong to the school district, which ran its own portfolio study in May. The 7th Avenue lot sits on the hill above the Inner Sunset's Irving Street strip and has been floated for years as educator housing. Middle Point Road is on the edge of Hunters Point near India Basin, where a typical home in Hunters Point runs about $724,000, one of the lowest figures in the city.

The honest catch: land is the cheap part

Here is where the report is refreshingly blunt. Land is only about 10 percent of what it costs to build an affordable home in San Francisco, roughly $121,000 of a total that now runs about $1 million per two-bedroom unit, according to Planning Department figures the analysts cite. More than 12,600 affordable units across 59 projects are already sitting in the pipeline waiting for funding. At the city's current pace of about 650 fully affordable units a year, clearing that backlog alone would take more than a decade.

So a free parcel does not build itself. That is why the report pushes the city to treat its land as a portfolio, selling or leasing sites it does not need and using the proceeds to pay for housing on the sites it keeps. It is also why a mixed-income building on a Muni lot in the Richmond, where market-rate units can carry some of the cost, may move faster than a fully subsidized one.

And the list is a starting point, not a plan. The analysts did not screen out sites with the wrong zoning, historic status or environmental limits. Some of the 147 are working fire stations and libraries. Housing over a rec center would need a charter amendment and a citywide vote.

What this means if you live near one of these sites

For most neighbors, nothing changes soon. This is a screening list, and every site still has to survive a department that may want to keep it, a feasibility study and, often, a rezoning.

But if you are buying on the west side, it is worth knowing where the public lots are, because they are among the few places in those neighborhoods where a five or six story building could appear without anyone losing a home. We think that is a good thing. San Francisco had reached only about 12 percent of its 82,069-home state target by 2025, per the report, and the city's own parking lots are about the least controversial place to start. You can see how each neighborhood scores on schools, safety and price on our San Francisco neighborhood rankings.

The next step is political. Sauter told the Chronicle he has called a hearing at the Land Use and Transportation Committee for early November, where departments will be asked how their property could be used for housing. The report's main recommendation is simple: put one agency in charge. As Sauter put it to the Standard, "Where there's no ownership, that's where things go to die."

Sources

• San Francisco Budget and Legislative Analyst's Office (Fred Brousseau and Daniela Estrada), Opportunities for Using Vacant, Underutilized, Surplus, and other Public Properties for Housing and other Public Purposes, policy analysis report to Supervisor Danny Sauter, September 29, 2026
• Maggie Angst, The San Francisco Standard, SF's housing solution may be hiding in plain sight: Land it already owns, September 29, 2026
• Alyce McFadden, San Francisco Chronicle, Map: These publicly owned properties could become S.F. housing, September 29, 2026
• GrowSF, Only 405 homes built in first half of 2026, July 31, 2026
• Houseberry, San Francisco neighborhood rankings, neighborhood prices as of September 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The most interesting piece of land on The Alameda right now might be a motel parking lot. It sits beside the Arena Hotel at 817 The Alameda, a short walk from SAP Center, and San Jose housing staff want it to become new homes for moderate-income households.

That is the quiet headline inside Item 8.2 on the Oct. 6, 2026 San Jose City Council agenda (File 26-1035). Staff recommend a three-year exclusive negotiating agreement with Episcopal Community Services of San Francisco to take over the city-owned Arena Hotel, keep the building as housing under its state Homekey rent limits, and explore building mixed-income housing on the parking lot for households at or below 120 percent of area median income. Propertyworx LLC, the second-highest scorer, is the backup. No unit counts or dollar figures appear in the legislation text, and I could not get the attached staff memo, so those numbers are not public in anything we have read yet.

What you need to know

• Staff pick Episcopal Community Services, a San Francisco nonprofit, as the Arena Hotel's next developer, with a three-year exclusive negotiating window plus two optional six-month extensions.
• The 90-room hotel stays housing at Homekey rent levels. The parking lot could get new homes for households earning up to 120 percent of AMI, which is $246,600 a year for a family of four in Santa Clara County in 2026.
• The city would cut its original partners, Urban Housing Communities and HomeFirst, out of the state Homekey contract and become the sole grantee.
• Nothing gets built from this vote. Any new construction needs its own CEQA review and a final deal back at council.

From budget motel to city-owned Homekey site

The Arena Hotel has been public housing infrastructure for three years, and it has not been a smooth three years. San Jose first lined it up in 2021 as an 89-unit Homekey project with developer Urban Housing Communities and operator HomeFirst. The state awarded $25.2 million in Homekey money for it in May 2022.

It opened as interim housing with a grand opening on Sept. 28, 2023: 90 rooms (80 singles and 10 doubles) at a total cost of $45.2 million, $25.2 million from Homekey and $20.8 million from the city's Measure E property transfer tax. That is roughly $502,000 per room, which is what hotel conversions cost when the building needs real work.

Then the problems. By late 2025, residents were describing broken elevators, food-poisoning incidents and uneven case management, and the city replaced HomeFirst with WeHope on Dec. 1, 2025, under a $2.7 million contract. To be fair to HomeFirst, it moved 52 residents into permanent housing in just over two years. In May 2026, the site switched again, becoming a waiting room for people who already hold federal housing vouchers, with the county Housing Authority paying up to $2.3 million over three years.

What Item 8.2 would actually set in motion

The resolution does two jobs at once: it picks the next team, and it cleans up the old one. On the first, it authorizes the Housing Director to negotiate with Episcopal Community Services on two pieces.

• The hotel. Convert the existing Arena Hotel into housing units with rents consistent with the project's Homekey agreement.
• The parking lot. The potential future construction of mixed-income housing on the parking lot portion of the site, serving households at or below 120 percent of AMI.

A right-of-entry agreement lets the nonprofit onto the property for due diligence. If Episcopal Community Services cannot sign or misses a material obligation, the same authority passes to Propertyworx LLC.

On the second job, the city would amend its Homekey standard agreement with the California Department of Housing and Community Development to remove A0729 San José L.P., Urban Housing Communities LLC and HomeFirst Services of Santa Clara County, and name the city as sole grantee. It would also unwind the ground lease, the acquisition and rehabilitation grant agreement and the operating agreements with the original partners. That is a lot of paperwork, and it is the paperwork that has to happen before anyone new can finance anything.

One shift is worth flagging. As recently as April 2026, the long-term plan reported for this site was to demolish the building for up to 200 permanent affordable apartments. The new recommendation talks about converting the existing hotel and adding housing on the lot instead. Keeping 90 occupied rooms while building next door is less disruptive for the people living there. It may also mean fewer total homes than a full rebuild, and we will not know until the proposal has numbers.

Why this nonprofit, and why the parking lot matters

Episcopal Community Services has done exactly this kind of project before. In San Francisco it acquired two hotels through Homekey, the 232-room Granada and the 130-room Diva, and it owns and operates 1064 Mission Street, a 256-studio supportive housing building. This would be a move south of its home turf, which is worth watching, but the hotel-to-housing playbook is familiar to it.

The parking lot is where I get genuinely optimistic. The hotel's rooms are deeply affordable by design. The lot is different: 120 percent of AMI is moderate income, the teachers, nurses, bus drivers and city staff who earn too much for most affordable housing and not nearly enough for a $1.6 million San Jose house. Almost nobody builds for them, because the rents do not cover construction costs without help. A city-owned parcel, where land cost is already paid for, is one of the few places it can pencil.

And this is a good spot for it. The site is a few blocks from SAP Center and walking distance to Diridon Station, where Caltrain, VTA light rail and the future BART extension meet. Putting homes on surface parking next to a regional transit hub is the most basic pro-housing move there is. I would rather see a four-story building on that asphalt than another decade of parked cars.

What 120 percent of AMI means in dollars in 2026

For a family of four in Santa Clara County, the moderate-income ceiling is $246,600 a year. That comes from the 2026 state income limits published by HCD, which put the county's four-person area median income at $205,500, up from $195,200 a year earlier. Here is the full ladder, with our own math on what 30 percent of that income works out to per month.

Read the table carefully. These are income caps, not rents. The right column is the most a household at the very top of the band would pay at the standard 30 percent affordability line. Actual rents on a 120 percent AMI building are set in the regulatory agreement and are usually calculated below the cap, so most tenants would pay less. The point is who qualifies: a two-earner household making just under $250,000 is still in.

What it means if you are buying near The Alameda

For nearby buyers, the practical takeaway is that the Arena Hotel is becoming more stable, not less. A city-owned site with a named nonprofit, a clear contract and a moderate-income building planned next door is a better neighbor than a troubled interim shelter whose operator and mission both changed within six months.

The Alameda also sits on a real price seam. Rose Garden, the older residential neighborhood around the Municipal Rose Garden off The Alameda, had a median of about $1.96 million in August 2026 on our data. College Park, between The Alameda and Highway 880, was at about $1.32 million. Buena Vista, the older flat grid stretching from the Bascom area toward downtown, was at about $941,500. The San Jose citywide median was about $1.63 million the same month.

That spread is the story of The Alameda as a whole. It is the seam between some of San Jose's priciest older blocks and some of its most attainable ones, and it is where the city keeps putting new housing, like the 18 townhomes proposed for a long-vacant Emory Street lot. Rose Garden scores 3.5 overall and 3.9 on safety in our San Jose neighborhood rankings, and nothing in this item changes that.

The honest caveat is time. This is a negotiating agreement, not a building permit. We have watched the Diridon area stay stuck between approved and actually built for years, and a four-year negotiating window plus CEQA review means shovels on the parking lot are realistically years out. If you are buying near The Alameda, price in what is there today.

When the next decision lands

Council takes up Item 8.2 on Oct. 6, 2026. The item sits in Council District 6, and the city has deemed the agreements themselves not a project under CEQA, since no physical change happens yet. The real test comes later, when Episcopal Community Services returns with a final proposal that has unit counts, a budget and a construction plan for the lot.

That is when we will know whether this becomes 60 homes or 200, and what the moderate-income rents look like. We will be reading that staff report when it lands. Until then, if you are weighing a home on either side of The Alameda, start with how the surrounding neighborhoods actually compare, then look at the house.

Sources

• City of San Jose, File 26-1035, Item 8.2 legislation text (Oct. 6, 2026 agenda)
• City of San Jose, File 26-1035 staff memorandum (attachment)
• Office of Mayor Matt Mahan, Arena Hotel grand opening (Sept. 28, 2023)
• San Jose Spotlight, San Jose pushes ahead to convert hotels to housing (Oct. 2021)
• San Jose Inside, San Jose to receive $25.2 million Homekey grant to buy Arena Hotel (May 2022)
• San Jose Spotlight, San Jose homeless hotel gets new management (Nov. 2025)
• San Jose Spotlight, San Jose shelter is a stop on the way to permanent housing (April 2026)
• California HCD, 2026 State Income Limits
• Episcopal Community Services of San Francisco, Supportive Housing Expansion
• Houseberry, Rose Garden, Buena Vista and College Park neighborhood pages and San Jose city page, price history through August 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

If you live in the Santa Teresa foothills, you have probably stood at the top of your street and looked south at all that flat farmland and wondered the same thing your neighbors do. Will somebody eventually build out there?

For another big piece of Coyote Valley, the answer just got closer to a permanent no. On September 28, 2026, the State Coastal Conservancy filed a notice of exemption for the Santa Clara Valley Open Space Authority to buy about 417 acres of Coyote Valley open space, to keep it farmed, keep wildlife moving between two mountain ranges, and later let Fisher Creek spread back into its floodplain. The Conservancy board approved $14.5 million for the purchase on September 17. We could not find anyone who has reported the filing yet.

What you need to know

• The Open Space Authority is set to acquire roughly 417 acres in Coyote Valley, the farm valley between South San Jose and Morgan Hill, using a $14.5 million state grant approved September 17, 2026.
• The deal has not closed. The filing names no seller, price or parcel numbers, and the next public record will be a deed at the Santa Clara County recorder.
• Peninsula Open Space Trust said in March 2026 that more than 1,600 acres of the valley were already protected. This purchase adds a meaningful chunk on top of that.
• For buyers in Santa Teresa, Almaden and Blossom Valley, it makes the green edge south of town a little more permanent. For the region, it is land that will never hold homes.

What the September 28 filing actually says

The document is short, and most of it is legal plumbing. The project is called the Coyote Valley Landscape Linkage Acquisition, and it relies on the state's categorical exemption for land transfers that keep property in its natural condition. That exemption only works because nothing is being built.

The stated purpose has three parts: let agriculture continue, protect the landscape linkage wildlife use to cross the valley, and set up a future restoration of the Fisher Creek floodplain and its streamside habitat. Fisher Creek is the small creek that runs north up the middle of the valley, straightened into a drainage ditch for farming in the early 1900s, according to the Open Space Authority.

What it does not say is just as useful. No seller, no sale price, no parcel numbers. The $14.5 million grant works out to roughly $34,800 per acre, but the grant may not cover the full price, so treat that as a ceiling on what the state is putting in, not the land's value. Until a deed records, this is a funded plan, not a finished purchase.

Coyote Valley, deal by deal

This is the latest step in a run of purchases that started in earnest in 2019, and the pattern matters more than any single deal.

Sources for the rows: the Protect Coyote Valley timeline on Measure T, POST on the 2019 deal, which used $46.3 million in Measure T money, San Jose Spotlight on the 2021 vote, and POST on Coyote Fields.

Read the table top to bottom and the direction is not subtle. The city stopped planning for Coyote Valley as a job center in 2021, and ever since, the buyers have been conservation agencies, not developers.

What it means if you live on the Santa Teresa side

For the neighborhoods that look down on the valley, this purchase mostly protects the view and the quiet you already have. It does not add anything overnight.

Santa Teresa Foothills East, the hillside streets on the slope of the Santa Teresa Hills with Santa Teresa County Park behind them, scores 4.2 out of 5 on schools and 4.5 on safety in our data. Its median sale price was about $1.53 million in July 2026, a little under the San Jose citywide median of about $1.63 million in August 2026.

That is the part we find most useful when we compare neighborhoods on Houseberry. You get foothill streets, a county park at the back fence and farmland in the view for roughly the city median, sometimes less. Almaden Valley, boxed in between the Santa Cruz Mountains and the Santa Teresa Hills, is the pricier outlier at about $2.03 million in August 2026. Edenvale, down on the flats near Hayes Mansion, sold at about $894,600.

Our Public Spaces score, which measures parks and open land close to home, follows the hills. Almaden Valley scores 4.3 and Santa Teresa Foothills East 4.0, while Blossom Valley, the flatter middle of South San Jose around Oakridge mall, sits at 3.0.

Does this raise nearby home values? Probably not in a way you could measure. The valley was already off the table for big development after 2021, so buyers were not pricing in a warehouse district next door. What the purchase does is lower the odds that a future council reopens the question on these particular acres.

Will it add trails? Not right away. POST said in March 2026 that it is looking at two possible public access options, near the Laguna Seca wetland and in the mid-valley farm area, without mileage or dates. The trail you can use today is the Coyote Creek Parkway, 15 paved miles along the creek from Hellyer County Park south to Anderson Lake County Park.

What about flooding? This is the benefit most people underrate. POST describes the plan as reconnecting Fisher Creek to its natural floodplain and restoring Laguna Seca, to hold water in the valley and ease downstream flood risk for San Jose. Anyone who remembers the 2017 Coyote Creek flood will understand why the agencies care about spare room for water upstream. Just do not expect a number, since no one has published how much risk this particular parcel reduces.

And the wildlife? Coyote Valley is the shortest workable route between the Santa Cruz Mountains and the Diablo Range. That means bobcats and mountain lions crossing, and it also means the wild pigs that tear up foothill lawns are not leaving. The Santa Clara Valley Transportation Authority board approved a design contract in January 2026 for wildlife crossings at Highway 101 and Monterey Road, Local News Matters reported.

The trade-off: land that will not hold homes

I am pro-housing, and I am not going to pretend a 417-acre purchase is free. Every acre set aside is an acre that will never hold a home in a region that badly needs them.

San Jose has to plan for 62,200 homes between 2023 and 2031 under its state housing target. That is the real pressure, and open space purchases do not relieve it.

But Coyote Valley is a bad place to meet it. Caltrain's Gilroy trains run through the valley without a stop between Blossom Hill and Morgan Hill, and homes out there would mean new roads, new sewers and long car commutes. The better answer is the one San Jose is already fumbling toward, which is more homes along corridors and in urban villages near jobs and transit. Our look at San Jose's general plan reset walks through where 4 to 10 homes could replace one. If you support protecting Coyote Valley, the honest follow-through is supporting those infill projects too, including the ones near you.

The deed is the next document to watch

The next thing to look for is a recorded deed at the Santa Clara County recorder with the Open Space Authority as buyer. That will finally name the seller, the price and the parcels. Until then, the September 28 filing is a strong signal, not a closing.

If you are shopping the southern edge of San Jose, open our San Jose neighborhood rankings and compare the foothill neighborhoods side by side. The Coyote neighborhood page covers the rural ranches around the valley itself, where a thin run of sales put the median near $2.45 million in August 2026. That figure describes a handful of big properties, not a trend. The view south is looking more settled every year. The housing math, sadly, is not.

Sources

• CEQAnet, Coyote Valley Landscape Linkage Acquisition, SCH 2026091176
• State Coastal Conservancy, September 17, 2026 grant awards
• Peninsula Open Space Trust, 937 acres in North Coyote Valley (2019)
• Peninsula Open Space Trust, Protecting What Protects Us (March 2026)
• Peninsula Open Space Trust, Coyote Fields acquisition (April 2026)
• Local News Matters, 71-acre Coyote Valley farm deal (April 2026)
• Santa Clara Valley Open Space Authority, public ownership of Coyote Valley (2025)
• San Jose Spotlight, council keeps Coyote Valley as open space (2021)
• Protect Coyote Valley, timeline
• Protect Coyote Valley, flooding and Coyote Valley
• Santa Clara County Parks, Coyote Creek Parkway
• San Jose Spotlight, San Jose's 62,200 housing unit question
• Houseberry, San Jose neighborhood rankings and neighborhood pages, September 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

A fire-gutted house with squatters inside and a hole in the roof just drew multiple offers at $450,000. That is what the cheapest house in San Francisco looks like in 2026. For anything you could actually move into, the cheapest houses in San Francisco sit in a short list of southeast and southwest neighborhoods, and only 9 of the 88 neighborhoods we price have a typical home price under $1 million.

Below is the full list, with safety and school scores next to each price, because the cheapest block and the best deal are rarely the same thing.

Quick take

• Only 9 of 88 priced San Francisco neighborhoods have a typical home price under $1 million, per our September 2026 figures.
• Four of those nine are condo-dominated central neighborhoods: the Tenderloin ($499,159), Van Ness/Civic Center, South of Market and Downtown.
• The cheapest mostly single-family neighborhoods are in the southeast: Bayview ($915,567), Bayview Heights and Little Hollywood.
• The best price-to-safety trade is in the southwest. Ingleside Heights, Oceanview and Crocker Amazon all score 4.4 out of 5 on safety and sit between $1.08 million and $1.19 million.
• The middle San Francisco neighborhood costs about $1.69 million, and every neighborhood in the chart below runs roughly 30 percent or more under that.

The $450,000 house that started this

The San Francisco Standard's Emily Landes reported on September 29 that a house on Farallones Street in the Ingleside area, badly damaged in a 2023 fire and still occupied by squatters, went into contract after drawing multiple offers at a $450,000 asking price. The listing agent told buyers not to go inside. They bid anyway.

That price is about 60 percent below the $1.13 million typical home price in neighboring Oceanview, and it is not a real benchmark for anyone who needs a place to live. Whoever closes has to get the occupants out legally, then rebuild most of a house. Bidders were buying a lot, a shell and a headache, and pricing all three. So what does a habitable house cost in the cheapest parts of the city?

San Francisco's 16 cheapest neighborhoods, ranked

The chart below ranks the 16 lowest-priced neighborhoods we track, with each one's safety score. Red bars are the ones that clear 4.0 out of 5 on safety.

Read the top of that chart carefully. The Tenderloin, Van Ness/Civic Center, South of Market and Downtown are cheap because almost everything that sells there is a condo, often a small one. If you are searching for a house with a yard, those four are not your list.

The house list starts in the southeast. Bayview, the neighborhood along Third Street and the T line south of Islais Creek, has a typical price of $915,567. Bayview Heights, the slope above it toward Candlestick, sits at $974,881. And Little Hollywood, the small pocket of older single-family homes tucked between Bayshore Boulevard and the 101 near the Brisbane line, comes in at $986,974.

Price, safety and schools side by side

Here is the same group with the other scores that matter when you are choosing a block. All figures are from our neighborhood pages in September 2026.

Two patterns stand out. First, the southwest corner near the Daly City line (Ingleside Heights, Oceanview, Crocker Amazon and Outer Mission) is where the price gets you the most safety. Four neighborhoods score 4.3 or better on safety for less than $1.3 million. In Bayview and Bayview Heights you save roughly $100,000 to $200,000 against that pocket, but safety scores drop into the 2s.

Second, school scores are low almost everywhere on this list. Ingleside Heights leads at 3.0 out of 5. That is the real trade-off of buying cheap in San Francisco, and it is worth pulling up the specific assigned school before you fall for a house.

Where prices are softening

This is the part most "cheap neighborhoods" lists skip. Our citywide price is up about 9 percent over the past 12 months, but three of the cheap house neighborhoods went the other way. Little Hollywood is down 8.3 percent, Oceanview is down 3.4 percent and Bayview Heights is down 2.3 percent.

Falling prices are not automatically a bargain. Small neighborhoods like Little Hollywood have few sales, so a handful of fixer sales can pull the number around. But a flat or falling market is where buyers get to ask for repairs, credits and time, which is exactly what you want when the house needs work.

What $1 million does and does not buy here

For about $1 million in these neighborhoods you are usually looking at an older detached or row house, often with a garage at street level and living space above, and often needing updates. What it does not buy is a short walk to a busy commercial strip. Every neighborhood in the table scores below 3 out of 5 on amenities except Crocker Amazon, at 3.3.

The upside is that these are real houses on real lots in a city where the middle neighborhood costs about $1.69 million. And the cheapest house in San Francisco, the Farallones Street wreck included, keeps attracting investors for a reason. The land is worth something even when the building is not.

Cheapest houses in San Francisco: common questions

What is the cheapest neighborhood to buy a house in San Francisco? Among neighborhoods made up mostly of houses, Bayview has the lowest typical price, about $915,600 as of September 2026. The Tenderloin is cheaper overall at $499,159, but nearly everything sold there is a condo.

Can you still buy a house in San Francisco for under $1 million? Yes, mainly in Bayview, Bayview Heights, Little Hollywood and parts of Visitacion Valley and Silver Terrace, where typical prices sit near or just under $1 million. Expect older homes that need some work.

What is the cheapest safe neighborhood in San Francisco? Little Hollywood is the only neighborhood under $1 million that scores 4.0 or better on safety. Just above $1 million, Ingleside Heights ($1,076,500) scores 4.4. Safety scores are relative context, not a guarantee.

Why did a fire-damaged San Francisco house get multiple offers? Because even a gutted house sits on buildable city land. At $450,000, buyers were paying well under half of what nearby homes sell for and planning to rebuild.

How we built this list

Houseberry's typical home price is built from recent sales inside each neighborhood's boundary, across all property types, and is designed so that one mansion or one teardown cannot swing it. We ranked all 88 San Francisco neighborhoods with their own price figure and left out four that borrow the citywide number (Presidio, Golden Gate Park, Sherwood Forest and Lincoln Park). Safety, school and amenity scores are our 0 to 5 neighborhood scores, and the 12-month change comes from each neighborhood page.

If you want to run the list yourself, sort every neighborhood on our most affordable San Francisco neighborhoods ranking, or compare them block by block on the San Francisco neighborhood map. Looking at the neighborhood before the house is the whole idea behind what we built.

Sources

• Emily Landes, The San Francisco Standard, Squatters, a burned roof, and multiple offers: A $450K SF house that's literally a steal, September 29, 2026
• Houseberry, Best neighborhoods in San Francisco and neighborhood pages, prices and scores as of September 2026
• Houseberry, Most affordable neighborhoods in San Francisco

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Thirty. That is how many net new homes were proposed in the 70 project applications filed with the San Francisco Planning Department from September 1 to 28, 2026. It is the smallest September total since 2012, when the same stretch of the month produced 11.

And it landed in the same week AMD agreed to pay $8.2 billion for World Labs, a San Francisco AI company, and Zumper reported one-bedroom rents up 25.4 percent in a year. The money is here. The hiring is here. The proposals for new homes are not. We pulled San Francisco housing filings straight from SF Planning's records to check, and the gap between the AI boom and the housing pipeline is wider than the headlines suggest.

The short version

• 70 project applications filed September 1 to 28, 2026 proposed 30 net new homes, the fewest for that window since 2012 (11).
• The same window proposed 276 net homes in 2025, 516 in 2024 and 1,316 in 2022.
• Year to date through September 28, 2026: 2,880 net homes on 879 applications, against 5,932 on 881 in 2025. That is down 51 percent, or 35 percent if you remove 2025's single 1,500-home filing.
• Fifteen large filings account for 2,587 of 2026's 2,880 proposed homes. The other 864 applications added 293.
• San Francisco's median sale price was $1,875,000 in August 2026, up 25 percent in a year, per the California Association of Realtors.

Thirty homes, and the biggest filing was five

The largest September application proposes five net new homes. It is at 62-74 Laidley Street in Glen Park, where three lots would be reshuffled into two with a pair of new buildings.

Behind it sits a three-home SB 9 and ADU project at 8 Orizaba Avenue in Oceanview, two-home infill projects on 20th Street in Potrero Hill and on 5th Avenue in the Inner Richmond, and then a long tail of one-unit filings. Some of those singles are brand-new ADUs. Several are legalizations of apartments that already exist, like the unit at 567-579 Lombard Street in North Beach, which puts one more home on the books and zero new roofs on the skyline.

The other 46 applications, about two-thirds of the month, proposed no net change in housing at all. Remodels, additions, storefront work.

How we counted: we took every project application (record type PRJ) in SF Planning's projects dataset on DataSF, grouped them by the date each was opened, and added up the dataset's net-units field, which is proposed homes minus existing homes on the site.

Two caveats before anyone reads this as final. September still has two days left, and applications sometimes get entered late, so the month could rise a bit. To keep the comparison fair, the chart below uses September 1 to 28 for every year. Full-month totals for past Septembers are nearly identical: 277 in 2025, 517 in 2024 and 1,319 in 2022. It would take a single filing of roughly 250 homes in the last two days just to get back to last September.

Monthly numbers are lumpy. One big filing can make a September look like 2015 or 2022. But the last time the month came in this low, San Francisco was still climbing out of the foreclosure era. February 2026 also produced only 30 net homes, so September is not a one-off blip in an otherwise busy year.

An AI boom on one side of the ledger, 30 homes on the other

Every demand signal that normally sends developers to Planning is flashing at once. AI companies are expanding, buying and leasing across the city. The Real Deal counted 413 AI tenants in San Francisco offices by August 2026, up from 23 before ChatGPT launched. We mapped where they cluster in our look at the AI office neighborhoods, mostly SoMa and Mission Bay.

Home prices followed. The California Association of Realtors put San Francisco County's median sale price at $1,875,000 in August 2026, up from $1,500,000 a year earlier. Our own city data, which tracks a different mix of sales, shows the San Francisco median at about $2.20 million in August 2026, up from $1.89 million in September 2025.

Rents moved even faster. Zumper's September 2026 report put the median one-bedroom at $4,400, up 25.4 percent, and the median two-bedroom at $6,340, up 26.8 percent, as CBS News reported.

In most cities, a 25 percent jump in rents is the starting gun for a building cycle. In San Francisco this September, it produced 30 homes. That is the story in one number.

Down 51 percent, or 35 percent, depending on one project

The year-to-date drop is real, but one 2025 filing inflates it. From January 1 to September 28, 2026, applicants proposed 2,880 net new homes on 879 applications. Same period of 2025: 5,932 homes on 881 applications. Nearly the same amount of paperwork, half the housing.

But 2025 had 88 Bluxome Street, the 1,500-home proposal on the Mission Bay edge of SoMa filed in May 2025. Take it out and 2025 still had 4,432 net homes, so 2026 trails by 35 percent. 2025 also had 10 South Van Ness Avenue at 1,019 homes. Remove both mega-filings and the gap narrows to about 16 percent. Honest version: 2026 is weaker, but 2025 was carried by a couple of giants.

There is a downside to the gloomy read, and it is worth saying plainly. 2026 is still ahead of 2022 (2,217) and 2023 (1,711) for the same months. What makes 2026 look thin is how concentrated it is. Four filings, the Seawall Lot 330 site at 555 Beale Street (568 homes), the Outer Richmond Safeway at 850 La Playa Street (562), 3350 Mission Street (379) and 38 8th Street (288), supplied 1,797 of the year's 2,880 homes. Without them, San Francisco's other 875 applications this year proposed about 1,083 homes combined.

Why is almost nobody proposing new homes in San Francisco?

Because the math on a new building still does not work for most sites, even with rents up a quarter. City Hall has cut costs on paper this year. Filings have not caught up.

• Rates. Freddie Mac's 30-year fixed average was 7.03 percent on September 24, 2026, up from 6.30 percent a year earlier. Construction loans price above that, and a project's buyers or permanent lenders are looking at the same numbers.
• Costs and feasibility. The City Controller's April 2026 study found market-rate development had slowed because of higher rates and construction costs, and warned that affordable requirements "significantly above 0%" would further threaten feasibility, per Mission Local.
• Fees and inclusionary. Supervisors voted 9-2 in July to cut the inclusionary requirement from 15 percent to 5 percent, which we unpacked in our explainer on the 15 to 5 cut. The SF Standard reports impact fees were also cut by two-thirds. Both are months old. A mid-size application takes longer than that to design.
• Zoning. The Family Zoning Plan, signed in December 2025, created room for about 36,000 homes on the north and west sides, though the city's chief economist projected about 14,600 over 20 years, per Multifamily Dive. The first project under it, an eight-story building in the Inner Richmond, was approved in July, the SF Standard reported.
• Rent threshold. Developers quoted by the SF Standard in August said rents need to climb another 15 to 20 percent before many more projects pencil, and projected a building wave 12 to 24 months out if they do.

Meanwhile the state target is not moving. San Francisco owes 82,000 new homes by 2031 and is about 12 percent of the way there, according to a Budget and Legislative Analyst report covered by the SF Standard on September 29, with at least 12,600 affordable units stalled for lack of funding. My read: the policy fixes were real and overdue, and the builders are telling us they are still not enough at 7 percent money.

Where are San Francisco housing filings still landing?

Almost all of 2026's proposed homes sit in a handful of big filings, and they cluster in three parts of town. The dataset has no neighborhood field, so we placed each 2026 filing of 18 or more net homes by its address.

The Richmond is the surprise. The Outer Richmond, the fog-belt blocks between Park Presidio and Ocean Beach, scores 3.9 out of 5 on our rankings with a 4.3 for safety and a median of about $1.87 million in August 2026, up 14.7 percent in 12 months. The 562-home Safeway rebuild near La Playa and Cabrillo would be the biggest thing to happen to that part of the west side in decades.

On the Mission side, 3350 Mission would replace the Safeway near 30th Street with a bigger store and 379 apartments, 76 of them affordable. That site backs onto Bernal Heights, where our median hit about $1.81 million in August 2026, up from $1.61 million a year earlier. And the Seawall Lot 330 filing sits in South Beach, the waterfront condo district next to the ballpark, where the median has been flatter at about $1.23 million.

Everywhere else, including the Sunset, Noe Valley, the Marina and Pacific Heights, no 2026 application has proposed more than 17 net homes. The filings there are ADUs, duplexes, lot splits and legalizations.

When will new supply actually show up?

Not soon. A project application is the first step, not the last. After it come entitlements, financing and a construction schedule of two years or more for a mid-rise. That is why San Francisco finished only about 405 homes in the first half of 2026, as we covered in our look at the 400-home pipeline. This is the next chapter of that story. The completions were thin, and the filings that would feed completions in 2029 and 2030 are thinning too.

For renters, that means the next two to three years of supply is mostly what is already under construction. Expect lease renewals to stay tense, especially near the AI office clusters in SoMa and Mission Bay.

For buyers, it means new-construction condos stay scarce and resale competition stays concentrated on the west side, where the Richmond and Sunset rank well on schools and safety. It does not mean prices can only go up. AI hiring can cool faster than a housing pipeline can refill, and a 25 percent year is not a trend line.

What I will be watching: October through December filings. If the developers quoted this summer are right that rising rents will pull projects off the shelf, applications should start climbing before spring. If they do not, the problem is bigger than interest rates. In the meantime, the block matters more than ever, and we keep our San Francisco neighborhood rankings current so you can compare the areas where the new homes are, and are not, coming.

Sources

SF Planning Department, Planning Department Records: Projects (qvu5-m3a2), DataSF, queried September 29, 2026 (analysis by Houseberry)

California Association of Realtors, County Sales and Price Activity, August 2026

CBS News Bay Area, San Francisco one-bedroom rents at record high (Zumper September 2026 report), September 29, 2026

DatacenterDynamics, AMD acquires AI research firm World Labs in all-stock deal worth $8.2bn, September 2026

The Real Deal, Before ChatGPT, San Francisco had 23 AI tenants. It now has 413, August 10, 2026

Freddie Mac, Primary Mortgage Market Survey, Mortgage Rates Average 7.03%, September 24, 2026

Mission Local, San Francisco drops affordable housing requirement from 15% to 5%, July 15, 2026

The San Francisco Standard, San Francisco developers project a building boom, August 24, 2026

The San Francisco Standard, SF's housing solution may be hiding in plain sight (Budget and Legislative Analyst report), September 29, 2026

Multifamily Dive, San Francisco's new zoning opens the door for more housing, December 2025

Houseberry, San Francisco city guide and neighborhood pages (Outer Richmond, Bernal Heights, South Beach), retrieved September 29, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The paperwork landed Monday morning. At 8:50 a.m. on Sept. 28, 2026, San Francisco's Department of Building Inspection logged permit 202609288684, a new eight-story apartment building with 214 homes at 3700 California Street. Three minutes later came permit 202609288686, a second building with 36 more.

That is 250 homes filed for the old CPMC California campus, the hospital block at California and Maple that has sat mostly dark since Sutter Health moved inpatient care to its new Van Ness campus in 2019. It is the first construction filing for the housing plan San Francisco approved in 2025, and the clearest sign yet that the project is moving from entitlement to actual building.

The takeaways

What the city's permit data shows as of Sept. 29, 2026:

• Two new-construction permits cover 250 apartments on the block bounded by California, Sacramento, Maple and Cherry streets.
• Both describe an eight-story, Type I-A concrete building over three basement levels, a style the approved plan anticipated for its tallest building.
• The approved plan totals 530 homes, including 158 independent senior units and 74 memory-care units, so this filing covers just under half of it.
• Eight single-building permits filed in December 2019 for the earlier, smaller plan were withdrawn this month.
• Nothing is issued yet. One permit sits in triage and the other shows filed, the first step of plan review.

What DBI received on Sept. 28

The two filings are labeled B1 and B2 and share the same basic description: a new eight-story-over-three-basement, residential mixed-use building of Type I-A construction. That is the concrete-and-steel construction class used for mid-rise buildings, not the wood-frame podium you see on most five-story projects.

The occupancy codes on both permits read R-2, S-1, A-3 and B. In plain terms that is apartments (R-2), a garage or storage level (S-1), assembly space such as a lobby lounge or amenity room (A-3), and a small amount of business space (B). Here is the filing side by side.

Read the cost column carefully. Both permits list the identical $93 million estimate, even though one building has six times as many homes as the other. The most likely reading is that the same project-level figure was entered on both applications, not that the two buildings each cost $93 million. DBI's estimated cost is also a fee-setting number, not a construction budget, and it tends to change during review.

Which block this is, exactly

The permits sit on Assessor Block 1016, the middle of the three campus blocks. Presidio Heights is the neighborhood of large pre-war houses running from the Presidio wall south toward California Street, and this block is its southern edge, where it meets Laurel Heights and the Laurel Village shopping strip a block or two east.

The filings attach to every lot on the block: 3700 and 3790 California Street, 3875 through 3891 Sacramento Street, and 406 and 420 Cherry Street, plus one parcel still marked "situs to be assigned." If you have walked past the hospital's California Street side, with the 1 California trolley wires overhead and the hospital's long, blank walls facing the street, you know the spot.

What is not in this filing matters too. The block east of Maple, where the 1939 Marshal Hale building will be preserved for the senior and memory-care program, has no new-construction permit in the city's data yet. Neither does the block west of Cherry toward Arguello.

From 273 homes to 530: how the plan grew

The campus has been approved twice, and the second approval nearly doubled it. In February 2020 the Planning Commission approved a TMG Partners plan for 273 homes, with 31 new buildings ranging from 35 to 80 feet, much of it large family-size units and single-family homes.

Prado Group then took the project over and filed an expanded plan in 2024. The version the city signed off on in 2025, as SF YIMBY reported in June 2025, totals 530 homes: 298 apartments, 158 independent senior living units and 74 memory-care units, on about 4.9 acres, with an 80-foot height cap. Handel Architects describes the senior block as 232 units anchored by the preserved Marshal Hale building.

I will say it plainly: the 2025 version is the better plan. A former hospital site a short walk from Laurel Village, on a trolley bus line, in one of the highest-scoring parts of the city, is exactly where San Francisco should be adding more homes, not fewer. Going from 273 to 530 on the same footprint is the kind of math the city's housing targets require.

The 2019 permits are being cleared out

The older plan left a paper trail, and it is being retired. On Dec. 24, 2019, the prior team filed a stack of one-building permits across the campus, each for a single house, a fourplex, or a small apartment building, which is how the 31-building plan had to be permitted.

In September 2026, eight of those new-construction permits switched to withdrawn in DBI's permit dataset: one for 420 Cherry Street on Sept. 4, and seven on the Marshal Hale block on Sept. 17, including six filed under 3773 Sacramento Street and one under 3698 California Street. Most of the other 2019 filings on the campus still show filed. Expect more of them to go as the new building permits replace them.

Eight stories and an 80-foot limit

The permits say eight stories, and the approval coverage says 80 feet. Those are not in conflict. SF YIMBY's April 2025 story on the environmental review described the tallest building in the plan as eight stories, alongside two seven-story buildings and a five-story one. Story counts and height limits measure different things, and eight floors of apartments can fit under an 80-foot line.

What the permits cannot tell you yet is the final massing. A filing at triage has not been through plan check, and drawings often change before a permit is issued. We could not reach Prado Group for comment before publication.

What 250 rentals mean for Presidio Heights and Laurel Heights buyers

The short answer: more neighbors, more foot traffic at Laurel Village, and very little effect on what single-family homes here sell for. The for-sale market around this site operates at a completely different price level from a new rental building.

On our San Francisco neighborhood rankings, Presidio Heights scores 4.2 out of 5 and ranks third of 92 neighborhoods, with a 4.4 amenities score. Its 12-month median sale price was about $8.03 million in August 2026, down from roughly $10 million in July. Jordan Park / Laurel Heights, the neighborhood just south and east of the campus, also scores 4.2, with a 4.6 safety score and a median of about $3.52 million in July 2026. The San Francisco citywide median was about $2.2 million in August 2026.

Those numbers deserve interpretation. Presidio Heights swings by a million dollars or more month to month because only a handful of homes trade, and one or two trophy sales, like the Benioff-linked mansion deal we covered in August, move the median on their own. That is not a market 250 apartments will touch.

What does change is who gets to live here. Right now the entry ticket to this part of the city is a $3.5 million house. A new rental building opens the same schools, the same Sacramento Street shops and the same Presidio access to people who will never buy at that price. When we compare neighborhoods, that is the part of the story we care about most.

The honest downside: construction on a full city block takes a long time. SF YIMBY's review coverage put full build-out at around 40 months across three phases. If you are buying on Maple, Cherry or the 3800 block of Sacramento, expect years of trucks, lane closures and noise, and budget your patience accordingly.

Next steps: plan check, a site permit, then excavation

The next milestone is plan check. A permit in triage has to be accepted for intake, routed through building, fire, planning and public works reviewers, and approved before DBI issues it. Issuance is what allows work to begin, and with three basement levels, excavation and shoring would come first.

Prado has run this sequence recently a few blocks east. At the former UCSF Laurel Heights campus at 3333 California Street, SF YIMBY reported first-phase building permits in February 2026 covering 152 homes, and Hoodline reported vertical construction there was expected to start in late 2026. If 3700 California follows a similar arc, the more telling date is not Monday's filing. It is the day these two permits flip from filed to issued. We will be watching the permit record for that.

Sources

• SF Department of Building Inspection, permit 202609288684 (DataSF)
• SF Department of Building Inspection, permit 202609288686 (DataSF)
• DataSF, Building Permits dataset
• DataSF, permits at 3773 Sacramento Street
• SF YIMBY, Project approved for redevelopment of California Pacific Medical Center (June 2025)
• SF YIMBY, Environmental review available for 3700 California Street (April 2025)
• SF YIMBY, Prado Group files updated plans for 3700 California Street (April 2024)
• Richmond Review/Sunset Beacon, Development project approved for CPMC property (April 2020)
• Handel Architects, Block C of 3700 California approved (May 2025)
• SF YIMBY, New building permits filed for first phase of 3333 California Street (February 2026)
• Hoodline, Presidio Highlands gets a date with the bulldozers (January 2026)
• San Francisco Chronicle, CPMC Van Ness hospital opens (2019)
• Houseberry, San Francisco neighborhood rankings
• Houseberry, Presidio Heights
• Houseberry, Jordan Park / Laurel Heights
• Houseberry, San Francisco city page

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Item 34 on Tuesday's Berkeley City Council agenda will decide what gets built along 2.35 miles of San Pablo Avenue for years, and the most important page in the packet is a supplemental added after the main report. It asks the council to require at least 145 homes per acre on any residential or mixed-use project along San Pablo Avenue. On a standard 5,000-square-foot lot, that works out to about 17 homes.

The San Pablo Avenue Specific Plan itself is built for 6,500 to 6,750 net new homes over the life of the plan, compared with about 3,500 if the city left today's zoning alone. The council meets at 6 p.m. on Sept. 29 and could adopt the whole package in one night. We will have the outcome Wednesday. Until then, here is what the math means block by block.

The short version

• The plan covers 2.35 miles of San Pablo Avenue, from the Albany border to the Oakland border, and replaces two zoning districts with one new San Pablo Commercial (C-SP) district.
• City consultants project 6,500 to 6,750 net new homes under the plan, roughly 86 to 93 percent more than the 3,500 expected under current rules.
• Staff's Supplemental 1 adds a minimum density of 145 homes per acre in C-SP. The Planning Commission's June 3 recommendation had no minimum at all.
• The corridor is expected to lose a net 225,000 to 320,000 square feet of commercial space as older one-story buildings turn into housing.

What the council is actually voting on Tuesday

Council is being asked to take five actions under one agenda item, according to the Sept. 29 agenda. It is a public hearing on the Action Calendar, at the School District Board Room at 1231 Addison St.

• CEQA. Adopt an addendum to the Housing Element environmental impact report instead of a new EIR.
• The plan. Adopt the San Pablo Avenue Specific Plan and its design standards.
• General Plan. Approve the land use redesignations that line the General Plan up with the new plan.
• Two old plans retired. Rescind the West Berkeley Plan and the San Pablo Avenue Public Improvements Plan.
• Zoning. First reading of the ordinance creating the C-SP district, Chapter 23.202.160, plus conforming code changes.

That last one matters for timing. An ordinance in Berkeley needs a second reading at a later meeting before it takes effect, so a yes vote Tuesday starts the clock rather than finishing it. The plan has been in the works since June 2023, with a public review draft in October 2025 and three Planning Commission sessions before the commission's final recommendation in June.

Where the new homes are supposed to go

The short answer is three intersections. The consultants' buildout memo (Appendix C, revised August 2025) assumes projects in the Tier 1 nodes reach about 250 homes per acre, far above anything else on the corridor.

The October 2025 public review draft put those Tier 1 nodes at University Avenue, Ashby Avenue and Gilman Street, with Tier 2 nodes at Cedar Street and Dwight Way. In that draft, Tier 1 sites could go to 85 feet and eight stories. Everywhere else on the avenue, the model assumes about 140 homes per acre.

Look at the chart and one thing jumps out. Outside the nodes, the plan's assumed density (140) is barely different from today's C-W baseline (135). The big gain comes from the height and density allowed at the intersections, plus more sites becoming feasible at all. That is why the node locations, not the corridor as a whole, are where buyers should pay attention.

The site inventory behind the numbers is modest. The memo counts 51 housing opportunity sites, including 12 already in the pipeline and 6 more flagged as likely to develop. That is not a lot of dirt for 6,500 homes, which is exactly why staff wants a floor on density.

The 145-homes-per-acre minimum, lot by lot

The minimum density is the one real disagreement between staff and the Planning Commission. Supplemental 1, submitted by Jordan Klein of the Planning and Development Department, adds a 145 dwelling-units-per-acre minimum for residential and mixed-use projects in C-SP. The commission's June 3 version had "no minimum." Neither version sets a maximum density.

The supplemental's C-SP table also sets a 4.25 floor-area ratio and 75 feet and seven stories for residential and mixed-use buildings, and 40 feet and three stories for non-residential ones. Here is what a 145 floor means on real San Pablo lot sizes, next to what the Tier 1 assumption would imply.

These are Houseberry's own calculations from the city's density figures, rounded, not project approvals. But they show what the rule does. A 10,000-square-foot corner that might otherwise get 12 townhomes would need about 33 homes. With a 4.25 floor-area ratio, that lot allows roughly 42,500 square feet of building, so 33 homes averages under 1,300 gross square feet each. The minimum fits comfortably inside the envelope. It just stops someone from building well below it.

I think staff has this right. San Pablo is a trunk transit line with a 2.35-mile supply of commercial lots, and every one that gets a low-density project is gone for 50 years. If you believe the corridor should carry a real share of Berkeley's housing, a floor is how you protect the sites. The draft plan itself notes that about 31 percent of the city's regional housing target, 2,023 units, falls in the plan area.

The honest counterargument is feasibility. Construction costs are high, and a small-lot owner who can pencil 10 units but not 17 may simply build nothing. The supplemental lists no carve-out for small lots or existing buildings, so watch whether council adds one Tuesday.

The commercial space trade-off

San Pablo will get more homes and less shop space. The buildout memo projects a net loss of 225,000 to 320,000 square feet of commercial space under the plan, a 15 to 21 percent drop, versus a smaller 150,000 to 215,000 square-foot loss under current zoning. The draft counts about 1.49 million square feet of commercial space on the corridor today.

That sounds worse than it is. Much of what disappears is one-story auto repair, storage and single-tenant buildings (anyone who has driven past Dwight on a Saturday knows the look). The draft requires ground-floor commercial at the nodes, so the shopping should concentrate where the foot traffic will be. Ground-floor vacancy on the corridor peaked at 15.5 percent in 2023 before falling to 7.1 percent in 2024, per the draft. Fewer, better storefronts is a reasonable bet.

Still, if a specific shop or service is why you like your block, it is worth knowing whether its building sits on an opportunity site.

What this means if you are buying near San Pablo

Buyers in the flats west of Sacramento Street get the most direct effect, good and bad. The plan is aimed at parcels fronting the avenue, with some changes reaching about two blocks off it, mostly around University, according to SF YIMBY's preview of the June hearing. Side streets of single-family homes are not being turned into seven-story zones.

Here is how the neighborhoods along the corridor look in our data right now.

• Berkeley Oceanview, the flatland neighborhood west of San Pablo toward I-80 where Berkeley was first settled, had a median of about $1.15 million in August 2026. That is roughly a third below the $1.73 million Berkeley citywide median the same month. Its 4.6 amenities score is the highest of Berkeley's 18 ranked neighborhoods, and its 2.7 safety score is below the city's middle.
• Westbrae, the northwest pocket along the Ohlone Greenway near the Gilman node, reached about $1.56 million in August 2026, up from $1.32 million in September 2025.
• West Berkeley, the mix of industrial blocks and cottages between San Pablo and Fourth Street, ran about $1.23 million in August 2026 and ranks 17th of 18 overall, with a 2.1 safety score.

The discount on the San Pablo side is real, and it is not because the area lacks things to walk to. The trade-off is that you are buying next to a corridor designed to change. A single-family house a block off the avenue near University or Gilman could, over the next decade, look onto seven- and eight-story buildings with ground-floor shops.

For some buyers that is the pitch: more neighbors, more storefronts, more reasons for AC Transit to run often. For others it is a dealbreaker. Neither is wrong. What matters is knowing which one you are before you write the offer.

A few practical checks: look up whether the parcels facing the avenue near you are on the plan's opportunity-site map in the draft, note which node you are closest to, and read the height standards for that node. Land in West Berkeley is still cheap by Berkeley standards, as the $14.5 million sale of the University Inn near University and Eighth showed this month. Cheap land and higher zoning are exactly the conditions that get projects started.

Does the plan change what my own house can become?

Probably not directly, unless your lot fronts San Pablo. The C-SP district applies to the corridor parcels, and residential side streets keep their own zoning, including the state rules that already allow ADUs and lot splits. The indirect effect is the one to think about: what gets built across the street from you, and how that changes traffic, parking and the shops you walk to.

After Tuesday: second reading and the first projects

If council adopts the plan Tuesday, the C-SP ordinance returns for a second reading at a later meeting before the new zoning takes effect. The minimum density question is the piece most likely to be amended from the dais, so the version that passes may not match either the commission's or staff's.

We will report the outcome Wednesday and update our Berkeley neighborhood rankings coverage as projects are filed. For anyone shopping in Oceanview, Westbrae or West Berkeley this fall, the useful habit is the one we always come back to at Houseberry: learn what is planned for the neighborhood before you fall for the house.

Sources

• City of Berkeley, City Council Regular Meeting eAgenda, Sept. 29, 2026, Item 34
• City of Berkeley, Item 34 Supplemental 1, Minimum Density, Sept. 2026
• City of Berkeley, Item 34 staff report, Adoption of San Pablo Avenue Specific Plan
• City of Berkeley, San Pablo Avenue Specific Plan Appendix C, Buildout Summary Memorandum, rev. Aug. 29, 2025
• City of Berkeley, San Pablo Avenue Specific Plan Public Review Draft, Oct. 2025
• City of Berkeley, San Pablo Avenue Specific Plan project page
• SF YIMBY, preview of the June 3, 2026 Planning Commission hearing
• Houseberry Berkeley neighborhood rankings and neighborhood price pages, accessed Sept. 28, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The question parents and longtime renters keep asking me this fall is a simple one. If a fire guts my rent-controlled apartment, do I get it back?

In San Francisco, the answer is mostly yes. Once the unit is repaired, the landlord has to offer it back to you on the same terms and at the same rent. What the law does not give you is money to live somewhere else while you wait, and that gap is where people get hurt. The San Francisco Standard reported on September 28, 2026 that one Nob Hill tenant went from a $975 rent to a $2,900 replacement apartment after a fire.

In brief

• San Francisco counted 87 fire displacements by late September 2026, versus 80 in all of 2025 and 52 in 2023, according to Fire Marshal figures cited by the Standard.
• Rent Board Rule 12.19 requires a landlord to offer the repaired unit back within 30 days of finishing repairs, at the same rent and terms.
• The Rent Ordinance does not require relocation payments after a fire the landlord did not cause. That is the biggest surprise for most tenants.
• The city's Temporary Assistance for Displaced Persons program can cover the rent gap for rent-controlled tenants for up to 12 months, after insurance runs out.
• San Francisco's median one-bedroom rent was $4,295 in September 2026, per Zumper, which is what a displaced tenant is shopping against.

Why fire displacement is suddenly a bigger San Francisco story

Fires in residential buildings are displacing more people than they were two years ago, and 2026 has already passed 2025 with a quarter of the year left. Fire Marshal Ken Cofflin's figures, shared with the Board of Supervisors and reported by the Standard, put 2023 at 52 displacements, which was itself a 10-year high. Last year hit 80. This year is at 87.

The buildings behind the recent headlines are older, rent-controlled ones. The Standard walked through a 24-unit building on La Playa Street in the Outer Sunset, the fog-belt neighborhood running along Ocean Beach between Golden Gate Park and Sloat Boulevard. A December fire in the Tenderloin uprooted about 142 tenants on its own.

I want to be careful here. A rising count is worth knowing about, but it is not a reason to avoid older buildings. The Outer Sunset still ranks second of 92 San Francisco neighborhoods in our citywide neighborhood ranking, with a 4.3 out of 5 safety score. What the trend changes is how seriously you should take the paperwork below.

Your right to return, step by step

Rent-controlled tenants keep their tenancy through a fire, and the landlord must bring them back once the unit is fixed. The rule is Section 12.19 of the Rent Board's Rules and Regulations, and the city's own summary lays out the clock.

Within 30 days after repairs are finished, the landlord has to offer you the same unit, under the same terms and conditions you had before. You then have 30 days to accept and 45 days from the offer to move back in. The offer has to go to the address you gave, the unit itself, or any other address the landlord knows about, and that explicitly includes email.

If the landlord re-rents your apartment to someone else instead, that is treated as a wrongful eviction under the Rent Ordinance, with exposure to actual and punitive damages. So the single most useful thing you can do on day one is put a forwarding email in writing to your landlord and keep a copy.

One real caveat. The right to return assumes there is something left to repair. When a building is demolished to the ground, whether the replacement counts as "repair" gets murky fast, as Mission Local found with the 22nd and Mission fire. Most fires never get there, but big ones can.

The relocation money gap nobody warns you about

San Francisco does not require landlords to pay relocation money when a fire they did not cause forces tenants out. City and state rules do require payments when a landlord moves tenants out for capital improvements or major rehab. A fire is treated as outside the landlord's control, so the relocation piece falls away. (If the landlord's negligence caused the fire, say a known-bad heater or dead alarms, that is a separate lawsuit with a two-year window, and worth a call to a tenant attorney.)

Here is what that gap looks like in real money. The tenant the Standard profiled paid $975 a month for her unit on Leavenworth Street in Nob Hill. Her replacement costs $2,900. Twelve months of that difference is about $23,100 out of pocket, before movers or storage.

And she found something cheap by today's standards. Zumper's September 2026 report puts the San Francisco median one-bedroom at $4,295, up 21 percent in a year. The Standard cites an average closer to $4,500. For a long-tenured renter, that is the true cost of a fire: not the lost furniture, but the market you get thrown back into.

Where the city steps in

The city fills part of the relocation gap, but you have to ask, and the order matters. The Red Cross (866-272-2237) handles the first nights of shelter, food and clothing, per SFFD's fire damage page. After that, the Human Services Agency's Temporary Assistance for Displaced Persons program pays rent-controlled tenants a monthly subsidy for up to 12 months, or until the unit is ready if sooner. Non-rent-controlled tenants get a one-time move-in payment instead.

The eligibility lines are strict. You must apply within three months of displacement, have income at or below 100 percent of area median income, and have already used up any renters insurance benefit. HSA's page still lists an asset limit of $30,000 per person, but the Board of Supervisors voted in January 2026 to raise it, and Hoodline reported the new single-person cap at $130,000. Ask HSA which figure applies before you rule yourself out.

Two more tools are worth knowing. A Good Samaritan tenancy lets a landlord (yours or another) house you in a rent-controlled unit at your old rent, or up to 10 percent above it, for up to 12 months, extendable to 24, without that rent becoming the unit's new base. And if you will be out six months or longer, the Displaced Tenant Housing Preference moves you up in the city's affordable housing lotteries.

Here is how the pieces fit together.

Renters insurance is the piece people skip

A renters policy with loss-of-use coverage is what actually pays for the hotel and the pricier sublet in the first months. The California Department of Insurance's residential claims guide describes additional living expenses as covering temporary housing, food above your normal budget, storage, furniture rental and extra commuting, usually with an advance up front.

Keep every receipt. Insurers will ask you to account for each dollar, and HSA will ask whether your insurance is exhausted before it pays. The fire report itself comes from the SFFD Bureau of Fire Investigation (415-920-2933), and your insurer will want it.

If you own the building

Small landlords carry real obligations after a fire, and the ones that trip people up are about timing and notice rather than money. You must offer the repaired unit back within that 30-day window, at the old rent, to every address you know, including email. Doing basic research to find a displaced tenant's current contact is expected.

You are not stuck eating every cost. Repair costs your insurance does not cover can be passed through to tenants with a capital improvement petition at the Rent Board, with proper notice under Civil Code Section 827. A loss-of-rents endorsement on your landlord policy is what replaces income while units sit empty. And if you have a vacant unit, a Good Samaritan tenancy lets you keep a tenant housed without resetting that unit's base rent.

Where owners get in trouble is folding fire repairs into a larger remodel. Moving tenants out for capital improvements brings in notice rules, permits before any temporary eviction, and relocation payments, the Standard notes. Fire does not waive those when the work goes beyond fixing the damage.

I understand the temptation, and the numbers explain it. Our price history puts the Outer Sunset median at about $1.69 million in August 2026, up from about $1.47 million in September 2025, and the San Francisco citywide median at about $2.2 million. A small building with a $975 tenant is worth less to a buyer than one with a market-rate unit. That is exactly why the city made refusing a return a wrongful eviction, and why stalling rarely pays.

When waiting for your unit stops making sense

The honest downside of the right to return is that it has no deadline for the repair itself. Big restorations can stretch past a year, and the city subsidy stops at 12 months. Some tenants stay out long enough that the landlord offers a buyout instead, and landlord buyout filings are up 33 percent this year for reasons that have nothing to do with fire.

A buyout can be the right call, especially for someone who was already planning to move. But price it against the rent you would pay for the rest of your time in the city, not against this month's hotel bill. At a $3,300 monthly gap between a $975 rent and today's median one-bedroom, five years is roughly $200,000. Rent-controlled rents only rise by the Rent Board's annual allowance, and banked increases can add up, but the gap usually stays wide.

If you are the one deciding whether to hold on, whether that is in the Outer Sunset or on Nob Hill, the neighborhood still matters as much as the unit. We built Houseberry around looking at the area first, and that habit helps here too: if you end up moving, knowing which neighborhoods fit your budget and your commute before you start touring saves weeks when you have the least time to spare.

Sources

• San Francisco Standard, "Burned out: Fires are raging in rent-controlled apartments" (Sept 28, 2026)
• SF.gov, Tenant displacement due to fire or other disaster (Rent Board Rule 12.19)
• SF.gov, Good Samaritan tenancy information
• SF Human Services Agency, Emergency services for fire damage
• Hoodline, SF hikes fire aid after Tenderloin blaze (Feb 2026)
• SF.gov, Displaced Tenant Housing Preference
• San Francisco Fire Department, Help for fire damage
• Zumper, San Francisco rent research (Sept 2026)
• California Department of Insurance, Residential property claims guide
• Mission Local, Right to return after the 22nd and Mission fire
• Houseberry, Outer Sunset and San Francisco price history (Aug 2026)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

$604 a square foot. That is what an affiliate of Zurich Alternative Asset Management paid for 410 Townsend, the brick SoMa building that once housed Yammer, Eventbrite and Zendesk, and it is more than double what the last buyer paid in 2024. Two days later, a nearly empty office pair on Howard Street sold for about $231 a foot.

Same neighborhood, same month, a 2.6x gap. The short answer for anyone shopping condos in South Beach, Yerba Buena or Mission Bay: SoMa office sales now show that well-located, leased buildings have bottomed and turned, while empty ones still trade near land value. And the condo market next door has not priced in the good half of that story yet.

The numbers that matter

Here is the whole story in six lines.

• 410 Townsend sold for $47.4 million in September 2026, about $604 per square foot, after trading for $22 million (about $280 a foot) in April 2024.
• In 2019, the same building sold for roughly $86 million, or close to $1,100 a foot. The new price is still about 45 percent below that peak.
• Hudson Pacific sold 875 and 899 Howard, about 284,000 square feet, for $65.5 million in September 2026. One building was 36 percent leased and the other was empty.
• San Francisco office vacancy was 30.1 percent in the second quarter of 2026, down from 33.8 percent a year earlier, according to Cushman and Wakefield.
• On our data, the median home price in South Beach was about $1.23 million in August 2026, up roughly 3 percent in twelve months, while the San Francisco citywide median rose about 16 percent to $2.2 million.
• South of Market ranks 90th of 92 San Francisco neighborhoods on Houseberry's overall score. South Beach ranks 66th and Mission Bay 59th.

Why a broker is calling 410 Townsend a round trip

The 410 Townsend sale is the first San Francisco office building to go from boom price, to distress price, and back up to a real recovery price, and that is why brokers are excited about it. Seth Siegel of Cushman and Wakefield, who ran the deal, called it the first "round trip" for a San Francisco office building.

The building sits on Townsend between Fourth and Fifth Streets, a short walk from the Caltrain terminal at Fourth and King. It is a 1912 brick and heavy timber structure of about 78,455 square feet, and it was roughly 65 percent leased at sale. New York Life Real Estate Investors and Bridgeton Holdings bought it for $22 million in April 2024, a deal CBRE brokered, about $64 million under what it sold for in 2019.

So the 2024 buyers more than doubled their money in about two and a half years. That is not a rounding error. It is a sign that at least some investors now think the bottom for San Francisco office is behind us.

The Howard Street sale is the other half of the picture

The Howard Street deal shows that empty SoMa office space is still priced like a problem. Hudson Pacific announced the $65.5 million sale on September 25. 875 Howard is a roughly 188,000 square foot office building, and 899 Howard is a roughly 96,000 square foot former retail building that was fully vacant. Hudson Pacific bought the pair in 2007 for $46 million, according to Hoodline's reporting.

These buildings sit between Fourth and Fifth on Howard, across from the Moscone Center, and a few blocks from 410 Townsend. The difference is not the address. It is the rent roll. A building with tenants and a story sells for $600 a foot. A building with a vacant shell sells for a price that starts to look like a bet on what the land can become.

That split matters for housing, because the empty buildings are the ones that eventually become something else.

What SoMa office buyers are betting on

Buyers are betting that AI companies keep filling San Francisco office space, and SoMa has the most room to fill. Cushman and Wakefield's second quarter 2026 report counted 2.9 million square feet leased by AI companies in the first half of 2026, which was 47 percent of all new leasing in the city. The city has now logged five straight quarters of positive net absorption.

SoMa is still the weak spot. The Real Deal reported SoMa office vacancy at about 47 percent in early 2026, compared with about 20 percent in Mission Bay. The San Francisco Standard's September look at SoMa describes Fifth Street as a dividing line, with recovery showing up to the east and harder street conditions persisting to the west.

One note on the vacancy numbers. Hoodline cited 29.2 percent for the second quarter while Cushman reports 30.1 percent. Brokers measure inventory differently, so both can be right. The direction, down about three to four points in a year, is what every source agrees on.

The condo market has not caught up

Here is the gap nobody else is talking about. While office buyers repriced 410 Townsend upward by more than 100 percent, the condo neighborhoods around it barely moved. Using our 12-month price history, South Beach went from about $1.19 million in September 2025 to about $1.23 million in August 2026. Yerba Buena, the blocks around Moscone Center and Yerba Buena Gardens, went from $1.12 million to $1.27 million, peaking at $1.37 million in May before easing back.

Over the same stretch, the San Francisco citywide median climbed from $1.89 million to $2.2 million. That citywide number is pulled up by single-family homes on the west side. So the right read is not that South Beach condos are a bargain in absolute terms. It is that the downtown condo market has been left behind by the rest of the city.

The South of Market neighborhood itself is harder to pin down. Its monthly median in our data swings from about $1.7 million in July to about $860,000 in August, which tells you sales volume is thin and the mix of lofts and towers changes month to month. The Standard reported a SoMa median condo price of $855,000 in June, up 15.7 percent from a year earlier. We would treat any single SoMa monthly number with caution.

How the four neighborhoods compare

The four neighborhoods around the office core score very differently, and the scores explain some of the price gap. Here is how they line up on our data.

South Beach, the waterfront strip along the Embarcadero from the Bay Bridge down to Oracle Park, pairs a 3.5 safety score with one of the better amenity scores in the city. Mission Bay, the newer district south of Mission Creek anchored by UCSF and Chase Center, scores highest on safety at 3.9. South of Market sits near the bottom of the citywide ranking, mostly because of its 1.9 safety score.

The way we read it: a buyer choosing between a South Beach two-bedroom and a Yerba Buena two-bedroom at similar prices is choosing between a 3.5 and a 2.2 on safety for roughly the same money. That is the kind of trade-off the listing photos never show.

What this does and does not mean

It does mean the office story in SoMa is no longer only about falling prices. At least one building has bounced hard, and AI leasing is doing real work in the vacancy numbers. When offices fill, the lunch crowd, the evening foot traffic and the retail on Third and Fourth Streets usually follow, and those are the things that move how a condo neighborhood feels day to day.

It does not mean South Beach condos are about to jump 16 percent. One building is one data point, and 410 Townsend is a small, characterful building that was already mostly leased. The honest downside is that the downtown condo market has lagged for years for reasons offices do not fix: high HOA dues, lots of competing inventory in the towers, and buyers who prefer a house in the Sunset to a view unit on Folsom. The Howard Street sale is a reminder that plenty of SoMa office space is still empty.

And I would argue that is the good news for housing. Empty office shells at $231 a foot are much closer to the math that makes conversions or new residential work, which is exactly what this part of the city needs. We looked at the neighborhood side of the AI office boom in our earlier piece on where AI offices are landing and what it means for nearby neighborhoods.

Questions people are asking

Did 410 Townsend really double in price?

Yes. It sold for $22 million in April 2024 and $47.4 million in September 2026, which is about 2.15 times the earlier price. It is still well below the roughly $86 million it sold for in 2019.

Who bought 410 Townsend?

An affiliate of Zurich Alternative Asset Management bought it from New York Life Real Estate Investors and Bridgeton Holdings. Cushman and Wakefield brokered the sale.

Is the San Francisco office market recovering in 2026?

The data says yes, slowly. Cushman and Wakefield put citywide vacancy at 30.1 percent in the second quarter of 2026, down from 33.8 percent a year earlier, with AI companies accounting for nearly half of new leasing in the first half.

Will office recovery raise SoMa and South Beach condo prices?

It can help, but it is not automatic. South Beach prices rose only about 3 percent over the past year on our data, while the city median rose about 16 percent. More office workers nearby supports demand, but HOA costs and tower inventory still weigh on downtown condos.

Which is safer, South Beach or SoMa?

On Houseberry's scores, South Beach rates 3.5 out of 5 on safety and South of Market rates 1.9. Mission Bay rates highest of the four at 3.9.

If you are weighing a downtown condo right now, the useful move is to compare the neighborhoods block by block before you compare the units. The office market just told you which parts of SoMa investors believe in. The neighborhood scores tell you which ones you would actually want to walk home through at 10 p.m.

Sources

• Connect CRE: Zurich affiliate acquires 410 Townsend
• Hoodline: Zurich investor pays $47.4M for SF's old start-up village
• CBRE: $22 million sale of 410 Townsend, April 2024
• Newsweek: 410 Townsend 2019 and 2024 prices
• Business Wire: Hudson Pacific sells 875 and 899 Howard
• Hoodline: Hudson Pacific sells nearly empty SoMa office complex
• Cushman and Wakefield: San Francisco Office MarketBeat, Q2 2026
• The Real Deal: Will spillover of SF office demand resurrect SoMa?
• The San Francisco Standard: SoMa is San Francisco's rise and fall, all at once
• Houseberry: San Francisco neighborhood rankings and price history

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

On Oct. 6, the San Francisco Board of Supervisors is scheduled to vote on a new name for Cesar Chavez Street, and the choice is down to two: “Sí Se Puede” or “Las Campesinas.” New signs could be up by the end of the year, according to SFist. Two hours up Interstate 80, Sacramento is on a parallel track. Its parks commission takes up a recommendation on Oct. 1 to rename downtown's Cesar Chavez Plaza as Union Plaza.

Both moves trace back to the same March 2026 investigation into Chavez's conduct. And if you own a home or run a business on the San Francisco street, the practical answer is simpler than the politics. Your house number stays, your parcel stays, and your mail will keep arriving. The paperwork is what moves.

Key takeaways

• San Francisco's working group narrowed the Cesar Chavez Street new name to Sí Se Puede, the farmworker slogan Dolores Huerta coined in 1972, or Las Campesinas, meaning women farmworkers. Neither honors an individual.
• The Board of Supervisors vote is set for Oct. 6, 2026. The city puts sign and sidewalk-etching replacement at about $107,000.
• Sacramento's recommended name for Cesar Chavez Plaza is Union Plaza, the top choice among more than 2,600 participants. The City Council makes the final call.
• A street rename changes the name on your address, not the parcel it describes. Expect a stretch of updating mail, accounts, listings and business records, not a change in what your home is worth.

Why Cesar Chavez's name is coming off the street

The renaming follows a New York Times investigation published March 18, 2026, which reported that Chavez sexually abused two girls in the 1970s, beginning when they were minors. The same day, co-founder Dolores Huerta said publicly that Chavez had sexually assaulted her in the 1960s, as NPR reported. The United Farm Workers called the allegations “crushing” and pulled out of that year's Cesar Chavez Day events.

California moved fast. The Legislature passed AB 2156 unanimously and Gov. Gavin Newsom signed it on March 26, renaming the March 31 state holiday Farmworkers Day. Los Angeles, San Diego and Denver are among the places that have since renamed or revisited their own Chavez observances.

What stands out in San Francisco's version is who is leading it. The working group took its cue from Huerta, who asked that places honor the movement's organizers, farmworkers and families rather than swap in another single name, hers included.

How San Francisco landed on Sí Se Puede and Las Campesinas

A 27-member working group assembled by Mayor Daniel Lurie's office, Assessor-Recorder Joaquin Torres and former Supervisor Susan Leal picked the finalists, Mission Local reports. Labor leaders, muralists and educators were on it. Leal, a former supervisor and longtime resident, reached out to the mayor's office right after the March revelations.

The group presented the two names at a community meeting at Leonard R. Flynn Elementary in the southern Mission. Reaction ranged from enthusiasm for Sí Se Puede to calls to go back to Army Street, and some neighbors worried about pronunciation, spelling and the cost of changing addresses. KQED described “fairly broad support” for Sí Se Puede in the room.

District 9 Supervisor Jackie Fielder is carrying the legislation. The process, as reported:

• Now. The city is taking public comment on both names through an online form.
• Oct. 6, 2026. Board of Supervisors vote, per SFist.
• By year end. New street signs and sidewalk etchings, budgeted at about $107,000. Unlike the 1995 signs, the new ones will not carry Chavez's name.
• Alongside. San Francisco Public Works coordinates with the U.S. Postal Service, and the city has pledged help for residents and businesses updating records, per Hoodline.

My honest read is that the $107,000 number covers city street signs, and I would not assume it covers every freeway guide sign on US-101 and I-280. The 1995 rename from Army Street was budgeted at $20,000 and ended up around $900,000, mostly because of state highway signage, The Frisc found. Still a rounding error in a city budget. Worth watching all the same.

Three miles, five neighborhoods, very different price points

Cesar Chavez Street runs about three miles from Douglass Street in Noe Valley to Pier 80 on the bayfront, crossing Supervisor Districts 8, 9 and 10. It is six lanes for most of that run, paved over the old Precita Creek and widened into a car route by 1950. Nobody would call it a pretty street. But it is the seam that stitches together several of the city's most distinct neighborhoods.

Going west to east, it starts in Noe Valley, the sunny residential pocket south of the Castro built around 24th Street. It then cuts through the southern Inner Mission, including La Lengua and the edge of the Latino Cultural District around Mission Street. Next it forms the northern border of Bernal Heights along Precita Park, before diving under the US-101 interchange locals call the Hairball and running between Potrero Hill and Bayview to the water.

Our neighborhood price data shows just how much range one street name covers.

Potrero Hill's median was about $2.96 million in August 2026, and Noe Valley's about $2.53 million. Bernal Heights sat near $2.40 million and the Inner Mission near $2.33 million the same month. Bayview, at the eastern end, was about $869,000 in July 2026. So the same renamed street will appear on the address of a Noe Valley single-family house and a Bayview starter home that costs a third as much.

The scores tell a similar story of range. Here is how the neighborhoods along the street compare in Houseberry's San Francisco neighborhood rankings, out of 92.

Read that table and the name on the signs stops looking like the variable that matters. A buyer choosing between Bernal and Noe Valley is weighing a 0.7-point safety gap and about $130,000 in median price. What the street is called barely registers.

What a new name changes for your address

The short version is that a street rename changes the words on your address and nothing underneath them. San Francisco identifies property by Assessor's Block and Lot, and that number does not change when a street is renamed. Your house number should not change either.

What does change is a long list of records keyed to the old name. The practical checklist, if you own, rent or work on the street:

• Mail. Don't file a USPS change-of-address form for a city rename, because you have not moved. Cities coordinate the update with the Postal Service directly. When St. Paul renamed Rondo Avenue in 2024, mail to the old name was forwarded for up to 18 months. San Francisco has not yet published its own window.
• Carriers. Amazon, UPS and FedEx keep their own address files and do not always update automatically. Change the saved address yourself.
• Deeds and title. A recorded deed generally stays valid after a rename, because California deeds describe the parcel by its legal description, not only its street address. Future documents will use the new name. If you are selling or refinancing in the transition, ask your title officer to reference both names. Our guide to checking your own title walks through the Assessor-Recorder lookup.
• The boring accounts. Lender, homeowners insurance, DMV, voter registration, utilities, bank and payroll. Most take five minutes each.
• Businesses. Business registration, permits, insurance, Google Business Profile, Apple Maps, Yelp, printed checks, letterhead and storefront signage. Google and Apple update street data on their own schedules, so expect a few months of maps showing both names.

The real cost falls on small businesses. Businesses along the street went through this once already in 1995, and some will do it again three decades later. The city's promise to help with that is the part of the plan I would hold it to.

Will renaming the street affect home values?

There is no good evidence a rename moves prices, and the local precedent points the same way. Army Street became Cesar Chavez Street in 1995, Noe Valley residents put Proposition O on the ballot that November to change it back, and it lost by nearly 10 points. The blocks on either side kept doing what San Francisco real estate did for the next 30 years.

Buyers price the neighborhood, the block and the house. Noe Valley still commands a premium over Bernal because of its housing stock and its reputation, not its cross streets. The main thing for a buyer in escrow on the street this fall is paperwork. Make sure your disclosures, loan documents and insurance binder agree on which name they use.

Sacramento's Cesar Chavez Plaza is heading toward Union Plaza

Sacramento's version involves a park, not an address. Cesar Chavez Plaza is the downtown square directly across from City Hall, part of the city's original 1849 plan, and the city has recommended renaming it Union Plaza, according to the city's own announcement.

The Youth, Parks, and Community Enrichment Department ran the process under the city's facility naming policy. An online survey opened in June 2026, and more than 2,600 people took part and submitted over 350 unique names. Union Plaza drew about 450 responses, nearly double the runner-up, per CBS Sacramento. The other finalists were Dolores Huerta Park, Farmworker Plaza, People Plaza, Plaza Park and Nisenan Plaza.

A council subcommittee appointed by Mayor Kevin McCarty reviewed the results. The Parks and Community Enrichment Commission considers the recommendation on Oct. 1, and the City Council makes the final decision on a date not yet announced. The Chavez statue in the plaza has been covered since the spring, and the city canceled its Cesar Chavez Day event this year.

For homeowners, the Sacramento change is almost entirely symbolic. No residential addresses sit on the plaza itself. The neighborhood around it, Downtown Sacramento, scores 2.8 overall and ranks 68th of 185 in our Sacramento rankings, with a 4.0 amenities score that says more about daily life there than any name on a park sign.

Quick answers on the Cesar Chavez Street new name

What will Cesar Chavez Street be called?

Either Sí Se Puede or Las Campesinas. The Board of Supervisors is scheduled to pick on Oct. 6, 2026.

What was Cesar Chavez Street called before?

Army Street. It was renamed for Chavez in 1995. Going back to Army Street is not one of the finalists.

Do I need to file a change of address?

Not with USPS, since you have not moved. You will want to update banks, carriers, insurers and online accounts once the new name takes effect.

Will Cesar Chavez Elementary be renamed too?

That is a school district decision, separate from the street. The San Francisco Unified School District had not said as of late September.

The next two weeks decide both names

Sacramento's commission meets Oct. 1. San Francisco's board votes Oct. 6, and if it passes, sign crews could be out before New Year's. In both cities the naming is a real statement about who a place honors, and it deserves to be taken seriously. For the people who live along these three miles, though, the thing that shapes their daily lives is still the neighborhood itself: the park at the end of the block, the school assignment, the commute across the Hairball.

That is where we spend our time at Houseberry. If you are weighing a move along the corridor, start with the neighborhoods and let the street sign sort itself out.

Sources

• SFist, “SF's Cesar Chavez Street to Be Renamed ‘Sí Se Puede’ or ‘Las Campesinas’,” Sept. 25, 2026
• Mission Local, “San Francisco unveils two new names for Cesar Chavez Street,” September 2026
• KQED, “Here Are San Francisco's Top Picks to Rename César Chavez Street,” September 2026
• Hoodline, “SF Narrows Cesar Chavez Street Renaming to Two Options,” September 2026
• NPR, coverage of the New York Times investigation and Dolores Huerta's statement, March 18, 2026
• CalMatters, “César Chávez Day renamed to Farmworkers Day,” March 2026
• The Frisc, “What Happens If SF Takes Cesar Chavez Street Off the Map,” March 22, 2026
• City of St. Paul, Rondo Avenue renaming FAQ (USPS forwarding precedent)
• Sacramento City Express, “Community input leads to recommendation to rename Cesar Chavez Plaza as Union Plaza,” Sept. 25, 2026
• CBS Sacramento, “Sacramento reveals top choice to rename Cesar Chavez Plaza,” September 2026
• Houseberry neighborhood pages and San Francisco and Sacramento rankings, accessed Sept. 28, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Did San Jose really get a third of its street homelessness off the streets in a year and a half? It is a fair question after last week's headline, and one plenty of San Jose parents and house hunters are asking.

Here is the short answer. The city says unsheltered homelessness fell 32 percent in 18 months, to about 3,700 people in August 2026. But the thing that changed first was the method. San Jose stopped leaning on the federal point-in-time count and built a new data model with the AI firm UrbanLogiq, and the 32 percent is measured on that new yardstick. Advocates say the number does not match what they see on the ground.

In brief

If you only read one part of this, read these.

• San Jose's new model estimated 3,702 unsheltered people in August 2026, down from 5,477 in February 2025, per KTVU. That is the 32 percent.
• The old yardstick, the county's point-in-time count, found 3,959 unsheltered people in San Jose in January 2025. The model's starting point sat about 1,500 people higher.
• The model blends 15 data points, including police and fire encounters, lived-in vehicle counts, and outreach records. The city plans to update it every quarter.
• Advocates, including Shaunn Cartwright and Bea Nasser, dispute that a drop of that size is real, pointing to homeless deaths on pace to match last year's and people cycling back to the street.
• What no one disputes: San Jose has added a lot of shelter. The sheltered count went from 980 in 2019 to 2,544 in 2025.

What changed: the way San Jose counts

The biggest change in this story is the measuring tape, not the street. For years, San Jose's official number came from the point-in-time count, a federally required survey where volunteers fan out over two nights every other winter and tally who they find outside.

Everyone, including the mayor, agrees that count misses people. "The frustration we hear with the Point-in-Time count, the inaccuracy, the likely undercount, the infrequency, is legendary," Mayor Matt Mahan said at the September 16 news conference, as KTVU reported.

So the city built its own estimate. According to San José Spotlight, the model combines police and fire encounters with homeless residents, counts of lived-in vehicles, the countywide Homeless Management Information System (the database where outreach workers log their contacts), and other sources. KTVU and ABC7 describe it as an algorithm running on 15 data points. ABC7 reports the city calls it "independently verified." As of September 28, we could not find a published methodology or the name of the verifier, which is the first thing I would want to see.

The model's upside is real. A quarterly number lets the city see whether a new shelter site or an encampment policy is working this season, not two years from now. Housing Director Erik Soliván framed it as understanding the "constant demand" of people falling into homelessness, per the Spotlight.

Look at the two lines. They are not one series. The blue line is the street survey, which peaked near 5,000 unsheltered people in 2022 and fell to 3,959 in 2025. The red line is a different instrument that starts in a different place.

Is the 32 percent a like-for-like comparison?

Only inside the model. As reported, both ends of the 32 percent are the city's own estimates, not point-in-time counts, so the model is at least being compared with itself. What you cannot do is line it up against the point-in-time counts San Jose has cited for years and call it a continuation of that trend.

Here is why that matters. Mahan said the model estimated about 1,500 more unsheltered people than the county's count. A tool that starts higher has more room to fall. If you set August's 3,702 next to the January 2025 street count instead, the gap is 257 people, or about 6 percent.

To be clear, that 6 percent is not a valid comparison either. It mixes two methods, which is exactly the problem. The honest reading is that the city's model shows a large decline by its own measure, and nobody yet has a like-for-like, independent number to confirm or refute it.

Here is how the two approaches stack up side by side.

The model may well be the better tool. Point-in-time counts really are blunt. But a better tool still needs a published manual before a headline number built on it can carry this much weight.

What the critics say

The sharpest objection is simple: the people who work the streets every day do not see a one-third drop. Shaunn Cartwright of the Unhoused Response Group put the question directly to reporters, per ABC7: does it seem like there is a 32 percent decrease? "No." She added that deaths are on pace to match last year's 160, and questioned whether the new metric exists to give the mayor "a more favorable number."

Bea Nasser, a homeless advocate, argues the drop reflects people being moved out of view more than out of homelessness. She told the Spotlight that of 16 Columbus Park residents she helped move into a Motel 6 program, 12 ended up back outside, many removed over belongings limits. Her read: the city is driving "a numbers-first approach with people who have lives."

Sandy Perry, also of the Unhoused Response Group and a longtime Mahan critic, gave the mayor credit for vastly expanding shelter capacity, per CBS News Bay Area, but warned that without truly affordable housing the system is "somewhat stuck."

City leaders have acknowledged the algorithm is not perfect, KTVU reported, and there is an open question about which number state and federal funders will use. Mahan's defense is speed. The city wants a number "not once every two years, like the PIT count, but at least once every quarter in near real time," he said, per ABC7.

The part nobody is arguing about: San Jose built a lot of shelter

Whatever you think of the 32 percent, the shelter numbers are not in dispute. The county's 2025 count found 2,544 sheltered people in San Jose, up from 980 in 2019, and the city says it has added about 1,000 interim housing units. The 2025 count still found 6,503 homeless residents in total, up slightly from 2023.

That is the pattern I would watch. More people are inside, in tiny home communities and converted motels, but the total has not gone down. The bottleneck is permanent housing. About 6,000 households are waiting for it in Santa Clara County, per the Spotlight, and the city's own scorecard says San Jose is short more than 3,000 shelter and interim units. Interim beds buy time. Only more homes, of every price, end the wait. (We looked at the state side of that funding pipeline in our piece on California's new housing and homelessness agency.)

What it means if you are house hunting in San Jose

For buyers, the practical takeaway is that a citywide homeless number, old method or new, tells you very little about a specific block. What you experience is local: a creek corridor, a freeway underpass, a stretch of RVs near a business park.

San Jose is a big, uneven city. Our San Jose neighborhood rankings cover 112 neighborhoods, and the city scores 3.5 out of 5 on safety overall. On our San Jose safety rankings, The Villages in the Evergreen foothills tops the list at 4.8, while downtown-adjacent areas score lower. We walked through that spread in Is San Jose Safe? What 112 Neighborhood Scores Show.

Prices have not moved on this news, and I would not expect them to. The San Jose median sale price was about $1.63 million in August 2026 on our San Jose city page, inside the same $1.59 million to $1.68 million band it has held for a year. If you are weighing a particular home, walk the surrounding blocks at a few different times of day, check the nearest creek trail, and read the city's quarterly updates as they come out. That tells you more than any single headline figure.

This does not make any neighborhood better or worse overnight. It changes what is worth checking, and that is how we try to look at every area on Houseberry: the blocks first, then the house.

The real test comes in early 2027, when the next federally required point-in-time count happens. If the street survey also shows a sharp drop from 3,959, the city's model will have earned a lot of trust. If it does not, expect this argument to get louder.

Sources

• San José Spotlight: New methodology reveals drop in San Jose street homelessness (Sept 19, updated Sept 24, 2026)
• KTVU FOX 2: San Jose reports 32% drop in homeless population with new tracking model (Sept 16, 2026)
• ABC7 News: San Jose sees 32% drop in unsheltered residents, according to data (Sept 17, 2026)
• CBS News Bay Area: San Jose mayor says city's homeless population has dropped by 32% (Sept 17, 2026)
• San José Spotlight: San Jose homeless population is still rising (July 2025)
• Local News Matters: San Jose homeless population is still rising despite more temporary, permanent housing (July 18, 2025)
• Office of Mayor Matt Mahan: 2025 point-in-time count release (July 14, 2025)
• City of San José: How the City is Reducing Unsheltered Homelessness (FY2025-2026)
• Houseberry: San Jose city page, neighborhood rankings, and safety rankings (September 2026)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

If you have a kid in Richmond who plays softball, you already know the question. Where is the nearest field with real grass, a working bathroom, and a fence that keeps the ball out of the street?

As of this month, one good answer is Wendell Park. The city finished about $2.3 million in Richmond park improvements there, with a new softball field, a new restroom, a new irrigation system, fencing and playground upgrades, and celebrated the reopening on September 15, 2026. It is the latest of several park projects that have landed in the same handful of central and north Richmond flatland neighborhoods, the ones where Houseberry's public spaces scores run lowest and home prices sit below the city median.

The takeaways

The short list, before the details:

• Wendell Park, at 26th Street and Wendell Avenue in the North and East neighborhood, reopened in September 2026 after about $2.3 million in upgrades paid for largely with federal American Rescue Plan money.
• Local News Matters reports Richmond has put about $37 million into park and recreation improvements in recent years. The four city projects we could document add up to roughly $27 million of that.
• The biggest checks went to Shields-Reid Park ($11.9 million, reopened July 2026) and Boorman Park on the edge of the Iron Triangle ($12 million, still under construction).
• The three neighborhoods where the big projects landed score 2.7 to 2.9 out of 5 on Houseberry's Public Spaces measure and had August 2026 medians between about $397,000 and $588,000, below Richmond's $662,000 citywide median.

What Wendell Park got for $2.3 million

Wendell Park is a neighborhood park, not a destination one. It sits in North and East, the grid of mostly 1940s cottages and bungalows between 23rd Street and San Pablo Avenue, north of Macdonald Avenue and up to Bush and Moran avenues. People walk there from their front doors.

The work was practical. A restroom that runs on electricity, renovated ballfields including the new softball diamond, an irrigation system so the turf stays green past June, and new fencing. When Richmondside previewed the project in April, it pegged the ARPA grant at about $2 million. Local News Matters put the finished job at about $2.3 million.

That deadline matters, by the way. Cities had to commit their federal rescue money by the end of 2024 and spend it by the end of 2026, as Richmond staff told the council back in 2023. Wendell Park beat the clock by about three months.

Where the rest of Richmond's park money went

The recent spending is concentrated in the flatlands, and it follows a clear logic. Richmond has around 50 city parks, and a 2024 Richmondside review found the city had about 2.4 acres of parkland per 1,000 residents, below California's 3-acre benchmark. The state grants that paid for most of this work were written for exactly that gap.

Here is what we could verify, neighborhood by neighborhood:

• Shields-Reid Park, Shields-Reid. $11.9 million total, including an $8 million Prop 68 state grant. New play area, bike pump track, performance stage, fitness zone, sports fields, lighting and restrooms. It reopened with a festival on July 11, 2026.
• Boorman Park, Iron Triangle edge. $12 million from Prop 68 and a Caltrans Clean California grant, for a trail, skating area, athletic field, basketball court, play areas and restrooms, on a site a 2019 county assessment flagged as a high priority. It broke ground in January 2025 and was still under construction this spring.
• Wendell Park, North and East. About $2.3 million, reopened September 2026.
• Dirt World Bike Park, on the Richmond Greenway. A $1.14 million Prop 68 grant rebuilt the BMX track and its stretch of Greenway trail. The state lists it as complete.
• Nicholl Park, Macdonald Avenue at 33rd Street. The 21-acre park a few blocks east of Wendell is just entering planning, with no budget published yet.

Add up the first four and you get about $27.3 million. We could not find a public breakdown for the other $10 million or so in the $37 million figure, so treat that total as the city's number, not ours.

Now put those locations next to Houseberry's own scores. Our Public Spaces sub-score rates the parks, plazas and open space people can actually use near home. The neighborhoods that got the checks all sit well below the ones that did not need them.

Point Richmond scores 4.2 because it has Miller/Knox Regional Shoreline and Keller Beach minutes away. Carriage Hills North, up by Sobrante Ridge, scores 4.0. North and East comes in at 2.9, the Iron Triangle at 2.8 and Shields-Reid at 2.7. That spread is a real difference in daily life. It is the gap between walking to a shoreline trail after dinner and driving somewhere for a playground with a working bathroom.

What Richmond park improvements mean for home prices here

The honest answer is that a new restroom and a softball field do not move a neighborhood median on their own, and anyone telling you otherwise is selling something. What they change is the daily-life column, which is where these neighborhoods have always been weakest.

Prices tell you who is buying. North and East had a median sale price of about $587,800 in August 2026, on its Houseberry page, and it has bounced between roughly $550,000 and $636,000 over the past year. Shields-Reid was about $576,400 in August, up about 6.6 percent from September 2025. The Iron Triangle was about $397,400, down roughly 19 percent over 12 months, though a neighborhood that small can swing hard on a few sales.

Richmond as a whole sat at about $662,000 in July 2026, per our Richmond city page. Point Richmond was about $1.35 million in August. So the park money is going where a family can still buy a single-family house in the Bay Area for under $600,000. That is the right place for it. And it is a strong argument for letting more homes get built within a few blocks of these parks, so more people get to use them.

Is North and East a better place to live now? It still ranks 27th of 32 overall in our Richmond neighborhood rankings, with a safety score of 1.4. A park does not change that. But it already ranks 9th on amenities, helped by a 4.3 merchants score along San Pablo and Macdonald, and a reliable park is the piece that was missing.

The catch is who mows the grass

Building a park is the easy part in Richmond. Keeping it is the hard part. The 2024 Richmondside review counted about 12 workers maintaining 45 parks on a roughly $2.9 million operations budget, and Council member Doria Robinson put it bluntly at the time: the city has a history of building new things it cannot maintain.

That is the real risk for these projects. New irrigation and new turf are only an asset if someone fixes the broken sprinkler head in year three. If you are buying near Wendell, Shields-Reid or Boorman, walk the park on a weekday afternoon a year from now. That tells you more than the ribbon-cutting did.

The shoreline piece still coming

The next big opening is not a city park at all. The East Bay Regional Park District started building a 2.5-mile Bay Trail segment at Point Molate in October 2025, running from the Richmond-San Rafael Bridge past Point Molate Beach to the old Winehaven district. The district expects it to open in early 2027. Richmondside put the combined budget at about $13 million.

That one mostly helps Point Richmond, which already leads the city on public spaces. It is a good project. It is just not the same kind of project as a bathroom at Wendell Park, and it is worth keeping the two straight when someone tells you Richmond is getting more parks.

How we would use this if we were shopping in Richmond

Start with the park, then look at the block. The same way we would with a school, we would check which park is actually within a 10-minute walk, whether it has a restroom that is open, and how it looks at 5 p.m. on a Tuesday, not on reopening day.

For the flatland neighborhoods, 2026 was the year the park column got better. That does not erase the safety and school scores, and it should not. But if you have been comparing North and East, Shields-Reid or the Iron Triangle on price alone, the parks are now worth a line in the spreadsheet too. Our Richmond pages break every neighborhood down on those scores if you want to do the side-by-side yourself, and if you are weighing the East Bay shoreline more broadly, our piece on the Golden Gate Fields park plan in Albany and Berkeley covers the next big open-space change a few miles south.

Sources

• Local News Matters (Bay City News), Wendell Park reopens with new softball field, restroom, irrigation system, Sept. 22, 2026
• Local News Matters, Richmond to celebrate Shields-Reid Park's $11.9M makeover, July 10, 2026
• Richmondside, Richmond park upgrades include soccer field fixes and modern restrooms, April 7, 2026
• Richmondside, Richmond's Boorman Park is getting $12M renovation, Jan. 27, 2025
• Richmondside, Richmonders want more parks. City struggles to maintain the 50 it has, Sept. 17, 2024
• Parks for California, Shields Reid Park Revitalization (Prop 68 grant)
• Parks for California, Dirt World Bike Park (Prop 68 grant)
• East Bay Regional Park District, San Francisco Bay Trail: Point Molate
• Richmondside, New bike trail to improve access to Richmond's Point Molate beach, Aug. 5, 2025
• Contra Costa Pulse, Richmond continues to consider uses for Rescue Plan funds, May 17, 2023
• Houseberry, Richmond neighborhood rankings, Neighborhood pages and City page, retrieved September 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

574. That is how many pre-buyout disclosures San Francisco landlords filed with the Rent Board between January 1 and August 31, 2026, according to the city's own buyout dataset. Same eight months last year: 432. The year before: 322.

So San Francisco tenant buyouts, or at least the paperwork that has to come before one, are up 33 percent in a year and running at their highest level since 2019. We have not seen it reported anywhere else. We pulled the Rent Board data ourselves, and the pattern lines up almost exactly with the rent spike.

The numbers that matter

• 574 pre-buyout disclosure declarations were filed January to August 2026, up from 432 in 2025 (+33 percent) and 322 in 2024 (+78 percent).
• That is the most for those eight months since 2019, when landlords filed 588. The all-time high was 691 in 2018.
• Through September 11, 2026, the count reached 609, already more than all of 2025 (604).
• The Sunset/Parkside area led with 72 filings. The Inner Sunset, Pacific Heights and Inner Richmond more than doubled.
• A disclosure is not a completed buyout. It is the form a landlord must file before making an offer.

What a buyout disclosure is, and what it is not

A pre-buyout disclosure is a landlord's sworn notice that it has told a tenant their rights before offering money to move out. Under the Rent Ordinance's buyout section (37.9E), in effect since March 2015, a landlord has to hand the tenant a Rent Board form and file a declaration with the Board before buyout talks begin, as the city's buyout page lays out.

That makes the disclosure count a measure of intent, not outcomes. Some tenants say no. Some negotiate for months. Some offers never get made. If a deal is signed, the tenant gets 45 days to rescind it, and only then does the landlord file the final agreement.

So when I say buyouts are up 33 percent, read it precisely: the number of landlords formally preparing to offer one is up 33 percent. For a leading indicator, that is about as clean as housing data gets.

Why landlords are offering buyouts now

Because the gap between a rent-controlled rent and a market rent just got a lot wider. San Francisco's median rent hit $3,844 in September 2026, up 25.6 percent from a year earlier, per Apartment List, a pace Axios called the fastest of any major U.S. city this year. The allowable increase on a rent-controlled unit is 1.6 percent for March 2026 through February 2027.

Every month that gap persists, a sitting tenant is worth more to the landlord gone than staying. State law lets a vacated unit reset to market rent for the next tenant. And a building delivered empty sells for more than one with long-term tenants, because the buyer is pricing the rent roll.

A buyout is the legal, negotiated route to that vacancy. The alternatives, an owner move-in or an Ellis Act withdrawal from the rental market, come with relocation payments, restrictions, and a lot more friction. Mayor Lurie's September 10 rent emergency package would raise Ellis Act relocation payments by $3,000, which, if anything, makes a negotiated buyout look relatively cheaper.

The chart also shows how the cycle works. Filings peaked in 2017 and 2018, the last time rents ran hot, then fell by more than half to 322 by 2024 as the market cooled. They have now climbed back most of the way in two years.

Where San Francisco buyout disclosures are rising fastest

The west side leads, and the jumps are sharpest in neighborhoods with lots of small, older rental buildings. The Sunset/Parkside area, the city's planning district covering the Outer Sunset and Parkside grids west of 19th Avenue, filed 72 disclosures in the first eight months of 2026, up from 57.

The bigger story is in the percentage moves. The Inner Sunset, the blocks around 9th and Irving near UCSF, went from 14 to 34. The Inner Richmond went from 9 to 24. Pacific Heights more than doubled, 13 to 30, which tells you this is not only a low-rent-neighborhood phenomenon. Some of the city's most expensive flats are being cleared too.

Here is how those filings line up with what homes are selling for, using our own neighborhood price history.

Two things jump out. First, the buyout surge is not concentrated where homes are cheapest. The Outer Sunset sold at about $1.73 million in August 2026, below the $2.2 million citywide median, while the Inner Sunset ran $2.23 million and Noe Valley $3.29 million. Second, the west side neighborhoods leading in filings also score well on our numbers. The Outer Sunset ranks second of 92 San Francisco neighborhoods overall, with a 4.5 school score and 4.3 for safety. Strong fundamentals plus a rental surge equals an expensive tenant to keep in place.

A caveat on the match-up. The Rent Board's areas are larger than Houseberry's neighborhoods, so Sunset/Parkside covers more than the Outer Sunset alone. Read the prices as the price level of the area, not as comps for a specific fourplex.

Are more tenants actually taking buyouts?

Not yet, on paper, but the dollars are up sharply. Landlords filed 233 signed buyout agreements dated January to August 2026, against 243 for the same months of 2025. The median payment on those 2026 agreements was $50,000, up from $35,000.

The flat agreement count is mostly a timing artifact. A signed deal cannot be filed until the 45-day rescission window closes, and a disclosure filed in July might not become a signed agreement until winter. If the usual pattern holds, the agreement count should climb through early 2027. The payments are the real signal. A $50,000 median tells you landlords see a lot of value in an empty unit right now.

One limit on the amounts: each agreement can cover more than one tenant in a household, so the median is per agreement, not per person.

What this does and does not mean

It does mean more rent-controlled tenants in San Francisco will get a knock on the door in the coming months, especially on the west side. It also means more of the city's older small buildings will come to market empty and priced accordingly.

It does not mean mass displacement by force. A buyout is voluntary, the tenant can walk away from the offer, and the city's annual buyout reports track every one filed. It also does not prove any single landlord plans to sell. Some are clearing a unit for family, some for a remodel, some to re-rent at market.

For renters with a long tenancy, the practical point is that your below-market rent is now a real asset with a price. The SF Tenants Union and the Rent Board both have guides worth reading before you answer an offer. Buyouts also carry a string for owners: two or more buyouts, or one involving a senior or disabled long-term tenant, can block a condo conversion for 10 years.

For buyers, a vacant pre-1979 two-unit building in the Sunset or Richmond is going to cost more than the same building with tenants, and part of that premium is somebody's buyout check. If you are weighing whether to rent or buy in this market, or watching the banked rent increase cap moving through City Hall, count this in. My read: a buyout boom is what a shortage looks like from inside a rent-controlled building. The fix is still more homes, and until the city permits them, the cheapest apartment in San Francisco will keep being the one somebody already lives in. When we compare neighborhoods at Houseberry, we look past the listing to what the block is actually doing, and our San Francisco neighborhood rankings are the place to start.

Sources

DataSF, Buyout Agreements dataset (wmam-7g8d), San Francisco Rent Board, data as of September 23, 2026 (analysis by Houseberry)

SF.gov, Buyout Agreements, San Francisco Rent Board

SF.gov, Rent Board Annual Buyout Reports, 2016 to 2026

San Francisco Tenants Union, Buyouts

Apartment List, San Francisco, CA Rent Report, September 2026

SF.gov, Learn about rent increases in San Francisco (allowable increase 1.6 percent, March 2026 to February 2027)

Axios San Francisco, Lurie declares San Francisco rent emergency as costs soar, September 10, 2026

Houseberry, San Francisco city guide, neighborhood pages and rankings, retrieved September 23, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

If you walk the fire roads above Berkeley or Oakland in late summer, you have probably met the goats. A herd works its way across a dry slope behind a white mesh fence, and a week later the hillside looks like a fresh buzz cut.

Somebody decides which slope gets the goats first. Starting now, that decision gets a map. The California State Coastal Conservancy is paying the East Bay Regional Park District to build a ranked plan for where fire work should happen first across all 126,000 acres of its parkland in Alameda and Contra Costa counties. It is a plan, not a chainsaw. Nothing gets cut because of this piece of the grant alone.

In brief

Here is the short version for anyone who lives near a regional park.

• The state logged the project on September 22, 2026. It pays the park district to rank every acre it manages by how urgently it needs brush clearing, grazing or tree thinning.
• The Conservancy's staff recommended adding $100,000 to a $2,075,000 grant from September 2024, for a total of $2,175,000.
• The older part of that grant already pays for hands-on work in Tilden, Anthony Chabot and Carquinez Strait, including a roughly 34-acre fuel break at Lake Chabot.
• Houses next to park land in Berkeley, Oakland, Orinda and Castro Valley sold for monthly medians from about $1.02 million to $1.94 million in summer 2026.
• Work on park land does not earn you an insurance discount. Work on your own lot can.

What the East Bay Regional Park District wildfire map actually is

Think of it as a ranked to-do list drawn on a map. The park district has far more hillside than it has crews, so it needs a fair way to decide what comes first.

The Conservancy's September 17 staff report calls it a “vegetation management prioritization framework.” In everyday words, that means computer mapping software that layers what could burn, where the rare plants and animals are, where the trails and access roads run, and how weather moves through each canyon. The park district's request for bids says the ranking should balance wildfire risk, the health of the land and public access. Consultant bids were due March 10, 2026.

The result will be an online map and app that park staff use to plan every future fire project on their land. The report does not say whether the public will get to see it. That is worth asking about.

The grant also pays for an online workshop, so other local agencies and nonprofits can learn to build their own versions.

You may see the phrase “CEQA notice of exemption” on the state filing. CEQA is California's environmental review law. A notice of exemption is a short form saying a project is small or harmless enough to skip a full environmental study. Making a map and holding a video call does not touch a single tree, so it qualifies.

Who is paying, and how much

State money covers almost all of it, and the new piece is small. The Conservancy is a state agency, and the extra $100,000 comes from a $20 million wildfire block grant it receives from the California Department of Conservation.

Two notes on that table. The 2024 amount comes from the Conservancy's own announcement, which named Tilden, Anthony Chabot and Carquinez Strait. And the September 22 filing lists no dollar figure at all. The $100,000 comes from the staff report the Conservancy board took up in Fort Bragg on September 17, and the notice filed five days later says the Conservancy is approving the grant.

What fire work looks like on the ground

Most of it is surprisingly low-tech. Fire crews call anything that can burn “fuel.” That means dry grass, brush, fallen branches and the peeling bark that piles up under eucalyptus trees. “Fuel work” just means getting less of it near the places fire likes to travel.

According to the park district's fuels management page, the toolkit includes hand crews, mowing, weed-eating, chipping, thinning crowded stands of trees, removing dead trees and some prescribed burns. Cattle, sheep and goats graze about 86,800 acres of park land every year. The district says crews and machines treat more than 1,000 acres a year on top of that.

The bigger jobs look more dramatic. In September 2025 the district finished a 667-acre project at Anthony Chabot Regional Park that removed eucalyptus on a large scale and turned the debris into biochar. This year's Tilden projects run from May through December 2026 near Golf Course Drive, Shasta Road, Park Hills Road and Wildcat Canyon Road, with some trails and group camps closed while machines are working.

That gap is the whole reason for the map. Grazing covers a lot of grass cheaply. But the careful work, like thinning trees near a ridge of houses, happens on about 1 percent of the land each year. Picking the right 1 percent is the job.

There is history here, too. The October 1991 Oakland Hills firestorm killed 25 people and destroyed 3,469 homes and apartment units. Hillside residents who were here for it do not need a reminder. Newer neighbors can think of this map as the slow, unglamorous follow-through.

The hillside neighborhoods that share a fence line with the parks

The parks run along the back of some of the East Bay's most sought-after streets, and the prices show it. We pulled six neighborhoods on our East Bay neighborhood rankings that sit right against district land, and checked each one's last 12 months of sales.

Berkeley Hills, the upper Berkeley streets off Grizzly Peak Boulevard that back onto Tilden Regional Park, had a median sale price of about $1.80 million in August 2026. It scores 4.5 out of 5 on our Public Spaces measure. A little farther south, Claremont Hills sits below Claremont Canyon Regional Preserve at about $1.94 million, with a perfect 5.0 safety score.

In Oakland, Skyline-Hillcrest Estates, along Skyline Boulevard with Redwood Regional Park trailheads on its northern edge, was about $1.68 million in August. Leona Heights, where the Leona Canyon trail starts, was about $1.02 million in July 2026. Across the ridge in Orinda, El Toyonal has trails into Tilden and a 4.9 school score.

Castro Valley is the one to watch for this specific grant. Lake Chabot Regional Park, home of the new 34-acre fuel break, is its backyard. Its median rose from about $1.15 million in September 2025 to about $1.22 million in August 2026, roughly 6 percent in a year.

One caution about those numbers. These are small neighborhoods with few sales, so a single big house can swing a month. El Toyonal went from $1.54 million in May to $1.11 million in August. Look at the range, not one dot.

Will this lower my home insurance?

Not directly. California's Safer from Wildfires rules, which took effect in October 2022, require insurers to offer discounts for specific steps. Those steps are about your house, your yard and your block, not the park behind it.

The list covers things like a fire-rated roof, ember-resistant vents, clearing the first five feet around your house, and joining a community program such as Firewise USA. Insurers also have to tell you your wildfire risk score and let you appeal it. Fuel work on public park land is not one of the listed discounts.

That matters because costs are still climbing. The California FAIR Plan, the insurer of last resort, raises rates an average of 29.1 percent on October 15, 2026. Some lower-risk urban Bay Area areas may see decreases, while high-risk hillsides can see much bigger jumps. We walked through how that is spreading beyond the hills in our piece on Bay Area home insurance costs.

So the honest trade-off is this. A better-run park next door lowers the odds of a bad day, and that is worth a lot. It just will not show up as a line on your bill.

What a hillside homeowner can do this fall

The useful moves are small and mostly free.

• Look over your back fence. If you see dead trees or piled brush on park land near you, the district's fuels line is (510) 881-1833.
• Clear your first five feet. Move wood piles, dry plants and anything else that burns away from the walls. This is the zone insurers ask about.
• Ask your insurer two questions. What is my wildfire risk score, and which discounts am I missing?
• Talk to your neighbors. A Firewise USA group on your street counts toward discounts for everyone in it.
• Ask when the map will be public. The staff report gives no finish date, so a note to the park district now keeps the question alive.

If you are shopping for a hillside home, add one step to your tour. Check the address on CAL FIRE's hazard zone map, then get an insurance quote before you write an offer, not after. The park next door is a big part of why these streets cost what they do. We think about it the way we think about everything at Houseberry: the neighborhood first, then the house.

Sources

• CEQAnet, State Coastal Conservancy Notice of Exemption, SCH 2026090931 (Sept. 22, 2026)
• California State Coastal Conservancy, staff recommendation, EBRPD Wildfire Resilience augmentation (Sept. 17, 2026)
• California State Coastal Conservancy, September 2024 grant awards
• East Bay Regional Park District, RFP: Framework for Prioritization of Vegetation Management
• East Bay Regional Park District, Fuels Management
• East Bay Regional Park District, 667-acre fuels reduction project (Sept. 2025)
• East Bay Regional Park District, Tilden 2026 Major Fuels Reduction Projects
• East Bay Regional Park District, The Oakland Hills Firestorm
• California Department of Insurance, Safer from Wildfires
• California Department of Insurance, wildfire safety regulation release (Oct. 17, 2022)
• KQED, California FAIR Plan announces 29.1% rate hike (Aug. 2026)
• CAL FIRE Office of the State Fire Marshal, Fire Hazard Severity Zones
• Houseberry neighborhood pages and East Bay rankings, September 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Sixteen listings. Twelve street addresses. Four buildings whose operators will still put their name on it in September 2026. That is San Francisco's data center count once you do it by hand, which, it turns out, nobody at City Hall has.

The San Francisco Standard reported on September 22 that the Planning Department does not track how many data centers the city has, and cannot say how many permits it has approved for them. So we built the list ourselves from operator websites, the commercial directories, and the one environmental filing on record. Every address we found sits in the eastern third of the city, and every neighborhood holding one scores below the San Francisco average on our data.

Key takeaways

The short answer to how many data centers San Francisco has, as of September 23, 2026.

• Published counts run from 5 to 16 depending on who is counting. The Planning Department's answer is that it does not track them.
• The directory listings collapse to 12 distinct street addresses, and only 4 buildings are confirmed on their operators' own websites today.
• The biggest are two Bayview-edge campuses across Paul Avenue from each other: Digital Realty's 495,000-square-foot SFO10 and Novva's campus, with 9 megawatts live and 36 planned.
• All 12 addresses sit in or beside South Beach, South of Market, Mission Bay, Bayview, Silver Terrace or Portola. None is in any of the 17 San Francisco neighborhoods scoring 4.0 or higher on Houseberry.
• San Francisco, Hayward and Oakland are all moving toward moratoriums. None of them started from a public inventory.

Why City Hall cannot tell you the number

The city cannot count data centers because its code has no category called data center. The closest thing is an internet services exchange, a use definition written in 2001 and approved case by case as a conditional use. A server hall gets filed under that label, or as an accessory to an office, or as a telecom building that has been one since the 1980s. There is no box to tick and so no list to pull.

"We do not track the number of data centers in San Francisco," Planning spokesperson Patrick Hannan told the Standard. Supervisor Shamann Walton's version was blunter: "We cannot continue to use rules from 2001 to define what data centers are."

The paper trail does exist in places, if you know the old name. The clearest one I found is a 2014 mitigated negative declaration for 320-400 Paul Avenue, filed on CEQAnet as an Internet Services Exchange. It describes a 243,000-square-foot facility on 7.3 acres of PDR-2 land bounded by Bayshore Boulevard, San Bruno Avenue and Paul Avenue, fed by two new underground 12-kilovolt circuits carrying 24 megawatts from PG&E's Martin substation. That site is Novva's campus today. It was on the public record twelve years ago, just not under a name anyone searches for.

We counted San Francisco's data centers address by address

The honest count depends on what you are counting, and the published numbers are counting different things. KQED put the figure at 12 data centers. Mission Local said 15 regional campuses. The Standard said at least six known sites with a dozen operators. The directories disagree with each other too: Compute Atlas lists 5, Datacentermap lists 10, Kadoa tracks 12, and Inflect shows 16 operator listings.

Here is how we got from 16 to 4. Many of the Inflect listings are tenants sharing one building. Four separate operators list 200 Paul Avenue alone, and two list 365 Main Street. Stripping the duplicates leaves 12 street addresses. Then we asked a harder question of each one: does the company running it still advertise it on its own website this month? Four clear that bar.

• 200 Paul Avenue. Digital Realty's SFO10, 495,000 square feet over five levels, the city's main carrier interconnection hub.
• 400 Paul Avenue. Novva's San Francisco campus, 7.5 acres with 182,000 square feet of data halls and a 36-megawatt design. Nine megawatts came online this summer, with the rest planned through 2028.
• 365 Main Street. Digital Realty's SFO12, 227,000 square feet on six levels in a building the company says was built as a military tank assembly plant.
• 274 Brannan Street. Fortress SF1, inside the 1923 Hawley Terminal Building, a telecom switching site since 1982 and a data center since 2015.

The other eight (185 Berry, 630 Third, 360 Spear, 60 Federal, 501 Second, 651 Brannan and 650 Townsend Streets, plus 375 Newhall Street) show up only in directories. Some are carrier rooms inside telecom buildings. 650 Townsend is a big office building whose server floor carries old brand names. Unconfirmed is not the same as gone. It means nobody is publicly vouching for it.

Where San Francisco's data centers actually sit

Every address in our count lands in one of three clusters, and all three are on the east side of the city. We matched each address to the Houseberry neighborhood it sits in or borders. The pattern is not subtle.

Two things jump out. Half the addresses are in the South Beach and Rincon Hill cluster, where the old warehouse and telecom buildings near the Bay Bridge now share blocks with high-rise condo towers. And the heaviest power users, the Paul Avenue pair, sit on the industrial seam where Bayview, Silver Terrace and Portola meet at the 101 interchange.

Be careful with what that chart does and does not say. Data centers did not lower these scores. A century of industrial zoning did, and data centers followed the same zoning, the same fiber routes along the old rail easements, and the same heavy power lines. What the chart does show is who absorbs the rooftop chillers and the diesel generator tests. It is never Presidio Heights or the Outer Sunset, and on the 92-neighborhood San Francisco ranking, none of the 17 neighborhoods scoring 4.0 or better holds a single address on our list.

A moratorium without an inventory is a pause on something you cannot see

Three Bay Area cities are reaching for the pause button at once. Walton introduced San Francisco's 45-day moratorium at the Board of Supervisors on September 15 with Connie Chan, Jackie Fielder and Chyanne Chen, extendable to 22 months and 15 days. Hayward's council voted 7-0 on September 15 to have staff start a moratorium, returning in October, after the 310,000-square-foot Stack project at 26203 Production Avenue was approved in May 2025. And Oakland's version, which we walked through when Carroll Fife proposed it, cleared a committee on September 23, the Oaklandside reported.

I get the impulse. But a pause does nothing about the buildings already running, and in San Francisco nobody can say with confidence how many that is. The fix is cheap and boring: define a data center by its electrical load, require a public registry of every site above that line, and put the megawatts, the backup generator count and the cooling type on it. San Jose, Santa Clara and Milpitas already have data center definitions in their codes. San Francisco is using a label from the dial-up era.

And from where I sit, the registry is also a housing tool. The big fights ahead are about land, the four acres at 1215 Michigan Street that we wrote about in our Dogpatch zoning piece being the obvious one. A city that knows exactly how much industrial land is already committed to server halls can make a much better argument about which parcels should become homes instead.

Questions people are asking

Is there a data center on my block?

If you live in the Richmond, the Sunset, the Marina, Noe Valley or anywhere else on the west side, none of the 12 addresses is near you. On the east side, check the address list above, then look up nearby parcels on the city's Property Information Map, where permits and conditional use records attach to each lot. Rows of boxy rooftop units on a windowless building are the giveaway.

Do data centers use a lot of power in San Francisco?

By city standards, yes. The 400 Paul Avenue campus was designed around 24 megawatts of new PG&E service in 2014 and now advertises 36. The proposed Dogpatch site was listed at 49.5 megawatts, which would make it the largest in the city.

Can San Francisco ban new data centers?

It can pause them. California's interim ordinance law caps a moratorium at about two years, which is why every Bay Area proposal pairs the pause with a promise of permanent rules. What it cannot easily do is undo a use already permitted.

What happens next

The next markers are close together. San Francisco's moratorium still needs a four-fifths vote of the Board, and Hayward's staff report is due back in October. Watch whether either one asks for the inventory first. If you are shopping on the east side of the city, pull up the neighborhood on our South Beach page or any other, then walk the block. Knowing what hums behind the blank wall down the street is exactly the kind of neighborhood-first homework we built Houseberry for.

Sources

The San Francisco Standard: The AI capital of the world doesn't know how many data centers it has, September 22, 2026

KQED: San Francisco supervisors push for moratorium on new data centers, September 2026

Mission Local: Amid rising fear over data centers, an S.F. supervisor quietly plans a moratorium, September 2026

CEQAnet: 320-400 Paul Avenue Internet Services Exchange Facility and Extension of 12kV Electrical Circuits, June 10, 2014

Digital Realty: SFO10, 200 Paul Avenue

Digital Realty: SFO12, 365 Main Street

Novva Data Centers: San Francisco facility

Fortress Data Centers: SF1, 274 Brannan Street

Inflect: data centers in San Francisco

Datacentermap: San Francisco data centers

Compute Atlas: data centers in San Francisco County

Kadoa: San Francisco data center sites

Local News Matters: Hayward City Council to consider data center moratorium in October, September 18, 2026

The Oaklandside: Data center moratorium clears key hurdle in Oakland, September 23, 2026

San Francisco Planning: Property Information Map

Houseberry: San Francisco neighborhood rankings, September 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The lender on the troubled Fisherman's Wharf office conversion is the school that sold it. Stockton Development LLC, the entity tied to restaurateur and investor Tony Garnicki that bought 2300 Stockton Street from the Academy of Art University in May, has been served a notice of default on a $3.4 million seller-financed loan, according to The Real Deal, citing the San Francisco Business Times and county recorder filings. The balance had grown to about $3.6 million by the time of the notice.

The short answer on what it means: nothing is lost yet. A default notice starts a statutory clock of at least three months and 20 days before any foreclosure sale, and Garnicki says he is refinancing. But this building is one of only four office-to-housing conversions San Francisco has seen filed in 2026, and the only one outside downtown. So it is worth reading the paperwork closely.

Key takeaways

• A notice of default is step one of a California nonjudicial foreclosure. Under Civil Code 2924, the earliest a trustee sale can happen is three months and 20 days after the notice is recorded, which puts any sale in 2027 at the soonest.
• The Academy of Art financed about 19 percent of the $18.3 million price itself. The loan has matured, so curing it realistically means a payoff or an extension, not a catch-up payment.
• This parcel has a past with City Hall. The city's 2016 settlement with the Academy, valued at $60 million, forced the school to withdraw its bid to legalize its use of 2340 Stockton Street, the same building.
• San Francisco finished no office-to-housing conversions in the four years before February 2026. Four filings this year add up to 322 homes, and the city's richest conversion subsidy does not reach the Wharf.

What a notice of default actually starts

A notice of default is a recorded warning with a legal timer on it, not a foreclosure. Most California commercial loans are secured by a deed of trust, which lets the lender foreclose without going to court. The recorded notice is the first public step.

The clock, as the statute sets it:

• Notice of default recorded. Reported this month from filings with the San Francisco Assessor-Recorder.
• Three months, minimum. Civil Code 2924 bars a notice of sale until at least three months after the default is recorded.
• Twenty more days. The sale date itself can be no earlier than three months and 20 days after the default notice, which lands in early 2027 at the soonest.
• Cure window. Civil Code 2924c lets a borrower reinstate up to five business days before the sale.

Here is the wrinkle most coverage skips. This loan has already matured, per Garnicki's comments to the Business Times. On a matured note, the whole balance is due, so “reinstating” is really a payoff. Garnicki said the company is in refinancing talks with the seller and aims to pay the debt off in the coming weeks. If that happens, expect a rescission or reconveyance to show up in the same recorder's index.

I would also bet on patience from the lender. The Academy is trying to sell its buildings, not take them back. Foreclosing on a vacant office it just unloaded would put it right back where it started, holding a building it does not want.

The Stockton Street parcel the Academy promised to give up

2300 Stockton is a three-story, 1970 office building on the block between Beach and North Point Streets, half a block from Pier 39. The county's 2025 tax roll lists it as Block 0018, Lot 004, with 43,366 square feet of building on a 37,812-square-foot lot (0.87 acres) zoned C-2, under the address 2340 Stockton. The Academy used it for administrative and academic space.

That address shows up in one of the more colorful chapters of San Francisco land-use history. In December 2016, then-City Attorney Dennis Herrera announced a settlement valued at $60 million with the Academy, which by then controlled about 40 properties. The city said at least 33 had zoning, permit or code violations, and that 160 rent-controlled and affordable units had been turned into student housing. Among the terms: the Academy had to withdraw its requests to legalize use at a set of outlying properties, and 2340 Stockton Street is named on that list.

In other words, the city told the Academy a decade ago this building did not fit its campus. Now it is on track to become housing. That is a better ending than most land-use fights get.

The sale is part of a broader exit. In October 2025 the school put 11 properties totaling about 375,000 square feet on the market for roughly $130 million. It has since sold the church at 2151 Van Ness Avenue for $4.7 million and 2300 Stockton, which went for $18.3 million, about 12 percent over its $16.4 million ask. Undergraduate enrollment is around 4,200, SFist reported, with many students taking classes remotely.

The math on an $18.3 million conversion

The price looked strong in May. It looks thinner with a default attached. The county's 2025 roll assessed the parcel at about $11.9 million, $7.47 million for land and $4.43 million for the building. The sale cleared that by 54 percent.

A few numbers worth sitting with:

• About $422 per square foot for the existing building, before a dollar of construction.
• About $261,000 per planned home for the building alone, across 70 units.
• Roughly $75,000 to $80,000 a year in added property tax once the parcel is reassessed at the sale price, our estimate at San Francisco's roughly 1.2 percent rate, on a building earning no rent while it is vacant.

For context, the developers of 150 Hayes told Axios that conversions can deliver apartments for $500,000 to $600,000 a unit, against roughly $1 million for new construction. At 2300 Stockton, nearly half of that budget is spent at closing. The plan to add two floors, taking floor area from about 62,300 to 90,160 square feet by the filing's own count per the preliminary filing, is what makes the unit count work, and new floors are construction, not remodel.

Where San Francisco's conversion incentives stop

San Francisco's conversion toolkit mostly covers downtown, and 2300 Stockton sits at the edge of it. Here is how the pieces line up for this site.

The Downtown Adaptive Reuse Program applies in C-3 districts and in C-2 districts east of Franklin and 13th Streets and north of Townsend. The parcel is C-2 and well east of Franklin, so on paper it qualifies for the program's code relief and the March 2025 inclusionary and impact fee waivers. We have not seen a project file confirming which path the developer is using.

Proposition C, passed in March 2024 with 52.8 percent of the vote, waives the transfer tax on the first sale after a commercial-to-residential conversion. It is capped at 5 million square feet and requires approval before January 1, 2030.

The biggest incentive is the Downtown Revitalization Financing District, signed this February. It hands back new property tax growth for up to 30 years, worth 10 to 20 percent of development costs by the Chronicle's accounting. It covers the Financial District, Union Square, SoMa and the Market Street corridor. The Wharf is not in it.

That gap matters. The one incentive that puts real cash into the capital stack stops short of this project. Meanwhile the Family Zoning Plan, the upzoning whose new height limits took effect January 31, 2026, supplies the extra height that makes the two added floors legal. The city has given this site permission. It has not given it money.

Four filings, 322 homes, zero finished

San Francisco's office-to-housing pipeline is real but tiny. The Chronicle reported in February that the city had no completed office conversions in four years, despite about 30 million square feet of vacant office space. This is what has been filed since.

Add it up and you get 322 homes in four buildings. The city estimates about 50 downtown buildings are good candidates, worth roughly 4,400 homes, per the mayor's office. The SF Standard points to an earlier conversion, Emerald Fund's 418 apartments at 100 Van Ness as the model, and notes the Humboldt Building conversion died last year when financing fell through.

We covered two of the others when they filed, the 136 homes planned at 901 Market and the 12-unit conversion at 575 Sutter near Union Square. 2300 Stockton is the first of the new wave to hit a financing wall in public. One default does not sink a program. But the Humboldt Building shows how quickly a stalled conversion can become a dead one.

Why the Wharf needs residents more than downtown does

Fisherman's Wharf is a district built for visitors, and 70 households would change its math more than 136 would on Market Street. Ripley's Believe It or Not, open on Jefferson Street since 1967, will close before the end of 2026. The Port's Fisherman's Wharf Forward plan is building a public plaza on Taylor Street at the old Alioto's site and an overlook on Al Scoma Way. Those help tourists. Year-round neighbors are what fill a grocery aisle on a foggy Tuesday in February.

The neighborhoods uphill are solid, not flashy. On our San Francisco rankings, Russian Hill, the cable car slopes west of Columbus Avenue, scores 3.6 out of 5 and ranks 44th of 92, with a 4.3 safety score. North Beach, the Washington Square and Columbus Avenue core just south of the site, scores 3.5 and ranks 48th. Telegraph Hill, around Coit Tower to the east, scores 3.6 with a 4.2 for amenities. The North Waterfront trails at 3.3, 60th of 92.

On price, Russian Hill's median was about $1.84 million in August 2026 on our price history, below the $2.2 million San Francisco citywide median. New rentals a few blocks down the hill would give people another way into that part of town without competing for a Russian Hill flat.

The honest downside: a Wharf apartment is a harder product than a Union Square one. Streets are crowded with visitors in summer and quiet in winter, daily retail is thin, and the nearest cable car turnaround at Taylor and Bay is a tourist line first. If rents here come in soft, the $261,000-a-home basis gets harder to carry. That is a real risk to the project, not a reason to skip it.

The next filings that decide 2300 Stockton

Three documents will tell us where this goes, and all of them are public. First, a rescission or reconveyance at the recorder would mean the seller loan was paid or restructured. Second, a formal project application at Planning would show whether the developer is using the adaptive reuse program and its fee waivers. Third, if neither shows up, a notice of trustee's sale could be recorded no earlier than about three months after the default notice.

My hope is the boring outcome. The Academy extends the note, the refinance closes, and 70 homes get built a few blocks from the Powell-Mason turnaround. We will update this post when any of those filings land. And if you are weighing the blocks between North Beach and the water in the meantime, our San Francisco neighborhood rankings are a good place to start comparing them.

Sources

The Real Deal, “Tony Garnicki affiliate hit with default notice for Fisherman’s Wharf offices,” September 23, 2026 (citing the San Francisco Business Times)

The Real Deal, “Academy of Art University continues SF real estate slimdown with $18M trade,” May 12, 2026

The Real Deal, Academy of Art $130 million portfolio listing, October 22, 2025, and 2151 Van Ness church sale, April 28, 2026

The Registry, Academy of Art sells 2300 Stockton for $18.3 million, May 2026

SF YIMBY, “Office-to-Housing Conversion Proposed For 2300 Stockton Street,” December 2025

San Francisco City Attorney, “Herrera secures landmark settlement valued at $60 million from Academy of Art University,” December 19, 2016

SFist, Academy of Art portfolio sale, October 21, 2025

San Francisco Assessor-Recorder, 2025 secured property tax roll, Block 0018 Lot 004, via DataSF

California Civil Code sections 2924 and 2924c

SF Planning, Downtown Adaptive Reuse Program

Ballotpedia, San Francisco Proposition C (March 2024)

SF.gov, Mayor Lurie signs Downtown Revitalization Financing District legislation, February 2026

Planetizen summarizing the San Francisco Chronicle, “Zero office conversions in four years,” February 2026

SF Standard, 150 Hayes office conversion, August 10, 2026

Axios San Francisco, office conversion incentives, July 13, 2026

CBS News Bay Area, Ripley’s Believe It or Not closure, September 2026

Port of San Francisco, Fisherman’s Wharf Forward concepts, September 3, 2025

Houseberry, San Francisco neighborhood rankings, neighborhood pages and city price history, retrieved September 23, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The lion picked a crawl space. On the morning of Friday, September 18, a young male mountain lion wedged himself under a house on 44th Avenue, on the block between Ulloa and Vicente streets, a few blocks up from the San Francisco Zoo, while police on bullhorns sent neighbors back inside. Animal Care and Control darted him. He bolted into the bushes. A second dart about half an hour later ended it.

By afternoon he wore a GPS collar and was back in the Santa Cruz Mountains. A mountain lion in San Francisco is still rare, and nobody was hurt. But it was the second sighting in two weeks, which makes it fair to ask what living on the city's wild edge involves.

Quick take

• A 2-year-old male mountain lion of roughly 78 pounds was tranquilized near 44th Avenue and Ulloa Street in the Outer Sunset on September 18, 2026, then collared as 187M and released in the Santa Cruz Mountains.
• A lion spent September 5, 2026 in a Panhandle tree at Fell and Baker streets. It may have been 187M, but the state says that cannot be proven.
• California has fewer than 50 verified mountain lion attacks on people since 1890. Coyotes, about 100 of them citywide, are the wildlife San Francisco pet owners actually plan around.
• Outer Sunset's median sale price was about $1.73 million in August 2026, well under the $2.2 million San Francisco citywide median that month.

How does a mountain lion end up on 44th Avenue?

It walks there, almost always as a young male looking for a home range of his own. Male lions leave their mothers around age two and can roam across 200 square miles looking for territory, according to Bay Area Puma Project founder Zara McDonald.

The Santa Cruz Mountains are boxed in by freeways and housing on every side. Chris Wilmers, who leads the UC Santa Cruz Santa Cruz Puma Project, said after the January capture that dispersing males wander and end up in places like San Francisco. The northbound ones tend to come up along the coast from the hills south of the city.

Nobody tracked 187M's exact route, so treat this as the likely path, not a documented one: up the Peninsula through the Crystal Springs watershed and Sweeney Ridge, past San Bruno Mountain and the Daly City bluffs, into Fort Funston, Lake Merced and the brush around the Zoo. From there, 44th and Ulloa is a short trot north.

San Francisco's mountain lion history is short, and it keeps ending the same way

Here is the baseline. In July 2015 the Presidio Trust logged sightings in Sea Cliff, the western Presidio and near Lake Merced. On November 10, 2017, a female was tranquilized near Diamond Boulevard and Duncan Street in Diamond Heights. In May 2021 a collared lion walked through Bernal Heights and left on its own.

Then 2026 happened. In January, 157M, a young male the Puma Project had collared as a kitten, was captured in Pacific Heights. The Panhandle cat slipped away overnight. And now 187M, whom the project collared and released the same day.

So roughly one episode every couple of years, then three in nine months. That is a real uptick, and it fits a boxed-in population pushing young males outward. Every one ended with the cat leaving or being moved, and none hurt a person.

The law shapes that outcome too. Mountain lions have been a specially protected mammal since Proposition 117 in 1990, and on February 12, 2026, the Fish and Game Commission voted to list the Central Coast and Southern California population as threatened under the state Endangered Species Act. That population includes the Santa Cruz Mountains. Dart, collar and release is the default for a reason.

The San Francisco neighborhoods on the wild edge

Nine San Francisco neighborhoods back onto the big, brushy open space a wandering lion or a resident coyote uses, and they are not all pricey. Here they are on our amenities score, which folds in parks, and current median price.

Outer Sunset, the fog-belt grid of stucco row houses running from Golden Gate Park south to roughly Taraval, scores 3.9 for amenities and 4.2 overall. Its median of about $1.73 million in August 2026 sits roughly 21 percent under the city. The capture block itself falls south of Taraval, in what realtor maps generally call Outer Parkside, where the median was about $1.74 million in July 2026.

The honest trade-off: Outer Parkside ranks 91st of 92 San Francisco neighborhoods for sunshine on our climate data. The ocean edge is cheaper partly because it is gray. Lake Shore, on Lake Merced, costs more ($2.22 million in August) with the list's lowest amenities score, 2.8. Diamond Heights and Presidio Heights medians swing hard on a handful of sales, so read those two as rough.

If you see one

Stand your ground and look big. That is the core of the California Department of Fish and Wildlife guidance:

• Do not run, and do not turn your back.
• Pick up small children right away. Keep dogs leashed and close.
• Face the lion, raise your arms, and make noise. Do not crouch or bend over.
• Never approach it, and leave it an escape route.
• Call San Francisco Animal Care and Control at 415-554-9400, or 911 if someone is in danger.
• If it ever attacks, fight back. Do not play dead.

For scale, CDFW counts fewer than 50 verified attacks on people in California since 1890, six of them fatal, and says you are a thousand times more likely to be struck by lightning.

Coyotes are the wild neighbor you will actually meet

If you have a cat or a small dog near these parks, plan around coyotes, not lions. San Francisco Animal Care and Control has estimated about 100 coyotes in the city, back since about 2002, with more than half of reported sightings in Golden Gate Park. In one sample, 35 percent of coyotes tested had eaten cats.

The fixes are boring and they work. Cats indoors. Dogs leashed in Golden Gate Park, Glen Canyon and McLaren, especially at dawn and dusk. If a coyote gets bold, yell and wave until it moves off.

How to think about the wild edge before you buy

Treat wildlife like fog: a real feature of the address, not a dealbreaker. The same brush that hid 187M for a morning is why Glen Canyon feels like a canyon.

On a tour, check whether the back fence opens onto a greenbelt and whether dense ivy crowds the deck. CDFW suggests clearing thick vegetation near the house and adding outdoor lighting.

The bigger fix is regional. A 2023 wildlife undercrossing on Highway 17 cut collisions there to zero, and connected habitat gives young males somewhere to go besides a crawl space in the Sunset. For the other big Outer Sunset question, see our look at Prop G and the Great Highway. To weigh open space against everything else, our San Francisco amenities ranking and the full San Francisco neighborhood rankings are the place to start.

Sources

ABC7 San Francisco, “Mountain lion captured in dramatic standoff after roaming San Francisco's Outer Sunset neighborhood,” September 18, 2026

KTVU FOX 2, “Mountain lion tranquilized, captured in SF's Outer Sunset,” September 18, 2026

NBC Bay Area, “Young mountain lion captured in San Francisco's Outer Sunset,” September 18, 2026

CBS San Francisco, “Video shows mountain lion in San Francisco's Outer Sunset District bolting away before being captured,” September 18, 2026

Local News Matters, “'Exploring' mountain lion captured in San Francisco's Outer Sunset, released into wild,” September 18, 2026

San Francisco Chronicle, “San Francisco mountain lion captured in Sunset, released back into wild,” September 2026

The San Francisco Standard, Outer Sunset capture (September 18, 2026), Panhandle sighting (September 6, 2026) and Santa Cruz Puma Project follow-up (September 22, 2026)

ABC7 San Francisco, 157M Pacific Heights capture, January 2026, and Smithsonian Magazine and Lookout Santa Cruz coverage of the same capture

Presidio Trust, “Mountain Lion Sightings Reported in San Francisco and the Presidio,” July 7, 2015

Hoodline, Diamond Heights mountain lion capture, November 2017, and NBC Bay Area, Bernal Heights sighting, May 2021

California Department of Fish and Wildlife, Fish and Game Commission CESA decision, February 2026, and Mountain Lions human-wildlife conflict guidance

Allen Matkins, “Mountain Lion Listed as Threatened Species in Central and Southern California,” 2026

Richmond Review/Sunset Beacon, “Learning to Coexist With at Least 100 Coyotes in San Francisco,” February 3, 2023

Houseberry, San Francisco city guide, neighborhood pages and rankings, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Item 8a on Thursday's San Francisco Planning Commission agenda is, for now, a parking lot. It sits at 334 Ellis Street, a few doors west of Taylor, and the September 24 agenda asks the Commission to let GLIDE replace it with a seven-story building for social services. No homes. Staff recommend approval with conditions.

Stand at Ellis and Taylor today and the church comes first: the 1931 Glide Memorial Church on the corner, GLIDE's six-story office building next door, and then a gap in the street wall where cars park. This is the Tenderloin, the dense grid of apartment buildings and residential hotels wedged between Union Square and Civic Center, and that gap is about to close.

The short version

• GLIDE is asking for a conditional use authorization and a variance for a building seven stories over a basement, 95 feet 5 inches tall, with about 41,173 square feet of social service space at 334 Ellis Street. The hearing is Thursday, September 24, 2026.
• The building replaces an existing surface parking lot. The agenda lists no housing units.
• Three lots merge into one 17,636-square-foot corner parcel holding three buildings and about 96,247 square feet in all.
• The plan is roughly 70 feet shorter than GLIDE's 2022 proposal, keeps the historic 330 Ellis building, and was found exempt from CEQA.
• Houseberry scores the Tenderloin 2.1 out of 5, 89th of 92 San Francisco neighborhoods, with a 0.5 for public spaces.

What GLIDE is asking for at 334 Ellis Street

The agenda packs the project into one dense paragraph. Unpacked:

• Parcels: 302 and 330 to 334 Ellis Street at the northwest corner of Taylor, Block 0324, District 5.
• New building: seven stories over a basement plus a mechanical penthouse, 95 feet 5 inches tall, on Lot 005 at 334 Ellis.
• Use: about 41,173 square feet of Social Service and Philanthropic Facility use. The design leaves room for future upper-floor connections to 330 Ellis to pull GLIDE's programs under one roof, per SF YIMBY's September 2025 report.
• Lot merger: one 17,636-square-foot corner lot, three buildings, about 96,247 square feet combined. By my math the new building is roughly 43 percent of that.
• Zoning: RC-4 high-density residential-commercial in the North of Market Residential special use district, with an 80-T-130-T height limit.
• Approvals sought: a conditional use authorization under Planning Code sections 209.3, 249.5 and 303, plus variances on ground-floor ceiling height, bicycle parking, and showers and lockers.
• Design: terracotta tile and sand-toned stucco, per SF YIMBY.

The more telling comparison is what the plan used to be.

Sources for the 2022 column: SF YIMBY, March 2022. Demolishing a historic building meant an environmental impact report. Building on the parking lot skipped that fight, the smartest move in the file.

Why does a services building need a special hearing in the Tenderloin?

Because the Tenderloin's zoning was written to protect housing, and anything else has to ask.

The site sits in the North of Market Residential Special Use District. The neighborhood's rules trace back to a 1985 rezoning that, as KQED has reported, capped new buildings at about 13 stories and put limits on non-residential uses above the second floor. So a building that is social services from the basement up needs a conditional use hearing.

The variances are smaller than they sound: ground-floor ceiling height, bike parking, and showers and lockers.

Could this corner hold housing instead?

Physically, yes. RC-4 is the city's high-density residential-commercial zone, and the height district runs from 80 to 130 feet. A couple of blocks away, 550 O'Farrell Street filed updated plans in August 2026 for a 20-story, 158-home tower on a lot of roughly 0.3 acres, using state density bonus law. GLIDE's merged parcel is about 0.4 acres, though the church and the 330 Ellis building already stand on much of it.

So the honest trade-off is this: a high-density lot two blocks from the Powell Street cable cars is getting 95 feet of offices, program rooms and service space, and zero homes, in a city where rental vacancy fell to 2.2 percent this summer (our rent and vacancy breakdown has the numbers). As a pro-housing writer, I would have loved to see a few floors of supportive housing stacked on top.

But GLIDE owns this land, and its job is services. Some 25 to 29 percent of the Tenderloin's homes are permanently affordable, per KQED's 2018 reporting, and many are residential hotel rooms without full kitchens. Meals, clinics, childcare and case management are part of what makes that housing livable. A services building is infrastructure for the homes already here. I can hold both thoughts.

What else is changing in the Tenderloin in 2026

On June 14, Sunday Streets closed Ellis Street between Mason and Larkin, straight past GLIDE's door, despite spring funding cuts. Private donors matched gifts up to $50,000, Mission Local reported, and a children's safe zone of more than a dozen blocks is in planning with the district attorney and the mayor's office.

GLIDE's own year has included a big fundraising night. On May 15, 2026, an anonymous bidder paid $9 million for lunch with Warren Buffett and Stephen and Ayesha Curry, with proceeds split between GLIDE and the Curry family's Eat. Learn. Play. Foundation. SFist tallied about $53.2 million for GLIDE from Buffett's lunches since 2000.

GLIDE is also still in its first years without the man who defined it. Cecil Williams, Glide's pastor from 1963, died in April 2024. He was born on September 22, 1929, the same year Lizzie Glide founded the church. Rev. Dr. Gina Fromer, a former GLIDE client, has led the organization as CEO since October 2023.

How the Tenderloin scores next to Nob Hill, Downtown and SoMa

On our data, the Tenderloin has everything within walking distance except a decent place to sit down.

Our Tenderloin neighborhood page scores it 2.1 out of 5 overall in September 2026, 89th of 92 San Francisco neighborhoods, with safety at 2.1, schools at 2.2 and amenities at 1.6. The merchant score is a perfect 5.0. Public spaces score 0.5. Curb appeal scores 1.0.

Compare that with Nob Hill, uphill to the north, which scores 3.3 overall with a 4.0 for public spaces, or Downtown at 2.5 overall and 3.0 on public space. South of Market lands at 2.1 overall, the same as the Tenderloin, but still scores 3.0 on public spaces.

Our Tenderloin page also shows a citywide price rather than a neighborhood median, most likely because so little housing here sells as individually owned homes. A services building will not move a sale price. It will show up on the sidewalk, in whether Ellis Street reads as a block instead of a parking lot.

What happens after the September 24 hearing?

The Commission meets at noon Thursday in Room 416 at City Hall, with 334 Ellis on the regular calendar, so expect a presentation and public comment.

If the Commission grants the conditional use, anyone can appeal it to the Board of Supervisors within 30 calendar days, according to the agenda. The variances are the Zoning Administrator's call, heard at the same time, and those decisions go to the Board of Appeals if challenged. After that come building permits. No public filing we found gives a construction start date.

My bet is a smooth approval. A reduced plan, a parking lot for a site, a CEQA exemption and a staff recommendation to approve is about as clean as San Francisco gets. We will keep watching the corner, and the rest of the city's neighborhoods, on our San Francisco neighborhood rankings.

Sources

San Francisco Planning Commission, agenda for September 24, 2026, Items 8a and 8b, Case 2025-008874CUA and 2025-008874VAR

SF YIMBY, “Reduced Plans Filed For Glide Foundation Expansion, San Francisco,” September 2025

SF YIMBY, “GLIDE Foundation Proposal Modernization Project, Tenderloin, San Francisco,” March 2022

SF YIMBY, “Updated Illustrations For 550 O’Farrell Street in Tenderloin, San Francisco,” August 2026

KQED, “Why Hasn’t the Tenderloin Gentrified Like the Rest of San Francisco?” 2018, updated 2023

Mission Local, Tenderloin Buzz, June 2026

SFist, Buffett and Curry lunch auction for GLIDE, May 15, 2026

GLIDE, press release announcing Dr. Gina Fromer as CEO, October 9, 2023

Wikipedia, Cecil Williams (pastor), and Glide Memorial Church

Houseberry neighborhood pages for the Tenderloin, Nob Hill, Downtown and South of Market, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Fifty thousand cubic yards. That is how much sediment and brush Marin Water now says it will dig out of Upper Nicasio Creek and Halleck Creek, roughly 5,000 ten-yard dump-truck loads by our count, as part of its Atmospheric River Capture Project. The number showed up in a revised Notice of Preparation filed September 17, 2026, and it tells you the pipeline plan is getting bigger, not smaller.

The short answer on Marin water supply: it is in far better shape than in 2021, when reservoirs fell to 32 percent, but it still leans on rain that falls in a few winter months. ARC is the district's bet on fixing that. Buyers should read it as a modest insurance policy with a real price tag.

Is Marin's water supply secure right now?

Yes, for this year. On September 16, 2026, Marin Water's seven reservoirs held 65,846 acre-feet, 82.8 percent of capacity, about 19 percent above the long-run average for that date.

The risk is two dry winters back to back, which is what happened in 2020 and 2021. About 75 percent of the district's water is local rain caught on Mount Tamalpais and the West Marin hills, and the rest comes from the Russian River.

The bottom line

Here is what the September filing and the district's own numbers add up to.

• Marin Water's ARC project would pipe surplus winter Russian River water 13.2 miles to Nicasio Reservoir, at an estimated $195 million (range $137 million to $293 million), with operation targeted for 2030.
• The best case is up to 3,800 acre-feet in a dry year. One October 2021 storm added 8,763 acre-feet to Marin's reservoirs in a single week.
• A typical Marin Water single-family bill rose from $138.66 to $218.80 every two months between mid-2023 and July 2026, up 57.8 percent.
• Novato is not a Marin Water town. It is served by North Marin Water District, which buys about 80 percent of its water from Sonoma Water.
• Comments on the revised plan are due by noon on Monday, October 19, 2026.

What the revised filing adds

The new piece is creek work, and a lot of it. The March 2026 version of the plan was a pipeline and two pump stations. The September supplement adds a long-term management plan for Upper Nicasio Creek, where the creek meets the reservoir in the unincorporated village of Nicasio.

Per the filing on CEQAnet (state clearinghouse number 2026030496):

• About 50,000 cubic yards of vegetation and sediment removal over at least two construction seasons.
• Roughly 3,150 linear feet of Upper Nicasio Creek and 1,200 linear feet of Halleck Creek.
• Lowering the sediment delta in Nicasio Reservoir by about 4 feet to create a bench about 100 feet wide along 2,000 feet of creek bank.
• Follow-up maintenance estimated at 10,000 cubic yards every 5 to 10 years, per the supplemental notice itself.

The filing does not say why the creek plan was added. The maintenance line is the tell: this is recurring work, not a one-time job.

A pipeline sized for storms, measured in weeks

ARC works like this. When the Russian River runs high in winter and its minimum stream flows are met, Marin Water would take extra water through the existing North Marin Aqueduct, then push it through a new 36-inch pipe to Nicasio for storage. The filing lists capacity at 30 million gallons a day.

Our math: 30 million gallons is about 92 acre-feet a day. Hitting the district's best-case 3,800 acre-feet would take roughly six weeks of full-throttle pumping. A critique published by CalSPORT in June 2026 argues the conditions for that (low reservoirs plus big Sonoma storms) rarely line up. The draft EIR should answer that with modeled years.

It is also the second try. Marin Water dropped its plan to raise the Nicasio spillway in August 2025 over rising costs and limited benefit. The district's Water Supply Roadmap counts ARC for 3,800 to 4,750 acre-feet a year toward a goal of 8,500.

October 2021 is the number that explains the whole project. Lake Lagunitas got less than half its usual 50-plus inches the winter before, storage fell to 25,772 acre-feet on October 18, 2021, and the district was planning an emergency pipeline across the Richmond-San Rafael Bridge. Then one atmospheric river hit and storage jumped to 34,535 acre-feet in a week. Weather bailed Marin out. ARC is a bet on needing that luck a little less.

Novato drinks from a different pipe

Marin Water serves central and southern Marin, about 191,000 people from San Rafael to Sausalito. Novato is served by North Marin Water District, which relies on Russian River water from Sonoma Water for about 80 percent of its supply and Stafford Lake for the rest.

Here is the local twist. ARC's pipeline would run under Novato streets, Wood Hollow Drive, Redwood Boulevard, San Marin Drive and Novato Boulevard, with a pump station near Stafford Lake. Novato gets the construction. The stored water goes to the other district.

The table puts water supplier next to Houseberry's August 2026 prices and scores.

Novato is the cheapest city in our Marin set at $1.19 million, and its water depends least on local rain. Santa Rosa sits on the same Russian River ARC wants to tap, while EBMUD's supply starts in the Sierra. None of that shows up in a listing.

Rates are the part buyers actually pay

Rates are already up, and ARC has not been built yet. Under the schedule Marin Water adopted in 2023, a typical single-family bill went from $138.66 to $218.80 every two months, the last step landing July 1, 2026.

That schedule is now used up. Marin Water plans to pay for ARC with bonds, and bonds get repaid from rates. Expect the next rate cycle to carry some of the $195 million.

What this does and does not mean

It does mean central and southern Marin is shoring up supply, and water costs there are rising faster than inflation. It does not mean Marin is running out of water. At 83 percent full, it is not.

And it does not change what drives a home's value in Mill Valley or San Rafael. A water bill that is up about $480 a year since 2023 is noise next to a $1.84 million median price. Where it matters is a repeat of 2021: mandatory cutbacks and brown lawns.

I am for ARC in principle. A district that almost ran dry five years ago should store more water. But I want the draft EIR to show how often the pipe would really run.

Questions buyers ask about Marin water

Does Marin have water restrictions right now?

No. Marin Water lifted most drought restrictions in early 2022 after October and December storms refilled reservoirs to about 95 percent. Storage is above average in September 2026.

Which Marin towns does ARC help?

Towns served by Marin Water: San Rafael, Mill Valley, Corte Madera, Larkspur, San Anselmo, Fairfax, Ross, Sausalito, Tiburon and Belvedere, plus unincorporated areas such as Kentfield. Novato is on North Marin Water District.

How do I comment on the Atmospheric River Capture Project?

Email arcproject@marinwater.org, or mail Marin Municipal Water District, c/o Lucy Croy, 220 Nellen Avenue, Corte Madera, CA 94925, by noon on October 19, 2026. The district's ARC page says the draft EIR is expected this fall.

Should water supply change where I buy in Marin?

Rarely by itself. Treat it like fire insurance: a line item to check, not a deal breaker.

That is the neighborhood-first habit we built Houseberry around, checking what sits under the address before falling for it. Our Marin neighborhood rankings and the earlier look at Marin's most affordable neighborhoods are good places to start that comparison.

Sources

CEQAnet, Supplemental Notice of Preparation, Atmospheric River Capture Project, SCH 2026030496, Marin Municipal Water District, received September 17, 2026

Marin Water, Atmospheric River Capture (ARC) Project page, retrieved September 22, 2026

Marin Water, Water Watch reservoir storage, as of September 16, 2026

Marin Water, Water Supply Roadmap, retrieved September 22, 2026

Marin Water, 2023 Water Rate Setting

Marin Water, Notice of Exemption for Temporary Urgency Change petition, August 30, 2021

National Drought Mitigation Center, Drought Impact Reporter, Marin County storage and pipeline reports, 2021 to 2022

CBS News Bay Area, “Drought-Stricken Marin Putting Into Place Tough Water Use Restrictions,” April 2021

Point Reyes Light, “Marin Water drops Nicasio spillway plan,” August 27, 2025

CalSPORT, “Marin Water’s Atmospheric River Capture (ARC) Project Unlikely to Live up to its Name,” June 5, 2026

North Marin Water District, rate increase and rate structure changes page

East Bay Municipal Utility District, water supply page

Houseberry, city guides for San Rafael, Novato, Mill Valley, Corte Madera, Larkspur, San Anselmo, Oakland and Santa Rosa, and Marin neighborhood rankings, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

“Does my Novato street qualify?” If you live where the backyards meet the hills, around San Marin, the Marin Country Club or along Atherton Avenue, the honest answer is probably yes. And most of the help that comes with it costs nothing.

On September 18, 2026, the Novato Fire Protection District filed a state notice of exemption for two homeowner programs that remove brush, ladder fuels and hazard trees within up to 200 feet of structures. Two days later the Chileno Fire burned 341 acres of ranchland off Chileno Valley Road west of Petaluma. Crews stopped its forward progress that evening and no mandatory evacuations were ordered. It was still a useful reminder of what dry September grass does.

What you need to know

• Novato Fire cleared the Novato Core Direct Assistance Program and the Novato Hazardous Vegetation Removal Project with no environmental review, filed as SCH 2026090835 on September 18, 2026.
• Crews can clear flammable vegetation and trees under 12 inches in diameter at chest height, up to 200 feet from homes.
• The filing says over half of the Novato Zone is in high or very high fire hazard severity zones. In-town streets we checked on CAL FIRE's 2025 map rated High, Moderate or non-wildland.
• Free help already exists: a Novato Fire home assessment, curbside Chipper Days, and Marin Wildfire resident grants that have run up to $2,500 per parcel.
• Novato's median sale price was about $1.19 million in August 2026, per Houseberry's city data, so a clean defensible-space record protects a large asset.

What the Novato filing actually clears

The notice approves work, not a mandate. It lets the district send help onto private lots without writing a separate environmental study for each one.

The Core Direct Assistance Program helps homeowners build defensible space that ties into fuel reduction projects already done next door. The Hazardous Vegetation Removal Project targets invasive and introduced trees and plants that pose a wildfire hazard or could block an evacuation route.

The legal hook is Public Resources Code section 21080.49, added by SB 131 in June 2025. It exempts defensible space and fuel break work within 200 feet of structures in High or Very High zones, capped at trees under 12 inches in diameter. That cap matters. This is brush and small trees, not clear-cutting a canyon of mature oaks.

What the filing does not say is just as useful. It names no cost to homeowners, no sign-up date and no age or income rules. The listed contact is Lynne Osgood, Novato Fire’s Battalion Chief of Prevention, at (415) 878-2693.

Novato fire hazard zones, neighborhood by neighborhood

Novato’s built-up core rates lower on CAL FIRE’s map than the filing’s “over half” line suggests, and both things are true at once.

The “Novato Zone” is Novato Fire’s slice of the Marin Wildfire Prevention Authority, the Measure C agency, and it includes a lot of open space and unincorporated ranchland. So we checked the actual streets. Using the CAL FIRE hazard map dated March 24, 2025, we looked within about 400 meters of one street named in each of our seven Novato neighborhood profiles.

Two cautions. Those are spot checks, not full neighborhood boundaries, so your address can differ from the street we sampled. And the August medians in Northeast, Southeast and Central Novato jumped on a handful of big sales, which is why we also show the typical month over the past year.

In a box drawn around the whole city, the map’s only Very High patch sits on state-responsibility land, outside the city-rated areas. The High-rated edges are where you would expect: West Novato against Indian Valley and O’Hair Park, the Atherton Avenue hills, and Southwest Novato around the Marin Country Club and the Ignacio and Pacheco Valley preserves.

Hazard does not sort by price here. The hills are part of what people pay for in Novato, and they are also the part that needs tending.

The free help, in the order I would use it

Start with an assessment, because every grant and most discounts ask what an inspector found.

• Book a free home assessment through Novato Fire’s StopWildfire site. A wildfire mitigation specialist walks the property and leaves a written list of fixes.
• Check the grant round. The most recent posted Marin Wildfire resident grant paid up to $2,500 per parcel, $15,000 lifetime, with no matching money required. It covered metal gutters, vents with 1/8-inch screening, noncombustible fence sections and clearing within 30 feet of the house. Confirm current amounts before you spend.
• Pile brush for Chipper Days. It is free, runs six weeks a year, takes branches up to 8 inches thick in piles up to 20 by 4 by 4 feet, and you register by the Friday before your week.
• Call Novato Fire about the new direct assistance programs, especially if you cannot do heavy clearing yourself. The filing does not define who qualifies, so ask.

The money is real. The Marin Wildfire Prevention Authority approved a $24.1 million budget on May 21, 2026, with more than $1.35 million for Chipper Days, over $3 million for reimbursements, resident grants and direct assistance, and roughly 30,000 home evaluations planned countywide.

Zone 0 probably reaches fewer Novato lots than you might think

California’s new five-foot ember-resistant rule applies inside city-rated areas only in Very High zones, and none of the in-town streets we checked carry that rating.

The Board of Forestry adopted emergency Zone 0 regulations on August 19, 2026, and they still need Office of Administrative Law approval to take effect. On state-responsibility land they apply to every structure, so unincorporated edges of the district are in. Existing homes get a three- to five-year phase-in after new construction. Novato Fire can still set local rules, so ask before assuming you are exempt.

I would do the five feet anyway. Clearing bark mulch and dead plants from against the house costs a weekend, and it is on the state’s insurance discount list.

What buyers should check before an offer

For a buyer, the practical question is whether the home comes with paperwork, not whether the hills look scary.

Since July 1, 2021, state law (AB 38) has required sellers in High or Very High zones to hand over documentation of defensible space compliance. If they cannot by closing, the buyer typically gets one year to get it done. Ask for the inspection report early. It tells you what the next year of weekends looks like.

Then call an insurance broker before you remove contingencies. Under the state’s Safer from Wildfires rules, insurers must offer discounts for items like a Class A roof, ember-resistant vents and a cleared five-foot zone. The FAIR Plan offers its own. The trade-off is plain: a tidy lot does not guarantee a cheaper policy, because each carrier weighs mitigation differently.

Context helps. The Novato citywide median rose from about $984,000 in September 2025 to $1.19 million in August 2026.

The citywide number does not split out hillside homes, so read it as context, not a verdict on any one street.

How to think about the hills

A hazard zone describes the land around a house, not the house itself. Two neighbors on the same High-rated street can carry very different risk depending on the roof, the vents and what is piled against the back fence.

That is why this filing matters more than it looks. It moves Novato from telling people what to clear toward sending help to clear it. If you are comparing Novato against the rest of Marin, our look at the most affordable neighborhoods in Marin is a good next read. Check the address, book the assessment, and read the hills as one more thing to plan for.

Sources

CEQAnet, Notice of Exemption, SCH 2026090835, Novato Core Direct Assistance and Novato Hazardous Vegetation Removal Programs, Novato Fire Protection District, posted September 18, 2026

CAL FIRE, Chileno Fire incident page, September 2026

CAL FIRE Office of the State Fire Marshal, Fire Hazard Severity Zones, 2025 maps, and the FHSZ layer dated March 24, 2025 as hosted by MTC

Fire Safe Marin, “Marin Wildfire Approves $24.1 Million Budget,” May 2026

Marin Wildfire Prevention Authority, Resident Grants FY 24/25

Novato Fire Protection District, Chipper Day program page, and StopWildfire.net residential assessments

Ascent, streamlining synopsis on SB 131 and Public Resources Code 21080.49

Best Best & Krieger, “California Board of Forestry Adopts Emergency Zone 0 Regulations Under AB 3074,” September 2026

California Legislature, AB 38 (2019), fire safety and defensible space disclosure

California Department of Insurance, Safer from Wildfires FAQ, and California FAIR Plan discount announcement

Houseberry, Novato neighborhood rankings, neighborhood pages and city price history, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Here is the question a lot of state workers in Elk Grove, Folsom and Natomas are asking this week: do I keep driving in four days a week, or do I finally move closer? Gov. Gavin Newsom answered half of it on Friday, September 18, 2026, when he vetoed AB 1729, the bill that would have loosened work-from-home rules for state employees. The four-day return-to-office order stays in place.

The other half, where to live now that the commute is permanent, is a neighborhood question. That is the part we can help with.

In brief

• Newsom vetoed AB 1729 on September 18, 2026, saying it “takes the state’s telework policy in the wrong direction.” The four-days-a-week office order that began July 1, 2026 is unchanged.
• The Sacramento-Roseville-Folsom metro had about 139,600 state government jobs in August 2026, so this lands harder here than anywhere in California.
• Downtown foot traffic rose 12 percent in the first days of July 2026 versus a year earlier, per the Downtown Sacramento Partnership.
• Sacramento’s citywide median home price was about $439,000 in August 2026 on Houseberry, roughly $193,000 below Elk Grove’s.

What AB 1729 would have done, and why Newsom said no

AB 1729 would have made each department put in writing why a job needs to be in the office, and it told the Department of General Services to build a public dashboard tracking what telework saves. Assemblymember Alex Lee, a South Bay Democrat, carried it with Folsom Republican Josh Hoover as a co-author, and the engineers’ union PECG sponsored it.

It passed easily, 67 to 7 in the Assembly and 31 to 6 in the Senate. Newsom vetoed it anyway, writing that in-person collaboration is “an essential part of a productive and effective workplace” and objecting that the bill would encourage telework “to be the default work arrangement.” Those margins clear the two-thirds needed for an override, but the Legislature has not overridden a California governor since 1979.

Newsom’s March 3, 2025 order set four office days a week starting July 1, 2025. Then the unions bargained. PECG traded a one-year delay for a raise largely offset by unpaid hours, and SEIU Local 1000 and the scientists’ union CAPS got similar delays. In August 2025 the State Auditor found that telework three days a week could save about $225 million a year and that the order went out without a full look at office space needs or costs. The delay ran out and the four-day week took effect July 1, 2026, covering roughly 108,000 hybrid workers.

The unions are not done. SEIU Local 1000, which represents about 96,000 state workers, is holding a strike authorization vote over pay and telework, and PECG President Brian Simon said in a statement after the veto that telework “should be embraced, not discarded.” For anyone planning a move, though, the working assumption for the next year is four days a week.

Why this is a Sacramento story first

Because the capital region holds about 40 percent of the state civil service. Federal labor data show about 139,600 state government jobs in the Sacramento-Roseville-Folsom metro in August 2026, and CapRadio puts roughly 90,000 state employees in Sacramento County alone. When their week changes, the grid downtown feels it.

And it has, a little. The Downtown Sacramento Partnership, using Placer.ai phone data, counted 12 percent more foot traffic from July 1 to 11, 2026 than in the same days of 2025. The district had about 100,000 workers before the pandemic and was running near 60 percent of that in February 2026. Office vacancy downtown was about 24 percent in early 2025. A 12 percent bump is real for the lunch counters on K Street. It is not a full recovery.

Parking is where the new week gets personal. The state runs more than 20 parking facilities on lotteries and waitlists, added lots at 450 N Street and at the May Lee State Office Complex, the 1.25 million square foot campus on Bannon Street in the River District, and private garages start around $10 a day. Four days a week, that adds up fast.

The Sacramento commute shed, neighborhood by neighborhood

The neighborhoods that make a four-day office week easiest are the ones inside about three miles of the Capitol, and they trade safety scores for walkability. Here is how the main state-worker commute sheds compare on Houseberry scores, with rough driving miles and the rail option each one actually has.

Midtown, the grid of Victorians and newer apartments between the Capitol and the Business 80 loop, is a bike ride from almost every state office, but it scores 1.6 of 5 on our safety measure. East Sacramento, the older neighborhood of 1920s and 1930s homes around McKinley Park, scores 4.5 on amenities and puts you on the Gold Line at 39th, 48th or 59th Street. Its safety score is 2.7. Curtis Park and Land Park, south of Broadway, sit near 4.3 to 4.5 on amenities with safety in the low to mid 2s. Our Sacramento safety rankings run from 1.1 to 4.8, so context matters more than any single number.

Natomas Park is the interesting outlier. It ranks fifth of 185 Sacramento neighborhoods at 3.6 overall with a 4.4 safety score, but there is no light rail across the American River yet. Every trip downtown is I-5, and the planned Green Line extension to the airport is still unfunded.

The workers who moved out during remote work

The people hit hardest by the veto are the ones who bought on the edge of the region when two office days a week felt permanent. Elk Grove, Folsom and Roseville buy you higher top-end scores. Elk Grove’s best-ranked area scores 4.1, and both Empire Ranch in Folsom and Johnson Ranch in Roseville score 4.0. But you pay in miles.

And in Elk Grove’s case, you also paid more. Houseberry’s city pages put Elk Grove’s median at about $632,000 in August 2026, against about $439,000 for Sacramento citywide. In no month of the past year did the gap narrow below about $170,000.

The citywide number hides a wide range, though. East Sacramento and Land Park have long sold well above the city median, so “closer” usually means a smaller or older house, or a condo in Midtown, more than a cheaper one. Davis is its own case, with a median of about $903,000 in August 2026 on Houseberry and a Capitol Corridor station, which makes it a train commute for people who can afford it.

Rail is the pressure valve, and it is uneven. SacRT light rail was running about 75 percent of pre-pandemic ridership heading into July, with Capitol Corridor ridership up 23 percent year over year. Folsom has the best of it now that SacRT runs 15-minute weekday Gold Line service out to the Historic Folsom end. Elk Grove drivers can park at Cosumnes River College and ride the Blue Line in.

One detail worth knowing before you list the house: CalMatters reported that exemptions are available for some workers who live 50 or more miles from their worksite. That mostly helps people who live well outside the region. It does not reach Elk Grove, which sits about 16 road miles from the Capitol.

How to think about a move right now

Treat the four-day week as the plan for at least a year, and pick the commute you can live with about 180 times a year, which is roughly what four office days add up to once holidays and vacation come out.

For some households the answer is a better commute: a Blue Line park and ride, a vanpool, a parking lottery spot. For others it is a smaller place inside the three-mile ring. Both are reasonable. What we would not do is buy on the assumption that the next governor, or the next contract, restores two-day hybrid. It might. It is not a plan.

If you are coming from the Bay Area for a state job, our guide to what money buys moving to Sacramento from the Bay Area covers the price side, and the full Sacramento neighborhood rankings let you sort all 185 neighborhoods by schools, safety and amenities, which is how we would start the shortlist. Downtown also has room for a lot more homes near its light rail stations, and every one that gets built makes this trade a little less painful for the next person.

Sources

CalMatters, “Newsom vetoes bill to loosen work-from-home rules for state workers,” Lynn La, September 18, 2026 (republished by Local News Matters, September 21, 2026)

CBS Sacramento, “Newsom vetoes bill aimed at preserving California state workers’ telework,” September 2026

The Sacramento Bee, “Gov. Gavin Newsom vetoes state worker telework bill in latest battle over RTO,” Corey Schmidt, September 19, 2026

Industry Insider California, “State Worker Telework Bill Clears California Legislature,” August 28, 2026

Office of the Governor, return-to-office announcement, March 3, 2025

CalMatters, “Newsom return-to-office order delayed for one state worker union,” Adam Ashton, June 24, 2025

CapRadio, “Audit finds California can save hundreds of millions if state workers stay remote,” August 14, 2025

CapRadio, “Union-backed bill would require telework options for California state workers,” February 23, 2026

Abridged by PBS KVIE, “Two weeks into return-to-office mandate, Downtown Sacramento sees modest signs of rebound,” July 16, 2026

The Sacramento Bee, “State workers returning to the office. Is Sacramento’s commute about to change?” June 30, 2026, and “Will RTO change downtown Sacramento?” July 1, 2026

CalMatters, “Gavin Newsom’s return-to-office rules are stricter than many states,” Lynn La, July 20, 2026

PECG, statement on the AB 1729 veto, September 21, 2026

U.S. Bureau of Labor Statistics via FRED, state government employment, Sacramento-Roseville-Folsom MSA, August 2026

California Department of General Services, May Lee State Office Complex

SacRT, Folsom 15-minute service

Houseberry, Sacramento, Elk Grove, Roseville, Folsom and Davis rankings and city pages, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The idea fits on an index card. Cap banked rent increases in San Francisco's rent-controlled apartments at 10 percent a year, and make landlords remind tenants every year what the limits are. Supervisor Danny Sauter has introduced it, Bay City News reported, as part of the tenant package Mayor Daniel Lurie rolled out on September 10 when he declared a rent emergency.

It does not touch the annual allowable increase, which is 1.6 percent for March 2026 through February 2027. It does not abolish banking. What it changes is the speed. A tenant who has not had a raise since 2015 can legally be handed a 21.1 percent increase today. Under the cap, the same total arrives in three steps.

Key takeaways

• Sauter's ordinance would limit how much banked rent a landlord can impose on a rent-controlled tenant to 10 percent in any year and require an annual notice of rent limits and tenant rights.
• Banking lets owners save up skipped annual increases with no expiration and no ceiling. A tenancy that began in March 2015 has a 21.1 percent bank as of September 2026.
• For a buyer of a pre-1979 two to four unit building, the cap delays the day-one rent bump more than it shrinks it. On two long-term units at $2,500, roughly $7,300 of rent shifts into later years.
• As of September 22, 2026 no committee hearing is on the calendar, and we could not find a public Legistar file for the ordinance.

What banking means under the San Francisco Rent Ordinance

Banking is the right to take an annual increase later instead of now. Each year the Rent Board sets an allowable increase at 60 percent of Bay Area inflation. If a landlord skips it, the skipped amount goes into a bank that the city says has no limit and does not expire during the tenancy.

A few rules shape how the bank behaves, and they matter for the math:

• Banking has been allowed since April 1, 1982. Skips before then do not count.
• Banked percentages add up. They cannot be compounded.
• The bank belongs to the tenancy, not the owner. A new buyer inherits the prior owner's bank, and it disappears when the tenant moves out.
• Any increase over 10 percent in a year already requires 90 days of written notice instead of 30.

Most owners who bank are not scheming. They just never raised rent on a good tenant. The trouble starts at sale, when a new owner finds a decade of unused increases on the rent roll and applies them at once. The San Francisco Public Press documented the pattern back in 2019, and noted that the Rent Board does not track when banks get used.

A tenant since 2015, worked out

Here is the stack for a unit rented in March 2015 at $2,500 where rent was never raised. The allowable increases on each March anniversary, from the Rent Board's historical table and its recent announcements, ran 1.6, 2.2, 1.6, 2.6, 1.8, 0.7, 2.3, 3.6, 1.7, 1.4 and 1.6 percent from 2016 through 2026.

Add them and you get 21.1 percent. Today the owner could raise that $2,500 to about $3,028 in one notice, an extra $527.50 a month. Under a 10 percent annual cap it goes to $2,750, then $3,000 a year later, then $3,028 in year three.

Longer tenancies stretch further, same $2,500 rent, same assumption that no raise was ever taken.

The cap barely registers for anyone who moved in after about 2020. It bites hardest on tenancies from the 2000s, which is exactly where the scariest stories come from. Think of a 44 percent notice landing on someone who has lived in the same Richmond flat for 26 years.

The ordinance, line by line

This is the reported version. The ordinance text was not public yet when we checked.

• Sponsor: Supervisor Danny Sauter of District 3, with Supervisor Matt Dorsey named on the tenant-notice piece, per SFist and the SF Standard.
• Covers: rent-controlled units under the city's Rent Ordinance, generally pre-1979 multi-unit buildings.
• The cap: banked increases limited to 10 percent in a year. Our three-step math assumes that limit sits on top of the regular annual increase, which the reporting does not yet settle.
• Notices: landlords must tell tenants each year about rent increase limits and their rights.
• Pass-throughs: Mission Local and the Richmond Review report the package also reins in capital improvement charges. Details are not out.
• Unchanged: the 1.6 percent annual rate, banking itself, inheritance of banks by new owners, and vacancy decontrol when a tenant leaves.

Sauter framed it as preventing “devastating surprise rent hikes.” Fair enough. It is a speed limit, not a rent cut.

What changes for buyers of 2 to 4 unit buildings

For buyers, the cap mostly moves income later rather than erasing it. Take a pre-1979 fourplex where two units are near market and two long-term tenants from 2015 each pay $2,500.

Under current rules, a buyer's pro forma can book both banks at close: $527.50 times two units times 12 months, about $12,660 a year in added income. At a 5 percent cap rate, which is our assumption for illustration and not a market quote, that line alone supports roughly $253,000 of value on paper.

With the cap, year one adds $6,000, year two $12,000, year three the full $12,660. The shortfall over the phase-in totals about $7,300. On a building in the low millions, that is a rounding error in the price and a real line in a first-year cash budget.

The bigger underwriting change is quieter. Some buyers count on a big banked notice nudging a long-term tenant out, resetting the unit to market. A 10 percent step is easier to absorb, so fewer will move. If your offer only works because someone moves out, the cap is aimed squarely at your spreadsheet. That is a policy choice, and I think an honest buyer should price it rather than argue with it.

There is a real downside for tenants too. Banking has been an informal courtesy, the landlord who skips raises because the tenant is reliable. Put a speed limit on collecting later and some owners will simply take the 1.6 percent every March.

Richmond, Sunset and Mission: where it lands

The cap matters most where the rent-controlled stock is flats and small apartment buildings. It matters least where it is single-family houses, which state law largely exempts from local rent limits. Our own neighborhood data tracks that split.

The Inner Richmond, the grid of Edwardian flats between Arguello Boulevard and Park Presidio around Clement Street, had a median sale price of about $2.61 million in July 2026. That is well above the $2.2 million San Francisco citywide median for August on Houseberry's price history. Scores are strong, 4.4 out of 5 for both schools and safety, and much of the housing is exactly the two to six unit stock the ordinance targets.

The Outer Richmond, out toward Ocean Beach along Balboa and Geary, ran about $2.01 million in July 2026. Same small pre-1979 buildings, lower entry price.

The Outer Sunset, the Judah and Noriega corridors running down to the Great Highway, sold at about $1.73 million in August 2026 and ranks among the top two of 92 San Francisco neighborhoods on our overall score. Its stock is mostly detached houses, so a banked-rent cap reaches fewer renters here than the headline suggests.

The Inner Mission, the Victorians and walk-ups around Valencia and the 16th and 24th Street BART stations, posted about $2.33 million in August 2026. It scores 2.1 overall on Houseberry, pulled down by school and safety scores, but holds some of the city's densest pre-1979 rental stock and a lot of very long tenancies.

Those medians cover every sale, mostly houses and condos. Read them as price levels, not fourplex comps.

And the cap would not reach the Richmond case that made headlines in June. The 90 percent increase at 6838 Geary Boulevard involved a 1996 building outside local rent control and relied on a state-law exemption. Banking had nothing to do with it.

Who is for it, who is against it

Support is broad, and opposition is aimed elsewhere so far. Gail Gilman of the Eviction Defense Collaborative praised the package, per the Richmond Review. Tenant groups want more: Anastasia Yovanopolous of the SF Tenants Union said it “falls seriously short,” and Meg Heisler of the Anti-Displacement Coalition called it incomplete.

Landlord groups have spent their fire on Supervisor Jackie Fielder's eviction measure. The San Francisco Apartment Association sent the Board a letter against it, the SF Standard reported. We found no formal landlord statement on the banking cap as of September 22.

My view, for what it is worth: a speed limit on banked increases is reasonable, and it will not house one extra person. Rents are up about 25 percent in a year because vacancy fell to 2.2 percent and the city finished only 489 homes this year, per the SF Standard. The fix for that is permits, not notices.

When the Board could take it up

Nothing is scheduled yet. Board rules generally hold a newly introduced ordinance for 30 days before a committee can hear it, so the earliest realistic hearing is late October, most likely at Land Use and Transportation, where Rent Ordinance changes usually go. We will update this post with the hearing date, and our San Francisco neighborhood rankings are there if you are weighing which block to buy or rent on in the meantime.

Sources

Local News Matters / Bay City News, “San Francisco supervisors move to limit steep rent increases landlords can bank,” September 17, 2026

SF Standard, “Lurie declares rent emergency as SF housing costs surge,” September 10, 2026

SFist, “SF Mayor Declares Rent Emergency, Announces Several Tenant Protection Proposals,” September 10, 2026

Mission Local, “Lurie declares ‘rent emergency’ as S.F. rents spike,” September 2026

Richmond Review/Sunset Beacon, “Mayor Unveils Tenant Protection Package as Rents Surge to Record Highs,” September 11, 2026

SF Standard, “SF landlords go on offense against Lurie’s rental emergency declaration,” September 15, 2026

SF.gov, Banked rent increases, and Learn about rent increases in San Francisco, retrieved September 22, 2026

SF.gov, Annual Rent Increase for 3/1/25 to 2/28/26 Announced

San Francisco Rent Board, Allowable Annual Increases table (1982 to 2021), via SF Tenants Union

Elke & Merchant LLP, SF Rent Board announcements of the 0.7 percent (2021 to 2022) and 2.3 percent (2022 to 2023) allowable increases

San Francisco Apartment Association, Rent Increases, Interest and Fees (3.6 percent for 2023 to 2024)

Hoodline, Richmond District 90 percent rent increase at 6838 Geary Boulevard, June 2026

San Francisco Public Press, “Cities Grapple With ‘Banked’ Rent Hikes,” September 16, 2019

Houseberry, San Francisco city guide, neighborhood pages and rankings, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Atherton's median home price was about $17.3 million in June 2026 on our numbers. That is roughly 14 times the $1.27 million Bay Area median the California Association of Realtors reported for August. It is also the town with the most at stake, in housing terms, when Californians vote on Proposition 40, the one-time 5 percent billionaire tax, on November 3, 2026.

The short answer on what the California billionaire tax could do to these neighborhoods: probably less than the campaign rhetoric on either side suggests. The houses themselves are excluded from the tax. The last wave of billionaire departures, in 2020 and 2021, did not dent Atherton prices. And if a trophy estate does sell, Prop 13 means the local property tax take goes up, not down.

The numbers that matter

• Proposition 40 would levy a one-time 5 percent tax on the net worth of Californians worth more than $1 billion who were state residents on January 1, 2026. The vote is November 3, 2026.
• Real estate held directly is excluded, so an estate in Atherton or Pacific Heights is not itself taxed.
• Fortune counted six billionaires, including Larry Page, Sergey Brin and Peter Thiel, who cut ties with California before the January 1, 2026 residency date.
• Atherton's 94027 median climbed from $7 million in 2020 to $7.9 million in 2022 on PropertyShark's rankings, straight through the Musk, Ellison and Oracle exits.
• Larry Ellison's Pacific Heights home, bought for $3.9 million in 1990, sold for $45 million in December 2025, which resets its assessed value under Prop 13.

What the Prop 40 billionaire tax actually taxes

It taxes stock, business stakes and other financial wealth. It leaves the mansion alone.

Per the Legislative Analyst's Office, billionaires who lived in California on January 1, 2026 would owe 5 percent of their net worth once, with "real estate, pensions, and retirement accounts" generally excluded. The sponsors put the number who would pay at about 200 and the take at roughly $100 billion over five years, with 90 percent going to health care. The LAO's own estimate is "tens of billions."

The mechanics, per Holland & Knight's September 2026 summary: net worth is measured December 31, 2026, the bill is due April 15, 2027, and payers can spread it over five annual installments with a 7.5 percent yearly deferral charge. Moving out later does not help. A person who was a resident on January 1, 2026 stays in scope "even if the person later fully left the state." Two rival measures, Props 41 and 42, would void Prop 40 entirely if either passes with more votes.

So where does housing come in? Through residency. California decides who is a resident by weighing many facts, and where you keep your home is one of the heaviest. When Ellison's San Francisco sale surfaced, tax adviser David Lesperance told the New York Post that Ellison was likely advised that keeping a California residence "was not worth the risk." That is the channel. A tax on stock portfolios can still show up as a for-sale sign.

Where California's billionaires actually live

Mostly in five Peninsula towns and a few blocks of San Francisco.

Forbes residence data, compiled by Stacker and published by KRON4, listed Page, Mark Zuckerberg and Laurene Powell Jobs in Palo Alto, Brin and Nvidia's Jensen Huang in Los Altos, Eric Schmidt, Jan Koum and George Roberts in Atherton, Charles Schwab and John Doerr in Woodside, and Brian Chesky, Dustin Moskovitz and Nathan Blecharczyk in San Francisco. Page and Brin have since cut ties, per Fortune. We keep it to the town level.

On the ground that means Atherton, the flat, oak-shaded town between Menlo Park and Redwood City with one-acre minimum lots. Woodside and Portola Valley, the horse-country foothill towns above Sand Hill Road. Palo Alto's Crescent Park, northeast of downtown along San Francisquito Creek, and Old Palo Alto, south of Embarcadero Road. Los Altos Hills, the no-downtown hill town that wraps around Foothill College. And in the city, the stretch of Broadway through Pacific Heights that locals call Billionaire's Row, plus neighboring Presidio Heights. Our colleague Elena Marsh walked several of these in where the top AI executives live.

Here is how those places score and what they cost right now.

Scores are Houseberry's 0 to 5 ratings. Prices are the latest month on each City or Neighborhood page. *Our Pacific Heights and Presidio Heights pages currently fall back to citywide data for price, so we left those two out rather than guess.

Two things jump out. First, billionaire money does not buy the best scores. Old Palo Alto and Crescent Park beat Atherton on schools and safety, and Pacific Heights scores 3.7 overall, 35th of 92 in San Francisco. People buy these places for land and privacy, not a spreadsheet. Second, even the cheapest enclave on the list, Portola Valley, runs about 3.5 times the regional median.

What happened the last time billionaires left

Atherton prices kept rising. That is the best precedent we have, and it points one way.

In 2020 Palantir moved its headquarters from Palo Alto to Denver, Elon Musk moved to Texas, and in December Oracle moved its headquarters from Redwood City to Austin. Days later Larry Ellison told staff he had moved to Lanai. Musk listed his 47-acre Hillsborough estate at $35 million in May 2020. It sat for about 18 months and closed for about $30 million in December 2021, per SocketSite's read of the recorded sale. One trophy estate took a haircut and a long wait.

The town-level median did not. PropertyShark's annual ranking put Atherton's 94027 ZIP at $7 million in 2020, $7.475 million in 2021 and $7.9 million in 2022. After the 2025 departures it hit $9.93 million in the first half of 2026, up 20 percent in a year and back to No. 1 nationally.

The honest caveat: 2020 and 2021 also brought record-low mortgage rates and a tech stock boom, and 2026 brought AI money. This August, ABC7 reported that an AI-industry buyer paid $70 million for a Hillsborough estate, double the town's old record. Departures were never the only force on these prices.

Why a few listings can swing Atherton's median

Because there are so few sales that each one moves the average.

On our Atherton city page, the monthly median ran from $9.94 million in July 2025 to $18.65 million in January 2026, then $15 million in March. Palo Alto, with far more transactions, stayed between $4.08 million and $4.54 million. PropertyShark's first-half 2026 Atherton median of $9.93 million and our June figure of $17.31 million disagree by more than $7 million, because they cover different windows and a handful of sales. That gap is the point.

What this does and does not mean. If three or four estate owners listed in the same slow quarter, as opponents warn could happen, Atherton's median could drop sharply on paper. That would not mean the typical Atherton house lost value. For buyers, land is the steadier yardstick. Our look at Peninsula teardown land values found Atherton lots trading around $188 to $258 per square foot.

The Prop 13 twist: a sale raises the tax base

When a long-held trophy home sells, the town's property tax take goes up.

Under Prop 13, a home's assessed value is its purchase price, rising no more than 2 percent a year until it sells, when it resets to the new price. Ellison bought his Pacific Heights home for $3.9 million in 1990 and sold it for $45 million on December 12, 2025. Thirty-five years of 2 percent growth caps a $3.9 million base at about $7.8 million, plus whatever the late-1990s remodel added. The new base is about $45 million. At the 1 percent general levy, that one parcel goes from well under $100,000 a year to roughly $450,000 (our estimate, before local bonds).

That matters most where schools live on local property tax. Palo Alto Unified says it is funded primarily through local property taxes, and its Measure B parcel tax failed in June with 60.8 percent, short of the two-thirds it needed. A reassessed estate helps a district like that. The revenue a departure costs is state income tax, which the LAO pegs at less than $1 billion a year. One more catch: an owner who leaves but keeps the house changes nothing on the property tax side.

What supporters and opponents argue

Both sides make a real case, and we are not taking one.

Supporters, led by sponsor SEIU-United Healthcare Workers West and joined by Sen. Bernie Sanders and the California Democratic Party, say the money would backfill federal health care cuts. Economists Emmanuel Saez and colleagues argue in an expert report on the measure that "the super wealthy do not generally move because of taxes." Nvidia's Huang told Forbes in January he was "perfectly fine" with it.

Opponents include Gov. Gavin Newsom, per CalMatters, the California Business Roundtable, the California Teachers Association, Planned Parenthood, Ron Conway and Chris Larsen. Building a Better California, co-founded by Brin, has put $82 million behind Props 41 and 42, per GV Wire. Their core argument is the one Fortune quantified in March: six departures took about $27 billion of potential revenue with them. San Jose State economist Matthew Holian told San José Spotlight that tax changes "can cause a host of unforeseen reactions."

Prop 40 billionaire tax FAQ

What is the California billionaire tax?

A one-time 5 percent tax on the net worth of Californians worth more than $1 billion who were residents on January 1, 2026, due in 2027 and payable over five years, with 90 percent going to health care.

When do Californians vote on Prop 40?

November 3, 2026. It needs a simple majority, and it is voided if Prop 41 or Prop 42 passes with more votes.

Where do billionaires live in the Bay Area?

Forbes residence data puts most of them in Atherton, Woodside, Palo Alto, Los Altos and Los Altos Hills, plus San Francisco, where Pacific Heights' Broadway corridor is known as Billionaire's Row.

Would home prices in Atherton fall if billionaires leave?

The last test says not by much. Atherton's median rose from $7 million to $7.9 million between 2020 and 2022 while Musk, Ellison and Oracle left. A cluster of estate listings could drag the monthly median down on paper, because so few homes sell there.

Does Prop 40 tax homes?

No. Real estate held directly or through a revocable trust is excluded, according to the LAO and Holland & Knight. Where you keep a home can still count toward whether you are a California resident.

Can a billionaire avoid Prop 40 by moving now?

No. Residency is fixed at January 1, 2026, and leaving after that date does not remove someone from the tax, per Holland & Knight.

Whichever way the vote goes, the housing math starts at the neighborhood, not the headline. If you are watching these towns, our Peninsula neighborhood rankings are the place to compare scores and current prices side by side.

Sources

California Legislative Analyst's Office, Proposition 40 ballot analysis, 2026

Holland & Knight, “California's Proposed Billionaire Tax: What You Need to Know,” September 2026

Galle, Gamage, Saez and Shanske, expert report on the California 2026 Billionaire Tax, December 2025

San José Spotlight via Redwood City Pulse, “Silicon Valley voters to weigh ‘billionaire tax’ ballot measure,” September 22, 2026

Fortune, “Only 6 billionaires left California over its proposed wealth tax,” March 17, 2026

GV Wire, “How to Decode the Minefield of California's Tax Propositions,” July 20, 2026

CalMatters, Newsom and the billionaire tax, July 2026

Forbes, Jensen Huang on the billionaire tax, January 6, 2026

KRON4 and Stacker, Forbes residence list of California's richest billionaires

SF Standard, Larry Ellison San Francisco home sale, January 5, 2026, and New York Post via AOL

SocketSite, Musk Hillsborough estate sale, December 2021, CNBC and Fortune on Oracle and Ellison, December 2020

PropertyShark most expensive ZIP code rankings, 2020 to 2026, via ABC7, CBS, SFGate, KTVU, NBC Bay Area and Inc.

ABC7, Hillsborough $70 million sale, August 2026

California Association of Realtors, August 2026 home sales report, September 16, 2026

LAO, Understanding California's Property Taxes, and Palo Alto Unified School District parcel tax page

Palo Alto Daily Post, Measure B results, June 2, 2026

Houseberry City and Neighborhood pages and Peninsula and San Francisco rankings, retrieved September 22, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

On an identical $1,000,000 house, the first property tax installment due November 1 is $7,120 in Richmond and $5,250 in Antioch. Same county. Same due date. Same house. The gap is $1,870 every six months, and it comes entirely from the rate.

County tax collectors are mailing secured property tax bills right now. The first installment is due November 1 and goes delinquent after 5 p.m. on December 10. Almost everything written about that deadline repeats the same wrong number, which is 1 percent. One percent is the floor. What you pay is the floor plus whatever bonds and pre-1978 pension obligations your corner of the county voted for, and across the fourteen cities we ran the math on, the total lands between 1.0492 percent and 1.4243 percent.

One warning first, because it is the mistake this whole genre makes. Every figure below describes a home bought at today's median price. It is not what the neighbor who bought in 1994 pays. Under Proposition 13 your assessed value is your purchase price, rising a maximum of 2 percent a year until the house sells again, so most long-time owners are taxed far below market. A median sale price is not an assessed value. These are new-buyer bills.

The numbers that matter

• Across 14 Bay Area cities, the total tax rate runs from 1.0492 percent in Antioch to 1.4243 percent in Richmond, a spread of 0.3751 points. Richmond's rate is 36 percent higher than Antioch's.
• On a home assessed at $1,000,000, that is $5,250 versus $7,120 due on November 1, or $3,751 a year in difference for the same house.
• Richmond has the second lowest median sale price of the fourteen at $662,000 in July 2026, and still owes more annual property tax than Concord, where the median was $804,560 in August 2026.
• Across the fourteen cities the correlation between median sale price and tax rate is negative 0.16, which is close to no relationship at all.
• Rates come from each county's FY2025-26 rate documents, the most recent published. Fixed parcel charges are excluded, because they are dollars per parcel and not a rate.

What the November 1 installment actually looks like, city by city

Here is the whole board. Median sale price from each city's Houseberry price history, the total rate from that city's primary tax rate area, and the arithmetic in between. The November 1 installment is half the annual figure.

Rate sources, in order: the Alameda County Auditor-Controller property tax rates dataset for FY2025-26, the Contra Costa County Detail of Tax Rates 2025-2026, the Santa Clara County Compilation of Tax Rates and Information 2025-2026, the Marin County FY2025-26 Tax Rate Book, and the San Francisco Treasurer and Tax Collector, which publishes one citywide rate for the consolidated city and county.

One percent is a floor, not a rate

Proposition 13 caps the general levy at 1 percent of assessed value. It does not cap what sits on top of it.

The Legislative Analyst's Office splits a property tax bill into three parts: the 1 percent charge, voter-approved debt rates for pre-1978 pension obligations and post-1986 bonds, and direct levies charged on something other than assessed value. The middle piece is what makes one city cost more than another.

Oakland is a clean example. Its main tax rate area, 17-001, stacks a county bond at 0.0084 per $100, Oakland Unified school bonds at 0.0900, Peralta community college bonds at 0.0432, BART at 0.0152, East Bay Regional Park at 0.0011, and the City of Oakland's own voter-approved levy at 0.1200. Add them to the 1.0000 and you get 1.2779, which on Oakland's August 2026 median of $974,220 is $12,450 a year. Every one of those passed at a ballot box, and none of them appears in the 1 percent people quote.

Why the cheapest cities often carry the heaviest rates

A bond rate is not chosen. It is arithmetic, and the arithmetic punishes places with less property value to spread the debt across.

The rate for a voter-approved bond is that year's debt payment divided by the assessed value of everything inside the district. Two districts that borrowed the same amount charge very different rates if one sits on a much larger assessment roll. A city of $600,000 houses needs a bigger number to raise the same dollars as a city of $2.9 million houses.

Richmond carries a second thing on top of that. Contra Costa's rate book lists a Richmond pension tax of 0.1400 per $100, one of the pre-1978 obligations the state describes, a promise made before Proposition 13 existed. That single line is larger than Concord's entire package of voter-approved add-ons, which totals 0.1023. Richmond owners are still paying for a pension deal older than the tax system itself, and it costs them more than every school bond, college bond, park levy and transit measure Concord has passed put together.

Concord houses cost about $142,560 more than Richmond houses. The tax bill is smaller.

Order these fourteen cities by price, then by annual property tax, and the two lists do not match.

Richmond sits 13th of 14 by median price and 12th by annual tax, passing Concord on the way. The Richmond median home owes about $9,430 a year against Concord's $8,870, even though the Concord house costs roughly $142,560 more. Alameda does the same thing to Walnut Creek, owing $17,220 a year on a $1.42 million median against $16,730 on Walnut Creek's $1.53 million. Two swaps out of fourteen is not chaos. It is enough to matter when you are comparing two offers, and it never shows up on a listing page. To see where the price line falls block by block, our full East Bay ranking by median price runs every scored neighborhood in the region from cheapest to most expensive.

Where the rate stops mattering

Rate differences are worth a few thousand dollars a year. Price differences are worth tens of thousands, and the price wins every time.

Mountain View has the fifth lowest rate of the fourteen at 1.1560 percent and by far the largest bill, $33,760 a year on a $2.92 million median in July 2026. That is $16,880 on November 1, five times Antioch's $3,170, and the rate is working in Mountain View's favor the whole time. So read the rate as a tiebreaker, not a strategy. It decides between two similar houses in two similar towns. It never decides whether you can afford the Peninsula.

What the rate leaves out

Everything above is the ad valorem part of the bill, the part charged as a percentage. The rest of your bill is not a percentage at all.

Parcel taxes, Mello-Roos charges, lighting and landscaping assessments, sewer and vector control lines: those are fixed dollars per parcel. A school parcel tax charges the same amount on a $600,000 cottage and a $3 million house on the next street, which is why folding them into a rate produces a number that is wrong for everybody. We left them out and labelled them out. Depending on your district they add a couple hundred dollars to a few thousand, itemized on your own bill under a heading like special assessments or direct charges. That is the only reliable place to read them.

The number on your bill will not match this table, and here is the main reason

Assessed value. Not the rate.

The LAO puts it plainly in its property tax primer: a property's market value is generally greater than its assessed value, because assessed values rise by at most 2 percent a year while market values tend to rise faster. An Oakland owner who bought in 2003 may be assessed at a third of today's median and pay roughly a third of our table's figure at the same 1.2779 percent rate. The rate is the same for everyone in a tax rate area. The base is not.

Three more caveats. Buy this fall and you will also get a supplemental bill for the gap between the old assessment and yours. The $7,000 homeowners exemption knocks about $80 a year off, and we did not apply it. And these are FY2025-26 rates, because the FY2026-27 rate books were not published as of September 21, 2026. None of this is tax advice, and the only authoritative number is the one printed on your own bill.

What moves these rates before next November

Two things, and they pull in opposite directions.

New voter-approved debt pushes a rate up and retired debt pulls it down, which is why a city's number wobbles year to year. The quieter force is the assessment roll. Because a bond rate is debt service divided by total assessed value, a roll growing faster than the debt service on it pulls rates down with nobody voting on anything, interrupted every time a new measure passes.

Richmond's 0.1400 pension levy is the exception, because it answers to a pre-1978 obligation rather than a bond schedule. It is the most durable line in this comparison. We will re-run the table when the FY2026-27 books post, and the number to watch is whether Richmond's gap over Antioch widens past 0.3751 points.

How we did the math

Fourteen cities across five counties: Alameda, Contra Costa, Santa Clara, Marin and San Francisco. For each we took the total rate printed for that city's primary tax rate area in the county's own FY2025-26 rate document, listed above: the 1 percent general levy plus voter-approved debt service. Tax rate areas vary inside a city, so treat each figure as that city's main rate, not a rate every parcel pays. Median prices come from our city price histories, retrieved September 21, 2026, with the month stated alongside every figure. We dropped one city, Los Gatos, whose twelve-month series sat flat at $4.01 million for seven straight months and then fell 35 percent in a single step, which reads as a small-sample artifact rather than a market.

Three exclusions. Fixed special assessments, because they are per-parcel dollars and not a rate. Santa Clara County prints its Valley Water State Water Project levy of 0.00390 separately from the tax rate area total and we followed the county, so add about $39 per $1 million of assessed value there. Contra Costa's printed totals likewise exclude a land-only water levy of 0.0018. And once more, because it is the error everyone makes: a median sale price is not an assessed value.

Questions people are asking about the November bill

When is the first installment due, and when is it late?

The first installment of the 2026-27 secured property tax is due November 1, 2026 and is delinquent after 5 p.m. on December 10, 2026, after which a 10 percent penalty attaches. The second installment is due February 1, 2027 and delinquent after April 10, 2027. Alameda County's tax collector notes that if a delinquency date falls on a weekend or legal holiday the deadline moves to the next business day. December 10, 2026 is a Thursday, so there is no extension this year.

Why is my property tax rate higher than 1 percent?

Because Proposition 13 caps only the general levy. Voter-approved school, community college, city, park and transit bonds are charged on top of it, as are pension obligations approved before 1978. Those add-ons range from 0.0492 points in Antioch to 0.4243 points in Richmond among the cities here.

How much are property taxes in San Jose?

There is no single San Jose number, which is why the answers online disagree. The city spans San Jose Unified, East Side Union, Alum Rock, Berryessa, Oak Grove and more, and each combination is its own tax rate area with its own bond stack. Look up your parcel's tax rate area in the Santa Clara County rate book rather than trusting a citywide average. For scale, the Santa Clara County cities in our table run 1.1258 to 1.2147 percent.

Does buying a house change the tax bill right away?

Yes. The county reassesses at your purchase price and sends a supplemental bill for the difference between the seller's assessment and yours, covering the rest of the fiscal year. That is the bill that surprises first-time buyers in the spring.

Which Bay Area city in this comparison has the lowest property tax rate?

Antioch, at 1.0492 percent in tax rate area 01001, then Walnut Creek at 1.0937 and Concord at 1.1023. All three sit in Contra Costa County, the same county as Richmond, which is the tell that rates follow districts and bond history, not county lines.

Before you compare two offers this fall

Run the rate before you run the mortgage calculator. A 0.3751 point spread is roughly $3,750 a year on a million-dollar house, about what buyers argue over when they argue about an eighth of a point on a mortgage, and unlike the mortgage it never refinances away.

Then go back to the part that matters more than either number. The rate is a line on a bill. The block is what you live on, and comparing blocks before anyone falls in love with an address is the whole reason we built Houseberry. Open our Richmond neighborhood rankings next to the same rankings for Concord and see what that extra 0.3220 points of tax rate is actually buying.

Sources

Alameda County Auditor-Controller, Property Tax Rates FY25-26 dataset, Alameda County Open Data, retrieved September 21, 2026

Alameda County Auditor-Controller, Property Tax Reports (2026-27 tax rate book listed as not yet available), retrieved September 21, 2026

Alameda County Treasurer and Tax Collector, property tax FAQs, retrieved September 21, 2026

Contra Costa County, Detail of Tax Rates 2025-2026, October 2025

County of Santa Clara Controller-Treasurer, Compilation of Tax Rates and Information, Fiscal Year 2025-2026

County of Marin, FY2025-26 Tax Rate Book, Distribution of Tax Dollars

San Francisco Treasurer and Tax Collector, Secured Property Taxes, retrieved September 21, 2026

California Legislative Analyst's Office, Understanding Your Property Tax Bill, EconTax Blog

California Legislative Analyst's Office, Understanding California's Property Taxes, November 2012

California State Board of Equalization, Homeowners' Exemption

Houseberry city price histories for the 14 cities cited, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The question I keep getting about this one, from people who already hike Garin most weekends, is whether they will finally be able to walk from Hayward across to Pleasanton Ridge. The honest answer is no. Not from this project, and not soon.

What the East Bay Regional Park District actually filed on September 15 is smaller than that and more useful than it sounds: about half a mile of new trail, a public trailhead on Palomares Road where none exists today, and walking access to 955 acres the district has owned since 2011 and kept closed. Public comments close October 13, 2026. Nobody has written about it, and it is the closest thing to a new front door these hills have been offered in a while.

In brief

• The park district filed a Mitigated Negative Declaration for the Garin to Pleasanton Ridge Public Access Project on September 15, 2026, state clearinghouse number 2026090573.
• The work is about half a mile of new multi-use trail, a trailhead and staging area on the former Chouinard property at 33853 Palomares Road, four seasonal creek crossings, a Stonybrook Trail realignment, and an at-grade pedestrian crossing of Palomares Road.
• Public comments are due to the district by 5 p.m. on October 13, 2026, and the board is scheduled to consider the project on December 15, 2026.
• Palomares Rd in Union City scores 4.5 out of 5 on our Public Spaces measure, the highest of any neighborhood Houseberry rates in Hayward or Union City.

Half a mile of trail, and somewhere to leave the car

The trailhead is the real news here, not the trail.

The district bought the Chouinard property, 10 acres at 33853 Palomares Road, in 2020. It had been a winery. The buildings come down, a public trailhead and staging area goes in, and roughly half a mile of new multi-use trail runs from there into the existing Garin and Dry Creek Pioneer network.

Those two parks run to more than 5,800 acres and 35 miles of trails, and almost everyone walks in from the end of Garin Avenue in Hayward. The canyon side of that ridge has no public entrance at all. Half a mile is a twelve-minute walk and nobody's idea of a hike. As plumbing, it is the whole project.

Palomares Canyon is closer than it looks and harder to reach

Palomares Road is a two-lane canyon road that leaves Castro Valley near Dublin Canyon Road and drops southeast to Niles Canyon Road near Sunol. No shoulder, blind curves, a favorite with road cyclists, and shut by slides often enough that the county tracks it.

The site sits roughly midway down, about 5.5 straight-line miles from Union City's civic center, 6.9 from Hayward City Hall and 7.9 from Castro Valley Boulevard and Redwood Road. By car it is a good deal farther, because you drive around these hills rather than through them.

Here is the part to be careful with. The filing creates public access along selected existing ranch roads on the Owen property, the 955-acre former ranch across Palomares Road that the district's own project page says holds over eight miles of interior roads. That is ground the public has not walked in fifteen years. It is not a signed, continuous route from Garin's gate to Pleasanton Ridge, and the filing never claims to be. A separate district project is pushing the Bay Area Ridge Trail from Garin south to Niles Canyon. This one opens the middle.

What our scores already say about this corridor

These hills already rate well on open space, which is why I read this as an addition and not a rescue.

Palomares Rd, the mostly rural Union City neighborhood that shares the road's name, scores 4.5 out of 5 on our Public Spaces measure. That is the highest of any neighborhood we rate in either city. Fairview, up in the hills east of Hayward, sits at 4.0, as do Hayward Highland and unincorporated Castro Valley.

Run those same neighborhoods against our Amenities score and the actual trade shows up.

Palomares Rd is 4.5 on open space and 3.1 on amenities. Downtown Hayward, six miles down the hill, comes out even at 4.0 and 4.0. That spread is the hillside bargain in numbers. You buy trail access and a view by giving up the ten-minute grocery run, which is fine as long as you chose it rather than discovered it in month three.

Fairview's median sale price was about $902,000 in August 2026 and Palomares Rd's about $1.39 million, against a Hayward citywide median near $935,000 in July 2026. Hill markets up here are thin, so read any single month as a data point rather than a trend.

The at-grade crossing is the part worth reading closely

One line in the project description is doing more work than all the rest of it: an at-grade pedestrian crossing of Palomares Road.

People will park on one side and walk across to the other. On a canyon road with no shoulder, sight lines shaped by the hillside, and a steady weekend load of cyclists coming downhill fast. I want this built. I also think that crossing is the single thing most worth writing a comment about before the 13th.

Two more belong in the same paragraph. A new staging area puts cars where there are none today, and a canyon road has nowhere to absorb the overflow on the first warm Saturday in March. And four crossings of seasonal tributaries in the Stonybrook Creek watershed, which drains to Alameda Creek, means streambed work that four state and federal agencies weigh in on, the Army Corps among them. The filing concludes those impacts come down to less than significant with mitigation. That is a finding, and the review window exists so people can test it.

How to think about a trail that does not exist yet

None of this is approved.

Comments go to Kim Thai, principal planner, at the district's Oakland office by 5 p.m. on October 13, 2026. The board is scheduled to take it up on December 15. If it clears, construction runs about four months in one dry season, so the earliest realistic public use is 2027.

Which is the right frame for what a filing like this does to a house. A proposed trailhead is not a reason to pay more for anything. A built trailhead with real parking is. We learned that the plain way when the Skyline watershed trail opened on the Peninsula and the binding constraint turned out to be 29 parking spaces rather than the permit everyone worried about. This filing does not say how many cars the Chouinard staging area will hold, and that number will shape daily life up here more than the half mile of trail will.

So if you are shopping Fairview or the Castro Valley hills this fall, treat it the way we treat any amenity: as something to research, not something to price in. Our Hayward neighborhood rankings already reflect what is open today. Put October 13 in the calendar, read the crossing section, and let the trail earn its way into the score once it is actually there.

Sources

CEQAnet, Garin to Pleasanton Ridge Public Access Project, Mitigated Negative Declaration, SCH 2026090573, East Bay Regional Park District, received September 15, 2026

East Bay Regional Park District, Notice of Intent to Adopt a Mitigated Negative Declaration, public review September 14 to October 13, 2026

East Bay Regional Park District, Garin-to-Pleasanton Ridge Regional Trail Public Access Project page

East Bay Regional Park District, Garin/Dry Creek Pioneer Regional Parks

East Bay Regional Park District, Bay Area Ridge Trail: Garin Regional Park to Niles Canyon

Houseberry, Hayward and Union City neighborhood pages and Hayward city guide, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

There is a house in the Outer Sunset, the grid of stucco row houses between Golden Gate Park and Sloat Boulevard, with a card table at the bottom of the back stairs. Three foil trays on it, a pickup window written on an index card and taped to the gate. Saturdays, four to seven.

That setup is legal in Hayward, in San Mateo, in Concord and in Petaluma. It is not legal in San Francisco.

Six of the nine Bay Area counties already permit microenterprise home kitchen operations, the state's term for a permitted food business run out of a private home kitchen. San Francisco is one of the three that do not. Supervisor Jackie Fielder, who represents District 9, introduces an ordinance on Tuesday that would move the city across that line. An ordinance introduced is not an ordinance passed.

In brief

What a Bay Area house is allowed to earn, as of September 2026.

• Home kitchens are a county decision. Alameda, Contra Costa, San Mateo, Santa Clara, Solano and Sonoma permit them. Marin, Napa and San Francisco do not.
• The caps are specific. No more than 30 meals a day and 90 a week, and gross annual sales of $100,000 as adjusted for inflation, set at $107,121 effective January 1, 2025.
• San Francisco's ordinance gets introduced on Tuesday, September 22, 2026. It is a proposal. Nothing in it is in force today.
• An ADU is the larger number and the larger bill. Bay Area ADU rents run $2,000 to $4,200 a month against $250,000 to $500,000 to build one.
• The rule almost nobody reads is the home occupation permit. Most cities bar customers at the door and non-resident employees. Piedmont allows one room, not the garage.

Six counties said yes. San Francisco is the hole in the map.

A home kitchen permit exists in state law and nowhere in practice until a county board votes. Health and Safety Code section 114367 lets a county, or a city running its own health department, authorize the permits by ordinance. Until that vote, the permit does not exist where you live.

The Bay Area voted in no particular order. Solano went first in March 2021, Alameda in May. San Mateo County piloted from July 2021 and made it permanent in September 2023. Santa Clara County approved a pilot in March 2023. Contra Costa began issuing permits in July 2024 and made the program permanent on May 12, 2026. Sonoma County adopted its ordinance in December 2024 and began permitting the following April.

That leaves three. Napa never opted in. Marin put an ordinance on its board calendar in May 2025 and then pulled it to a later date, where it has stayed. And San Francisco has never had a program at all.

Fielder's ordinance would close the gap. The San Francisco Standard reported that it comes with the usual machinery attached: a health department inspection, food handler certification, and written landlord approval for anyone renting. The Standard puts the statewide count at roughly 1,200 permitted operations, 97 percent of which drew no complaint between 2019 and 2024.

The caps are the whole design

These programs pass because the ceiling is low on purpose. Health and Safety Code section 113825 limits a home kitchen to 30 individual meals a day and 90 a week, and to one full-time-equivalent employee who is not a household member. Food goes directly to the person eating it, never to a wholesaler or retailer.

The money cap is the one people get wrong. The statute says $100,000 in gross annual sales, adjusted yearly for inflation, and the California Department of Public Health set the adjusted figure at $107,121 effective January 1, 2025. That is gross, not profit, and it is why 90 meals a week binds before the dollar limit does. Sell 90 meals at $18 and you gross about $84,000 a year before you buy a single onion.

Renting does not disqualify you, but your landlord gets a vote, because the programs require written permission. The permit is not free either. Alameda charges $696 a year, San Mateo $1,108.

The ADU is the bigger number, and the harder one

A second unit beats a kitchen by an order of magnitude, and costs an order of magnitude more to start.

We have been through the current rules in detail, so here is the short version. California's 2026 ADU laws cut fees on small units, put a 15-business-day clock on permit completeness reviews, and dropped owner-occupancy for junior ADUs with their own bathroom. AB 956 would push cities to ministerially approve two detached units on a single-family lot instead of one. It was on the Governor's desk as of early September.

The catch is the same one we ran into in Mill Valley: a detached ADU runs about $450 a square foot in Marin, roughly $360,000 for 800 square feet. At $3,000 a month that is a ten-year payback before maintenance, vacancy or the loan. A home kitchen costs a permit fee and a weekend. An ADU is a construction project with a mortgage attached.

The Bay Area home business rule almost nobody reads

Underneath both sits an older layer, and it is the one that trips people up: the home occupation rules in your city's planning code.

San Francisco's version is typical. Its guide to home-based businesses says you cannot have clients coming to the home, cannot have employees who do not live there, cannot advertise on the building, and cannot use more than one third of the floor area commercially. Piedmont is stricter still. Its home occupation permit costs $100, allows exactly one room and not the garage, and bans outside employees and customer visits outright.

Read those lists against a home kitchen and the tension is obvious. A permitted home kitchen is, by design, a business where a stranger comes to your door to collect dinner. Any city adding one has to square that with its own planning code, a separate question from the health permit.

Three routes, three very different ceilings. Only one is a construction project.

How to think about this if you actually own the house

Start with what the house physically allows, then with what your county and city allow. A kitchen with a back entrance and somewhere for a car to pull over is a candidate for one thing. A flat rear yard is a candidate for another.

Where this shows up in a purchase decision is smaller than the headlines suggest. A permitted home kitchen will not move an appraisal. What it changes is how many ways a block can work for the person living on it, which is part of how we look at neighborhoods at Houseberry. The Outer Sunset is a fair example. Median sale price about $1.73 million in August 2026 against a citywide median near $2.2 million, and a 4.2 out of 5 overall score that ties for the top of the 92 San Francisco neighborhoods we rank. It is also block after block of single-family stock with real kitchens and real back stairs, the precondition for any of this.

Here is the caveat I would want a friend to hear first. A home business is not a mortgage strategy. Ninety meals a week at a realistic margin is grocery money, not a second income that changes what a lender will hand you. Underwriters generally want a two-year track record before self-employment income counts toward a loan at all. Buy the house you can afford on the income you already have, and treat the kitchen as an option it happens to come with.

And if you live in San Francisco, Marin or Napa, the honest answer today is that you cannot do it legally at all. Fielder's ordinance gets introduced Tuesday, then goes to committee, then to the full board. Watch the committee calendar, not the press release.

Sources

The San Francisco Standard, “Selling tacos from your backyard could soon be legal in San Francisco,” September 21, 2026

California Health and Safety Code section 113825, definition and limits for microenterprise home kitchen operations

California Health and Safety Code section 114367, local authorization of microenterprise home kitchen operations

California Department of Public Health, Microenterprise Home Kitchen Operation Adjusted Gross Annual Sales Limit, effective January 1, 2025

County of Santa Clara Department of Environmental Health, microenterprise home kitchen operations permit page

San Mateo County Health, Microenterprise Home Kitchen Operations program page

Sonoma County Department of Health Services, Environmental Health, MEHKO program page

Alameda County Health, Environmental Health Department, MEHKO program page

Local News Matters, “Contra Costa County expands opportunities for home-based cooks, food entrepreneurs,” May 12, 2026

County of Marin news release, “Microenterprise Home Kitchen Ordinance Proposed to Assist Individual Entrepreneurs,” May 1, 2025

Solano-Napa Small Business Development Center, MEHKO Solano program page

SF.gov, “Guide to opening a home-based business”

City of Piedmont, home occupation permit page

Houseberry, San Francisco neighborhood rankings, city guide and Outer Sunset neighborhood page, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Richmond's annual rent cap reset to 1.5 percent on September 1, and if you rent a covered apartment there, that is the whole increase your landlord can take this year. Drive fifteen minutes and the number changes. Oakland is at 2.3 percent. Berkeley is at 1.0. Alameda is at 2.7. Hayward allows 5. Anything the local ordinances do not reach falls to California's AB 1482 cap, at 8.8 percent.

Five East Bay cities, five different ceilings, and a state backstop almost nine times the strictest of them. Below is every East Bay rent cap we confirmed against the city's own rent program, each with the window it covers. A cap quoted without its dates is wrong inside a year.

The numbers that matter

Every figure here was checked against the agency that sets it, on September 21, 2026.

• Richmond's Annual General Adjustment is 1.5 percent for increases taking effect between September 1, 2026 and August 31, 2027, or up to 6.5 percent with banked increases.
• Oakland allows 2.3 percent from August 1, 2026 through July 31, 2027. Berkeley allows 1.0 percent for calendar 2026. Alameda allows 2.7 percent from September 1, 2026 through August 31, 2027.
• Hayward is the outlier at a flat 5 percent in any 12-month period, doubling to 10 percent with a banked increase.
• Units no local ordinance covers fall under AB 1482, capped at 8.8 percent for the San Francisco, Oakland and Hayward region from August 1, 2026 through July 31, 2027.
• Emeryville, Fremont and Union City have no local cap and rely on AB 1482. San Leandro adopted one in February 2026, effective January 1, 2027.

East Bay rent caps, city by city

The construction date decides more than the percentage does. Every ordinance below draws a line at a year built, and a building on the wrong side of it is not covered at all.

One honesty note. Oakland's 2.3 percent comes from the California Attorney General's rent cap listing, because the city's own 2026 information sheet would not load when we checked. The other four come straight off the city pages. Confirm any of these with the city before you act on it.

Why one building gets 1.5 percent and the one next door gets 8.8

The spread is not politics. It is six formulas reading the same inflation data in opposite directions.

AB 1482 is the only one of the six that adds to inflation. Five percent plus the regional cost-of-living change, capped at 10. Every local ordinance discounts inflation instead: Berkeley takes 65 percent of Bay Area CPI, Alameda 70 percent, Oakland the lower of 60 percent of CPI or a flat 3 percent.

So when inflation runs hot, the two systems pull apart fast. The state cap went from 6.3 percent last year to 8.8 this year. Richmond went down to 1.5. Same inflation, opposite directions, and 7.3 points now separate two units on either side of a 1995 permit date.

Berkeley voters also have Measure X on the November ballot, a rewrite of the Rent Stabilization Ordinance that leaves the 65 percent formula alone and adds a hard 10 percent annual ceiling. Worth its own piece.

What a cap protects, and what it does not

A rent cap protects a tenancy, not a unit. When you move out, Costa-Hawkins lets the owner reset the rent to whatever the market pays, and the next tenant starts from that number with the same 1.5 percent on top.

That is the trade-off worth naming. These ordinances keep a sitting tenant in place and do almost nothing for the person searching today. Coverage also clusters in the oldest stock, since every cutoff above is a construction date.

Richmond's citywide median sale price was about $662,000 in July 2026, the most reachable entry point on the list, and Oakland's was roughly $974,200 in August 2026. The strictest caps and the cheapest for-sale housing landing in the same city is not a coincidence. Both track age of stock. Oakland's 1983 line is not even settled law, as a federal challenge to it showed in August.

Four rent cap questions we get asked constantly

What is a banked rent increase?

An increase the landlord skipped in an earlier year and saved. Richmond, Oakland, Berkeley and Hayward all allow it, each with its own ceiling: Richmond 6.5 percent in one year, Oakland three times the current rate, Hayward 10 percent. Alameda publishes none. A jump well above the posted cap is usually banking, and usually legal.

How do I know if my unit counts as covered housing?

Two questions decide it almost every time. How many units sit on the parcel, and what year the building got its certificate of occupancy. Two or more units plus a date before the city's cutoff and you are probably covered. A detached house or a condo is not.

What happens if my unit is exempt from the local ordinance?

You are not unprotected. AB 1482 caps you at 8.8 percent through July 31, 2027, and requires a stated just cause to end the tenancy after twelve months. A much weaker ceiling than 1.5 percent, but a real one.

How do I check my own unit?

Pull your building's year built and unit count from the county assessor record, free and public, then match it to the row above. If the increase you were handed does not fit, call the city's rent program before you pay it. Richmond, Oakland, Berkeley and Alameda all staff counselors.

How to use this page

Start with two facts about your building, then read one row.

Find the year built and the unit count on the parcel. Match them to the city row. Check today's date against the window in column three, because these numbers roll over on a schedule and Berkeley's resets every January. Then run the arithmetic. On a $2,400 Richmond apartment, 1.5 percent is $36 a month. Under AB 1482, $211.

That $175 a month comes down to a permit date, which is worth knowing before you sign a lease rather than after. Same habit we push at Houseberry: look hard at the block before you fall for the listing photos. Our rankings of the 258 East Bay neighborhoods we score are a reasonable start, and the cap is one more column to carry into that comparison.

We update this page as these numbers reset. Berkeley's 2027 adjustment is next.

Sources

City of Richmond Rent Program, Rent Control and Annual General Adjustment, 1.5 percent effective September 1, 2026, https://www.ci.richmond.ca.us/3376/Rent-Increase

California Attorney General, Limits on Rent Increases, local ordinance table and AB 1482 regional caps, https://oag.ca.gov/rentcaps

Berkeley Rent Board, Annual General Adjustment, 1.0 percent for 2026, https://rentboard.berkeleyca.gov/rights-responsibilities/rent-levels/annual-general-adjustment

City of Alameda Rent Program, Annual General Adjustment effective September 1, 2026, https://www.alamedarentprogram.org/News-articles/Annual-General-Adjustment-announced-effective-September-1-2026

City of Hayward, Summary of the Residential Rent Stabilization and Tenant Protection Ordinance, https://www.hayward-ca.gov/sites/default/files/RRSO-Summary-English_0.pdf

City of Oakland Rent Adjustment Program, Allowable Annual Rent Increase Information Sheet, https://www.oaklandca.gov/Community/Housing-Programs-Support/For-Landlords/Allowable-Rent-Increases/Allowable-Annual-Rent-Increase-Information-Sheet

City of San Leandro Rent Program, Ordinance 2026-001, effective January 1, 2027, https://www.sanleandro.org/1483/Rent-Program

City of Emeryville, Tenant Protection Resources, no local rent cap, https://www.emeryville.org/Services/Housing/Tenant-Resources/Tenant-Protection-Resources

City of Berkeley, November 2026 ballot measures, Measure X, https://berkeleyca.gov/your-government/elections/ballot-measures

Houseberry, Richmond and Oakland city guides and East Bay neighborhood rankings, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Seventy-five percent. That is the share of San Francisco single-family homes above $6 million that sold entirely in cash between January and September this year, per Old Republic Title's read of public records. The median sale price in Bayview was $869,480 in July 2026. The cash wave is real, and it is breaking almost entirely at the top of the market.

Here are eight San Francisco neighborhoods where the median still starts with a zero or a one, cheapest first, and what our scores say you get for it.

First, how big is the cash problem actually

The San Francisco Standard reported on September 20 that about one in three city homes sold for cash this year, the highest share in two decades, with agents saying a financed offer no longer wins without a large premium. Redfin data reported by NBC Bay Area puts Bay Area cash sales at 30 percent from April through June.

Then there is the number nobody led with. Realtor.com's cash-buyer report, published August 18, 2026, put the San Francisco-Oakland-Fremont metro at 24.4 percent cash for the first four months of 2026, below the 31.4 percent national figure. Three of four buyers here still used a mortgage.

Both are true. Cash dominates the trophy market and thins out fast as prices fall. Realtor.com found more than 40 percent cash above $1 million nationally and a majority above $2 million. Under $1.4 million, financing is still the default.

The eight, cheapest first

• Bayview, $869,480, July 2026
• Silver Terrace, $990,300, July 2026
• Excelsior, $1.05 million, July 2026
• Visitacion Valley, $1.10 million, August 2026
• Crocker Amazon, $1.12 million, July 2026
• Oceanview, $1.15 million, July 2026
• Mission Terrace, $1.23 million, August 2026
• Ingleside, $1.33 million, August 2026

What the eight cost and how they score

Prices come from each neighborhood's own 12-month history, not the snapshot on the ranking page. Scores are out of 5.

The middle of that list costs $1.11 million, almost exactly half the $2.2 million citywide median we recorded in August 2026. The eight average 3.7 out of 5 on safety against a 3.8 citywide average, so the trade here is amenities and schools, not safety. To sort the whole city this way, our San Francisco neighborhoods ranked by median price page does it live.

1. Bayview, San Francisco

Bayview is the cheapest San Francisco neighborhood we track with a usable price history, at a median of $869,480 in July 2026. It scores 2.4 overall and 1.9 on amenities, the lowest number here. The T Third light rail runs the length of Third Street. The caution is volatility: the median moved between $869,480 and $1.05 million inside twelve months.

2. Silver Terrace, San Francisco

Silver Terrace had a median of $990,300 in July 2026, the only other neighborhood here still under seven figures. It scores 2.8 overall and 3.3 on safety, sitting on the hill between Bayshore Boulevard and Highway 101, which buys freeway access and costs quiet. Its twelve-month range ran from $824,530 to $1.04 million, a thin market that cuts both ways in a negotiation.

3. Excelsior, San Francisco

Excelsior came in at a median of $1.05 million in July 2026, down from $1.14 million a year earlier. That is one of the few genuinely falling medians in the city, which is the condition a financed buyer wants. It scores 2.6 overall and 3.4 on safety, anchored by the Mission Street strip below Silver Avenue. Schools score 2.0, the joint lowest here.

4. Visitacion Valley, San Francisco

Visitacion Valley sat at a median of $1.10 million in August 2026, up from $1.05 million a year before. It scores 2.6 overall with a 3.7 on safety, the best safety figure under $1.1 million in the city. McLaren Park forms its western edge and the T Third terminates at Sunnydale. Schools score 1.8, the lowest number in this article.

5. Crocker Amazon, San Francisco

Crocker Amazon posted a median of $1.12 million in July 2026 and scores 4.4 out of 5 on safety, 62nd of 92 overall at 3.1. It is the cheapest San Francisco neighborhood in our data that clears 4.0 on safety. Crocker Amazon Playground and the Daly City line define its south edge. Schools score 2.0, and the median swung $265,000 this year.

6. Oceanview, San Francisco

Oceanview had a median of $1.15 million in July 2026, with a 4.4 on safety and a 3.1 overall, the second neighborhood here under $1.2 million to clear 4.0 on safety. The M Ocean View line puts downtown within a one-seat ride. Amenities score 2.2, and an October 2025 print of $910,010 shows how thin some months get.

7. Mission Terrace, San Francisco

Mission Terrace finished August 2026 at a median of $1.23 million, down from a $1.46 million peak in May. It scores 3.0 overall, 68th of 92, with 3.6 on safety and a 3.2 on amenities that is the strongest here, and Balboa Park BART is the reason. A median down $230,000 in three months is the clearest negotiating window in this group, though your appraisal may not land where you expect.

8. Ingleside, San Francisco

Ingleside is the highest-scoring neighborhood on this list at 3.6 overall, 42nd of 92, and still closed August 2026 at a median of $1.33 million. Safety scores 4.2 and schools 3.5, the best school figure here by eight tenths of a point, with the Ocean Avenue corridor and City College on its northern edge. The median ran to $1.64 million in April before falling back, so this is the entry most likely to price itself out of reach first.

One band up, and still under the city median

Four more are worth a look from anyone who can stretch past $1.7 million. Outer Sunset is the headline: San Francisco's second-highest-scoring neighborhood at 4.2 out of 5, with a median of $1.73 million in August 2026, roughly $470,000 under the city figure. Outer Parkside sat at $1.74 million in July with a 4.5 on safety, Parkside at $1.99 million with a 4.6, and Outer Richmond at $2.01 million with a 4.1 on schools. None of the four scores above 3.9 on amenities.

What the cash number does and does not mean

It does mean that at the top of the market a financed offer competes against a two-week close. If you are shopping Pacific Heights or Presidio Heights, a loan is a real handicap.

It does not mean financed buyers are locked out of San Francisco. A 24.4 percent metro cash share is a market where three in four transactions run through an underwriter, and the national share is falling, from 32.3 percent to 31.4 percent year over year.

What none of it tells us is the cash share inside any one neighborhood. Nobody publishes that. This piece infers competition from price band, which is reasonable and still an inference.

How we built this list

Scores are Houseberry's own. We rate 92 San Francisco neighborhoods from 0 to 5 on overall quality, schools, safety and amenities. The full set sits on our San Francisco neighborhood ranking page. Everything here was retrieved September 21, 2026.

Prices are harder. The price column on a ranking page is a periodic snapshot and can run badly stale. Cole Valley/Parnassus Heights shows $773,609 there while its own history put the July 2026 median at $4.04 million. So every price here comes from a neighborhood's 12-month history or our San Francisco city page, with the month stated.

We pulled histories for 16 neighborhoods and threw out four as unusable: South of Market, which drops from $1.7 million to $860,000 in one month, South Beach, flat at $3.9 million all year, Bayview Heights, which prints $445,840 between a $695,900 and a $962,900, and Candlestick Point, which repeats $900,000. Four more share one identical $2,195,630 figure, the city-level fallback. We ignore the appreciation and value fields.

Questions buyers keep asking about cash offers here

Do I need cash to buy a house in San Francisco in 2026? No. Realtor.com put the San Francisco-Oakland-Fremont metro at 24.4 percent all-cash for January through April 2026, so roughly three in four buyers financed. The share climbs above $2 million and hits 75 percent above $6 million.

What is the cheapest San Francisco neighborhood right now? Bayview, at a median sale price of $869,480 in July 2026, is the lowest of any San Francisco neighborhood we track with a usable price history. Silver Terrace is next at $990,300, also July 2026.

Which affordable San Francisco neighborhood scores best on safety? Crocker Amazon and Oceanview both score 4.4 out of 5, the highest of any San Francisco neighborhood priced under $1.2 million. Their July 2026 medians were $1.12 million and $1.15 million.

Why does the price on your ranking page not match this article? The ranking page price is a periodic snapshot that can lag the market by about a year. Every figure here comes from a neighborhood's own 12-month price history, dated by month.

The practical move

If your budget is under $1.4 million, the cash story in the headlines is mostly somebody else's problem. Pick two or three names off this list, watch how their medians move, and write on the one whose comps are falling rather than the one everyone is talking about.

Sort the city yourself: open our San Francisco neighborhood rankings and put price next to safety and schools before you spend a Saturday on open houses.

Sources: The San Francisco Standard and Old Republic Title, September 20, 2026. Realtor.com cash-buyer report, August 18, 2026. Redfin via NBC Bay Area. Houseberry San Francisco rankings and price histories, retrieved September 21, 2026.

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

San Francisco rewrote the rules inside one rectangle, bounded by Market, Townsend, 2nd and 6th, so the blocks in the middle could finally get built. The Central SoMa Plan passed the Board of Supervisors unanimously on December 4, 2018. Six years on, one of its key sites is on Thursday's Planning Commission agenda, and it is not there to break ground. It is there to ask for eighteen more months.

The sponsor at 725-765 Harrison Street wants its Large Project Authorization pushed to March 24, 2028, carrying a 981,000-square-foot approval with it. Nothing is decided. This is a request at items 7a, 7b and 7c on Thursday's September 24, 2026 Planning Commission agenda. Staff recommends approval with conditions. The vote is at noon in Room 416.

Attached to that approval is a roughly 15,000-square-foot slice of the site promised to the city for affordable housing. It has been waiting since 2019, and it waits another 18 months with everything else.

Key takeaways

• Items 7a, 7b and 7c on the September 24, 2026 Planning Commission agenda, Case Nos. 2005.0759ENX-03, CUA-03 and VAR-04, ask to extend the 725-765 Harrison entitlement 18 months to March 24, 2028. It is a request, not an approval.
• The approved project is 14 stories, 185 feet and about 981,000 gross square feet, 810,000 of it office. Approved December 12, 2019, modified June 16, 2022.
• A roughly 15,000-square-foot land dedication to the Mayor's Office of Housing rides on the same entitlement. It was meant to produce 144 affordable homes.
• San Francisco office vacancy was 29.2 percent in the second quarter of 2026, per CBRE, against roughly 3 percent when the plan was adopted.
• We score South of Market 2.1 out of 5 overall, 90th of the 92 San Francisco neighborhoods we rank, against a citywide 3.5.

The plan was written for a city that had run out of office space

Central SoMa was the city's bet that the blocks between the Financial District and the ballpark could absorb a generation of growth.

The adopted plan was projected to deliver close to 16 million square feet of housing and jobs, 33 percent affordable, and more than $2 billion in public benefits including $500 million for transit. The Central Subway was going in underneath it.

Office vacancy then was about 3 percent, as Anne Taupier of the Office of Economic and Workforce Development reminded the San Francisco Examiner in March 2025. It is now 29.2 percent citywide, per CBRE's second-quarter 2026 figures. Better than it was, and the market absorbed almost 964,000 square feet net that quarter, mostly on AI leasing. We wrote about where those tenants are landing. But 29 percent does not finance a new 810,000-square-foot tower, and the Board of Supervisors conceded as much on March 18, 2025, when it revised the plan to stop requiring that most of a large project's floor area be commercial.

What 981,000 square feet was supposed to hold

The approved building is a plan-era artifact and the proportions show it. Office is 83 percent of the floor area. Everything the neighborhood uses day to day fits in what is left.

That table is the deal Central SoMa struck. A key site got 185 feet and 14 stories, well above the old zoning, in exchange for PDR space, child care, a plaza and land for housing. The height was the carrot, the rest the price, and neither has been collected.

The entitlement expires the day of the hearing

Do the arithmetic. Eighteen months back from March 24, 2028 lands on September 24, 2026. The authorization runs out the day the Commission meets to consider extending it.

The original approval came December 12, 2019, with an initial 505,000-square-foot office allocation. Boston Properties, the sponsor then, said in a December 19, 2019 release that "construction could commence in the second half of 2020." It did not. The approval was modified June 16, 2022 to add 40,000 square feet of office, and the site went quiet.

The 15,000 square feet nobody is talking about

The part of this item that matters most to housing gets one clause in the agenda. The site carries a land dedication of roughly 15,000 square feet to the Mayor's Office of Housing, and that parcel was supposed to become 144 affordable homes.

The city already picked developers. In a July 8, 2021 announcement, the Mayor's Office of Housing and Community Development awarded 725 Harrison to HomeRise and the Tenderloin Neighborhood Development Corporation, for permanent supportive housing for formerly homeless San Franciscans in the Filipino Cultural Heritage District. TNDC's project page still lists it as planned, last updated July 2021.

Five years in predevelopment, because the land under it belongs to an office project that never broke ground. I could not confirm the exact trigger that transfers the parcel, and the agenda does not spell it out, so treat that as Thursday's open question for staff. What is clear is that the dedication is a condition of this entitlement, and this entitlement is what expires.

I want the office building. More jobs in Central SoMa is the point of the subway running under it. But if the tower is years from financing, the better outcome is not another extension in 2028. It is getting the 144 homes loose from the office they are hostage to.

What a frozen pipeline looks like from the sidewalk

People live in Central SoMa now, and the stall is their daily experience, not a capital-markets abstraction.

South of Market, the industrial-turned-mixed-use district the Central SoMa Plan sits inside, scores 2.1 out of 5 with us, 90th of the 92 San Francisco neighborhoods we rank. Safety is 1.9, schools 2.0, and only amenities at 3.2 comes near the citywide line. South Beach, the waterfront stretch just east of it around the ballpark, scores 3.1 overall with a 3.8 on amenities. Two neighborhoods, one street grid apart, a full point between them.

The gap is about what got finished. South Beach got its towers twenty years ago. South of Market got a plan. A 1.9 on safety describes surface lots, roll-up doors and blocks with nobody on them after 6pm, the exact condition more housing would change.

Meanwhile the citywide median sale price reached about $2.20 million in August 2026, up from roughly $1.89 million in September 2025. The city keeps getting more expensive while its largest planned district produces nothing. A low score with a big approved pipeline behind it is a very different bet from a low score with nothing coming, and Central SoMa has been the first kind for six years.

Four questions people are asking about the Harrison Street extension

Does this mean 725 Harrison is cancelled?

No. An extension request is the opposite of a cancellation. It is the sponsor keeping a valid approval alive instead of letting it lapse, which is what walking away looks like.

What happens if the Commission says no?

The authorization expires and the project loses its entitlement. Rebuilding it means going back through Planning under today's rules, which allow far more housing here than the 2019 approval did. Staff recommends approval, so this branch is unlikely.

Are the 144 affordable homes at 725 Harrison dead?

Not dead, not moving. HomeRise and TNDC were selected in 2021 and the project is still listed as planned. The dedication is tied to this entitlement, so 18 more months means 18 more months waiting.

Would a new proposal here have to be office?

No longer. The Board removed the commercial-square-footage mandate for large Central SoMa projects on March 18, 2025, which is why a rethink of this block could come back mostly residential.

What to watch on Thursday

The vote is the least interesting part. Staff recommends approval, the Commission usually follows on extensions, and 725-765 Harrison will most likely have until March 2028.

Listen for whether a commissioner asks what the agenda does not: what has to happen before the housing parcel transfers. The same agenda carries GLIDE's new seven-story, 41,173-square-foot facility at 330-334 Ellis Street, replacing a parking lot. One item is a building somebody is ready to build. The other is a promise asking for more time.

If you are shopping SoMa or South Beach, the read is simple. Approved square footage is not built square footage, and the gap here is years wide. Our San Francisco neighborhood rankings show how the finished parts of the city compare while the planned parts wait. The hearing starts at noon Thursday in Room 416 at City Hall.

Sources

San Francisco Planning Commission, Notice of Hearing and Agenda, September 24, 2026, Items 7a, 7b and 7c, Case Nos. 2005.0759ENX-03, CUA-03 and VAR-04

San Francisco Planning Commission, agenda for December 12, 2019, Case 2005.0759ENX/OFA/VAR, 725 Harrison Street

BXP (Boston Properties), "Boston Properties Receives Approval From San Francisco Planning Commission for 4th and Harrison Development Project in Central SoMa District," December 19, 2019

SF Planning, Central SoMa Plan program page

San Francisco Examiner, "SF revamps Central SoMa Plan amid new economic reality," March 18, 2025

CBRE, San Francisco Office Figures, Q2 2026

HomeRise and Tenderloin Neighborhood Development Corporation, "HomeRise and TNDC to build permanent supportive housing in SoMa," July 8, 2021

Tenderloin Neighborhood Development Corporation, 725 Harrison property page, last updated July 15, 2021

San Francisco YIMBY, "Modifications Proposed for 725 Harrison Street, SoMa, San Francisco," May 2022

Houseberry, San Francisco neighborhood pages, city guide and rankings, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Start with the division. Palo Alto owns a 0.67-acre surface parking lot at Hamilton Avenue and Waverley Street, two blocks south of University Avenue. Five blocks northeast, on a 0.43-acre city lot at Lytton and Kipling, the city has already approved 72 apartments. That is 167 homes per acre, a density Palo Alto set itself, on its own land, this year.

Run it across downtown. 0.67 acres times 167 homes per acre is about 112 homes.

On September 14, 2026, the City Council approved something else for that parcel: a six-story public parking garage with roughly 265 stalls, at a cost of about $41 million.

There is a housing piece and it is small. A temporary plaza on the east end of the site is set aside for 15 or 16 future affordable apartments, none of them funded, designed, or committed. Call it 16 homes where 112 fit.

The easy version of this is that Palo Alto chose cars over homes last Monday night. That version is wrong, and the real reason is stranger.

Key takeaways

• The Palo Alto City Council unanimously approved a six-story, roughly 265-stall public garage at 375 Hamilton Avenue on September 14, 2026, on a 0.67-acre city-owned lot, parcel 120-15-086.
• The garage costs about $41 million, which is roughly $155,000 per parking stall.
• At the 167 homes per acre Palo Alto approved at 450 Lytton Avenue in February 2026, the same 0.67 acres would hold about 112 homes.
• The city used an addendum to its 2019 environmental impact report rather than new review, finding no new significant effects.
• Palo Alto had permitted 680 of the 6,086 homes in its 2023 to 2031 housing element as of December 2024. University South, the neighborhood around the lot, scores 3.5 out of 5 with us, 26th of 27, at a median sale price near $4.49 million in August 2026.

Where 167 homes per acre comes from

The density figure is not mine. It is Palo Alto's, from a project the city approved seven months ago on nearly identical land.

Lot T is the 0.43-acre city-owned surface lot at Lytton Avenue and Kipling Street, five blocks northeast. Its planning file at 450 Lytton Avenue puts the parcel at 19,700 square feet and the project at 72 deed-restricted homes in six stories and about 76,400 square feet, replacing a 52-space public lot. Alta Housing is building it, and we covered it when a county judge threw out all seven causes of action in the lawsuit against it.

So: 72 divided by 0.43 is 167.4 homes per acre. Multiply by 0.67 and you get 112.2.

Check it a second way. Lot T averages about 1,061 square feet of building per home. The Hamilton garage is about 116,450 square feet. Divide and you get roughly 110 homes.

The garage is the bigger building, roughly 40,000 square feet larger than the housing five blocks away, and it holds zero bedrooms.

What the council actually approved on Monday

The Notice of Determination filed with the state clearinghouse under SCH number 2017052040 describes a six-story garage of approximately 265 stalls with accessible spaces and EV charging, plus bicycle parking, landscaping, stormwater features, and a public plaza.

The rest, from the city project record and the September 14 council coverage: 52 bike spaces, a roof built to carry solar later, a pedestrian paseo, public art, and a narrowing of Hamilton from 50 feet to 43 to fit bike lanes both ways. The existing lot holds 86 cars, so the net gain is 179 spaces. Construction runs 18 to 24 months from late this year.

One wrinkle in that table. The stall count moves by document. The state filing says approximately 265, the plan set says 266, and the Public Works page still says 274. I use 265 because it is in the approval filing.

Addendum sounds like paperwork and is doing real work here. Palo Alto certified a full environmental impact report for this garage in 2019. An addendum says the project changed, but not enough to create any new significant effect the old report missed, so no new review is required. Legal, common, and also how a project designed in a different decade gets approved in 2026 without anyone re-asking whether a garage is the right use of the land.

The case for the garage is better than housing advocates admit

I am going to argue against this project shortly, so here is the strongest case for it first.

Downtown parking really is tight at peak, and the commercial core really does depend on people arriving by car from places Caltrain does not serve. The lot is asphalt. No home is demolished, no tenant relocated, no canopy lost. The plaza is about 4,048 square feet of real public space on a block that has none.

Then the argument that gives me the most trouble. Structured parking is what makes surface lots politically available for housing. Lot T removes 52 public spaces. The next lot removes more. Downtown merchants went along with Lot T partly because a garage was coming. Kill the garage and you may lose the coalition that let 72 affordable homes survive a lawsuit.

That is the honest trade. I still think 112 homes beat 179 net parking spaces here, but anyone calling the garage pure waste has not sat through a downtown parking meeting.

The money could only buy parking, and that is the actual story

Nobody on the council had to weigh homes against stalls, because the money involved cannot legally become homes.

The garage is paid for out of Palo Alto's parking in-lieu fee fund, set up in 1995. Downtown developers who did not want to build parking on their own sites paid into it instead. Under California's Mitigation Fee Act, that money must be spent on what it was collected for, and promptly.

Palo Alto learned the second half expensively. In Hamilton and High LLC v. City of Palo Alto, the Sixth District Court of Appeal found the city had held in-lieu fees more than five years without the findings the statute requires, and ordered it to refund $906,900 to developer Chop Keenan in March 2023. Build parking or hand the money back.

So the September 14 vote was barely a land-use decision. It was made in 1995, collected one permit at a time, and enforced by an appellate court in 2023. The land came along for the ride.

$41 million divided by 265 stalls is about $155,000 per parking space. If a fix exists it is upstream of Palo Alto, in what in-lieu parking money is allowed to pay for in a state that now wants homes on exactly these parcels.

Palo Alto is 11 percent through a housing element with six years to go

The timing is what makes this parcel sting. Palo Alto carries a 2023 to 2031 housing element target of 6,086 homes, 1,556 of them very low income, and had permitted 680 in total as of December 2024, about 11 percent, with three-quarters of the cycle left.

One hundred and twelve homes on one downtown lot would have been 1.8 percent of the entire eight-year obligation, from a parcel the city already owns, with no land to buy and nobody to relocate.

Meanwhile the state keeps pointing at these blocks. When SB 79 briefly overrode local zoning within a half mile of Caltrain stations this summer, developers filed seven applications for about 253 apartments in a two-week window, clustered in north Palo Alto rather than the El Camino corridor the city planned around. Downtown is where the transit is, so downtown is where the pressure goes. The same gravity put 98 apartments at Cal Ave in August.

What this changes for anyone buying near University Avenue

Not the schools, and not the price. What changes is the next two years of that block, and the skyline after that.

375 Hamilton sits in University South, the downtown neighborhood covering the blocks south of University Avenue between Alma and Middlefield. It rates 3.5 out of 5 overall with us, 26th of the 27 Palo Alto neighborhoods we rank, with 4.7 on schools, 3.7 on amenities, and 2.3 on safety. Top of the city on schools, near the bottom on safety. That split is the most useful thing to know before shopping here.

Prices have not softened the way the citywide line has. University South closed August 2026 at a median sale price near $4.49 million, while the Palo Alto citywide median fell to about $4.08 million that month after peaking near $4.54 million in May.

For people comparing neighborhoods, the questions are narrow. Within two blocks of the site, expect construction noise into 2028, then a six-story wall where an open lot used to be, which matters if you are buying a unit facing west. Expect parking to get easier, which counts for something if you drive.

The way we look at it when we compare neighborhoods at Houseberry, a block like this is a set of trade-offs you can see coming. None of it makes University South better or worse than the top-ranked Palo Alto neighborhoods. It changes what you go look at before you write an offer.

The bottom line: two city lots, five blocks apart

Put the two parcels side by side.

Lot T is 0.43 acres. It cost the city 52 parking spaces, a four-year fight, a lawsuit from a downtown property owner, and an August court ruling to survive. It produces 72 permanently affordable homes, earliest construction 2028.

375 Hamilton is 0.67 acres, half again as large. It cost nobody anything politically, passed unanimously, and breaks ground within months. It produces 179 net parking spaces, about 16 homes if a future council funds them, and a $41 million bill.

Per acre, Lot T delivers 167 homes. Hamilton and Waverley delivers 24, and only on paper. Same downtown, same owner, same year, same council.

The lot Palo Alto had to fight for became housing. The lot nobody fought over became parking. That tells you where the pressure to build in this city actually comes from, and it is not from inside City Hall.

Sources

CEQAnet, Notice of Determination, SCH 2017052040, City of Palo Alto Downtown Parking Garage at 375 Hamilton Ave, approved September 14, 2026, posted September 18, 2026

City of Palo Alto, 375 Hamilton Ave project record, applications 26PLN-00050 and 25PLN-00277

City of Palo Alto, Downtown Parking Garage Project, Public Works Engineering Services

City of Palo Alto, 450 Lytton Ave project record, application 26PLN-00016

Palo Alto Daily Post, “Council approves 6-story parking garage at Waverley and Hamilton,” September 14, 2026

Hoodline, “Palo Alto advances $41 million garage after parking fee dispute,” September 2026

Palo Alto Daily Post, “Court orders city to return $906,900 in parking fees to developer Chop Keenan,” March 20, 2023

Hamilton and High LLC v. City of Palo Alto, California Court of Appeal, Sixth Appellate District

SV@Home, City of Palo Alto RHNA allocation and permitting progress, 2023 to 2031 cycle

Houseberry, Palo Alto neighborhood rankings, University South neighborhood page, and Palo Alto city guide, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The entire environmental review for 16 new homes in Santa Clara is one state form. A few boxes, a statute number, a planner's signature, filed to the California clearinghouse on September 15, 2026 under the name 860 Civic Center Townhouses. No environmental impact report sits behind it. No mitigated negative declaration. Nothing does, and all of that is legal.

I went looking for coverage and found none. Not a story, not a newsletter, not a comment thread. For a Santa Clara CEQA exemption that clears a whole block of housing, the filing is the only record there is.

Key takeaways

• Santa Clara approved 16 townhouses at 860 Civic Center Drive with no environmental review, under Public Resources Code section 21080.66. The notice was posted September 15, 2026 as clearinghouse number 2026061275.
• Three buildings hold eight, six and two homes: 42,007 square feet of housing plus 715 of commercial on 0.6 acres, about 27 homes per acre.
• The site is in Santa Clara East Central, the grid around Santa Clara University, which we score 4.0 of 5 on amenities, best of the city's seven.
• Fourteen months ago a nine-home townhouse project on a half acre in the same city had to produce a mitigated negative declaration.

Three buildings, and the smallest one is a duplex

The filing does not describe one building. It describes three, wildly different in size. Three narrow parcels, APNs 224-29-009, 224-29-010 and 224-29-023, were assembled into a single 0.6-acre project, and the buildings follow the old lot lines rather than one clean footprint.

Look at the smallest of the three. That building holds two homes. California's biggest change to environmental law in fifty years was applied, in part, to a duplex.

The density is in the lot, not the floor plans

These are not small homes. Divide 42,007 square feet of residential floor area by 16 and you get about 2,625 square feet per home, garages and stairwells included.

So the 27 homes per acre comes from the land, not from shrinking the units. The buildings run about 1.6 times the square footage of the lot they sit on, a genuinely urban ratio on a street of one-story houses. That is the part a neighbor notices.

Why 16 townhouses in Santa Clara need no CEQA review

Because the site clears a checklist, and clearing it deletes the environmental document rather than shortening it.

Section 21080.66 arrived with AB 130 in June 2025. A project has to sit on 20 acres or less inside an already developed urban area, match the local zoning, stay off farmland and fault zones, and beat a density floor set at half what the state expects of the city. Per Holland & Knight's reading of the statute, that underlying figure runs from 10 to 30 homes per acre. At about 27, this one clears the top of the range. We walked through the same statute when it cleared 15 condominiums in Menlo Park earlier this month. The filing does record tribal consultation with the Tamien Nation and the Muwekma Ohlone Tribe, completed July 31, 2025, the one piece of process the exemption does not erase.

Lafayette and Civic Center Drive, block by block

The site is a block off El Camino Real and a 12-minute walk from a train platform, which is most of the reason it qualifies.

860 Civic Center Drive sits in Santa Clara East Central, the college-town grid of 1920s bungalows and mid-century houses around Santa Clara University, with Lafayette Street and El Camino Real as its working arterials. Walk Score rates the address 68 for walking and 53 for transit. The Santa Clara Transit Center, where Caltrain and Capitol Corridor call, is half a mile off.

East Central is the strongest amenities neighborhood in Santa Clara on Houseberry's scores, 4.0 out of 5, and 3.6 overall, third of seven. Its median sale price was about $1.81 million in August 2026, just under the $1.86 million citywide median that month.

Sixteen homes will not move that number and I am not going to pretend otherwise. What a filing this size tells a buyer is direction, not magnitude, and University Station's 408 homes are going in across the tracks on the same corridor.

Nine homes got a study. Sixteen got a form.

In July 2025 Santa Clara finished a mitigated negative declaration for nine townhomes on half an acre at Cheeney Street and Agnew Road. Same city, same product, a lot within a tenth of an acre of this one. That project got a study, mitigation measures and a monitoring program. Fourteen months later a bigger and denser townhouse project on a comparable lot got a checkbox.

I think that is the right trade. A nine-unit environmental study in a built-out Silicon Valley city was never going to find anything, and the months it burned came out of whether small infill pencils at all.

But notice what actually changed. The exemption did not unblock a tower. It deleted a study from a project small enough that nobody was watching in the first place. That is what this law looks like on a Tuesday in Santa Clara: 16 homes on a lot you could walk across in twenty seconds, recorded on a page no one will read. The filings are the whole story now, which is why our Santa Clara neighborhood rankings sit next to them.

Sources

CEQAnet, Notice of Exemption, SCH 2026061275, 860 Civic Center Townhouses, City of Santa Clara, posted September 15, 2026

CEQAnet, Notice of Determination, SCH 2024120531, Cheeney Street Townhomes Project, City of Santa Clara, July 16, 2025

Holland & Knight, “California Legislature Enacts Major CEQA Reforms for Housing-Rich Infill Projects,” July 2025

Walk Score, 860 Civic Center Drive, Santa Clara, retrieved September 21, 2026

Houseberry, Santa Clara city guide and neighborhood rankings, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Eighty-two bills signed. Twenty-four vetoed. One hundred and six pieces of California law settled on a single Thursday, September 18, and nobody we can find has pulled the housing out of that list. So we counted it ourselves, off the Governor's own published update. Four of the 82 signings change how a Bay Area home gets inspected, financed or foreclosed on. Two of the 24 vetoes tell builders where the state's help stops.

Earlier this month we listed the eleven housing bills sitting on Newsom's desk until September 30. Here is the first thing worth knowing about Thursday: none of those eleven moved. What went through instead was a quieter set that never got a press release, and one of them matters to anyone who has taken a Tuesday off to wait for an inspector.

Key takeaways

• 82 signed, 24 vetoed, all on September 18, 2026. We counted both lists ourselves off the published update.
• Four signings are housing bills: SB 1072 (housing omnibus), AB 1738 (remote inspections), AB 1802 (mitigation land endowments) and AB 1957 (foreclosure bidding).
• AB 1738 is the one homeowners will feel. Most California cities must offer remote inspections on water heaters, heat pumps, reroofs, small solar and detectors. Bay Area cities have until January 1, 2028.
• Two vetoes hit production: AB 2166, a state backstop for offsite housing factories, and SB 954, a CEQA exemption for advanced manufacturing.
• Nine days left, and all eleven bills from our September 2 list are still on the desk.

AB 1738 is the bill you will actually feel, and not until 2028

AB 1738 requires almost every California city to let a homeowner or a contractor ask for a remote inspection instead of waiting for a van. It is now Chapter 271 of the Statutes of 2026. It does not bind your city until January 1, 2028.

The covered work is small and common: heat pump water heaters, heat pump HVAC, reroofs, rooftop solar at 15 kilowatts or less with battery storage, and smoke and carbon monoxide detectors. A department can work by videoconference or from photos the homeowner sends, and the inspector picks which. The remote option has to come at no greater cost and with no greater delay than the in-person one, the clause that stops a city from complying on paper and burying it in practice.

The inspection itself takes five to fifteen minutes. The wait around it takes hours. SPUR, which backed the bill, put contractor standby cost at $600 to $1,000 per inspection in February 2026, and found that only 19 of California's roughly 540 permitting jurisdictions offer remote video inspections today. Four are here: Berkeley, San Mateo, San Ramon and Santa Rosa.

Santa Rosa already offers it, and with a citywide median sale price of about $742,200 in June 2026, a thousand dollars of standby time is not a rounding error. In Bennett Valley, the Sonoma County neighborhood southeast of downtown that we score 3.7 out of 5 overall and 5.0 on safety, the August 2026 median was about $775,000.

The deadlines are tiered. Cities over 50,000 people comply by January 1, 2028, smaller ones by July 1, 2028. Cities under 5,000 and counties under 150,000 are exempt, so no Bay Area county escapes and only the smallest towns do.

The honest objection came from the building officials. CALBO opposed AB 1738 on the ground that a camera cannot find a gas leak or a bad splice the way a person standing in the room can. Not a frivolous complaint. It is why the method stays at the inspector's discretion.

The six bills, side by side

The whole housing slice of Thursday, in one place.

Everything signed Thursday is January 2027 law at the earliest, and the one people will notice is 2028.

SB 1072 and AB 1957 are the quiet ones with teeth

Two of the other three do more than housekeeping.

SB 1072 is the Housing Committee's annual cleanup bill, and most of it reads like one. Buried in it is a real requirement. An owner of subsidized housing must now notify affected public entities at least twelve months before ending a subsidy contract or letting rent restrictions expire. Twelve months is enough time to assemble a preservation deal. Ninety days is not.

AB 1957 reopens the post-foreclosure bidding rules California wrote in 2020 to keep investors off the courthouse steps. It narrows the preference to tenant bidders and community buyers, drops prospective owner-occupants, extends the window before a trustee's sale is final, and adds civil penalties. Foreclosures cluster when they happen, and this rule decides who ends up owning the block.

AB 1802, by San Francisco's Catherine Stefani, is plumbing. It deletes a January 1, 2027 repeal date so agencies and nonprofits can keep holding endowments for mitigation land, which any project buying habitat offsets needs somebody to do.

The vetoes tell modular builders to find their own backstop

AB 2166 would have put the state's credit behind surety bonds for offsite housing factories, and Newsom said no.

Factory-built housing keeps stalling in California for one unglamorous reason. A factory needs payment and performance bonds to win multifamily contracts, sureties will not write them for a young manufacturer with no completed projects, so the track record never starts. AB 2166 would have had the California Infrastructure and Economic Development Bank stand behind those bonds, subject to appropriation. Newsom also vetoed AB 2012 the same day, which would have eased hauling manufactured homes on state highways.

We would like to give you his reasoning. The veto messages for AB 2166 and SB 954 are posted as scanned images with no readable text, so we cannot quote either and will not guess.

Nine days, eleven bills, one desk

The deadline is September 30, and everything we flagged on September 2 is still sitting there.

AB 956 on two detached ADUs. SB 1117 on impact fees. SB 1116 on starter home subdivisions. AB 2074 on transit hub districts. The two insurance bills and the three HOA bills. Eleven in all, none touched. A bill the Governor neither signs nor vetoes by the deadline becomes law anyway, so silence is not rejection, but it is a strange way to find out.

For anyone weighing a project before January, the read is short. Nothing signed Thursday speeds up a permit this year, and the remote inspection you want is fifteen months out unless you live in one of those four cities. Permitting friction is the layer people skip when they compare places at Houseberry, because it stays invisible until you are the one living in it. The Santa Rosa neighborhood rankings are one place that comparison starts.

Check back in ten days. The bigger list resolves itself either way.

Sources

Office of the Governor of California, “Governor Newsom issues legislative update 9.18.26,” September 18, 2026. Signed and vetoed bills counted directly from the published list.

CalMatters Digital Democracy, bill records for AB 1738 (Chapter 271), SB 1072 (Chapter 322), AB 1802 (Chapter 274) and AB 1957 (Chapter 279), Statutes of 2026.

California State Assembly, Committee on Local Government analysis of AB 1738, April 15, 2026.

California Building Officials (CALBO), position statement on AB 1738 (Carrillo).

SPUR, Colleen Corrigan, “How Virtual Construction Inspections Can Address California’s Housing Crisis and Meet Clean Energy Goals,” February 18, 2026.

LegiScan, text of AB 1738 as amended July 2, 2026, Health and Safety Code section 17970.9.

CalMatters Digital Democracy, bill record for AB 2166 (vetoed) and SB 954 (vetoed), September 18, 2026.

Houseberry, Santa Rosa city guide, Bennett Valley neighborhood page and Santa Rosa neighborhood rankings, retrieved September 21, 2026.

Houseberry, “Eleven Housing Bills Now Sit With Newsom Until September 30,” September 2, 2026.

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The document the Port of Oakland's board took up at its September 10 meeting does not authorize a stadium. It authorizes a conversation. Commissioners voted 6 to 0, with one member absent, to open a one-year exclusive negotiating window over Howard Terminal, the 50-acre former marine terminal at 1 Market Street on the estuary edge of Jack London Square. What the Oakland Roots got is twelve months of a locked door and the right to come back with a real plan. The A's held the identical piece of paper in April 2018.

The useful question is what quietly changed about the site since the last time this happened. One thing did, and it barely made the coverage. The roughly 3,000 homes that came with the ballpark plan are gone. Not trimmed. Gone, and gone before the Roots ever filed a bid.

Key takeaways

• Port of Oakland commissioners voted 6 to 0 on September 10, 2026 to enter one-year exclusive negotiating agreements covering Howard Terminal, the 50-acre site next to Jack London Square.
• The board approved agreements with both finalists, not one: the Oakland Roots and Soul Sports Club and the Los Angeles industrial developer Industrial Realty Group. Each pays the Port $25,000.
• The Roots have 90 days to file a refined project description, site plan and proposed acreage, then the balance of the year for an entitlement strategy and a financial feasibility report.
• The A's 2021 plan for the same ground carried about 3,000 homes. The Port's September 2024 solicitation did not ask for housing, and this round has none.
• Prescott, the West Oakland neighborhood inland of the site, had a median sale price of about $630,000 in July 2026, roughly 31 percent under Oakland's citywide median of about $919,000 that month.

What the commissioners actually voted on

An exclusive negotiating agreement is a promise not to shop the site to anyone else while a developer does homework. It is not an approval, a lease, an entitlement, or a building permit.

The deliverables are the interesting part. Inside 90 days, each party owes the Port a refined project description, a site plan, the acreage it wants, and parking and access plans. Over the rest of the year they owe an entitlement strategy, a financial feasibility report, option terms and a capital improvements schedule. SFist reported the 6 to 0 vote and the $25,000 fee on September 19.

Financial feasibility is the line to watch. That is where waterfront stadium plans usually stop, and where the last one stopped.

The word exclusive is carrying less weight than the headlines suggest

Both finalists got one. The Port is running two incompatible ideas on the same 50 acres for a year to see which survives contact with a spreadsheet.

The Roots want a stadium. Industrial Realty Group wants maritime commodity handling and logistics, which is the use the Port has quietly preferred here for most of the last decade. One of the year-one deliverables is confirming how much acreage each party actually gets, which is a polite way of saying the Port has not decided. About four acres in the southwestern corner are spoken for regardless, reserved for the Inner Harbor Turning Basin widening.

The vote still matters. But "Roots land exclusive deal" describes a starting gun, not a finish line.

The A's held this same piece of paper in 2018

This is the second exclusive negotiating agreement at Howard Terminal in eight years. The first one got much further.

Port commissioners approved a one-year exclusive negotiating agreement with the A's on April 26, 2018. The plan that grew out of it was enormous: a 34,000-seat ballpark, about 3,000 homes, 1.5 million square feet of commercial space, and a performance venue. Oakland's City Council approved a term sheet 6 to 1 to 1 in July 2021 and certified the environmental impact report 6 to 2 in February 2022. The Bay Conservation and Development Commission then voted 23 to 2 to lift port priority use from the site.

Then it died anyway. The A's bought land in Las Vegas in April 2023, Oakland ended negotiations, the agreement lapsed that May, and the team is now playing in Sacramento on the way to Nevada. The wall was never zoning or CEQA. It was who pays for the roads, the utilities and the grade separations on a site that has none of them. The A's cleared harder procedural hurdles than the Roots have even reached, and the ground is still asphalt.

The 3,000 homes came off the table before the Roots ever bid

Housing is missing from the Roots proposal because the Port never asked for it. The club did not leave it out.

When the Port reopened the site in September 2024, the solicitation said it would consider maritime, light industrial, commercial, recreational and water-oriented uses. Read that list again. Residential is not on it.

The reason is a piece of nineteenth-century property law that governs most of the Oakland shoreline. Howard Terminal is public trust land, granted by the state to the City and administered by the Port, and the trust limits it to commerce, navigation, fisheries and related public uses. Housing does not qualify. The A's did not argue their way around it. They got a statute. AB 1191, carried by then-Assemblymember Rob Bonta in 2019, authorized a State Lands Commission trust exchange for the Howard Terminal property so housing could legally go there.

Nobody has asked the Legislature for a second one. Until somebody does, the answer to "will there be housing at Howard Terminal" is no, and it is a legal no, not a political one.

I think that is a genuine loss and I am not going to soften it. Three thousand homes on 50 flat acres where nobody lives, half a mile from West Oakland BART and walking distance to the Amtrak platform at Jack London Square, is close to the easiest housing argument anyone in the East Bay will ever get to make. The state already wrote the law once.

The counterargument deserves its due. A privately financed soccer stadium is something a mid-sized club can actually build, and Oakland has spent fourteen years learning what the alternative looks like. We have written about a West Oakland site approved for 1,032 homes that now wants to be a parking lot while it waits for financing. Approvals are not buildings.

What people are paying on either side of the site right now

Both neighborhoods flanking Howard Terminal sit near the bottom of our Oakland ranking, for different reasons. Here is where they stand before anything is built.

The blank in that table is a finding. Produce and Waterfront, the strip between the Embarcadero rail tracks and the estuary that contains Jack London Square, does not sell enough homes in a year for us to publish a neighborhood median. There is almost nothing there to buy, because there is almost nothing there to live in. That is the condition a stadium would be dropped into.

The scores tell the rest. Produce and Waterfront rates 3.3 of 5 on amenities, which is Yoshi's and Drake's and the Sunday farmers market doing real work, and 1.6 on safety, the lowest number on this page. Prescott, the West Oakland neighborhood most people call the Lower Bottoms, inverts it: 2.6 on safety, 2.9 on amenities, Victorians on flat, bikeable blocks. Same overall score, one rank apart, almost nothing in common.

West Oakland's discount is the number a stadium would move first

West Oakland has spent the last year trading between 21 and 31 percent below the Oakland median, and that gap is the thing worth watching, not the stadium renderings.

Prescott closed July 2026 at about $630,000, roughly 31 percent under the citywide figure. Oak Center, West Oakland's historic preservation district and the largest in the city with something like 500 restored pre-1925 houses, sat at about $729,000. Oakland citywide bottomed at roughly $785,000 in April 2026 and recovered to about $974,000 by August.

West Oakland is not tracking the city. It is its own market, discounted for reasons a soccer stadium does not touch: freeway and rail proximity, soil history, schools scoring 1.5 of 5, and a century of industrial land use. A 25,000-seat venue a mile away changes none of that.

What it could change is the block-by-block edge closest to the water, the way Brooklyn Basin did on the other side of Jack London Square. That is a narrow strip, and it is where speculation lands first. Treat a stadium headline as a reason to look harder at the specific block, not the zip code.

The honest case for not getting excited

The Roots have a stadium problem before they have a site problem, and the calendar is unkind.

Their Coliseum lease runs out on December 21, 2026. The published concept is two phases: a modular 15,000-seat venue first, estimated at roughly $129 million, then a permanent 25,000-seat stadium designed by HOK. Club leadership has said it would be paid for principally through private sources of equity, with some site infrastructure possibly eligible for public money. A more honest financing posture than the A's ever took, and a much smaller building.

Even so, a modular stadium on Howard Terminal by 2027 is not happening. The negotiating window alone runs into late 2027. After that comes a lease or disposition agreement back before the board, then city entitlements, then CEQA, then a BCDC permit for shoreline work. None of those are formalities on a former marine terminal with contaminated soil and an active rail corridor past the gate.

The dates that will actually tell you something

The next real checkpoint is about 90 days out, in mid to late December 2026, when the Roots owe the Port a refined project description, a site plan and a stated acreage.

That filing is the first time anyone outside the negotiation sees how much of the 50 acres the club wants and where the stadium sits. Whether Industrial Realty Group's logistics concept is still standing next to it will tell you as much as the drawings do. After that, the financial feasibility report in 2027 is the document that decides this. The window closes around September 2027 either way.

Until one of those lands, this is a real estate story about a piece of paper. The way we look at it when we compare neighborhoods at Houseberry, a 2.1 overall score and a $630,000 median in West Oakland are facts you can act on this month. A stadium might exist in 2030. If you are shopping either side of the estuary, start with how Oakland's neighborhoods actually rank today. The asphalt has been there since 2013. It can wait a little longer.

Sources

The Oaklandside, "Oakland Roots land exclusive deal with the Port of Oakland to develop Howard Terminal," Jose Fermoso, September 18, 2026

SFist, "Oakland Roots Reach Deal to Develop Howard Terminal Site," September 19, 2026

NBC Bay Area, "Oakland Roots reach agreement to develop Howard Terminal site," September 2026

Port of Oakland, Board of Port Commissioners meeting agenda and resolutions, September 10, 2026

Port of Oakland, "Port of Oakland seeks development concepts for Howard Property," September 6, 2024

CBS News Bay Area, "Port Approves Exclusive Negotiations With A's For Howard Terminal," April 27, 2018

California Legislature, AB 1191 (Bonta, 2019), State Lands Commission trust lands exchange, Howard Terminal property

Oakland City Council and Bay Conservation and Development Commission records on the Waterfront Ballpark District at Howard Terminal, 2021 and 2022

San Francisco YIMBY, "Oakland Roots Releases Plans for Howard Terminal Soccer Stadium," January 2025

KQED, "Years After A's Deal Died, Port of Oakland Looks to Soccer Club for Waterfront Stadium," 2026

Houseberry, Oakland neighborhood rankings, city guide and neighborhood price history, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

One sentence in a Santa Cruz planning filing posted this morning explains why 102 affordable apartments had to be approved twice. A condition meeting the labor standards in AB 130, California's newest housing law, “now cannot be met,” the city wrote. The fast-track approval the project won in April is void. Santa Cruz rescued the building the same day by re-approving it under a different, older piece of environmental law that carries no labor conditions at all. The apartments are fine. The statute is what broke.

If you have never heard of AB 130, that is normal, and you are going to keep hearing about it. It is the law that ended environmental review for most urban housing in California. This looks like the first documented case of its labor rules costing a real project its exemption, and what the city did next is the more interesting half of the story.

What you need to know

• On September 21, 2026, the City of Santa Cruz filed a Notice of Exemption for Driftwood Flats, 102 all-affordable apartments at 407 to 415 Pacific Avenue, under state clearinghouse number 2026090868.
• The filing states that a project condition meeting the labor standards in Public Resources Code section 21080.66(d) cannot be met, so the April 2026 CEQA exemption under AB 130 is no longer valid.
• The city re-approved the same building under Public Resources Code section 21083.3 and CEQA Guidelines section 15183, an older path with no labor conditions. The unit count and the design do not change.
• At about 95 feet, Driftwood Flats sits 10 feet above the 85-foot line where AB 130 borrows the strictest labor standards in the statute.
• We score Downtown Santa Cruz 2.6 out of 5 overall, 20th of the 34 Santa Cruz neighborhoods we rate, with a 4.6 on amenities and a 2.0 on safety. The citywide median sale price was about $1.47 million in August 2026.

What AB 130 actually changed, for people who have never heard of it

AB 130 let most urban housing in California skip environmental review. Not shorten it. Skip it.

Since 1970, the California Environmental Quality Act has required a public agency to study what a project will do to traffic, air, noise, water and the rest before approving it. For a big apartment building that study runs hundreds of pages, takes a year or more, costs six figures, and opens a window in which anyone can sue over whether the study was thorough enough. The lawsuits are the part that matters. A project does not need to lose one to die. It just needs to be tied up long enough that the financing goes away.

Governor Newsom signed AB 130 on June 30, 2025 as part of the state budget, and it took effect immediately. It created a new statutory exemption at Public Resources Code section 21080.66. Clear its checklist and the environmental document does not get faster. It stops existing.

The checklist, roughly: the site is 20 acres or less, it is inside a city or a census-defined urban area, it was already developed or is surrounded by development, the project matches the local zoning or general plan, it hits at least half the density the state expects there, and it is not on farmland, wetland, high fire, fault zone or protected habitat. We have now watched it work twice in our own coverage area, starting with 264 apartments on Winchester Boulevard in San Jose in August.

AB 130's labor standards live in the part nobody reads

AB 130 is not a free pass for every building. Subdivision (d) attaches labor conditions, and they get heavier as the building gets taller.

The reason is political. Construction unions were never going to hand the Legislature a CEQA rollback for nothing. The State Building and Construction Trades Council wanted skilled-and-trained workforce rules on anything using the new exemption. The California Conference of Carpenters, working in what it called a virtually non-union industry, wanted production and a wage floor instead. CalMatters covered the fight as it landed in the budget on June 24, 2025. The compromise survives in the statute as three tiers.

Most projects under 85 feet get no labor conditions at all. A project that is 100 percent affordable to lower income households pays prevailing wage whatever its height. And anything 85 feet or taller has to pay prevailing wage and use a skilled and trained workforce, the standard the law borrows from SB 35 at Government Code section 65913.4(a)(8).

That last phrase is where a layperson usually stops reading, so here is what it means. A skilled and trained workforce is not a wage rule. It is a hiring-pipeline rule. Every craftworker on the job has to be a registered apprentice or a journeyperson, and depending on the trade, 30 to 60 percent of the journeypersons must hold a graduation certificate from a state-approved apprenticeship program. The contractor files monthly reports proving it, and those reports are public.

Meeting that in Oakland or San Jose is a phone call. Meeting it in a county with one mid-sized city in it is a different exercise.

Driftwood Flats landed in both hard tiers at once

Driftwood Flats is a 102-unit, 100 percent affordable building planned for 407 to 415 Pacific Avenue, the low-rise south end of Pacific Avenue near Laurel Street in downtown Santa Cruz, a few blocks inland from the Boardwalk. The developer is CRP Affordable Housing and Community Development. Eight stories, about 95 feet, 34 two-bedroom and 68 three-bedroom apartments over 1,490 square feet of ground floor commercial space. It replaces the Neptune Apartments, whose tenants had to vacate by July 27, 2026 with relocation help and a guaranteed spot in the new building. The Planning Commission approved it unanimously on April 16, 2026.

Read the tiers against those facts and the problem is obvious. The project is 100 percent affordable, so it owes prevailing wage. It is also 95 feet, so it owes the skilled and trained workforce standard on top. Two tiers, one building.

The filing does not say which condition failed, and I am not going to pretend it does. But the odds point one direction. A 100 percent affordable building is financed with federal low income housing tax credits, and prevailing wage is already part of that world. Paying it is routine. Finding general contractors and subcontractors in Santa Cruz County who can certify apprenticeship graduation ratios trade by trade, every month, on a 95-foot building, is not routine. That is my read of the most likely wall, not the city's statement, and the staff report on the modification is the document that will settle it.

The workaround was a law from a different era

Santa Cruz did not fight the problem. It swapped statutes. The new Notice of Exemption leans on Public Resources Code section 21083.3 and CEQA Guidelines section 15183, and those carry no labor conditions whatsoever.

In plain language, section 15183 says that if a city already did a full environmental study for its general plan, a project that fits inside what that study already analyzed does not need its own. Santa Cruz has a certified General Plan 2030 EIR and a Downtown Plan Expansion supplemental EIR. Driftwood Flats sits inside the buildout those documents already examined, the city found no new impacts peculiar to the project, and the exemption held.

Same building. Same 102 homes. Same block. Different statute, and a very different set of obligations attached to it.

That table is the whole finding. AB 130's labor deal only binds projects that use AB 130. A city that did its general plan homework has another door, and walking through it costs a developer nothing but paperwork.

So are the 102 apartments stalled or not?

The homes are not cancelled and never were. What stalled was the approval, not the project.

The practical cost is time and risk. Five months after a unanimous approval, the city had to redo the environmental finding and reopen the conditions, and the modification also asks for an additional density bonus concession. Tenants are already out. Driftwood Flats also carries a second building on its back. It is the off-site affordable housing required for Lincoln Property Company's 245-unit market-rate project at 201 Front Street, so roughly 347 homes in one small downtown ride on this paperwork holding together.

One detail worth noticing without reading intent into it. The market-rate building at 201 Front Street is capped at exactly 85 feet. The affordable one is 95.

What 102 deed-restricted homes do in a $1.47 million city

Santa Cruz passed the point of an affordability problem on the horizon a long time ago. The citywide median sale price was about $1.47 million in August 2026, down from roughly $1.56 million in June. A softer year at that level is still a market where an ordinary Santa Cruz salary does not buy a house, which is why 102 permanently deed-restricted apartments matter more than the number suggests.

The location is the interesting part. Downtown Santa Cruz rates 4.6 out of 5 on amenities, second of the 34 neighborhoods we score in the city, and 2.0 out of 5 on safety, which puts it 24th. The blended score lands at 2.6, which ranks it 20th, just under a citywide 2.7.

Read those two numbers together and you have the actual trade these 102 households make. They get the most walkable address in the county, groceries and transit and work within a few blocks, no car required. They also get the part of town the police blotter knows best. When we compare neighborhoods at Houseberry, that split is the thing we want people to see before they sign anything, because a strong overall score and a strong daily-life fit are not the same question. For a household that would otherwise commute in from Watsonville or over Highway 17, a 2.0 on safety and a 4.6 on amenities may be an easy trade. For someone with a kindergartner, less so.

The honest cost of the workaround

Someone did lose something here, and it was not the developer or the future tenants. It was the construction workers.

The Legislature attached a labor floor to 95-foot buildings on purpose. It was the price the trades extracted for giving up CEQA leverage, and it is the reason a lot of Democrats voted for the bill at all. A 95-foot building is now going up in Santa Cruz without that floor, through a door the deal never contemplated.

I am firmly on the side of building this thing. One hundred and two permanently affordable homes in downtown Santa Cruz is worth more to this county than a procedural point, and a labor condition that would have killed affordable housing in one of the most expensive markets in America is a condition doing the opposite of its job. But being pro-housing does not require pretending the trade was free. The cleaner fix is for the Legislature to decide whether the 85-foot line should apply to 100 percent affordable buildings at all, rather than letting the answer depend on whether a city happens to own a certified general plan EIR.

Because that is what it depends on right now. A well-planned city can offer the exit. A city that never certified a program EIR cannot, and its projects eat the labor conditions. That is a strange way to decide who gets paid what.

Where this shows up next

The next document to watch is the Santa Cruz staff report on the modification, which has to name the condition the filing only gestures at. After that, the interesting question is how many other projects have quietly done the same thing.

Any 100 percent affordable building over 85 feet approved under section 21080.66 since mid-2025 is sitting on the same tripwire, and any of them in a city with a certified general plan EIR has the same escape hatch. Menlo Park's 800 Oak Grove Avenue project, which we covered this week, came in at 72 feet and never touched the line. The ones to check are the towers.

If you are shopping downtown Santa Cruz, none of this changes what you should look at. It just means that a headline about a project being delayed is usually a story about paperwork, not about the neighborhood. The block is the thing that will still be there in ten years, and that is the part worth researching first. Our Santa Cruz neighborhood rankings are a reasonable place to start.

Sources

CEQAnet, Notice of Exemption, SCH 2026090868, Driftwood Flats, City of Santa Cruz, posted September 21, 2026

City of Santa Cruz, 407 Pacific Avenue project page, project CP26-0076, major modification to CP25-0156

Santa Cruz Local, “Design revealed for 8-story Pacific Ave. apartments in Santa Cruz,” January 23, 2026

Santa Cruz Local, “Tenants set to be displaced by new affordable housing in Downtown Santa Cruz,” April 25, 2026

City of Santa Cruz, 201 Front Street mixed-use project use narrative, CP25-0116

Holland & Knight, “California Legislature Enacts Major CEQA Reforms for Housing-Rich Infill Projects,” July 2025

Buchalter, “Effective Immediately: California Adopts Consequential Reforms to CEQA and Other Laws to Speed Housing Production,” July 2025

Associated General Contractors of California, skilled and trained workforce requirements

CalMatters, “California lawmakers weigh construction wage change in budget,” June 2025

Houseberry, Santa Cruz neighborhood rankings and city guide, retrieved September 21, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

A Catholic high school that never got built just became as many as 900 homes. On August 5, 2026, the City of Livermore filed the Notice of Exemption for Development Agreement 24-004, the paperwork confirming it had executed the agreement covering 3658 Las Colinas Road, a 122-acre property on the city's northern edge above Springtown. The council had voted for it eight days earlier. The filing is what turned that vote into a signed deal.

Nobody is pouring foundations. The agreement grants vested rights and sets a unit ceiling, and that is all it does. But in a city where two voter-approved growth boundaries decide most of what can be built at all, a signed ceiling on 50 developable acres is the part that actually matters.

The takeaways

What the August 5 filing changes, in short:

• The deal is executed, not pending. The council approved Development Agreement 24-004 unanimously on July 27, 2026, and the city filed the CEQA exemption on August 5 confirming the agreement is in force.
• The ceiling is 605 to 900 homes on roughly 50 developable acres of the 122-acre site, at 8 to 30 units per acre, per the city's July 7, 2026 staff report.
• It replaces a 2005 agreement that approved a private Catholic high school on the same land. The Diocese of Oakland decided that use was no longer viable.
• Nothing is entitled yet. A Neighborhood Plan, Planned Development zoning, tentative maps, site plan review and project-level environmental review all still have to happen.

What the August 5 filing actually did

It moved the project from authorized to negotiate to signed. That distinction is the entire story, because the earlier coverage in July described a council vote, and a council vote on a development agreement is not the same thing as an executed one.

The notice cites CEQA sections 15183(b) and 15183(d) along with the common sense exemption at 15061(b)(3), and it is careful to say the ordinance proposes no specific development and authorizes no physical change. City senior planner Shannon Pagan is listed as the contact. The applicant is Adventus, a nonprofit corporation tied to the Diocese of Oakland that has held the land since the 1990s, with the Diocese's chief financial officer signing as the applicant contact.

The 2005 version of this agreement was for a high school. The Real Deal reported in February 2025 that Adventus had given up on the school and asked for residential zoning instead, initially floating up to 500 homes. Eighteen months of negotiation later the ceiling came in considerably higher, and Adventus agreed to put $6 million into a community benefit fund for Springtown-area infrastructure and amenities.

The agreement at a glance

The specifics worth writing down, because the next round of hearings will be argued against these numbers:

605 to 900 homes is a large number for a city this size

At the high end this one site would deliver about a fifth of everything Livermore is supposed to build in eight years. The city's 2023-2031 Housing Element carries a Regional Housing Needs Allocation of 4,570 units, and 900 homes is roughly 20 percent of that. Even the 605-unit floor is about 13 percent.

That is unusual. Most Bay Area housing arrives in 40-unit and 130-unit increments, which is why the pipeline math rarely works. A single parcel that can absorb a double-digit share of a city's entire obligation is the kind of site planners quietly build their numbers around.

The catch is that a development agreement ceiling is permission, not production. Livermore has been here before. Our earlier look at the Wente family's 1,500-acre land sale found a property roughly twelve times this size that could legally hold about fifteen houses, because the growth boundaries and a 100-acre minimum lot size sit in the way. Las Colinas is the rarer case where the land is actually inside the line.

This lands on the cheaper side of Livermore, and that matters

Springtown is not where Livermore's expensive homes are, which makes it the more interesting place to add several hundred of them. The neighborhood, the planned community on the city's north side off Springtown Boulevard above the I-580 corridor, had a median sale price of about $892,700 in August 2026, according to our Springtown neighborhood page. Livermore's citywide median that same month was about $1.25 million.

That is a 29 percent gap inside one city. For anyone priced out of South Livermore or the vineyard edge, Springtown is already the answer, and it is about to get a few hundred more of whatever gets built next to it.

Springtown ranks 31st of the 42 Livermore neighborhoods we score, with an overall 3.0 out of 5. The interesting part is where that number comes from.

Schools are not the weak link. Springtown scores 3.9 for schools against a citywide 3.8. The gap is safety, at 2.7 against 3.3, which puts it 26th of 42 on our Livermore safety ranking. Fifteen Livermore neighborhoods score a 4 or better on safety. Springtown is not one of them.

A 0.6-point safety gap is not a reason to write off a neighborhood. It is a reason to ask what 605 to 900 new homes, a neighborhood park, new internal streets and a trail network do to it. Streets with more eyes on them and better lighting tend to move that number, slowly. That is the wager the $6 million benefit fund is making.

The school capacity question has an answer most people will not guess

Livermore's school district has been shrinking, not straining. Livermore Valley Joint Unified cut staff positions for the 2025-26 year after a $1.3 million drop in outside funding driven partly by declining enrollment, and it was running reserves of 3.9 percent on a $206 million budget.

So the usual hearing argument, that new homes will overwhelm the schools, does not fit the district's actual finances. Homes delivered over a decade into a district losing students look less like an overcrowding event and more like enrollment backfill, and enrollment is what funds California schools.

The schools nearest the site are on Livermore's north side, including Altamont Creek Elementary and Junction Avenue TK-8, with Granada High serving the north end of town. Attendance boundaries get drawn at the Neighborhood Plan stage, not now, and that document is where the real school analysis will live. Read it when it drops.

The honest downside is the calendar

This is years out, and anyone telling you otherwise is selling something. Nine of the 13 people who spoke at the July 27 hearing raised traffic, safety and access road concerns, which is a reasonable thing to raise about a site reached by two-lane roads. Labor representatives asked for apprenticeship and wage standards and got a non-binding recommendation instead of binding language.

Before a single home is permitted, Adventus needs a Neighborhood Plan, Planned Development zoning, tentative maps, site plan review, project-level environmental review and development impact fees. Livermore is also still working through its General Plan 2045 update, which is the document this land use change rides on. Call it three to five years to first shovels if everything goes smoothly, and Bay Area entitlements rarely go smoothly.

What to do with this if you are shopping north Livermore

Treat it as a reason to look at Springtown now rather than a reason to wait. The pricing gap against the rest of Livermore is real today, the school score is already above the city average, and the thing that would most plausibly change the area, a few hundred new homes with a park and trails attached, is now legally on the table rather than hypothetical.

The way we look at it when we compare neighborhoods, a signed development agreement is a data point about the next ten years, not the next ten months. Watch for the Neighborhood Plan hearing. That is when the 605-to-900 range narrows to an actual number, and when the street layout, the park siting and the school boundaries stop being adjectives. You can see how Springtown currently stacks up against the other 41 neighborhoods on our Livermore rankings while you wait.

Sources

CEQAnet: Notice of Exemption, Development Agreement 24-004, Las Colinas Road Property, City of Livermore, filed August 5, 2026

Livermore Vine: Livermore OKs 900-unit guardrail for potential residential development near Springtown, August 4, 2026

The Real Deal: Diocese scraps school plan in Livermore for up to 500 homes, February 27, 2025

City of Livermore: 2023-2031 Housing Element

The Independent: Livermore school district board reaches decision on final cuts for 2025-26 year

Houseberry: Springtown, Livermore neighborhood data

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

The lot at the northeast corner of The Alameda and Emory Street in San Jose has been empty since 1965. On September 18, 2026 the city filed the environmental review that could finally end that run: a General Plan Amendment and rezoning to allow 18 townhomes in five buildings on 0.99 acres. Public comment closes October 8, 2026.

No newspaper has written a word about it. That is normal for a project this size, and it is also the reason small filings like this one are worth reading. Eighteen homes will not move San Jose's housing numbers. What the paperwork says about the land underneath them is a different story.

The bottom line

The short version before the details:

• San Jose is processing a General Plan Amendment and rezoning at The Alameda and Emory Street to allow 18 townhome units in five three-story buildings on a 0.99-acre site, a density of 18.2 homes per acre.
• The site has been vacant since 1965 and is still zoned A(PD), agricultural planned development, in the middle of Central San Jose.
• The environmental document is a Mitigated Negative Declaration, SCH #2026090816, posted September 18 with comments due October 8, 2026.
• Twenty trees come out, 17 of them ordinance-size. That is the part of this project neighbors will actually argue about.
• The two neighborhoods the site sits between price about $640,000 apart as of August 2026.

What is actually proposed at The Alameda and Emory

The proposal is small, specific, and unusually easy to picture. Five buildings, three stories each, two with three units and three with four, on an L-shaped parcel of just under an acre. Two 26-foot driveways off Emory Street, nothing off The Alameda. Roof ridges at 36 to 37.5 feet. The applicant is Dan Mountsier of Emory Holding LLC, and the city is carrying it under file numbers GP25-002, C26-003, H26-007 and ER25-231.

Here is the whole thing on one page:

Nothing in there is exotic. Three-story townhomes with driveways off the side street is about the most conventional thing you can build on a corner like this. The fight, to the extent there is one, is over the trees and the rezoning, not the building.

A commercial designation that produced no commerce for 61 years

This is the detail worth stopping on. The site is designated Neighborhood/Community Commercial on the General Plan and zoned A(PD), agricultural planned development. It has produced neither a shop nor a crop since Lyndon Johnson was president.

Sixty-one years is long enough to call the experiment. When land sits empty that long under a designation meant to deliver storefronts, the designation is not protecting anything. It is just paperwork that has outlived the market it was written for. San Jose has thousands of parcels like this, and the city is currently rewriting its own rules about them.

We wrote about the bigger version of that rewrite in San Jose's General Plan reset, where staff proposed allowing four to ten homes where one sits now. The Emory Street filing is the retail-scale version: one parcel, one amendment, 18 homes.

Four empty lots, about 304 homes, one corridor

The Alameda is quietly converting. Not through a big specific plan or a headline rezoning, but one long-vacant parcel at a time, each with its own file number and its own small hearing.

Count the ones currently in the pipeline within roughly a mile of this corner: 174 modular apartments at 2165 The Alameda on a half-acre vacant lot, about 100 units at 945 West Julian Street near Diridon, 12 condos at 1038 McKendrie Street on another vacant lot, and these 18.

Three hundred and four homes is not a wave. But it is 304 homes on land that was producing nothing, in the most central, most walkable part of a city that spent decades building outward. The Emory Street project is the only one of the four that needs a General Plan Amendment, which is a fair measure of how out of date its designation is.

The trees are what the hearing will be about

Twenty trees come out. Seven are street trees, 13 are on the site, and 17 of the 20 are ordinance-size, which in San Jose means they are large enough to require a permit and a replacement plan to remove.

That is a real cost and it deserves to be named plainly. Mature trees on a Central San Jose corner take decades to replace, and the replacement plantings will be saplings for most of the next 20 years. Anyone who tells you this is a free trade is selling something.

The rest of the mitigation package is the standard Central San Jose list: Tier 4 construction equipment to cut diesel particulate by 60 percent, a vibration monitoring plan because there are historic structures within 200 feet, a Historical Resources Protection Plan, and archaeological monitoring with a Native American representative present during ground disturbance. All of it is in the Initial Study and Mitigated Negative Declaration.

What the blocks on either side of The Alameda actually cost

This corner sits on a seam. To the southwest is Rose Garden, the Central San Jose neighborhood built around the 5.5-acre San Jose Municipal Rose Garden and Rosicrucian Park. To the northeast, between The Alameda and Highway 880, is College Park, the older pocket anchored by Bellarmine College Prep and a Caltrain platform that sees only a few trains a day.

The gap is about $640,000 across one arterial. Rose Garden's median sale price was roughly $1.96M in August 2026, down from about $2.08M a year earlier. College Park's was about $1.32M in August 2026, essentially flat. San Jose citywide sat at about $1.63M, right between them.

The scores tell the same story from a different angle. Rose Garden ranks first of the 112 San Jose neighborhoods we score on amenities at 4.5 out of 5, and 61st overall at 3.5. College Park ranks 10th on amenities at 3.7 and 104th overall at 2.5, dragged down by a 2.1 school score and a 2.7 safety score.

So 18 townhomes are going in between a neighborhood that scores near the top of the city on what is walkable and one that does not clear the middle on much of anything. That is not a warning. It is the reason the corner is worth building on: the amenities are already there, a five-minute walk away, and nobody has to create them.

Questions people are asking about the Emory Street project

Is the rezoning approved?

No. As of September 2026 the project is in environmental review. The Mitigated Negative Declaration was posted for public comment on September 18, and a Planning Commission hearing and City Council action on the General Plan Amendment and rezoning would follow after the comment period closes.

How do I comment before October 8?

Comments go to San Jose's Department of Planning, Building and Code Enforcement, and the city's public notice for the draft MND carries the current contact and submission details. Reference file number ER25-231.

Will these be for sale or for rent?

The environmental filing describes 18 townhome units and does not state the tenure. Townhomes on a site this small are usually built for sale, but until a tract map or a marketing plan surfaces, treat ownership as unconfirmed.

What does a Mitigated Negative Declaration mean?

It means the city found no significant environmental effect that cannot be reduced with specific, enforceable conditions, so no full Environmental Impact Report is required. It is the middle tier of CEQA review, and it is the outcome a small infill project on a vacant lot usually gets.

October 8, then the hearing

The comment deadline is the only fixed date on this project right now. After it passes, the file moves toward a Planning Commission hearing and then a Council vote on the amendment and rezone, and those dates will post on the city's agenda calendar.

If you are shopping this part of San Jose, the practical move is to stop treating The Alameda as a boundary and start watching it as a pipeline. Four vacant parcels, about 304 homes, all of them inside the walkshed of the Rose Garden. When we compare neighborhoods at Houseberry, that is the kind of thing that moves a score over five years, well before it shows up in a listing description: what is being permitted two blocks away, not what is on the market this week.

A good starting point is the full San Jose neighborhood ranking, where Rose Garden and College Park sit 43 places apart with one street between them.

Sources

CEQAnet, Emory Street Townhomes Project, SCH #2026090816

City of San Jose, Initial Study and Mitigated Negative Declaration, Emory Street Townhomes

City of San Jose, public review notice for the draft MND (GP25-002, C26-003, H26-007, ER25-231)

San Francisco YIMBY, new modular housing proposed on The Alameda

San Francisco YIMBY, new housing proposed for 945 West Julian Street

San Francisco YIMBY, 1038 McKendrie Street, San Jose

Houseberry neighborhood data for Rose Garden, San Jose (accessed September 2026)

Houseberry neighborhood data for College Park, San Jose (accessed September 2026)

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Ask a Peninsula planner what the AB 130 CEQA exemption actually does and you usually get a hypothetical. Here is a real one. On September 18, 2026, Menlo Park filed a Notice of Exemption for a six-story, 15-condominium building at 800 Oak Grove Avenue, claiming a full statutory exemption from environmental review under Public Resources Code section 21080.66. No study, no comment period on traffic counts, no mitigation package to negotiate. The application landed on December 21, 2025. The Planning Commission approved it 7-0 on September 14.

That is 267 days for a six-story building in a city that has spent decades being careful about six-story buildings.

The bottom line

• Menlo Park filed the Notice of Exemption, SCH 2026090808, on September 18, 2026, citing the AB 130 infill exemption at Public Resources Code section 21080.66.
• A surface lot at Oak Grove Avenue and Crane Street becomes 15 for-sale condominiums, one of them deed-restricted for very low income households. The vacant Comerica Bank building next door becomes offices with a roof deck.
• Application to a 7-0 Planning Commission vote took 267 days. The City Council takes up the subdivision map on September 22, 2026.
• It is the second AB 130 filing we have covered in a month, and the first on a small for-sale project rather than a large rental one.
• We score Downtown Menlo Park 4.4 out of 5 overall, third of the 15 Menlo Park neighborhoods we rate. The citywide median sale price was about $3.13 million in June 2026.

A parking lot, 0.75 acres and no environmental document

The site is the corner lot at Oak Grove and Crane, one block off Santa Cruz Avenue and under a ten-minute walk to the Menlo Park Caltrain station. Anyone who has banked downtown knows it: a flat beige two-story Comerica branch, now vacant, next to a surface lot that has been some of the most valuable asphalt in San Mateo County for twenty years.

The lot gets the new building, about 72 feet tall. The bank gets an interior gut and a roof deck. The applicant is Rob Zirkle of Prince Street Partners, and the city project record puts the combined floor area at 61,562 square feet with 34 shared spaces on 0.75 acres.

We walked through the mechanics of section 21080.66 last month when San Jose used it on 264 apartments at 741 S. Winchester Boulevard, so the short version here: AB 130 was signed on July 1, 2025 in the state budget package, and it hands qualifying infill housing a complete statutory exemption instead of the narrower categorical ones cities had been stretching for years. Sites under 20 acres, in an urban area, consistent with zoning, over a minimum density floor, off protected and hazardous land. Meet the list and the analysis does not happen.

The statute was written for 264 apartments. It just cleared 15 condos.

That is the genuinely new thing here. Winchester was the case the CEQA reform debate was actually about: a big rental building on a dead commercial parcel, stuck since 2023. Oak Grove is 15 for-sale homes on three quarters of an acre in one of the most expensive downtowns in California. Same statute. Wildly different projects.

The seven-day notice on file with the city shows the real pattern. Oak Grove did not lean on one law. It stacked five: SB 330 to freeze the rules at application, the Housing Accountability Act because 72.5 percent of the floor area is residential, State Density Bonus Law for a 20 percent bonus plus a concession and waivers, AB 2097 to cut parking near transit, and AB 130 to delete environmental review at the end. Any one of those a city could have survived. Together they leave almost nothing to argue about, which is the point.

Fifteen condominiums in a $3.13 million market

Fifteen homes will not move the Menlo Park median. They add ownership product in a price band the city barely builds.

The citywide median was roughly $3.13 million in June 2026, down from about $3.72 million in July 2025, per the 12-month history on our Menlo Park city page. That is detached-house money. A two-bedroom condominium off Santa Cruz Avenue is a different product, and downtown Menlo Park has produced almost none of it in a decade.

The scores are why the location matters. Downtown Menlo Park rates 4.4 out of 5 overall, third of the 15 Menlo Park neighborhoods we cover, with a 4.7 on schools, 4.6 on amenities and a perfect 5.0 on nearby merchants. Safety sits at 3.7, the softest number on the card and an ordinary one for a walkable commercial core. Daily life here is already excellent. The entry price is the barrier.

Fifteen units do not fix that. They do put a few addresses in the best-scoring part of Menlo Park that are not a $3 million single-family house. The way we look at it, the neighborhood was never the problem on this block. The product type was.

One unit out of fifteen

Here is the part worth being honest about. The project deed-restricts exactly one home for very low income households, about 6.7 percent of base density, with an in-lieu fee covering the fraction left over. The Housing Commission recommended that package 6-0 on September 2. In a city where the median is over $3 million, one very-low-income condominium is not an affordability program, and nobody at the dais pretended otherwise. The other half of the site is not housing at all.

The faster path costs something procedurally too. Environmental review handed neighbors a document, a comment window and a hearing where parking and shade and the construction schedule got argued in public. AB 130 removes that venue along with the study. The trade is worth making, because the old process killed good projects about as often as bad ones. But it is a trade, and saying so is how you stay honest with the people who live across the street.

Tuesday night, and the filing after that

The City Council takes up the vesting tentative map on September 22, the last discretionary approval between this and a building permit. InMenlo previewed the Planning Commission hearing and the commission moved it 7-0, so the map is unlikely to be where this one dies.

The better question is the one this filing starts to answer. When San Jose used section 21080.66 in August, the fair response was that one exemption is not a trend. A month later a second Bay Area city has used the same section on a project a seventeenth the size, in a downtown with a long memory for fighting height. If the next few Peninsula applications come in this way, the real effect of AB 130 will not be that big projects got faster. It will be that small ones stopped being impossible to pencil, because a 15-unit building can now carry an entitlement cost that used to take 200 units to absorb.

That is worth watching on Santa Cruz Avenue, on El Camino, and in every downtown between here and Redwood City.

Sources

CEQAnet, Notice of Exemption, SCH 2026090808, 800 Oak Grove Ave. Project, City of Menlo Park, posted September 18, 2026

City of Menlo Park, 800 Oak Grove Avenue project page, project description, hearing dates and votes

City of Menlo Park, 800 Oak Grove Avenue mixed-use housing development seven-day notice, application filed December 21, 2025

InMenlo, "Menlo Park Planning Commission to consider 800 Oak Grove Ave. project September 14," September 14, 2026

The Almanac, "New six-story residential building proposed for downtown Menlo Park," February 3, 2026

San Francisco YIMBY, "Housing Planned for 800 Oak Grove Avenue, Menlo Park," February 2026

Ascent, "AB 130's Game-Changer for CEQA Streamlining of Housing Development, The New Infill Exemption"

Houseberry, Menlo Park city and Downtown Menlo Park neighborhood data, retrieved September 20, 2026

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.

Farmers asked California for a 6.99 percent homeowners rate increase. The state approved 1.5 percent. That change reaches about 915,000 California homeowners at each policy's first renewal after September 15, 2026, which means the envelopes are landing in mailboxes right now. If you bundle home and auto with Farmers, your bill probably goes down.

The part almost nobody has written about is in the same filing. Farmers also removed the cap on how many new homeowners policies it will write in this state, and it says it is going looking for business in the places private carriers spent five years backing out of.

The bottom line

Five numbers, for anyone holding a Farmers renewal notice in a Bay Area hills ZIP code.

• About 915,000 California homeowners are covered by the new plan, at an approved statewide average of 1.5 percent, applied at the first renewal after September 15, 2026.
• Farmers requested 6.99 percent in its November 21, 2025 filing. The Department of Insurance approved 1.5.
• The home-and-auto bundle discount rises from 15 percent to 22 percent, which is worth far more than the rate change takes away.
• Farmers dropped its cap of 9,500 new homeowners policies a month and plans to market into state-designated wildfire distressed areas.
• The California FAIR Plan goes up 29.1 percent on October 15, 2026. In the hills, that gap is the real story.

What 1.5 percent actually does to a hillside premium

Very little, and that is the point. Start with a high-fire-hazard Bay Area hillside home paying $2,400 a year with a 15 percent home-and-auto bundle discount already baked in. The approved 1.5 percent adds about $36.

Then the discount moves. Seven extra points of bundle discount beats 1.5 points of rate every time. Run the same policy through the new plan and it lands near $2,235, roughly $165 cheaper than it is today. For a bundled customer, this filing is a price cut wearing the word increase.

The alternative is where the money is. A California FAIR Plan policy, the state's insurer of last resort, pays for fire, smoke, lightning and in-home explosions and close to nothing else, so most owners bolt on a wrap-around difference-in-conditions policy to get liability, theft and water damage back. In high wildfire country that stack has been running roughly $5,500 to $9,000 a year. Against a private policy near $2,235, that is not a premium difference. That is a car payment.

The cap is the part that matters in Orinda and the Oakland hills

Farmers had been rationing California homeowners policies at 9,500 a month, up from 7,000 in 2023. Its November 21, 2025 announcement scrapped the limit and said the company would market directly to about 300,000 consumers in areas the state calls distressed. Fox Business confirmed the old cap figures.

Distressed is a defined term, not a mood. The Department of Insurance flags a ZIP code when FAIR Plan penetration reaches 15 percent or higher and CAL FIRE rates the area high or very high fire hazard. Larger insurers writing there commit to covering at least 85 percent of the properties in those areas.

Hold that test against a Bay Area map and the list writes itself. CAL FIRE's 2025 local responsibility area maps put very high fire hazard zones inside twelve Contra Costa cities, Orinda, Moraga and Lafayette among them. The Oakland and Berkeley hills have been in that category since well before the 1991 firestorm. None of these are cheap markets an insurer can afford to ignore.

Four Bay Area hillside markets where a returning carrier would actually matter. City medians come from our city pages. Montclair is a neighborhood median. Most recent month shown, not an annual average.

Read down that median column. Lafayette sat at about $2.12M in August 2026 across the 31 neighborhoods we rank there, and Montclair, the wooded Oakland hills neighborhood spread along Mountain Boulevard, was about $1.56M the same month. These are seven-figure homes whose owners have been quoted like uninsurable risk. A carrier that can write 20,000 policies a month instead of 9,500 has room to look at them again.

The honest caveat: removing a cap is permission, not a promise. Farmers expects several thousand new policies over two years in distressed areas. The FAIR Plan had 696,562 policies in force and $768 billion of exposure as of June 2026, up 157 percent in policy count since September 2022. Several thousand does not reverse that. It is a first trickle back up the hill, and it will show up address by address rather than town by town.

Asked versus got is the whole California story right now

The interesting comparison is not Farmers against last year. It is Farmers against everyone else who filed in 2026.

Farmers took the smallest ask and the smallest approval of the three, and it was the first filing under the state's Sustainable Insurance Strategy, the framework that lets carriers price wildfire risk with catastrophe models in exchange for writing in distressed areas. State Farm's 17 percent came out of a settlement in March 2026. The FAIR Plan asked for 35.8 percent and was approved for 29.1 effective October 15.

Two envelopes are arriving a month apart with opposite messages. A Farmers renewal after September 15 moves about 1.5 percent. A FAIR Plan renewal after October 15 moves about 29.1 percent, and KQED reports that owners in genuinely high-risk areas could see premiums roughly double while some urban Bay Area policyholders see reductions. The spread between staying private and falling back on the FAIR Plan just got wider, in one season, by design.

What I would do with that envelope

Three practical moves, none of which take an afternoon.

Check the bundle line, not the headline. If your Farmers declarations page still shows a 15 percent home-and-auto discount, ask when the 22 percent applies to you. That single line is worth more than the rate change.

If you are on the FAIR Plan, get requoted before October 15. Carriers re-entering distressed ZIP codes do not send letters announcing it. Farmers began pushing new-business marketing into those areas in early 2026, and an independent broker will know which addresses are being quoted again.

If you are shopping a hillside home, quote the address before you write the offer. Two houses half a mile apart in the Berkeley hills can carry very different answers, and the answer that matters is whether a private carrier will write it at all.

We built Houseberry around comparing the area before the address, and insurability has quietly become one of those area-level facts, sitting next to schools and safety rather than next to closing costs. It is now the fastest-moving one. When we wrote in June that the insurance problem was spreading past the hills, the direction of travel was carriers leaving. This is the first filing in a while pointed the other way, and it is worth watching whether it reaches the hillside blocks inside the 258 East Bay neighborhoods we rank or stops at the flats. Ask your agent in the spring, not next September.

Sources

• Farmers Insurance newsroom: Farmers to remove cap on writing new homeowners policies in California (November 21, 2025)
• Carrier Management: Farmers in California, HO rates up 1.5%, bundlers get 22% discount (May 12, 2026)
• Newsweek: Nearly 1 million Californians face home insurance change in coming months
• Fox Business: Insurance giant lifts cap on California homeowners policies
• California Department of Insurance: FAQ on insurer commitments for wildfire distressed areas
• KQED: California FAIR Plan announces 29.1% rate hike for homeowners this fall
• California FAIR Plan: Key statistics and data (June 2026)
• Local News Matters: State says 12 cities in Contra Costa County have very high fire hazard severity areas
• Real Cost Report: California FAIR Plan costs vs. private market 2026
• Houseberry: Lafayette, Orinda, Moraga city pages and the Montclair, Oakland neighborhood page

This article was originally published by the Houseberry Research Team at Houseberry. See the neighborhood data on Houseberry.