It seems like just yesterday that downtown San Francisco’s main industry was the production of “doom loop” headlines. Now SF is switching from doom to boom, thanks to an unprecedented flood of venture cash for AI startups.
A ton of that money is pouring into downtown leases. Get your key cards and happy hours ready. Office life is back.
The office vacancy rate has dropped from a high of 31 percent to 28 percent, according to commercial real estate company Colliers — still high, but it seems poised to keep dropping. In the first half of 2026, companies leased 8 million square feet, says Colliers, surpassing 2024’s total, and could challenge the all-time annual high of 13.9 million in 2019.
(Other firms measured SF’s top vacancy rate in recent years as high as 36 percent, but all analyses agree that the rate continues to decline from all-time highs.)
Two years ago, anticipating the current tide change, Mayor London Breed asked voters to lift some limits on new office construction at a time when the idea seemed ludicrous. (Voters said yes.)
Now, all the arrows are pointing up, which means the city should expect higher tax revenues, not to mention a boost for public transit and small businesses that cater to downtown workers.
“The AI boom has fuelled demand for top-tier space and I expect that will continue until the end of the current boom,” says the city’s chief economist Ted Egan.
Wait — the end? Why mention the end when the party has just gotten going?
Because it’s inevitable. San Francisco’s previous upcycle lasted a decade, from 2010 through early 2020. Some pundits are already predicting the end of this one. Some are already pointing out signs of an AI bubble. It’s hard to say when, but it will pop one day.
“When we put all of our eggs in the employment uses basket, there’s a lot of risk to city revenues, the economic life of the city in general, and transit ridership” when things go bad, says Ben Grant, director of planning at urban planning firm SITELAB.
When COVID arrived, cities with more diverse land uses like Austin, Tampa, and Denver were more insulated against hard times. SF eventually launched a variety of policies, exercises, and experiments to craft a different downtown. Some have stuck, such as street parties and concerts, and some haven’t — notably attempts to add more homes downtown through tax breaks and more.
But with office demand rising again, and with it the promise of lucrative taxes for city coffers, there’s less short-term incentive to encourage downtown housing. The next year or so could determine if downtown becomes a different place than it was for the past century — or if the city ends up reliving 2019 again.
Bohemian rhapsody?
Since World War II, SF has hewed to a standard model for U.S. cities, concentrating as many jobs as possible into a few downtown blocks.
Tom Radulovich, policy lead for the nonprofit Livable City, says urban monocultures are “easy for planning because they’re all one thing, and they can be intensely profitable. But they’re not resilient.”
Downtown SF needs deep pockets, big companies, and lots of commerce. But it also needs to add housing and nonprofits and “to sprinkle in enough bohemians to make things interesting,” says Radulovich.
As soon as office doors locked and parking lots emptied in early 2020, there was no shortage of visions of a different downtown. One was to turn all those empty office towers into apartments. But it was soon clear that for many buildings, conversion would be difficult or totally unfeasible.
“It is always a good idea for San Francisco to diversify its downtown,” says Karen Chapple, UC Berkeley professor emerita of city and regional planning. “That said, only a small share of buildings can be converted, and no big new economic uses like UC Berkeley are planning to move downtown.”
To make that “small share” as large as possible, lawmakers have passed a raft of new rules. The aforementioned Prop C in 2024 was one, waiving key taxes on office-to-housing projects. Even Aaron Peskin, then Board of Supervisors president, put aside his arch-preservation reservations with a 2023 bill to renovation rules in the Union Square area.
More recently, a new city law waived a long list of fees for new conversions within a specific downtown boundary, and a state law makes permits faster and easier for conversions that include certain amounts of affordable homes.
There have been a handful of successes. One old Mason Street building may become as many as 15 new homes, and the onetime Nordstrom Rack on Market Street is slated to turn into 136 units.
But most building owners have been content to wait for office workers to come back. And with median rents now topping $68 per square foot for high-end buildings, it seems their patience could be rewarded.
Top-dollar offices generate more tax revenue per property than homes do, but SF should still take a more holistic approach to these buildings, says Patrick Hannan, spokesperson for the planning and building inspection departments. “Commercial tax revenue is critically important, but we will not be able to build our way out of our housing shortage without adding new homes in every San Francisco neighborhood,” he says. Downtown doesn’t have to shut down at 6 p.m.”
One avenue for new downtown homes is to build them from scratch, and there is recent precedent.
The other avenue is office conversion. Despite the reluctance in recent years to make that happen, some observers feel the current office revival will actually make more conversions possible.
It’s a counterintuitive take, and it’s based on this premise: not all office buildings are created equal.
Lower class buildings
Several months into the AI-led boom, the new generation of startups are so far mostly interested in Class A properties: new buildings in prime locations like the East Cut.
The lion’s share of new leases favor Class A buildings, and they also create incentives to build more of them.
That might devalue older buildings (Class B and C) and make them better prospects for residential conversions. “Different building stock compete for different uses,” says Bennett Faloni, a broker with Maven Properties.
Some buildings outside the downtown core could also attract conversion. Sup. Danny Sauter, who represents the Financial District and Union Square, cited three examples not downtown but in Fisherman’s Wharf (also part of his district). “We’re seeing residential conversions happen in older buildings that have seen less commercial demand or find themselves in areas that are less attractive as office settings,” Sauter says.
The reheated market — and SF’s resurgent buzz — could also simply draw more developers willing to gamble on residential projects. Boomtowns are naturally more attractive than ghost towns.
City economist Egan throws some cold water on the class distinction, however. “Increased office demand means more tenants wanting office space. How does that make Class B buildings less valuable?”
Egan says the more obvious incentive for Class B office owners is to renovate them to look more like Class A and capture that office demand.
Despite his skepticism, Egan also agrees that the office-to-housing genie is not going back into its bottle, albeit for different reasons. “The idea will certainly stick around,” Egan says, in part because while office rents are rising, home prices are shooting up even faster. The city’s imbalance of housing supply and demand “creates a window of opportunity for office to residential. If they start to happen, that will be the reason.”
Conversion policies were relatively easy to write and pass when downtown real estate was in the doldrums. But with cash spigots now open, redesigning downtown to include more residents — and insulate against future shocks — will be an uphill battle.
And at some point, those shocks will come. Because no matter what AI startups and their backers want to believe, nothing goes up forever.
