The city written off after the pandemic has suddenly become one of America’s fiercest luxury housing markets—and artificial intelligence is rewriting the value of proximity.
For several years, San Francisco was held up as the cautionary tale of post-pandemic America. Empty office towers, remote workers, struggling retailers, falling commercial-property values and headlines about an urban “doom loop” created the impression that the city’s great technology-era real estate premium had finally broken.
In 2026, that narrative looks increasingly obsolete.
San Francisco has become ground zero for the artificial-intelligence economy, and the consequences are beginning to spill far beyond venture-capital offices and data centers. They are now showing up in residential bidding wars, luxury-home prices, office leasing and the extraordinary amount of private-company wealth accumulating among AI founders and employees.
The most remarkable number may be this: more than 140 San Francisco homes sold for at least $1 million above their asking prices between January and June 2026.
During the same six months of 2025, there were just eight such transactions.
And 44 of the 2026 million-dollar overbids occurred in June alone. San Francisco Chronicle
San Francisco hasn’t merely recovered.
Parts of the city are beginning to look like a land grab.
The Million-Dollar Overbid Returns
San Francisco has always been somewhat unusual when it comes to asking prices. Sellers and agents frequently employ a “price-to-entice” strategy, deliberately listing desirable homes below the price they expect to receive in hopes of creating a competitive auction.
So an over-asking sale does not necessarily mean the property unexpectedly appreciated by that amount.
But the scale of what is happening in 2026 is difficult to dismiss.
In May, the average San Francisco home reportedly sold for nearly 16% above asking, more than twice the comparable figure a year earlier. And unlike past cycles, when the greatest percentage premiums often occurred lower in the market, luxury properties have become one of the centers of the bidding frenzy. San Francisco Chronicle
Redfin reported that San Francisco’s median sale price had climbed above $1.7 million by March 2026—more than $1 million higher than the city’s post-financial-crisis bottom in 2012. Redfin
Inventory is part of the equation. But this is increasingly a story about a new concentration of money.
And the money has two letters attached to it:
AI.
AI Has Replaced the Old Tech Boom
The first technology boom transformed San Francisco through companies such as Google, Facebook, Salesforce, Uber, Airbnb and Twitter.
The artificial-intelligence boom may prove even more geographically concentrated.
OpenAI, Anthropic and an expanding ecosystem of AI startups, venture investors, infrastructure companies and specialized service firms have clustered in and around San Francisco. Workers who might once have regarded Silicon Valley suburbia—or remote work from Lake Tahoe, Austin or Miami—as perfectly acceptable are once again discovering the economic advantages of being physically close to founders, investors and other engineers.
That matters enormously to residential real estate.
Housing near the center of an economic network is not merely shelter. It becomes a scarce participation asset.
The more valuable the network becomes, the more valuable proximity to that network can become.
The office market offers evidence that something substantial is changing. Colliers reported 2.1 million square feet of positive net office absorption during the first half of 2026, San Francisco’s strongest first-half performance since 2018. Overall vacancy dropped to 28%, down from a 31.2% peak a year earlier. Colliers
CBRE says technology and AI companies leased more than 14 million square feet across San Francisco and Silicon Valley in 2025, representing 55% of total leasing activity. CBRE
More recently, AI company Cognition agreed to take approximately 180,000 square feet at 333 Brannan Street, dramatically expanding its local footprint. San Francisco Chronicle
The office towers are not full again. Vacancy remains historically high.
But the direction of travel has changed.
Return-to-Office Becomes a Real Estate Catalyst
The pandemic created one of the largest geographic experiments in corporate history.
If technology employees could work from anywhere, why pay millions of dollars for a house in San Francisco?
That argument looked persuasive in 2021 and 2022.
It becomes considerably weaker when employees are expected to spend more time in offices—and when careers increasingly depend upon participating in the informal networks surrounding the world’s most heavily financed AI companies.
Cresa reported that San Francisco office visits were up 8.2% year over year during the second quarter of 2026 as return-to-office policies became more widespread. Cresa
This doesn’t mean everyone is returning five days a week.
It means location matters again.
And when highly compensated workers decide they need to be within 15 or 20 minutes of an office, laboratory, investor dinner or founder meeting, San Francisco’s famously constrained housing supply suddenly becomes extremely important.
Artificial intelligence may be digital.
Its talent economy remains surprisingly physical.
Then Comes the IPO Wealth Effect
The most fascinating part of the San Francisco property boom may involve wealth that hasn’t fully arrived yet.
The market is beginning to price in the possibility of enormous future liquidity events at major private AI companies.
A Redfin analysis cited by Axios estimated that employees of OpenAI and Anthropic could collectively possess enough post-tax equity, under certain projected IPO valuations, to theoretically purchase a remarkable portion of the San Francisco housing stock. Those numbers are hypothetical rather than predictions, but they illustrate the extraordinary scale of private equity potentially sitting inside two companies alone. Axios
This produces a fascinating psychological phenomenon.
Buyers aren’t simply competing against today’s millionaires.
They may be trying to buy before tomorrow’s millionaires become liquid.
Employees holding valuable private shares can sometimes access wealth through secondary transactions or borrowing, while conventional buyers understand that major IPOs could release considerably more purchasing power into an already supply-constrained market.
That expectation itself can influence behavior.
The result resembles a real estate version of front-running.
Buy now, the thinking goes, before the AI equity becomes cash.
The Great San Francisco Contradiction
There is an important twist.
Not every statistic says San Francisco real estate is exploding upward.
Realtor.com’s July 2026 luxury report found that the threshold for entering the top 10% of San Francisco metro listings actually fell 8.6% year over year to about $2.49 million.
At first glance, that sounds bearish.
But the same report found that million-dollar properties were selling in a median 37 days, the fastest pace among the luxury metros it studied, while the number of million-dollar listings dropped 20.9% from a year earlier. Realtor.com characterized the market as one in which inventory was being cleared rather than simply stagnating. Realtor
That distinction matters enormously.
San Francisco isn’t experiencing a uniform property boom.
It is experiencing a highly selective repricing of scarce, desirable assets.
Move-in-ready houses in neighborhoods such as Pacific Heights, Noe Valley and the Richmond can attract intense competition. Meanwhile, less desirable condominiums, troubled commercial properties and buildings dependent on the old downtown economic model can tell a very different story.
There isn’t one San Francisco real estate market anymore.
There are several.
The Global Investor’s Opportunity—and Warning
For international investors, San Francisco presents an intriguing reversal.
The bearish thesis was easy to understand: high taxes, regulation, homelessness, office vacancies, remote work and a declining downtown.
The emerging bullish thesis is equally powerful: the world’s most important AI cluster is forming inside a city where adding housing remains notoriously difficult.
That creates scarcity.
Scarcity surrounding rapidly growing concentrations of income and equity wealth is one of the oldest real-estate investment formulas in existence.
But investors should resist the temptation to interpret the current bidding frenzy as proof that every San Francisco property has become a bargain.
Three risks remain especially important.
First, AI valuations themselves are extraordinarily high. A major repricing of private AI companies would quickly reduce the perceived wealth effect.
Second, mortgage rates and affordability remain restrictive, leaving much of the market dependent upon unusually wealthy purchasers.
Third, San Francisco’s post-pandemic problems have not vanished. Office vacancy remains above levels that would have been considered alarming before 2020, and rising housing costs are once again intensifying concerns about displacement. Recent reporting puts the city’s median home price near $1.6 million while rents have risen sharply. San Francisco Chronicle
The boom is real.
So are the structural problems.
From Doom Loop to Boom Loop
San Francisco’s reversal offers a larger lesson about global real estate.
Cities can be written off remarkably quickly.
And they can be repriced just as quickly when a new industry changes the economic geography.
Five years ago, the dominant question was whether remote work had permanently destroyed the premium attached to living near a technology headquarters.
Today, the better question may be whether artificial intelligence is creating a new urban premium around proximity to people rather than simply proximity to offices.
The most valuable asset in San Francisco may ultimately not be the Victorian house, the condominium or even the office tower.
It may be access to the network.
That network now includes some of the world’s most valuable private companies, enormous pools of venture capital, thousands of highly paid engineers and potentially hundreds of billions of dollars of future equity wealth.
When that much capital collides with one of America’s most constrained housing markets, strange things happen.
Like paying $1 million over asking.
The “doom loop” hasn’t disappeared completely. But in the neighborhoods benefiting most directly from the AI economy, another cycle has clearly begun.
San Francisco is becoming a boomtown again—and this time, the gold rush is artificial intelligence.
Leave a Reply