You have probably seen the headlines. A seller agrees to take pre-IPO OpenAI stock for a house. A Peninsula listing sells in a week over asking. The story writes itself: AI money is flooding the Bay Area and pricing everyone else out.
The real picture is more specific, and more useful if you are buying or selling here. AI wealth is real, and it is moving prices, but it is moving them in one part of the market, not across the board. The metro-wide average hides what is actually happening.
This piece walks through what the numbers show, how the money reaches housing before any company goes public, what past tech booms actually did to prices, the honest counterarguments, where the demand lands on the map, and the question in every online thread: will it crash?
The K-shaped market: what the numbers actually show
Start with the split the averages hide. Analysts call this a K-shaped market: an economy where the top tier pulls away and climbs while the lower tier slips, the two lines diverging like the arms of a letter K. Bay Area housing has been tracing that shape.
Per Redfin, as reported by Fortune, the divergence lines up with the arrival of ChatGPT in November 2022. Since then, Bay Area luxury homes (roughly the $3.1M to $7.6M band) are up 13.4%, while affordable homes (roughly the $535K to $615K band) are down 3.8%. Before ChatGPT, from 2020 to 2022, every price tier grew at about the same rate. The gap opened after.
The spring 2026 figures sharpen the point. Per Redfin, the San Francisco metro luxury median sale price (the top 5% of the market) reached about $6.8 million, with luxury sales up 22.2% from a year earlier. The non-luxury median (the middle of the market) was essentially flat, up 0.1%. Luxury homes there sold in a median of 12 days, down from 28 a year earlier. Money at the top is moving fast; the middle is barely moving at all.
The split is not only luxury versus affordable. It is also geographic. Per C.A.R.'s (the California Association of Realtors) May 2026 figures, single-family county medians tell different stories from one county to the next: San Francisco at $2.2M, up 22.2% from a year earlier; San Mateo at about $2.4M, up 9.1%; Santa Clara at about $2.1M, down 3.3%; Marin at about $1.81M, down 4.0%. Some expensive counties rose sharply while others slipped.
Now the number that ties it together. The S&P Cotality Case-Shiller index is a repeat-sales index, a gauge that tracks the same homes selling again over time, so it reflects the whole market rather than just whatever happened to sell at the top this quarter. Per that index, the broad San Francisco metro gauge was up only about 1.3% year over year in spring 2026, and it still sits below its 2022 peak. That is the tell. AI money is lifting the top tier, not the whole market.
How AI money becomes a down payment before any IPO
To understand the pressure at the top, follow the money from stock to down payment. Most people assume it starts with an IPO (initial public offering), the moment a private company first sells its stock to the public. But AI wealth is already moving well before any of these companies goes public. (If that is your situation, see the guide to buying a home after your liquidity event.)
One path is a tender offer: an organized sale in which a private company lets employees cash out some of their shares to outside investors at a set price, without waiting for an IPO. In an October 2025 tender offer, more than 600 current and former OpenAI employees sold about $6.6 billion in shares. Anthropic ran its own large tender offer in spring 2026, though many employees chose to hold, betting on a future IPO. By press tallies, employees at OpenAI and Anthropic together have sold on the order of $14 billion in private shares as both companies head toward the public markets.
Another path is already fully liquid. NVIDIA, headquartered in Santa Clara, is public, so its employees' stock can be sold today. Its RSUs (restricted stock units, company shares granted as pay that vest, meaning they become the employee's to keep and sell, on a schedule over several years) have risen enormously alongside the stock. Thousands of NVIDIA staff work in the Bay Area.
The timing is why this is in the news now. According to press reports, both OpenAI and Anthropic confidentially filed for IPOs in June 2026, Anthropic around June 1 and OpenAI around June 8.
Then there are the stock-for-homes headlines. In 2026, sellers of at least three multimillion-dollar Bay Area homes publicly said they would accept pre-IPO OpenAI or Anthropic stock as payment. Here is the honest version: as of this writing, none of those homes had actually closed in a stock deal. These transfers are possible but constrained. Private-company shares usually need board approval to change hands, and selling them is a taxable event with a value that is hard to pin down. Treat these stories as a sign of how much paper wealth is in the system, not as a working way people are buying homes.
As of July 13, 2026: confirmed, reported, unknown
Confirmed. Anthropic confirmed that it confidentially submitted a draft S-1 to the SEC on June 1, 2026. OpenAI also submitted a confidential draft S-1, but says it has not decided on timing and that it may take a while. Employees at both companies have already received liquidity through company-approved tenders and secondary sales: more than 600 current and former OpenAI employees sold about $6.6 billion of shares in 2025, and Anthropic completed an employee tender in spring 2026.
Reported, not announced. An October 2026 listing for Anthropic has been reported as a possibility, not announced by Anthropic. April 2027 is only a possible conventional-lockup scenario if an IPO occurs in October 2026 and uses a roughly 180-day restriction.
Unknown. The share count and price are not set, and actual lockup and trading-window terms will not be known until public offering documents disclose them. This section is updated as facts change.
What history says
None of this is the first tech boom to meet Bay Area housing, and the history is a useful check on the hype.
The San Francisco Fed studied the dot-com era and found that a 10% rise in Bay Area tech company valuations was associated with roughly a 1% to 2% rise in local house prices over about two years. It also found that stock prices tended to lead house prices, not the other way around. Wealth showed up first; homes followed.
The 2012 Facebook IPO offered a cleaner natural experiment. Zillow researchers found that for every 10 Facebook employees living in a given census tract at the IPO, home values in that tract grew about 1.6 percentage points more over the following year than elsewhere in the Bay Area. It is a correlation, not proof of cause, but the pattern is hard to miss.
A UCLA Anderson study called Cash to Spend, published in peer-reviewed form in Real Estate Economics (Hartman-Glaser and coauthors, covering about 4,500 U.S. IPOs), found that home prices near a company's headquarters rose about 0.7 to 0.9% around the filing and listing dates, concentrated in higher-priced homes within about two miles. The crucial detail is the condition attached to the later gain. Home-price growth picked up again after the lockup expired, and that pickup depended on the company's stock holding its value after listing. The lockup is a period, usually about six months after an IPO, during which employees are barred from selling. In other words, anticipation moves prices first, and the cash-driven leg arrives only when the stock performs.
The 2019 IPO wave is the cautionary tale. Uber, Lyft, Pinterest, Slack, and others went public in a stretch that was widely predicted to mint thousands of new millionaires and send prices soaring. It underwhelmed. By one economist's analysis (Issi Romem, then at Trulia), IPO-enriched buyers amounted to only about 2% of Bay Area home purchases over the surrounding years. The wave reinforced trends that were already in place rather than causing a sudden jump.
The honest counterweights
If AI wealth were the only force here, prices would be simpler to predict. It is not, and the strongest objections deserve a straight answer.
Decades of under-building are the baseline. The Bay Area has not built enough homes for a very long time, and that chronic shortage sets the floor under prices no matter who is buying. It is the deeper story beneath every tech-wealth headline.
Interest rates have moved prices more than any single equity event. When the average 30-year mortgage rate climbed from about 3% to about 7% in 2022, San Francisco home values fell roughly 13% from their 2022 peak. That rate shock overwhelmed the tech-wealth effect in the same window it hit.
Where the wealth lives matters as much as how much of it exists. In 2020 and 2021, remote work pulled the two apart: tech companies and their stock were booming while San Francisco prices and rents actually fell, because remote workers dispersed across the region and beyond. Money and demand are not the same thing.
Finally, the money is smaller and slower than the headlines suggest. Equity typically vests over about four years, windfalls are taxed, and private-share sales can require board approval, so the cash someone can actually deploy is far below the paper value of their holdings. Economists measure this with the wealth effect, the extra spending that follows a gain in wealth, and estimate it at only about 3 cents per year for each additional dollar of stock wealth. A dollar on paper is not a dollar into the housing market.
Where it lands on the map
Wealth this concentrated shows up in specific places, and it helps to know where.
The demand is anchored by where the companies sit. Anthropic now leases roughly 1 million square feet clustered on a few blocks of Howard Street in San Francisco, an area the press has nicknamed AI Alley. OpenAI leases more than 1 million square feet in Mission Bay and has said it plans to roughly double its San Francisco headcount by the end of 2026. That concentrates high-earning buyers in and around the city.
The Peninsula feels it too. Per The Real Deal's data analysis, off-market luxury deal volume rose sharply year over year in parts of the Peninsula and the East Bay, for example around Menlo Park and Palo Alto. An off-market sale is one that never appears in the public listings; it is handled privately between agents, and it is common at the high end, where buyers and sellers value their privacy.
The quieter movement is north. South Bay and San Francisco buyers are increasingly closing in Marin, including Tiburon, often off-market at the very top of the market. If you want the mechanics, here is how off-market sales work in Tiburon. The C.A.R. Marin county median sat at about $1.81M in May 2026, down 4.0% from a year earlier, a reminder that even in a sought-after county the county-wide number and the top-tier action can point in different directions.
Will it crash?
This is the question in every online thread, so let's take it evenly.
What the evidence supports is that concentration risk is real. A regional economy leaning hard on a handful of AI companies is exposed if those valuations fall. When so much local buying power is tied to the paper value of a few firms, a sharp drop in those valuations would be felt in the housing market too.
What the evidence does not support is a mechanical crash triggered by an IPO. Because the price response is largely anticipatory, as the UCLA study showed, the IPO day itself often disappoints: the expectation is already baked into prices before the bell rings. And the wealth effect is small per dollar, about 3 cents a year on the dollar, so even large paper gains translate into modest added housing demand. The 2019 wave is the reminder that an IPO does not automatically equal a boom.
The honest bottom line is that no one can predict a top or a crash, and anyone who tells you otherwise is guessing. What sets this market apart from most is the deeper shortage described above, which keeps a floor under prices that other regions do not have. That does not rule out a downturn. It does mean the Bay Area rarely behaves like a simple boom-and-bust story.
What this means for you
So what does this mean for you? It depends heavily on which part of the market you are standing in.
If you are buying, the tension is real and worth naming. The top tier is competitive and moving fast on scarce listings, while the middle and lower tiers have been steadier. Which tier are you actually shopping in? How much of the price strength you keep reading about applies to your band and your neighbourhood, and how much is happening in a segment you will never bid in? Those questions change the math entirely, and the answers look different for a $600K condo than for a $3M house.
If you are selling, whether your home sits in a segment AI wealth is lifting or one it is passing over depends on its price band and its location. Honest local comps (the recent sale prices of similar nearby homes) are what tell you which side of that line you are on, not the metro-wide headline.
Every situation is different. If you want to talk through where your price band and neighbourhood actually sit, call or text me at (415) 910-3958.
Lily Garipova
Lily Garipova, Realtor, in real estate since 2007, California licensed since 2016 (Cal DRE #02010731).
Email: lilyagaripova@gmail.com
Phone: (415) 910-3958
Web: lilygaripova.com
Fremont, CA
FAQ
Is AI money really what is pushing Bay Area home prices up?
It is a real force, but mostly at the top of the market. Per Redfin, Bay Area luxury homes are up 13.4% since ChatGPT launched in November 2022, while affordable homes are down 3.8%. The metro-wide Case-Shiller repeat-sales index, which tracks the same homes reselling over time, was up only about 1.3% year over year in spring 2026. The bigger baseline force is decades of under-building, a chronic shortage that sets a floor under prices regardless of who is buying. AI wealth sharpens the split at the top; it did not create the underlying scarcity.
Will AI wealth cause a housing bubble or crash in the Bay Area?
No one can predict a crash, and anyone who claims to is guessing. The evidence does show real concentration risk: an economy leaning on a few AI companies is exposed if their valuations fall. But it does not support a mechanical crash triggered by an IPO. A UCLA Anderson study found the price effect is largely anticipatory, already in prices before the offering, and the wealth effect is small, about 3 cents a year in extra spending per additional dollar of stock. The 2019 IPO wave was predicted to send prices soaring and underwhelmed instead.
How do tender offers and RSU vesting turn into home purchases before any company goes public?
A tender offer is an organized sale that lets employees of a private company cash out some shares to outside investors at a set price, without waiting for an IPO. In October 2025, more than 600 current and former OpenAI employees sold about $6.6 billion in shares that way. RSUs (restricted stock units) are shares granted as pay that vest, or become yours to sell, over several years; at a public company like Santa Clara-based NVIDIA, they can be sold immediately. Both routes turn stock into cash that can become a down payment before any public offering.
Which Bay Area areas feel AI wealth the most?
The demand clusters around where the companies sit. Anthropic leases roughly 1 million square feet on Howard Street in San Francisco, nicknamed AI Alley, and OpenAI leases more than 1 million square feet in Mission Bay. That concentrates high-earning buyers in and around the city. Per The Real Deal's data analysis, off-market luxury deal volume also rose sharply year over year in parts of the Peninsula and the East Bay, around Menlo Park and Palo Alto. There is a quieter spillover north into Marin, including Tiburon, often off-market at the very top of the market.
What usually happens to prices when a company like OpenAI or Anthropic goes public?
History says less than the headlines predict, and the published research is precise about where and when. The peer-reviewed version of the study (Hartman-Glaser and coauthors, Real Estate Economics, covering about 4,500 U.S. IPOs) found home prices near a company's headquarters rose about 0.7 to 0.9% around the filing and listing dates, concentrated in higher-priced homes within about two miles, with a further boost after the lockup (the roughly six-month post-IPO period when employees cannot sell) expired. That post-lockup boost depended on the stock holding its value after listing. The San Francisco Fed similarly found stock prices tend to lead home prices. The IPO day itself often disappoints, because the expectation is already in the price.
I do not work in tech. Should I buy now or wait for the IPO wave?
That is a decision only you can make, and it depends on which part of the market you are in. The top tier is competitive and moving fast on scarce listings, while the middle and lower tiers have been steadier, so waiting for an IPO wave to move your specific band may or may not matter. History also cautions against straight-line thinking: the 2019 IPO wave was expected to send prices soaring and underwhelmed. The most useful next step is to look at honest comps for your price band and neighbourhood. If you want to talk it through, call or text (415) 910-3958.
Are people really buying homes with pre-IPO stock?
You have seen the headlines, and they are real headlines, but read them carefully. In 2026, sellers of at least three multimillion-dollar Bay Area homes publicly said they would accept pre-IPO OpenAI or Anthropic stock as payment. As of this writing, none of those homes had actually closed in a stock deal. Private-company shares usually need board approval to change hands, and selling them is a taxable event with a value that is hard to pin down. Treat these stories as a sign of how much paper wealth is in the system, not as a common way homes change hands.
Why is the average Bay Area price up only a little if luxury is soaring?
Because the average blends tiers that are moving in opposite directions. Per Redfin, the San Francisco metro luxury median (the top 5% of the market) reached about $6.8 million in spring 2026, with luxury sales up 22.2% from a year earlier, while the non-luxury median was essentially flat, up 0.1%. The Case-Shiller repeat-sales index, which tracks the same homes reselling across the whole market, was up only about 1.3% year over year. When the top races ahead and the middle stalls, the metro-wide number lands somewhere in between and hides the split.
Is an Anthropic IPO in October 2026 confirmed?
No. Anthropic confirmed that it confidentially submitted a draft S-1 on June 1, 2026, but said the offering depends on SEC review, market conditions and other factors; the share count and price are not set. An October listing has been reported as a possibility, not announced by Anthropic. OpenAI also submitted a confidential draft S-1, but says it has not decided on timing and that it may take a while.
Will Anthropic or OpenAI employees first be able to cash out in April 2027?
No. Employees have already received liquidity through company-approved tenders and secondary sales: more than 600 current and former OpenAI employees sold about $6.6 billion of shares in 2025, and Anthropic completed an employee tender in spring 2026. April 2027 is only a possible conventional-lockup scenario if an IPO occurs in October 2026 and uses a roughly 180-day restriction. Actual lockup and trading-window terms will not be known until public offering documents disclose them.
Will an IPO tell us how many AI millionaires live in the Bay Area?
Not precisely. A prospectus may identify executives, directors, major owners and approved selling shareholders, but it will not provide every employee's grant, vesting, strike price, taxes, prior sales and residence. Redfin modeled about $198 billion of combined after-tax OpenAI and Anthropic employee equity, but described the exercise as hypothetical, used an estimated Anthropic employee-ownership share and did not publish a verified millionaire headcount.
Should I buy a San Francisco home before a possible AI-company lockup expires?
Not because of the calendar alone. Peer-reviewed IPO research found a modest local price response around filing and issuance, while the effect at lockup expiration depended on the company's later stock performance. That means anticipation can matter, but an April jump is not guaranteed. For a two- or three-year purchase, underwrite the property with zero IPO appreciation, include all carrying and resale costs, and preserve the ability to hold at least five years if the market disappoints.
Can a Bay Area home seller accept Anthropic or OpenAI stock as payment?
Only in an exceptional, fully approved transaction. Anthropic says an unapproved transfer of its shares or an economic interest in them is void. OpenAI says direct or indirect transfers, pledges and similar dispositions require its written consent. A proposed stock-for-home transaction would also need agreement on valuation and review by securities and real-estate counsel, tax advisers, title and escrow. Cash from an approved tender or verified financing is the ordinary route.
Is a two-year San Francisco home purchase a good way to trade the AI IPO thesis?
Usually not without a margin of safety. In an illustrative all-cash model, buyer closing costs, two years of property tax, insurance and maintenance, selling costs and San Francisco transfer tax require roughly 10.8% gross appreciation just to break even below a $5 million resale price, before income or capital-gains tax and the opportunity cost of cash. Financing raises the hurdle. A short hold becomes more defensible when the purchase is below conservative comps, has a separately costed value-add plan or provides meaningful owner-use value.
