AI executives are pricing San Francisco families out of their own neighborhoods. The math is not complicated. Anthropic and OpenAI are producing millionaires the same way the 1849 gold rush produced millionaires — by letting the people who got there first sell shovels to the people who arrived after, and the housing market is the machinery that makes sure the rest of us pay for the privilege of living in the same city. Six hundred OpenAI employees sold locked shares last fall and split $6.6 billion in tender offers. Roughly 75 of them walked away with $30 million each. Meanwhile, the median house in the city they just priced out their classmates sells for $2 million — an 18 percent jump from the year before — selling in 29 days. A one-bedroom apartment rents for $4,000 a month, a two‑bedroom for $5,500, and a family of four pulling in six figures now qualifies as low‑income. The structural math we’ve been tracking since these companies began testing investor appetite is already playing out on the street.
The Guardian piece documents what real estate agents are calling a pre‑IPO buying frenzy. Drew Wilkerson at Sotheby’s put the logic on the record: “If somebody’s thinking about it wisely, they’ll be thinking: ‘Well, I have this large sum of money coming my way. What is a large purchase that I may need to acquire at some point?’ And the home is on that very short list.” That home is not primarily a place to raise children in this formulation. It is a place to park capital before the next wave of capital makes parking capital more expensive. The arithmetic is rational, and it is also a description of a housing market that has stopped being a housing market and become a wealth‑storage mechanism for people whose wealth was generated somewhere else entirely. You bring a bag of cash to the open house half an hour early because the open house is not a marketplace anymore. It is an auction for the IPO windfall.
I grew up in a Lansdale rowhouse my parents paid off in 2007 on a single Postal Service supervisor’s salary, and I recognize my own privilege in the union pension and the two‑decade housing lock that made it possible. What I do not have is a city that operates on the same contract. I live in Philadelphia, not San Francisco. My mortgage is not $2 million. But I know what it feels like to run the math and watch the math not add up, and the math in San Francisco right now should make every millennial parent who has ever opened a daycare invoice and a mortgage statement on the same morning feel something between rage and recognition.
This is not the first time the city has been through the cycle. Ken Rosen at UC Berkeley’s Fisher Center noted that the dotcom era produced a millionaire class that sent house prices soaring, and the early‑2010s Facebook‑and‑Twitter IPOs did the same thing. “Booms are always followed by busts. Always,” Rosen told the Guardian, and after the dotcom crash, “house prices corrected downward for the next four years, five years. Easy come, easy go.”
Easy come, easy go for the people whose wealth was tied up in the boom. But the people who got priced out of the rental market during the boom do not get four or five years of downward correction on their rent. They get evicted. They move farther out. Their kids change schools. The housing market corrects; the household budget does not. Anne Helen Petersen spent a decade documenting the millennial burnout trap — the promise that if you optimized your time, managed your cognitive load, and treated yourself as a human‑capital asset, you would eventually buy the stability your parents had. The AI boom is the final audit of that promise: the reward for doing everything right in the knowledge economy is not a mortgage you can write down. It is a stock option you hope to cash out before the rent goes up again. Brigid Schulte calls the fragmentation of a working parent’s schedule “time confetti”; San Francisco is now generating “equity confetti,” but it only falls on the people who already own the building. When the broader investment frenzy finally hits the revenue wall, the bust will correct the asset prices, but the families paying $5,500 a month for a two‑bedroom in Mission Bay do not get a correction. They get an eviction notice.
The Joint Center for Housing Studies has been documenting the numbers for years: half of all U.S. renters were cost‑burdened as of 2022, and the fastest‑growing burden cohort was middle‑income renters making $45,000 to $75,000. In San Francisco those numbers look almost quaint. A family of four making $100,000 — a number that would read as solidly middle‑class in Lansdale, where I grew up — is low‑income in a city where the average two‑bedroom apartment costs $5,500 a month. That is $66,000 a year in rent. Before food. Before childcare. Before the ER bill that arrives eleven weeks after the visit.
Taylor Swift put the structural reality of American care infrastructure into the title of “You’re On Your Own, Kid” — not as a motivational poster, but as a receipt for what happens when a housing market answers only to IPO windfalls. The vertical institutions are not coming for you. That is why we need the real pro‑family policy architecture we already know how to draft. It looks like a municipal housing trust fund that pulls land off the speculative board before the tender offer turns the block into a private auction. It looks like rent stabilization that outpaces the wave of IPO cash so working tenants aren’t priced out by pre‑IPO anticipation. It looks like a universal child allowance that provides a structural floor regardless of equity‑lottery outcomes, shielding families from a market that only recognizes capital gains. The United States is one of a handful of countries — with Papua New Guinea, Suriname, and Tonga — without a national paid‑leave guarantee. We spend a quarter of the OECD average on public childcare. We fund schools through local property taxes, which means the school my child attends is funded at a fraction of the school the senator’s child attends. Those are policy choices made by named actors at named moments, and the fact that a handful of AI companies are about to produce the largest IPO in history while the city where they are headquartered cannot house the people who work there is not an accident.
The AI boom will produce a bust. Rosen is right about that. Booms are always followed by busts. The question is whether the people who get washed out in the bust are the same people who got rich in the boom, and the answer, in housing markets, is almost never yes.