The San Francisco real estate market is experiencing a unique phenomenon: sellers are increasingly turning away from traditional dollars and accepting shares of private AI companies — OpenAI and Anthropic — as payment. In my view, this is one of the clearest signals that the "artificial intelligence bubble" is beginning to spiral out of control.
The paradox of the situation is that neither OpenAI nor Anthropic have gone public yet. Their shares have no public market value and cannot be freely sold or exchanged. Nevertheless, real estate sellers in the Bay Area are already actively accepting such "paper" in transactions.
How homes are traded for shares: three illustrative cases
Investor and builder Nima Gabbay listed his home for $2.995 million and explicitly stated he was willing to accept shares of OpenAI or Anthropic as payment. An OpenAI employee immediately offered him $1 million above the asking price, valuing his own shares at an inflated rate. The buyer simply "stretched" the potential value of his paper to make the offer appear larger — since there is no real price for them.
Tech entrepreneur Vijay Chattoi went even further: he offered a $500,000 discount on his $2.5 million home if the buyer paid with Anthropic shares. The deal has not yet closed.
And Storm Duncan, head of a tech investment bank, created a separate LinkedIn page for his $8 million property and directly contacted Anthropic employees with an exchange proposal.
Why the housing market has gone vertical
The reason for the frenzy lies in the anticipation of upcoming IPOs. According to my data, the public listings of OpenAI and Anthropic could make millionaires out of over 16,000 people in San Francisco and its surroundings. Future millionaires are rushing to buy homes before that money hits the market and drives prices even higher.
The numbers confirm the overheating. This year, 144 homes were sold for at least $1 million above the asking price — compared to just 8 such deals in the first half of 2025. The number of transactions over $10 million has doubled in the last six months. In one case, a home listed for $7.9 million sold for $15 million.
A supply shortage exacerbates the dynamics: fewer than 600 homes remain on the market — roughly 40% below the ten-year average.
Echoes of the 2007 crisis: but in reverse
Analysts draw parallels to the 2007 crisis, but with a twist. Back then, people took out loans based on inflated home valuations and spent that "paper" value as real money. When the valuations turned out to be wrong, the debt remained — unlike the value. Now, the same pattern is unfolding in reverse: people are paying for homes with overvalued shares of private companies.
My expert opinion: We are witnessing a classic sign of a bubble, where assets begin to be used as a means of payment in the real economy before the market has objectively valued them. The risks here are enormous: if the valuation of AI companies corrects, real estate sellers will be left with devalued "paper," and buyers with overpaid homes. This reminds me of the situation with tokenized assets in the crypto sphere, where the illusion of liquidity often turns into harsh reality.