San Francisco's AI boom has pushed housing costs so high that even couples earning a combined $365,000 a year are struggling to find an apartment, according to a new report. Median rent in the city has climbed to $3,827 and the median home price has hit $1.7 million, driven in large part by AI companies concentrating enormous wealth among a relatively small pool of employees.
Six-Figure Salaries, No Apartment
The New York Times profiled a recruiter earning $180,000 a year and her partner, a software engineer making $185,000, who spent three months searching for an apartment under $5,000 a month and came up empty. The engineer eventually relocated to Lake Tahoe, while the recruiter still lives with roommates, paying $1,650 for her share. Their situation illustrates a broader shift: salaries that would count as upper-middle-class almost anywhere else in the country no longer guarantee decent housing in San Francisco.
Why Rents Are Climbing So Fast
Vacancy rates in sought-after neighborhoods like the Marina District and Pacific Heights have collapsed from about 13 percent in 2020 to roughly 3 percent today, a tightening that tracks closely with the AI industry's growth in the city. As AI labs hire aggressively and pay well above typical tech salaries, competition for a fixed supply of housing intensifies, pushing prices up for everyone, including workers in fields outside AI who aren't seeing comparable pay increases.
A Widening Wealth Gap in Tech
Venture capitalist Deedy Das of Menlo Ventures has described the emergence of a new AI elite: roughly 10,000 people now worth more than $20 million each, concentrated heavily in the Bay Area. OpenAI alone reportedly created about 75 multimillionaires through an internal share sale last fall, with each person cashing out around $30 million. That kind of sudden, concentrated wealth creation is a major factor behind San Francisco's housing squeeze, since a relatively small number of extremely high earners can outbid a much larger pool of conventionally well-paid tech workers for the same limited housing stock.
What Happens if OpenAI and Anthropic Go Public
The pressure could intensify further if OpenAI and Anthropic complete initial public offerings, since both companies are reportedly valued at close to a trillion dollars each. An IPO would let early employees and investors cash out equity at scale, likely creating another wave of newly wealthy residents competing for the same housing market that's already squeezing six-figure earners. For workers outside the AI boom's inner circle, that raises the odds that San Francisco's affordability problem gets worse before it gets better.
A Familiar Pattern for San Francisco
San Francisco has been through concentrated tech wealth booms before, from the dot-com years to the mobile and social-media era that brought Twitter, Uber, and Airbnb into the city's core. Each wave tended to follow a similar script: a surge of well-funded hiring drives up demand for a housing supply that barely grows, since new construction in the city has consistently lagged behind job growth for decades. What distinguishes the current AI boom is the sheer concentration of wealth at the very top, with a comparatively small number of employees and early investors capturing outsized gains, rather than the broader base of newly minted millionaires seen in some earlier tech cycles.
San Francisco Housing and the AI Boom: Frequently Asked Questions
Why is San Francisco rent so high right now?
A combination of falling vacancy rates and aggressive hiring and pay from AI companies has driven median rent to $3,827, squeezing out even well-paid tech workers who aren't part of the AI wealth boom.
How could OpenAI and Anthropic IPOs affect housing?
Both companies are valued near a trillion dollars each. A public offering could create another wave of newly wealthy employees, adding more competition for San Francisco's already limited housing supply.
For more on how AI wealth is reshaping the industry, visit our AI news section, and browse our business coverage for related market stories.
