In 2023, three friends who had competed together on the speech and debate team at Bellarmine College Preparatory dropped out of college in their sophomore year. Adarsh Hiremath left Harvard. Brendan Foody and Surya Midha left Georgetown. They were later awarded Thiel Fellowships, and the company they built — Mercor, which began by matching engineers with US companies and moved into supplying expert contractors to AI labs for training data — reached a reported $10 billion valuation and, as of July 2026, was reported in talks to raise $500 million at a $20 billion valuation.
All three were still under 23.
That story is being told constantly in San Francisco right now, and it is worth examining carefully — both because something real is happening, and because the lesson most people draw from it is the wrong one.
The structural shift underneath
The Mercor story is not an isolated anomaly. The pipeline into American higher education is genuinely contracting, for reasons that have nothing to do with startups.
Beginning with the fall 2026 semester, enrollments are absorbing the demographic effect of the birthrate decline during the 2007–2009 recession. Projections show a roughly twelve-percentage-point drop in the number of eighteen-year-olds entering college across 2025–2030. In a survey of more than 10,500 recent high-school graduates, 67% of those who chose not to enroll cited cost — the sharpest recorded rise in cost-driven avoidance in years. And the romantic framing is fading too: the share of non-enrollees saying they wanted time off first fell from 39% in 2024 to 26% in 2026. They are not taking a gap year. They are making a different choice.
Meanwhile the stigma attached to not holding a degree has largely dissolved in technology hiring, and the alternative routes have institutional backing. The Thiel Fellowship pays young people $200,000 to leave or skip college for two years and build; its alumni include Figma's Dylan Field and Scale AI co-founder Lucy Guo, and one analysis puts total value creation from the program above $750 billion with a unicorn rate near 14%.
So two things are true at once: fewer eighteen-year-olds are going to college, and the most ambitious slice of those who do not now has a destination.
What the destination actually is
The destination is a house in San Francisco.
This is the part that gets missed. The alternative to a university is not "working from your parents' home." For the people this works for, the alternative is a dense residential environment full of peers doing the same thing — which is, structurally, the one thing a university is unambiguously good at.
Think about what a campus actually provides a founder-shaped eighteen-year-old:
| What a campus provides | Does a founder house provide it? |
|---|---|
| Dense peers at the same life stage | Yes — often denser, and self-selected for ambition |
| Structure and deadlines | Partly — demo days, launches, investor meetings, runway |
| Credential | No |
| Formal instruction in fundamentals | No — this is the real loss |
| A safety net if it fails | No — this is the other real loss |
| Reputation and network | Yes, though narrower and more concentrated |
The two "no"s matter enormously, and we will come back to them. But the first row is why the model works at all. Nineteen-year-olds in a house with eleven other people building companies get the thing freshman year is actually prized for — the friendships, the ambient competition, the late-night argument that changes what you believe — without the four-year schedule.
The social education nobody plans for
The most underrated part is not technical. It is social, and it happens because of living together rather than working together.
A founder who lives in a house with ten others has to negotiate the kitchen, absorb criticism from someone they will see at breakfast, resolve a conflict they cannot leave the building to escape, and hold a reputation in a small group over months. That is a training environment for exactly the skills the job later demands at scale: hiring, firing, co-founder disagreements, investor relationships, telling someone their work is not good enough while keeping them.
Remote work does not teach this, and neither does a lecture hall. An office teaches some of it, but an office has an exit at 7pm. A house does not. The friction is the curriculum.
There is a legitimate research basis for taking the social dimension seriously rather than treating it as a nice side effect. Raj Chetty and colleagues' 2022 Nature work on social capital found that economic connectedness — the extent to which people form friendships across socioeconomic lines — is among the strongest known predictors of upward mobility, with estimates that children from low-income families would see adult incomes roughly 20% higher growing up in high-connectedness environments. Friendship structure, in other words, is not decoration. It is one of the mechanisms by which outcomes change.
A well-mixed founder house is a connectedness machine: a nineteen-year-old from a state school in the Midwest, a thirty-four-year-old second-time founder, a researcher who moved from Lagos, and a Stanford dropout, around one table, every night, for four months.
The honest part
If this article stopped there it would be propaganda, so here is the other side, stated plainly.
The sample is survivor-selected. You hear about Mercor because it worked. You do not hear about the hundreds of people who left school, spent two years on a company that did not find a market, and now have neither the degree nor the company. The Thiel Fellowship's near-14% unicorn rate is extraordinary for a venture portfolio and is also, read the other way, a reminder that most participants do not produce a unicorn — and they were selected from an exceptional applicant pool before they started.
The Mercor founders were not undifferentiated teenagers. They were admitted to Harvard and Georgetown, had been training together competitively for years, and had a specific, timely insight about AI labs' demand for expert data. The dropping out was downstream of having something worth dropping out for. It was not the cause.
Leaving school is reversible for some people and not others. A student with family capital, a strong network and an offer to return has a cheap option. A first-generation student on financial aid has an expensive one. Anyone celebrating the trend without naming that asymmetry is selling something.
The defensible version of the claim is narrow: if you already have real traction and a team, the marginal year of undergraduate coursework may be worth less than a year in a house in San Francisco shipping. That is a conditional, and the condition is doing most of the work.
What we see from the housing side
We run furnished houses for founders in San Francisco, which gives an odd vantage point on this. A few patterns are consistent.
The youngest guests are the most affected by the quality of the house, because for them it is not just accommodation — it is their entire social and professional environment. A twenty-two-year-old living alone in a studio in a city where they know nobody is in a genuinely bad situation, however good the studio is. The same person in a well-run house with eight other builders has a different year.
They also need the logistics removed more than anyone, because they have never done them. First US lease, no credit history, no idea how to get internet installed, parents on another continent. The all-inclusive format exists for exactly this: one monthly price covering rent, utilities, internet and cleaning, 30-night minimum, nothing to set up, nothing to assemble, no twelve-month commitment.
And they stay longer than they plan to. A one-month booking becomes four. The company changes shape. Someone they met at the kitchen table becomes a co-founder.
That is what hyperstay is for — furnished homes across San Francisco, private rooms or an entire house with six or more bedrooms, for founders in the city for a month or more. If you are eighteen and just got funded, or you are running a program that brings young builders to San Francisco and needs them housed together, write to team@hyperstay.com.
A campus is a place where people your age take the same thing seriously at the same time. That is all it ever was. It turns out you can build one in a house.