Where Founders Actually Come From

I've spent the last few months researching where to apply to university. I run a design agency and a couple of other ventures, so "best school for entrepreneurship" seemed like a simple question with a simple answer.

It isn't. The three most-cited rankings disagree so completely that at least two of them have to be measuring something other than what they claim to measure.

Here's what I found underneath them.

#Three rankings, three different answers

PitchBook ranks universities by counting alumni who founded venture-backed companies over a rolling ten-year window. Their 2025 edition analysed more than 173,000 founders. The undergraduate top five: UC Berkeley, Stanford, Harvard, Penn, MIT. Berkeley took first for the third consecutive year with 1,804 founders across 1,650 companies and $68.9B raised.

US News puts Babson College first — as it has for 29 consecutive years — followed by MIT, Michigan, Berkeley, and Penn.

The Princeton Review and Entrepreneur put the University of Houston first, then UT Austin, Babson, Michigan, Washington, and WashU.

Notice what's missing from that third list. Stanford. MIT. Harvard. Penn. The five schools that produce more founders than anyone else on earth are almost entirely absent from the ranking that claims to identify the best entrepreneurship programmes.

There's a reason for that, and it's the single most useful thing I learned.

#The participation problem

The Princeton Review ranking is survey-based. Schools submit data about their courses, faculty, mentorship, and funding. Schools that don't participate are ineligible to be ranked.

Elite universities largely don't bother. So the list isn't "the best entrepreneurship programmes" — it's "the best entrepreneurship programmes among schools that filled in the form." Those are very different things, and nobody says so on the front page.

Once you see that, the other rankings start looking shakier too.

US News is driven heavily by a peer-reputation survey of academics. It measures what professors think of other programmes, which lags reality by years and structurally rewards incumbents. Babson defined entrepreneurship education for three decades and deserves enormous credit for it — but a ranking where the leader doesn't change for 29 years is measuring reputation, not outcomes.

PitchBook is the most honest of the three because it counts something real. But it counts volume, and volume flatters size.

#Volume is not the same as quality

Berkeley has roughly 40,000 students. Producing the most founders in absolute terms is partly a function of pushing more people through the door.

Run the same data per founder and the picture inverts. Stanford raised $102.2B across 1,519 founders — about $67M per founder. Berkeley raised $68.9B across 1,804 — about $38M. Stanford produces fewer founders and substantially more valuable companies.

MIT is the extreme case. Roughly one in ten MIT students goes on to found a venture-backed company. That's a density no large public university comes close to.

Neither number is the right one. But if you only read the headline ranking, you'd conclude Berkeley is twice the school Stanford is for founders, and that isn't what the data says.

#What I think actually matters

After going through this properly, I stopped caring much about ranking position and started caring about five concrete things. These are harder to find and much more predictive.

#1. Can undergraduates actually get money?

Not "does the university have a venture fund" — can a first-year student get a cheque?

• Dorm Room Fund writes $40,000–$200,000 cheques to student founders, and has funded 400+ students across 250+ companies.
• Berkeley's House Fund now manages $330M. Its managing partner has said publicly they write first cheques up to $2M and will go as small as $100,000 into a recent graduate or a dropout.
• MIT's delta v accelerator raised its grant to $75,000 equity-free per team in 2026, up from $20,000, funded by a $6M gift from the Klaviyo founders. About one in five applicants gets in.
• Penn's Weiss Tech House gives ~20 non-dilutive grants a year of $500–$5,000, and its President's Innovation Prize awards $100,000 plus a $50,000 living stipend.
• Waterloo's Velocity takes zero equity and distributes roughly $390,000 a year in non-dilutive grants.

That last one is worth pausing on. Velocity claims 1,200+ founders and $2.4B+ raised, takes nothing, and almost nobody outside Canada talks about it.

#2. Accelerator access, not accelerator existence

Most universities have an incubator. Far fewer have one an undergraduate can actually join.

Stanford's StartX takes roughly 8–10% of applicants, charges no equity, and has produced 20 unicorns from 1,300 companies — about one in twenty-eight portfolio companies becomes a unicorn. MIT's delta v is similarly selective and open to undergraduates. Munich's UnternehmerTUM was ranked Europe's best university startup programme by the Financial Times two years running and produced Celonis, Lilium, and Personio.

#3. Leave-of-absence policy — the factor nobody discusses

This one surprised me most.

If a company starts working, you need to be able to pause your degree without torching it. Universities differ enormously here and almost never advertise it.

Harvard's leave-of-absence is effectively indefinite. You can go build, and you can come back. MIT and Yale cap leaves at four semesters. Stanford caps at two years cumulative, though the culture around taking a quarter off is relaxed.

Harvard's policy is, in financial terms, a free option on your own upside. For a founder, it may be worth more than several places of ranking.

#4. Who is going to build it with you?

Babson is the best-designed entrepreneurship curriculum in the world and has the strongest founder alumni network per capita. It also has about 2,800 undergraduates, all studying business.

If you're building anything technical, the pool of people on campus who can write the code is small. That's not a criticism of Babson's teaching — it's a structural fact about co-founder availability that a curriculum ranking will never surface. MIT, Stanford, Berkeley, Waterloo, and Cornell win here for reasons that have nothing to do with their business schools.

#5. Geography, which is doing more work than anyone admits

Berkeley and Stanford's dominance is substantially a Bay Area effect. Boston does the same for MIT and Harvard. London for Imperial and LSE. Tel Aviv, Singapore, and the Waterloo–Toronto corridor all punch above their institutional weight for the same reason.

Good teaching in a city with no capital does not produce companies. This is, I think, the real explanation for why the outcome rankings and the programme rankings disagree so violently.

#The genuinely underrated ones

Three that came out of this looking much stronger than their public profile:

University of Waterloo. Mandatory co-op, four incubators, Velocity taking zero equity, and a culture where building is the default. Alumni companies include Kik, North, and Vidyard. US venture firms recruit there aggressively. It's probably the highest founder-ROI university in the English-speaking world relative to how hard it is to get into.

Berkeley M.E.T. A dual engineering-and-business degree that explicitly selects for people who have already built things. Sits inside the best ecosystem on earth with a $330M alumni fund behind it. Test-blind, so standardised scores don't help or hurt.

NUS, via the Overseas Colleges programme. Six to twelve months embedded in a startup abroad — Silicon Valley placements come with Stanford coursework — with a stipend. 5,000+ alumni, 1,200+ startups, three unicorns including Carousell. Cheap, global, and almost invisible in Western rankings.

#How to read any of this

If you're doing the same search, the method I'd suggest:

1. Use PitchBook for network density, not quality. It tells you how many builders you'll be surrounded by.
2. Compute per-capita yourself. Founders divided by enrolment. It takes two minutes and changes the ordering completely.
3. Ignore any survey ranking that requires opt-in participation. Check the methodology page for the word "participate."
4. Read the funding pages, not the ranking pages. Cheque sizes and eligibility criteria are public and far more informative.
5. Email the registrar about leave policy. Nobody does this. It's a five-minute email about the thing that matters most if you succeed.

#The uncomfortable conclusion

The strongest predictor of whether someone starts a valuable company at university appears to be who else is there and what's within driving distance — not the curriculum, not the ranking, and not the entrepreneurship centre's brochure.

Which means the honest advice is less satisfying than a top ten list: pick the densest ecosystem you can get into, make sure you can pause if it works, and understand that the institution is a network and an option, not a programme.

The programme is the part they rank. It's also the part that matters least.

---

Sources: PitchBook University Rankings 2025; The Princeton Review & Entrepreneur Top Schools for Entrepreneurship Studies 2026; US News Best Undergraduate Entrepreneurship Programs 2026; MIT Sloan; Berkeley House Fund; Waterloo Velocity; Stanford StartX; The Harvard Crimson.