Let me be blunt: the one piece of advice two generations of founders grew up on — "don't go it alone" — is now outdated. And I say that as someone who both builds companies and backs other people's with money.
For twenty years, one line was the gold standard. In 2006, Paul Graham, co-founder of Y Combinator, ranked Single Founder as mistake number one in "18 Mistakes That Kill Startups" — above a bad niche, above launching too slowly. The verdict: "To start with, it's a vote of no confidence" — you simply couldn't talk a single friend into joining you. And the finisher: "Starting a startup is too hard for one person."
I've heard that rule a hundred times. And I think it's already dead. The only question is why it held on so long.
The dogma ran on numbers. The numbers weren't entirely honest
The "you need a co-founder" camp didn't pull its case out of thin air. Of Y Combinator's top 100 companies, only 4 had no co-founder. Among unicorns, just ~20% were solo-founded (Defiance Capital, 845 companies). First Round Capital, across ~300 investments, measured teams outperforming solo founders by 163%. And Noam Wasserman, studying nearly 10,000 founders, produced the classic scare stat: 65% of failures come from people problems, not the market.
Here's where it clicks for me. Reread that last number. The leading cause of startup death isn't that one person lacked the hands — it's that people can't get along. The cure for "people problems" was sitting inside the disease the whole time.
And the data caught this before AI ever showed up. Jason Greenberg (NYU) and Ethan Mollick (Wharton), in a paper with a sly title — "Sole Survivors" (2018) — delivered heresy for its time:
"Ventures started by solo entrepreneurs generally outperform teams of co-founders, particularly two-person teams." — Greenberg & Mollick, "Sole Survivors", 2018
Solo founders were 2.6x more likely to keep a live business. Greenberg himself admitted what he hadn't believed: "It sounds like two heads should be better than one." In the data — they weren't.
AI settled the argument
Then came the thing that closes the question entirely. In early 2024, Sam Altman described a bet in his group chat of tech CEOs:
"…a betting pool for the first year that there is a one-person billion dollar company, which would have been unimaginable without AI and now will happen." — Sam Altman, 2024
A one-person unicorn. A year later, Anthropic's Dario Amodei put the odds of that scenario at 70–80% by 2026, quipping on stage: "maybe it's a two-person company instead of a one-person company, but we'll get close."
And this isn't visionary talk. The average headcount of a company under a year old fell from 7–9 people fifteen years ago to 3–4 today. AI solo-founder registrations hit roughly 2,600 in Q1 2026 — nearly double six months earlier (Stripe/a16z). AI startups themselves took 61% of global venture capital in 2025. The market is voting with money for small teams and big leverage.
Solo founders always built giants
What always annoyed me is that the "impossible" was in plain sight. Bezos incorporated Amazon alone in 1994. Omidyar launched eBay by himself in 1995 — profitable in month one, its first sale a broken laser pointer for $14.83. Dell built his company out of a dorm room on $1,000. Blakely built Spanx on $5,000, took no outside money, kept 100% — and became the youngest self-made female billionaire.
Naval Ravikant put the "why" best: "Code and media are permissionless leverage… software and media that works for you while you sleep." Code is leverage that doesn't ask permission. His answer to the skeptics: "Minecraft and Bitcoin were both 1-person projects."
Now that lever has gone electric. Base44 — one founder, Maor Shlomo — sold to Wix for $80M cash six months after launch, with a team of about 8. That's the whole argument.
But I'm not selling you a fairy tale
It would be dishonest to skip the other half. Solo is not a cheat code.
Y Combinator still keeps solo founders at ~10% of a batch (11% in the Winter 2026 class, versus 64% two-person teams) and demands stronger traction from them — the "bus factor" hasn't gone anywhere. Carta showed the skew: solo founders were already 30% of 2024 startups but got only 14.7% of the money. The cohort is growing faster than the capital chasing it.
And loud stories can lie. Medvi — which The New York Times in 2026 framed as a "$1.8B company with two employees," and after which Altman wrote he "would like to meet the guy" — fell apart: an FDA warning, a disputed valuation, hype instead of AI. So filter the headlines. I do.
My bet
The "solo vs. team" debate was framed wrong from the start. What matters isn't the number of people on a slide — it's leverage and decision speed.
Graham was right for his world: in 2006, building something took hands, and hands alone were the bottleneck. But the bottleneck, as the data showed, was more often the friction between hands than the hands themselves. AI removed that too: hands now rent by the batch, on demand, with no equity and no boardroom fights.
So my position is simple. Being a solo founder is no longer a vote of no confidence in yourself. It's a vote of confidence in leverage. Not in every niche — fintech and heavy R&D stay team sports. But if you've spent years waiting for "the one" co-founder as permission to start — stop. Permission no longer comes from a partner. It comes from the tool.
Graham wasn't wrong. It's just that since 2006, what changed isn't founders' courage. It's the leverage — and with it, the rules.