No, AI Will Not Make Everyone a Founder

Over the past few weeks I have been to a string of venture capital events, and I met a lot of VCs and founders. One idea came up in almost every conversation: technology is no longer a barrier to launching a start-up. More than once, I even heard that AI is about to turn everyone into a founder.

The argument goes like this. If one person can now ship in a weekend what used to take a team six months, the barrier to entry is gone, and the world is about to fill up with companies. Dario Amodei, Anthropic's CEO, put 70 to 80% odds on the first one-person billion-dollar company appearing in 2026. Carta's Solo Founders Report shows the share of new startups with a single founder rising from 23.7% in 2019 to 36.3% in the first half of 2025.

I believe the first half of that argument. I build with these tools every day at Quanthome, and they have changed how fast we go from an idea to something a client can touch. What I don't believe is the conclusion. AI makes it much cheaper to start a company. I have not seen it make it any cheaper to keep one going, and that is where most of the work is. I also think this matters more to the investors in those rooms than to anyone else, because it changes what a pitch can prove.

What AI really changes

I agree with the optimists on one thing: the building part of a startup has never been cheaper. In March 2025, Y Combinator said that for a quarter of its Winter 2025 batch, 95% of the code was written by AI. The partners were careful to add that these founders were highly technical and could have written it all themselves. They simply no longer had to.

I see it in my own work. A mockup that used to take us weeks now takes an afternoon, and a prototype a client can try takes a few days. The search for product-market fit, which used to burn through a seed round, can now run on a fraction of the money and the time.

It also removes a signal. A few years ago, a working product in a first meeting told an investor something real: this team can build, and it cared enough to spend a year doing it. Today a polished demo tells you much less, because it can be put together in a weekend. I noticed at these events how many decks still lead with the product, as if the product were still the hard part.

The evidence was never about the idea

The data has been saying this for a long time. When Bill Gross of Idealab analysed 200 companies for his 2015 TED talk, the idea ranked third among the factors that explained success. Timing came first, accounting for 42% of the difference between successes and failures. Team and execution came second.

A good part of timing is chance, and no tool helps with that. What a founder can control is whether the company is still alive when the market is finally ready.

The age data points the same way. An NBER study of 2.7 million founders in US Census data found that the founders of the fastest-growing new ventures were, on average, 45 years old, and that prior experience in the same industry was one of the strongest predictors of success. I read that as a reason to take industry experience seriously. Understanding a market and the people in it takes years, and I suspect some of the patience comes from having been wrong before. I would give that more weight than the speed at which someone can produce a prototype.

The part nobody can generate

The investors who have seen the most founders tend to agree. Paul Graham, after years of watching Y Combinator companies, concluded that determination was the most important predictor of success, ahead of intelligence. He was precise about it: being strong-willed is not enough, you also have to be hard on yourself.

Ben Horowitz gave the darker version in his essay "The Struggle". He describes the moment when the plan has failed, the product is broken, the team is tired, and there is no guarantee of getting out. One line has stayed with me since I first read it: "The Struggle is when people ask you why you don't quit and you don't know the answer."

Every founder I know has had those months. A client who was supposed to sign goes silent, or someone you counted on leaves. Nothing you built is wrong exactly, but nothing is working either. A faster mockup is no help at that point. You have to find a reason to get up the next morning and do the same unglamorous work again, usually without any evidence that it will pay off.

AI does not carry that for you. I use it to draft investor updates and to explore product directions. I still have to decide which direction deserves five years of my life, and to keep saying no to the others when one of them starts to look easier. I think that kind of steadiness takes a long time to develop, and you mostly find out whether you have it when there is little sign of progress.

What changes if I'm right

If I'm right, AI will produce many more starts and roughly the same number of founders. We will see an explosion of projects and weekend launches, and that is a good thing. But as building gets cheaper, it says less about a company, and the question of whether the founders will still be there in five years says more.

For investors, I think this makes the job harder. Deal flow will grow much faster than the number of companies worth backing, and the product will say less and less about the team behind it. If I were on the other side of the table, I would look for what cannot be produced in a weekend: how many years the founders have spent inside their market, whether clients pay and renew, what data or relationships the company owns that a newcomer would need a decade to rebuild, and how the team behaved the last time things went badly. I would also expect capital to change roles. As building costs fall, more of the money should go to distribution and to the time it takes to earn clients' trust.

For anyone thinking about starting something, my advice would not be "learn the tools", though you should. "Can I build this?" is no longer a useful question, because the answer is almost always yes. I would ask whether I will still want to build it in year three, if it is not working yet.

AI has made the first step much lighter. I don't think it has made the road any shorter.

Sources

• Amodei, D., quoted by CO/AI, 2025: Anthropic CEO predicts billion-dollar solo startups by 2026
• Carta, 2025: Solo Founders Report 2025
• TechCrunch, 2025: A quarter of startups in YC's current cohort have codebases that are almost entirely AI-generated
• Gross, B., TED, 2015: The single biggest reason why start-ups succeed
• Azoulay, P. et al., NBER, 2018: Age and High-Growth Entrepreneurship
• Graham, P., 2009: The Anatomy of Determination
• Horowitz, B., a16z, 2012: The Struggle