Paul Graham: The proportion of single-founder companies at YC rises from 9% to 18%
Y Combinator co-founder Paul Graham noted that the proportion of single-founder companies in the recent YC batch has significantly increased, rising from 9% last summer to 18% this summer.
He explained that the main reason is that more single founders are applying, speculating that AI tools have greatly increased the workload that one person can handle. PG himself still does not believe this is a good idea, as the value of co-founders lies not only in sharing the workload but also in jointly bearing the pressures of entrepreneurship.
Historical data shows that YC has long preferred teams of 2-3 people. In early batches like S16, only about 8.5% were single founders, and only 4 out of the top 100 YC companies entered as single founders. The Winter 2026 batch has accepted 22 single founders, accounting for about 11% of 199 companies, with a particularly higher proportion in the development tools and infrastructure sectors.
Data from Carta and Stripe shows that since the emergence of AI, the growth rate of single-founder startups has surpassed that of multi-founder teams, with single founders now accounting for more than one-third of new companies, and in some quarters, the number of solo registrations is nearly several times that of multi-founder teams.
AI productivity tools have lowered the early engineering and operational barriers, making it easier for technical or highly capable individuals to independently advance products to a fundable stage, thus increasing application numbers and acceptance rates.
Market mechanisms are shifting from traditional "team complementarity" screening to evaluating "individual leverage capabilities." Driven by AI, investors are more willing to provide seed funding to single founders with a complete tech stack and high agency, benefiting individual entrepreneurs who can quickly validate products, while traditional team models that rely on labor division and still prefer strong co-founders face pressure.
Source: Public Information
ABAB AI Insight
Paul Graham lists "single founder" as the number one mistake in his book "The 18 Mistakes That Kill Startups," arguing that the lows of entrepreneurship require a shared "esprit de corps" for support, and that single founders are often seen as a signal of being unable to persuade friends to join. Early YC data reinforces this stance: only 4 out of the top 100 companies entered as single founders, and in the S16 batch, single founders accounted for only 8.5%.
In terms of capital pathways, YC has historically concentrated resources on selected teams through its batch model and Demo Day, emphasizing co-founder matching to reduce failure rates; now, as AI lowers the execution costs for individuals, the supply of single founders applying has increased, and funding is beginning to tilt towards individuals with a complete tech stack who can independently advance to a fundable stage, with the strategic motive being to capture the productivity leverage brought by AI rather than simply diversifying risk.
A similar case can be seen with Dropbox: Drew Houston initially applied as a single founder but was required by YC to find a co-founder before being approved, ultimately becoming one of YC's most successful cases; the industry is currently transitioning from "team forced complementarity" to "AI-enabled individual execution," with the proportion of single founders in the development tools and infrastructure sectors significantly higher than average.
The structural judgment belongs to a technology-driven reconstruction of the industrial chain: AI tools have replaced early engineering and operational labor division, making individual leverage capabilities a new screening variable, with the mechanism being that as productivity boundaries shift outward, the smallest effective unit of entrepreneurial organization shrinks from "2-3 people complementing" to "1 person + AI agent," thereby changing the risk pricing logic of accelerators and early capital.