YC says nearly one in five companies is solo-founded as startups move into hardware

Why it matters

YC's data points to a barbell startup market: AI lets one founder ship software with less labor, while defensibility increasingly moves into capital-heavy physical products.

Y Combinator (@ycombinator) says nearly one in five companies it works with has a solo founder, part of a wider shift toward smaller teams, physical technology and startups generating revenue earlier in their lives.

The September 17th post compresses three changes into one snapshot: founders are building with fewer people, more YC companies are working on physical products, and some are converting that work into sales faster. YC presented the observations as patterns drawn from the thousands of founders it works with each year, rather than statistics from a named batch or a defined period.

The solo-founder figure is still a substantial departure from YC's earlier cohorts. In a September 2016 breakdown, YC reported that 8.5% of its Summer 2016 companies had one founder. Two-founder teams accounted for 61.3% of that batch.

An independent review of YC's Spring 2026 batch also put the solo-founded share at 19%, among more than 190 companies. The review counted a median core founding team of three people and found that 60% of company descriptions mentioned AI or agents.

Those numbers describe a new founder archetype taking shape inside YC: one person can write, test and distribute software with AI systems handling work that previously required early engineering, design or operations hires. A solo founder still carries the concentration risk that comes with having no partner, but the cost of finding out whether an idea works has fallen.

AI pushes software costs down and ambition outward

YC's accompanying video, titled "Startups Are Moving From Bits to Atoms", places the solo-founder trend beside a rise in companies building robotics, industrial systems and other products tied to the physical world.

The pairing matters. Cheap code gives founders leverage inside software companies. It also lets technical teams put more of their time and capital into hardware, manufacturing, energy systems and scientific work. The software layer around a robot, factory or laboratory instrument can be built and changed faster, leaving founders to focus on the physical bottlenecks that remain difficult to copy.

YC has funded hard-tech companies for years. In 2016, it described its method as helping founders isolate a smaller, cheaper project within a capital-intensive plan, giving them a way to iterate before attempting the full system. The current move toward atoms extends that approach into a period when AI can reduce the software labor surrounding the physical product.

YC's Fall 2026 requests for startups make that direction explicit. The program says it wants companies applying AI to physical-world systems, including healthcare, defense, infrastructure, finance and work. That interest puts YC closer to businesses with longer sales cycles, regulatory exposure, manufacturing constraints and larger capital requirements than the software products historically associated with accelerators.

The result is a barbell inside early-stage company building. A single founder can launch an AI software product with little payroll, while another small founding team can use the same tools to attempt a technically difficult physical business. Team size may converge even when capital needs do not.

Revenue arrives earlier, with fewer answers about durability

YC's revenue observation is the least precise of the three. The September 17th post says companies are reaching "meaningful revenue," without attaching a dollar threshold, growth period or share of the cohort. Revenue can mean annual contracts, usage-based sales, pilots or recurring subscriptions, each carrying different implications for retention and margins.

Earlier sales still change the financing equation. Founders who can reach customers before hiring a conventional team have more room to delay fundraising, negotiate from a stronger position or keep a larger ownership stake. For investors, the same trend raises the standard for what counts as seed-stage traction. A working product and initial revenue are becoming cheaper to produce, so neither proves that a company has durable distribution or technology.

YC's snapshot captures both sides of the AI startup cycle. Company formation is concentrating around smaller founding teams, while some of the most ambitious products are moving into markets where capital, regulation and physical execution still matter. The founder can increasingly start alone. The business that follows may be anything but lightweight.