YC Solo Founders Reach 19% of Batch as Startups Push Into Hardware

A record share of YC companies now have one founder

Y Combinator has made public something its recent batches already suggested. The accelerator says nearly one in five companies it works with now has a single founder, and it presents this as one of several shifts happening together: teams are getting smaller, more startups are building physical products, and revenue is showing up earlier.21 YC's September 17 post described these as patterns it sees across the thousands of founders it works with each year. It did not tie the figures to a specific batch or time period.21

Outside counts support the number. Investor Jared Heyman analyzed YC's Summer/Spring 2026 cohort and found that 19% of companies had a single founder. The five-year average is 8.1%, so this is the highest solo share in five years and the third high batch in a row.27 The longer history makes the change clearer. In 2016, YC reported that 8.5% of its Summer 2016 companies had one founder and 61.3% had two.21

For years the standard advice was that YC wanted co-founders and that applying alone was a handicap. Guides for applicants still say solo founders make up about 10% of a typical batch and face a higher bar on traction.22 On YC's own numbers, that advice is now out of date.

The rise was gradual

The jump to 19% followed a smaller step up. Winter 2026 had 22 solo founders out of 199 companies, or about 11%.30 That batch included one-person companies outside pure software, such as a lunar hotel project, redesigned dive gear and robotic cattle herding. In developer tools, the solo rate reached 22%.30 So the pattern built over several batches. A single selection committee changing its mind does not explain it.

The reporting splits on how the solo trend relates to hardware. YC paired the solo-founder number with a video called "Startups Are Moving From Bits to Atoms." One reading of that pairing is a barbell: AI cuts the labor needed to build software, and the hardest-to-copy businesses move toward capital-heavy physical products.21 YC's Fall 2026 requests for startups back the physical-world theme. They ask for companies applying AI to healthcare, defense, infrastructure and other real-world systems.21

Heyman's data weakens the idea that lone founders are leading the hardware push. He expected solo founders to be roboticists working alone on thesis projects. He found the reverse: 64% of solo founders built software, compared with 52% of the batch overall.27 He also found that 55% of the solos were under 25, the same share as founders on two-person teams.27

The two findings fit together once you separate them. More of the batch is building hardware. Solo founders, as a group, lean toward software. Heyman put the open question directly: a young first-time founder can now get a software company far enough alone to earn YC's investment, but nobody knows yet whether one person can carry a hardware company.27 The headline pairs the two trends, but the data says solo founding is mostly a software story. Hardware is growing in the same batch and is still mostly built by teams.

YC's company directory points the same way. Many recent hardware entries are small teams of two to five people working on drones, actuators, inspection systems and semiconductors. A few list a single employee.2426

Why one person can now get this far

All the coverage gives the same explanation: AI coding tools have taken over much of the early engineering, design and operations work that used to require hiring.21 The independent review of the Spring 2026 batch found that 60% of company descriptions mentioned AI or agents.21

The tools behind this have become fairly standard. Indie developers converge on Next.js for the frontend, Supabase or a similar database, Stripe for payments, Vercel or Railway for hosting, and the OpenAI or Anthropic APIs for AI features.1 One analysis of more than 200 public case studies found that a stack of Lovable or Cursor, Supabase, Lemon Squeezy, Plausible and Resend appears in over 60% of solo businesses earning at least $1,000 a month. The same analysis puts total tooling costs around $39 a month before any revenue.3 Another survey says 75% of indie hackers use Claude Code and 42% use Cursor.6 These figures come from vendor and community surveys rather than audited data, so they are best read as signs of a dominant pattern.

Payments have changed as well. Merchant-of-record services like Lemon Squeezy, which Stripe now owns, and Paddle collect from customers and handle VAT and sales tax in many countries. For a single operator, that removes a big compliance task.4 There is now a real gap between the cheapest stacks: some guides say a full product can run for $0 to $50 a month on Cloudflare Workers and self-hosted tools, while broader estimates put a typical indie hacker's software bill at $100 to $500.85

The AI-assisted share is rising. Among products launched in 2024 and 2025, the split between traditional coding and AI-assisted or no-code building is close to even.3 One widely cited example, Formula Bot, reportedly peaked at $226,000 a month in subscription revenue and was built in Bubble by someone with no coding experience.3

What solo founders actually earn

The revenue figures are where the coverage disagrees most, and where the trend is most often oversold.

The best-known cases are real. Pieter Levels runs a portfolio of NomadList, RemoteOK and PhotoAI that brings in over $3 million a year with no employees, on deliberately simple PHP infrastructure.13 Maor Shlomo built the AI app builder Base44 alone, and Wix bought it for $80 million about six months after launch.14 App developers have their own versions. Sebastian Roehl's habit tracker HabitKit made about $600,000 in 2025 and reached roughly 25,100 paying subscribers at $1 to $2 a month. It has no backend server and stores data locally on the device.2 Its stack was Flutter, RevenueCat for subscriptions, and Cursor for coding. When iOS reached about 75% of his revenue, he moved his newer app to native SwiftUI.2 His three other apps made almost nothing, which goes against the common indie belief that a portfolio of apps is the safer bet.2

The overall numbers are much lower. One estimate says about 70% of solo founders earn under $1,000 a month and only 2% to 3% pass $1 million in annual recurring revenue.11 Another, based on Founder Reports data, says just 0.2% of solopreneurs reach $1 million in revenue.18 Community guides say the median active indie hacker earns under $1,000 a month and the top 10% pass $10,000 in monthly recurring revenue.5 Where the estimates disagree, the more cautious ones are probably closer to the truth. Every source agrees the distribution is steeply skewed toward a small number of winners.

Some of the best stories also need qualifying. One analysis notes that Base44 had eight employees when it was acquired, that 41% of solo founders report burnout, and that the one-person model tends to break between $50,000 and $150,000 in monthly recurring revenue, when the operational load gets too heavy.16 At least one outlet says solo-founded startups make up 36.3% of all new companies in 2026.19 That figure is far above YC's own numbers and should be treated with caution.

The billion-dollar one-person company is still a prediction. Anthropic CEO Dario Amodei put 70% to 80% odds on it happening in 2026.13 Some trackers say it has not happened yet.11 Telehealth company Medvi is often cited as the closest example, with reported first-year revenue of $401 million and a projection of $1.8 billion for 2026.18 Its description as having no traditional employee base is itself contested.16

Why it matters

YC's earlier-revenue claim is the least specific of its three observations. The post refers to "meaningful revenue" but gives no dollar threshold, time period or share of companies.21 The direction still matters for investors. When a working product and early sales are cheap to produce, neither one shows that a company has lasting distribution or a technical advantage.21 One summary of Winter 2026 says three times as many companies reached $1 million ARR as in Winter 2025, though the causes it lists are broad.30

The clearest conclusion from the data is that starting a company alone, as both an indie developer and a YC applicant, is now normal for software, and founders can validate ideas cheaply. The top of the market is concentrating in hardware and physical systems, where capital, regulation and manufacturing still require teams. Solo founding has become a real option for getting started, but the evidence does not yet show that one person can build a lasting company without eventually hiring. Founders are now more likely to begin alone, and their companies may still need a team later.