At 2 a.m. in Tel Aviv last spring, a former data-company CEO named Maor Shlomo woke to his phone alarm, walked to his laptop, and confirmed his servers were still running. He did this every two to three hours for months. He had no team. He had built a software company called Base44 by himself in four months using AI coding tools, and the alarm was the only thing between his platform and a six-hour outage. On a night when the site went down under a traffic spike, Fortune reported , he caught it in ten minutes. In June 2025, Wix acquired his company for $80 million.

This story isn’t an anomaly. Every founder in my network has a story like it. At a small dinner in San Francisco last month, a founder I’ve known since 2018 described her company to the table like this: "It's me, my dog, and about 40 AI agents." She was not being cute. She had exited her last startup, taken six months off, and quietly built a compliance software product that now serves close to 300 customers. Payroll is one person. Her team, functionally, is a wall of subscriptions to Claude, ChatGPT, Replit, and a dozen agentic tools I'd never heard of.

Every founder at the table nodded like they knew someone doing the same thing. Half of them were.

The dominant coverage still frames AI as a headcount-reduction story: cuts at Meta, layoffs at Oracle, cuts across engineering, marketing, and middle management. That story is real. What it misses is what is happening on the other side of the ledger.

According to the U.S. Census Bureau’s Business Formation Statistics , Americans filed 1.56 million new business applications between November 2025 and January 2026, the largest three-month stretch on record since the series began in 2004. January alone saw 532,319 filings, 36.8 percent above the same month a year earlier . Monthly small business formations now average above 478,000, a rise of more than 435 percent from the roughly 90,000-a-month pace of 2004. The number of people listing "founder" on LinkedIn is up 69 percent year over year .

The AI job story is not just a layoff story. It is a redistribution story, and the winners look nothing like the corporate incumbents most of the headlines focus on.

What Does This Founder Wave Actually Look Like?

The clearest read on it came on July 6, when Nathaniel Whittemore devoted an episode of The AI Daily Brief to a phenomenon he called the rise of one-person million-dollar companies. His framing was direct: AI is not just changing jobs, it is changing the risk-reward calculus of building a company. When one operator with AI agents can do what previously required a Series A team, the math for a would-be founder shifts. So does the decision for a laid-off senior manager weighing whether to keep chasing a director role or ship something of her own.

AI isn't just changing jobs. It's changing the risk-reward calculus of building a company. Nathaniel Whittemore, host, The AI Daily Brief

Fortune’s Beatrice Nolan traced the mechanics in a May 18 feature . Alongside Shlomo, she profiled Dana Snyder, a nonprofit consultant who used Replit's AI coding tools to build a software platform that walks small nonprofits through building monthly giving programs, targeting the roughly 93 percent of U.S. nonprofits too small to afford a human consultant. Snyder is still her company's only full-time employee.

"If we can use AI for the manual, repeatable tasks," Snyder said in an interview with Fortune, “we then have more brainpower to spend on ideating, which is the only thing that, as humans, we should really be spending our time on.”

The data catches up to the anecdote. J.P. Eggers, a professor of entrepreneurship at NYU's Stern School of Business, told Fortune that the average headcount of one-year-old companies has been shrinking for two decades. Twenty years ago, a one-year-old company employed seven to nine people. Two or three years ago, three or four. AI is compressing that number toward one.

Why Should The Layoff Numbers Be Read Differently?

The World Economic Forum’s Future of Jobs Report 2025 projected 92 million jobs displaced globally by 2030 and 170 million new ones created, a net gain of 78 million. Those forecasts have been widely reported and just as widely doubted, because the layoffs are concrete and the new jobs are theoretical.

But the founder data is the new jobs, showing up early. When a laid-off product marketer starts a consulting firm with an AI research stack and one contractor, that job does not appear in any large-company hiring report. It appears in the Census Bureau’s EIN filings. That is not a rebrand. It is a labor market restructuring in real time.

The Wall Street Journal has been tracking the same trend at the extreme end, reporting on teen founders using vibe-coding and social media to build real companies with no VC and no headcount. The point is not that a 15-year-old can now do it. The point is that the same tooling is sitting on the desk of every laid-off knowledge worker in America.

What Skills Actually Matter Now?

If the shape of the winning career is shifting from employee to operator, and from big-company specialist to small-team generalist, the skills that pay off shift with it. Three matter more than the rest.

The first is creativity, specifically the ability to prompt with a question no one else has thought to ask. A generative model gives you what it thinks the average person wants. The founders pulling ahead are the ones prompting with a constraint, a specific customer pain, or an unexpected combination of two disciplines. Ask a boring question, get a boring company. The World Economic Forum’s workforce research now flags creativity as one of the skills becoming more valuable, not less, as AI absorbs routine work.

The second is skepticism. The single most expensive habit any AI user can build is trusting the first answer. Large language models predict the most plausible next word, not the most correct one, and the Stanford AI Index has documented hallucination rates that have plateaued rather than disappeared as models have scaled. Treat every AI output as a first draft, never a final answer. That muscle is what separates a founder from a bot operator.

The third is curiosity. The half-life of a technical skill is now under 2.5 years, according to workforce research summarized in the 2026 Global AI Jobs Barometer from PwC . The founders who will compound over the next decade will not be the ones who mastered one tool. They will be the ones who assume every tool they use today is obsolete in eighteen months and treat learning the next one as the actual job.

The AI economy is quietly redrawing where jobs live, who owns them, and what it takes to make one. Would-be founders should stop reading the layoff coverage as bad news and start reading the Census filings as a map of where the next decade's work is going. The founders who build the muscles to prompt with creativity, question with skepticism, and learn on a two-year clock will not just survive this economy; they will be the ones staffing it.