No engineering team. No office. No HR department. Just one person, a laptop, and a stack of AI tools doing the work that used to require a payroll. And a 7-figure business.
That used to be the exception. Now it’s a measurable trend. As The Wall Street Journal reported in its recent feature on one-person, million-dollar companies, the number of solo operators at Stripe platform, generating more than $1 million in revenue roughly doubled between 2023 and 2025 — and the number crossing $10 million nearly tripled over the same span. Stripe’s chief economist, Ernie Tedeschi, told the Journal: “Now, AI can be a built-in business partner.”
The shift shows up in government data too. U.S. Census Bureau figures found that 117,060 nonemployer businesses — companies with zero paid staff — crossed $1 million in revenue in 2023, roughly double the number from two years earlier, before ChatGPT’s public launch. That’s out of tens of millions of solopreneurs nationwide, which tells you something important before we get to the success stories: even now, crossing seven figures alone is still the exception, not the rule. We’ll come back to that.
The Bright Examples
Base44 — Maor Shlomo
In December 2024, Maor Shlomo, a former member of Israel’s elite Unit 8200 intelligence corps, started building an AI tool that could turn a plain-language description into a working app. He launched Base44 as a solo founder, writing almost no front-end code himself — he let AI models handle it while he focused on product direction and growth. The results were startling: $1 million in annualized revenue within three weeks of launch, and more than 400,000 users within six months, all without raising a dollar of outside funding. In June 2025, Wix acquired Base44 for roughly $80 million in cash. Shlomo has since described restructuring his codebase specifically to make it easier for AI models to write and maintain — treating the AI less like a tool and more like the primary engineer on the team.
HeadshotPro — Danny Postma
Danny Postma builds AI products from Bali, and his breakout hit is HeadshotPro, a tool that generates professional-looking headshots from a handful of selfies using AI image models. Working solo, Postma took it past $1 million in annual recurring revenue within a year, later scaling the business toward $300,000 a month at its peak. It wasn’t his first success — an earlier AI writing tool, Headlime, sold for $1 million just eight months after launch. Postma’s pattern across both products is the same: identify a task AI models had just become good enough to do commercially, then ship a narrow, well-designed product around it before the market gets crowded.
The Pieter Levels Portfolio
No one embodies the “one person, many products” model better than Pieter Levels. Levels runs a portfolio that includes Nomad List, RemoteOK, and PhotoAI — an AI headshot and avatar generator — from levels.io, generating more than $3 million a year combined with zero full-time employees. In a widely shared experiment, he built fly.pieter.com, a browser-based flight simulator, largely with AI coding tools and pushed it to $1 million in annualized revenue within 17 days of launch. Levels has spoken about running his businesses on “hundreds of automation scripts” he affectionately calls his robots — a deliberately unglamorous, boring tech stack that keeps maintenance overhead low enough for one person to manage.
ShipFast, CodeFast, DataFast — Marc Lou
Marc Lou builds tools for other indie founders. His flagship, ShipFast (sold through marclou.com), is a code boilerplate that helps developers launch a SaaS product in days instead of months; CodeFast teaches people to code with AI assistance, and DataFast handles analytics. Across the three, Lou reported crossing $1.03 million in revenue in 2025 — with no employees. The products are designed to feed each other’s customer base, so a user of one is a natural prospect for the next, a flywheel Lou built almost entirely through AI-assisted coding and public building on social media.
Chatbase — Yasser Elsaid
Yasser Elsaid was a computer science student in Toronto, interning at large tech companies, when he built the first version of an idea in six weeks: let people upload a document and chat with its contents using AI. He put up a pricing page for Chatbase and got his first paying customer within 30 minutes. Five months later, the company had crossed $1 million in annualized revenue, entirely bootstrapped. It has since scaled well past that as an AI customer-support platform, eventually growing a small team as the business outgrew what one person could run. Looking back on the decision to skip the traditional playbook of pitch decks and co-founder searches, Elsaid said simply: “I just saw the idea and I started building.”
Miss Excel — Kat Norton
Not every AI-era success is a piece of software. Kat Norton, known online as Miss Excel, built a business teaching Excel and spreadsheet skills — now including AI-assisted workflows — through short, high-energy TikTok and Instagram videos. She had no influencer background when she posted her first video in June 2020: “I didn’t know how to be any sort of influencer,” she has said of those early days. Within months she had a viral hit and a rapidly growing following; within about a year and a half, her course sales passed $1 million, run as a true one-person operation. Her business, run through miss-excel.com, has since grown toward roughly $2.9 million a year.
Formula Bot — David Bressler
David Bressler had a full-time analytics job and no coding background when, during six weeks of paternity leave in 2022, he built an AI tool that converts plain-English instructions into Excel formulas. He used the no-code platform Bubble to build the entire front end and back end himself. Formula Bot went viral on Reddit within weeks of launch and has since grown to more than 900,000 users, generating over $2.8 million a year — built and still run by one person with no formal software engineering training.
TypingMind — Tony Dinh
Tony Dinh left a job as a Big Tech software engineer to build products on his own. His best-known one, TypingMind, is a premium interface for ChatGPT and other AI models — better prompt libraries, plugins, and design than the free, official apps. Alongside it he runs BlackMagic.so, a Twitter/X analytics tool, and DevUtils, a developer utility app. Combined, Dinh’s solo portfolio generates well into six figures annually, and he’s known for posting his revenue numbers publicly rather than staying quiet about the ups and downs.
Justin Welsh
Justin Welsh spent over a decade helping build two companies past $1 billion in valuation before burning out as a SaaS executive. He went solo in late 2019 with a laptop and a LinkedIn account. There’s no AI product to point to here — his business is content, digital courses, and sponsorships — but AI has become part of how he researches, drafts, and repurposes the volume of writing his business runs on. On the fifth anniversary of that decision, in May 2025, he posted that his one-person business had just crossed $10 million in cumulative revenue: “I ran zero ads & I operate at a ~89% margin.” Everything, including that milestone post itself, was published through his own channels at justinwelsh.me — no agency, no media buy, no team.
For Every Winner, a Graveyard of Quiet Failures
Here’s the part the highlight reels leave out: for every Base44 or Formula Bot, there are thousands of solo founders who built something, launched it, and heard nothing back.
The numbers are blunt. Industry trackers of small AI-built software products estimate that roughly 82% of micro-SaaS tools never reach $5,000 a month in revenue, and 93% never reach $10,000. A survey of solo founders found a burnout rate above 50%, with roughly three-quarters reporting anxiety episodes tied to the pressure of running everything alone. And recall that Census figure from the introduction: 117,060 nonemployer businesses crossed $1 million in 2023 — out of tens of millions of solopreneurs operating in the U.S. That’s a tiny fraction. The $1 million mark is real, and more people are reaching it than ever before, but it remains a needle in a very large haystack.
Part of this is simple math: AI has collapsed the cost of trying. A landing page that once took a developer a week now takes an afternoon. That’s wonderful for the people whose ideas land — and brutal for the visibility of everyone whose ideas don’t. Nobody writes a case study about the app that got 40 signups and died quietly. The success stories above are real, verified, and impressive. They are also a curated, self-selected sample of a much larger, mostly invisible population.
How They Marketed It
Building the product with AI turns out to be the easy half of this story. Every founder above still had to solve the older, harder problem: getting strangers to notice and pay. None of them hired a marketing department to do it.
Base44 grew almost entirely for free. Shlomo has described taking the product from three friends to 400,000 users without spending money on paid marketing. The mechanism was the product itself — an app-builder produces a shareable, working app as its output, which is inherently more viral than a landing page or a blog post. Every user who built something worth showing off became a distribution channel.
HeadshotPro rode a wave instead of creating one. Postma launched squarely into the AI-avatar trend at its peak, aiming at a buyer with an immediate, obvious need — a professional photo, today, without a photographer. He didn’t have to educate anyone about why AI headshots might matter; the market had already decided that days before he shipped. That timing was backed by years of an existing personal following built from prior public projects.
Pieter Levels turned transparency into a marketing channel. Levels has published his revenue, traffic, and decision-making in public on X for over a decade, which built a large, trusting audience before most of his individual products existed. His directory-style sites, Nomad List and RemoteOK, are also structurally built for organic search — they rank for the exact high-intent queries people type when they’re already looking to buy, which does the work a paid acquisition team normally would.
Marc Lou built his products to sell each other. ShipFast, CodeFast, and DataFast share one audience — indie developers — and each product is positioned to introduce its users to the next one. Combined with daily public build-logs on X, this turned Lou’s customer base into his primary acquisition channel, with very little spent on outside advertising.
Chatbase let timing be the entire funnel. Elsaid launched a bare pricing page during the earliest, most intense weeks of public interest in ChatGPT and had a paying customer within 30 minutes — before he’d built any formal marketing at all. The lesson wasn’t a clever campaign; it was recognizing a 4-to-6-week window where demand was arriving faster than most companies could physically serve it, and moving inside that window instead of after it.
Miss Excel won by being entertaining in a boring category. Kat Norton posted a new Excel tutorial video every day, but set them to music and choreography rather than the flat screen-recording format that dominated the category. One video reached 3.6 million views within 24 hours and pushed her past 100,000 followers in three weeks — organic, unpaid distribution built entirely on differentiating the format, not the underlying subject matter.
Formula Bot combined grassroots seeding with a real newsjacking stunt. Bressler’s early growth came from posting directly in relevant Reddit communities, including a post that became the top post of the week on r/Excel. But his sharpest marketing move came over a year later: during the height of March Madness, he signed an NIL deal with viral college basketball player Jack Gohlke and turned it into a single X post. It generated 2.7 million views, was reshared by Barstool Sports’ Dave Portnoy, and drove a 150–200% spike in signups and traffic in the days that followed — a paid-media-sized result from what was, functionally, one well-timed DM.
TypingMind sold a better front door to demand that already existed. Dinh didn’t have to convince anyone that ChatGPT was useful — millions of people already believed that. His entire marketing motion was demonstrating, in public, that his interface was a meaningfully better way to use it, and letting his transparently shared revenue numbers double as social proof.
Justin Welsh built his own distribution instead of renting it. Welsh’s strategy has stayed consistent for years: post daily on LinkedIn and X, use each post to earn attention rather than sell directly, and route interested readers toward his own newsletter and website — channels he owns and controls, rather than platforms that can change the rules overnight. He has summarized the mindset behind that discipline on his own LinkedIn page: “Lack of direction, not lack of time, is your true enemy.”
What Actually Separates the Winners
Strip away the survivorship bias, and a few real, repeatable skills show up across almost every founder in this piece.
Ship fast, then refine in public. Nearly all of them treat the first version of a product as disposable. Levels built fly.pieter.com and pushed it live within days; Postma’s pattern with both Headlime and HeadshotPro was speed over polish; Bressler built Formula Bot’s entire MVP in a handful of weeks on paternity leave. The skill isn’t perfectionism — it’s building a process where a wrong guess costs a weekend instead of six months, and then using real user reactions, not a roadmap document, to decide what to build next.
Build the audience before the product needs one. This is the most underrated skill in the group, and the one AI can’t shortcut. Dinh, Levels, and Welsh spent years posting publicly before their biggest hits. Norton had a TikTok audience forming in real time as she built her course business. When the product launches, these founders aren’t starting distribution from zero — they’re pointing an existing audience at something new. Elsaid’s Chatbase is the exception that proves the rule: he had no pre-built audience, but he compensated by shipping directly into a moment of demand so intense it found him within 30 minutes.
Pick a problem that already exists — don’t invent one. Formula Bot (turn a sentence into an Excel formula), Chatbase (chat with your own documents), HeadshotPro (a professional photo without a photographer) — none of these required convincing anyone the problem was real. Each one took a known, well-understood pain point and wrapped a newly viable AI capability around it at almost the exact moment that capability became good enough to trust.
Act as a generalist, and use AI as the staff you don’t have. Every founder here is doing the work of a small department — support, marketing copy, pricing, infrastructure decisions — and using AI tools as leverage rather than delegating to people. That’s a genuinely learnable operating skill: knowing which decisions require a human and which can run on an automated workflow, then building the discipline to actually let the automation run instead of micromanaging it.
Timing and domain instinct. This is the part that’s closest to luck, but it isn’t pure chance. Shlomo, Postma, and Elsaid all built almost exactly when a specific AI capability crossed a usability threshold. Being embedded enough in the AI space to notice that threshold — and moving before the opportunity became obvious to everyone else — is a skill in its own right, even if it rewards people who were already paying close attention.
None of this erases the role of luck. Plenty of equally disciplined, equally fast-moving solo founders picked the wrong niche, launched a week too late, or simply didn’t go viral. But the pattern across nine very different businesses — app-building, headshots, chatbots, code boilerplates, flight simulators, Excel formulas, Excel courses, and a one-person content brand — is too consistent to write off as coincidence. AI lowered the cost of building something. It did nothing to lower the cost of being noticed, disciplined, and fast. That part is still, stubbornly, on the humans — and on whoever helps them market it.