One Person Unicorn Examples: 12 Companies Already Doing It

The one person unicorn used to be a thought experiment: a company run by one or two founders with AI doing the work of a department. Two 2026 stories — Medvi and OpenClaw — moved it from thought experiment to the headline case study everyone in this space now cites. This list leads with both, then moves to the accessible, currently-verified end of the spectrum: real companies on this leaderboard generating $500K to $3.6M in ARR with one to three employees, followed by the established, larger-scale cases that proved the pattern before it had a name. Every entry includes what's actually known: founder, start date, what was built, tools used where reported, revenue, team size, and revenue per employee where it's calculable. What separates a real example from a wish is covered in more detail in what a one person unicorn actually is.

The two defining 2026 cases

Medvi

Founder: Matthew Gallagher, with his younger brother Elliot. Start date: September 2024, from a Los Angeles apartment, with $20,000 in starting capital. What they built: a GLP-1 telehealth company — prescribing and shipping weight-loss medication through an AI-run operation. Tools used: ChatGPT, Claude, Grok, Midjourney, and Runway, across code, ad creative, copywriting, and customer service. Revenue: $401 million in 2025 (its first full calendar year), across 250,000 customers, at a 16.2% net margin — roughly $65 million in profit. 2026 trajectory: $1.8 billion. Team: 2 people on payroll, though the company relies on contractors, outsourced medical staff, and partner pharmacies to deliver care — a distinction Forrester and other critics have raised directly. RPE: roughly $200 million per employee on the 2-person headcount, though that figure should be read alongside the "technically two employees" caveat, not instead of it. The New York Times profiled Medvi on April 2, 2026; the FDA has since issued a warning related to its GLP-1 compounding practices, and the company's advertising claims have drawn scrutiny.

OpenClaw

Founder: Peter Steinberger, a solo Austrian developer. Start date: November 2025, shipped as a side project after roughly four years away from building following the 2021 sale of his previous company, PSPDFKit, for more than €100 million. What he built: OpenClaw, a self-hosted, autonomous coding agent that runs against a user's own codebase and can be directed remotely through a Telegram bot. Tools used: the project itself is the tool — built with AI coding assistants rather than around them. Revenue: not applicable — OpenClaw is open source and was never sold as a product. Growth: past 200,000 GitHub stars within roughly three months, with reports of over a million autonomous agent instances running and thousands of code commits in a single month. Team: 1, until OpenAI hired Steinberger on February 15, 2026, in an acqui-hire — not a product acquisition — on the condition that OpenClaw stay open source under an independent foundation. Both Meta and OpenAI made competing offers, reportedly including a mid-nine-figure package from Meta.

Companies on the leaderboard right now

The Medvi and OpenClaw stories are compelling, but neither is independently verifiable the way a listing on this site is. These six are — self-reported, checkable today.

HeadshotPro

Founder: solo, based in the Netherlands. What it does: turns a handful of selfies into studio-quality corporate headshots using AI. Revenue: $3.6 million ARR. Team: 1. RPE: $3.6 million — the highest confirmed figure on this leaderboard for a single-employee company. Bootstrapped, no outside funding.

Photo AI

Founder: Pieter Levels. What it does: generates AI photos of yourself outside a corporate context — travel shots, portraits, lifestyle images — without booking a studio. Revenue: $1.6 million ARR. Team: 1. RPE: $1.6 million. Bootstrapped and confirmed profitable — one of the only companies on this leaderboard with profitability explicitly verified rather than unknown.

Polsia

Founder: solo, US-based. What it does: positions itself as "the AI that builds and runs your company while you sleep" — an agent-run approach to company operations rather than a single-function tool. Revenue: $1.0 million ARR. Team: 1. RPE: $1.0 million. Bootstrapped.

PDF.ai

Founder: solo, US-based. What it does: lets users chat with their PDF documents using AI instead of reading them line by line. Revenue: $591,700 ARR. Team: 1. RPE: $591,700. Bootstrapped.

TypingMind

Founder: Tony Dinh, based in Vietnam. What it does: gives users one interface for ChatGPT, Claude, and Gemini instead of juggling separate subscriptions and tabs. Revenue: $817,300 ARR. Team: 3. RPE: $272,433. Bootstrapped.

KNOWIDEA

Founder: solo-led, Canada. What it does: a predictive intelligence platform advising executives on real-time business decisions. Revenue: $500,000 ARR. Team: 3. RPE: $166,666. Bootstrapped.

The established cases that proved the pattern first

Midjourney

Founder: David Holz, founded 2021. What it built: an AI image generation product that scaled almost entirely through word of mouth. Tools used: its own proprietary diffusion models — Midjourney is the model, not a wrapper around someone else's. Revenue: reported figures vary by source and date — some 2026 estimates put ARR in the $200 million range, others report $500-600 million. Team: grown from a founding handful to somewhere between 40 and roughly 60-plus employees by most 2026 accounts, with some estimates higher. RPE: even at the more conservative revenue figures and higher headcount estimates, still in the multi-million-dollar-per-employee range — exceptional by any traditional SaaS benchmark, though no longer in the single-founder tier. Never took outside venture funding.

Pieter Levels' portfolio

Founder: Pieter Levels (also the founder of Photo AI, above). Start: Nomad List launched in 2014; RemoteOK and a string of other solo products followed over the next decade. What he built: a portfolio of profitable, single-founder software products, run without a conventional team, years before "AI-native" existed as a category. Revenue: individually smaller than the other entries here, but combined and sustained over 10+ years — Levels has been public about running his entire portfolio solo. His public, build-in-public approach is a large part of why this model now has a documented playbook rather than a handful of unexplained successes.

Cursor

Founder: the Anysphere team, led by Michael Truell. Start: 2022, publicly launched 2023. What it built: an AI-native code editor that predicts and writes multi-file changes rather than single-line completions. Team: stayed in the single digits through its earliest, highest revenue-per-employee stretch before raising venture funding and scaling headcount to support enterprise growth. It's included here not as a current one-person example but as the clearest case of a coding-tool company that proved the "small team, AI-assisted development, outsized revenue" pattern at the tool layer itself, before growing past it.

Base44

Founder: Maor Shlomo, an Israeli developer. Start: built largely alone. What he built: an AI-powered app builder that let users describe an app in natural language and get a working version. Outcome: sold to Wix for a reported $80 million in mid-2025, without ever assembling a large team to get there. Included here as a preview of the acquisition outcomes this model can produce when a solo build catches a large enough wave at the right moment, rather than a steady-state leaderboard entry.

The pattern across all twelve

Line them up and the throughline isn't the product category — telehealth, coding agents, AI photos, document tools, and image generation have almost nothing in common as markets. It's what didn't happen. None of them hired a department the moment revenue grew past what a small team could obviously handle. The companies that did take funding and add headcount — visible elsewhere on this site in pieces like bootstrapped AI startups — consistently show lower revenue per employee than the ones that stayed lean, which is a data point worth taking seriously rather than dismissing as coincidence.

Medvi and OpenClaw get the attention because they're the most extreme and the most recent. HeadshotPro, Photo AI, and the rest of this site's leaderboard get less attention and deserve more, because they're the ones you can check today rather than take on faith from a press profile. How solo founder revenue compounds without the traditional headcount curve is worth reading in full if this list raised more questions than it answered.

Finding the next one

Twelve is not a ceiling — it's what's verified and public right now. The next one person unicorn probably looks boring from the outside: a narrow, specific, recurring problem solved well enough that people pay without asking who's behind it. If that company is being built right now, the leaderboard is where it gets recognized before a magazine profile finds it. Submit your company and let the revenue numbers make the case instead of a pitch deck.