The One-Person Billion-Dollar Company Just Happened

How a solo founder with $20,000 and a dozen AI tools built a company on pace for $1.8 billion in revenue and what that means for how we think about startups

Author

Justin Wright
April 06, 2026 • Est. Reading Time: 9 minutes

Happy Monday!

In September 2024, Sam Altman told Reddit co-founder Alexis Ohanian that he and his tech CEO peers maintained a betting pool for the year the first one-person billion-dollar company would emerge. He called it "something that would have been unimaginable without AI and now will happen."

Eighteen months later, it happened.

Matthew Gallagher launched Medvi, a GLP-1 telehealth company, from his house in Los Angeles with $20,000 and more than a dozen AI tools. In its first full year, the company posted $401 million in sales, served 250,000 customers, and produced a 16.2% net profit margin. It's now tracking toward $1.8 billion in 2026 revenue. The total headcount: Gallagher and his brother Elliot.

For comparison, Hims and Hers reported $2.4 billion in revenue last year with 2,442 employees and a 5.5% net profit margin. Gallagher is running nearly three times that margin with a headcount roughly 1,200 times smaller.

This story has caveats worth mentioning, and we'll get to them. But the structural lesson is real regardless: AI tools have compressed the distance between idea and scaled business so dramatically that the traditional assumptions about what it takes to build a company at this scale no longer apply.

Matthew Gallagher built Medvi, a GLP-1 telehealth startup, with $20,000 and AI tools including ChatGPT, Claude, Grok, Midjourney, and Runway. First-year revenue was $401 million and the 2026 projection is $1.8 billion. Total employees: two. AI has compressed the build-test-iterate cycle so dramatically that solo founders can now compete with companies employing thousands.

TL;DR

How He Actually Built It

Gallagher didn't build a novel AI product. He built a traditional telehealth business using AI as the entire operational layer. ChatGPT, Claude, and Grok wrote the code powering his platform, generated website copy, and built AI agents that connected his software systems. Midjourney and Runway produced images and videos for advertising. AI handled customer service.

The clinical side was entirely outsourced. CareValidate, a telehealth-in-a-box platform, and OpenLoop Health handled doctors, prescriptions, pharmacies, shipping, and compliance. Gallagher's role was orchestration: using AI to build the customer-facing layer, the marketing engine, and the operational glue that connected outsourced services into a seamless product.

The model is more like a one-person holding company than a traditional startup. Gallagher didn't replace 2,000 employees with AI. He designed a business that never needed them, choosing to outsource everything AI couldn't handle and automate everything it could.

Medvi vs. Hims & Hers

Metric

Medvi

Hims & Hers

2025 Revenue

$401M

$2.4B

2026 Projected Revenue

$1.8B

~$2.8B

Employees

2

2,442

Net Profit Margin

16.2%

5.5%

Startup Capital

$20,000

$197M raised

Time to $400M Revenue

~12 months

~6 years

The Caveats Matter

The Medvi story is compelling, but it's not clean. The FDA sent Medvi a warning letter in February 2026 for misbranding violations, specifically for claims that falsely suggested the company was the compounder of the semaglutide it sold and implied FDA approval. That said, the FDA issued similar warning letters to over 30 telehealth companies in March 2026 for comparable violations. This was an industry-wide enforcement action, not a targeted investigation.

OpenLoop Health, Medvi's clinical partner, disclosed a data breach in January 2026 affecting at least 68,000 patients. Consumer reviews are mixed: 4.4 stars on Trustpilot across 11,400 reviews but recurring complaints about billing issues, surprise charges, and difficult refund processes.

These are real concerns, but they're also concerns that exist at established telehealth companies with thousands of employees and dedicated compliance teams. Medvi is not a perfect company, but the model of using AI to build and operate a business at this scale with minimal headcount is replicable. On that point, the evidence is increasingly clear.

The Broader Pattern Is What Matters

Medvi is the most dramatic example of a trend that's accelerating across the startup landscape. The build-test-iterate cycle that once took 12 to 18 months from MVP to market now takes days in some cases.

The numbers are striking. According to Designli's 2026 field report, 66.7% of founders now optimize for speed to market as their dominant strategy. Tools like Cursor, Lovable, and Bolt let non-technical founders generate full-stack applications directly in the browser. Simform reports that AI-assisted MVP development has compressed timelines from 12 months to three or four months, and for simpler products, even faster.

The result is a massive change in what founders can test, and how quickly they can do so. When building a prototype takes two days instead of two months, you can validate ten ideas in the time it used to take to validate one. The economics of experimentation have changed. Failure is cheaper, iteration is faster, and the feedback loop between idea and market signal has compressed from quarters to days.

Midjourney hit $200 million in annual recurring revenue with roughly 11 employees. Pieter Levels runs a portfolio of products generating over $3 million per year entirely solo. According to Crunchbase, 36.3% of new ventures in 2026 are solo-founded, up significantly from historical averages. The one-person company isn't an anomaly anymore. It's becoming a category.

What This Means for Practitioners

The Medvi story isn't really about Medvi. It's about the collapse of the traditional relationship between headcount and output.

If you're a founder, the playbook has changed. You no longer need to raise millions to hire a team to build and test an idea. You can prototype in days, validate with real customers, and iterate before most companies would have finished onboarding their first engineer. The risk profile of starting a company has fundamentally shifted: lower capital requirements, faster time to market, and the ability to test ideas that would have been too expensive to explore two years ago.

If you're at an established company, the competitive implications are uncomfortable. A solo founder with AI tools and outsourced operations just built a business running at three times your profit margin with a fraction of your headcount. The operational leverage AI provides to small teams is a direct threat to organizations whose competitive advantage was scale.

If you're building AI tools, the Medvi story is your market validation. The most valuable AI products right now aren't the ones with the highest benchmark scores. They're the ones that let a single person do what used to require a department. The tool builders who understand that are building the picks and shovels of this gold rush.

The Bottom Line

Sam Altman predicted the one-person billion-dollar company. Matthew Gallagher built one with $20,000 and a dozen AI tools in 18 months. The story has real caveats, from FDA warnings to data breach concerns with a partner, but the structural lesson stands: AI has compressed the distance between idea and scaled business to a degree that significantly changes what's possible for small teams.

The traditional startup model assumed you needed capital to hire people to build things. The emerging model assumes you need taste, judgment, and the ability to orchestrate AI tools and outsourced services into a coherent product. The bottleneck has shifted from resources to resourcefulness. And the founders who will lead the next wave of successful companies have plenty of the latter.

In motion,
Justin Wright

If a single founder can reach $1.8 billion in revenue with two employees and outsourced operations, what's the actual defensible advantage of a company with 2,000 people doing the same thing, and how quickly will the market force that question on every industry?

Food for Thought

If you haven’t listened to my podcast Mostly Humans: An AI and business podcast for everyone yet, new episodes drop every week!

Episodes can be found below - please like, subscribe, and comment!