Are We Already at the One-Person Billion-Dollar Company?

Matthew Gallagher started Medvi in September 2024 with $20,000 and more than a dozen AI tools. From his house in Los Angeles, the 41-year-old used AI to write code, produce website copy, generate ad images and videos, run customer service, and analyze business performance. He outsourced the rest.

In its first month, Medvi, a telehealth provider of compounded GLP-1 weight-loss drugs (patients fill out an online questionnaire, a licensed clinician reviews it remotely, and if approved, a pharmacy ships the medication to their door), signed 300 customers. In its second month, 1,000 more. In 2025, its first full year, the company generated $401 million in sales with a 16.2% net margin. Gallagher then hired his only employee: his younger brother Elliot. This year, they are on track to do $1.8 billion.

The New York Times wrote it up on April 2. Tyler Cowen declared Sam Altman's prediction fulfilled. The internet reacted the way the internet does: with awe, skepticism, and memes, roughly in that order.

The story spread fast because it confirmed something a lot of people wanted to believe. AI makes teams optional. One person with the right tools can build something enormous. The future of the firm is a founder, a laptop, and a stack of subscriptions.

But Gallagher's own description of Medvi tells a more complicated story.

The prediction everyone's testing

In October 2023, Sam Altman told Alexis Ohanian at the Robin Hood Investors Conference that his tech CEO group chat had a "betting pool for the first one-person billion-dollar company." In May 2025, Dario Amodei said at Anthropic's Code with Claude event that he gave it 70 to 80% odds of happening by 2026, most likely in areas with less "human-institution-centric" work, like proprietary trading. Medvi is the first company that looks like it might have crossed the line.

What the founder actually claims

Here is where the story gets interesting. On his own LinkedIn post celebrating the NYT article, Gallagher wrote that Medvi is "not an AI company." It was built, he said, with "AI-first fundamentals." He thanked CareValidate and OpenLoop by name.

Even Gallagher frames this as a lean operation powered by AI tools, not an AI product. The real question is simpler: can two people, using AI to compress the work that used to require a full team, coordinate a billion-dollar operation? And if most of that operation is outsourced to hundreds of people at partner companies, does it still count?

What Medvi actually is

Think of it like Booking.com. You search for a hotel, pick a room, and pay through Booking's interface. But Booking doesn't own the hotel, employ the staff, or wash the sheets. It owns the demand layer: the search, the storefront, the checkout, the customer communication. The hotel does the actual hospitality.

Medvi works the same way, but for weight-loss medication. Medvi's own terms and conditions describe it as a "patient management platform." It does not directly provide medical or pharmacy services. It does not produce compounded medications. Completing an assessment on the site does not create a doctor-patient relationship with Medvi. The clinicians who review your questionnaire and decide whether to prescribe are not Medvi employees. They work through partner networks and retain full prescribing discretion.

So what does Medvi actually do? It runs the ads that bring you to the site. It owns the website where you fill out the questionnaire. It operates the chatbot that answers your questions (with human contractor backup and a nursing line). It handles billing and subscriptions. The pricing starts at $179 per month for semaglutide injections, cash-pay only.

Everything else, the doctors, the prescriptions, the drugs, the shipping, sits with partners.

Here is how the operation flows:

1. You see a Medvi ad and land on the website
2. Medvi (2 employees) collects your information, runs the storefront, handles billing
3. CareValidate (11-50 employees) routes your case through its platform. Their LinkedIn says it plainly: "Medvi runs on CareValidate." They handle clinical workflows, provider access, pharmacy coordination, and compliance. They market a GLP-1 profit calculator on their website.
4. OpenLoop Health (201-500 employees) provides the licensed clinicians who review your questionnaire and decide whether to prescribe. They also handle legal, regulatory, and patient support.
5. Partner pharmacies compound the medication and ship it to your door. Medvi never touches the drugs.

Three companies, hundreds of employees across them, one brand on the bottle.

Where AI fits (and where it doesn't)

What AI does at Medvi: software development (Replit, confirmed by its CEO Amjad Masad after Gallagher responded to his post), website copy, ad images, ad videos, a customer service chatbot, and business analytics.

What humans and vendors do: all clinical consultations. All prescribing. All drug compounding. All dispensing and shipping. All compliance and regulatory work. Some customer service (contractors and a nursing line).

AI dominates the front office and customer acquisition. Humans dominate the entire regulated medical and pharmacy chain. If you removed every AI tool, the business would probably still exist. You would just need a web developer, a graphic designer, a copywriter, media buyers, an analytics team, and a support staff. AI gave Gallagher speed, low fixed headcount, and faster iteration. It did not give him the underlying medical capability.

Gallagher is not new to this playbook. Before Medvi, he founded Watch Gang, a subscription watch-box service that scaled to millions in revenue. But Watch Gang hired aggressively and stayed unprofitable. The lesson Gallagher seems to have drawn: keep the team small, outsource the operations, own only the demand.

The structural analogy is dropshipping. Gallagher built the storefront with AI. He runs the ads with AI. He handles some customer service with AI. A third party handles everything else: the doctors, the prescriptions, the drug compounding, the shipping. He never touches the product.

What the headline leaves out

The revenue is real, and it is large. The New York Times reported that Medvi generated $401 million in sales in 2025 with a 16.2% net profit margin, roughly $65 million in profit. The company claims 500,000 patients and says it is tracking toward $1.8 billion in 2026. Those figures were journalist-verified (the Times reviewed financials and interviewed business partners), though no public audit or SEC filing exists. The $1.8 billion number is a management projection, not a confirmed result.

But the story is incomplete without context.

Customers are frustrated. Medvi carries a BBB F rating with 368 complaints. The BBB says Medvi failed to respond to 5 of them. Common themes: double billing, being charged after cancellation, medications never arriving, and refund requests denied or ignored. Medvi's own terms say consult fees are generally nonrefundable and cancellation requires 72 hours before the next billing cycle. For comparison, Hims has an A+ rating and is BBB-accredited, with similar complaint categories but far better resolution rates.

Patient data may have been exposed. In January 2026, OpenLoop Health disclosed a data breach after discovering unauthorized access to its systems on January 7 and 8. OpenLoop said the breach did not involve medical records, Social Security numbers, or financial account data, but did involve other personal and medical information. Reporting tied the event to 1.6 million affected individuals. A class-action lawsuit (Allen v. Openloop Health, Inc.) was filed in February. Medvi customers are plausibly affected because Medvi routes care through OpenLoop's network. OpenLoop's own breach notice says it powers telemedicine platforms "made available through other companies." This is the exact kind of risk that a lean, outsourced model inherits: Medvi doesn't control the infrastructure, but its customers are the ones exposed.

The FDA flagged Medvi's marketing claims. In February 2026, the FDA issued Warning Letter #721455 after reviewing Medvi's site. The agency objected to claims like "same active ingredient as Wegovy and Ozempic" and to branding that made Medvi appear to be the compounder when it was not. No public resolution has been posted. Medvi sits inside a broader regulatory crackdown: the FDA sent 30 telehealth warning letters in March 2026 alone.

Before the NYT story, there were already questions about Medvi's advertising. In May 2025, Futurism reported that Medvi's website featured before-and-after weight-loss photos that appeared to be fabricated. The investigation traced the original photos to unrelated sources on the internet (including a 2018 Daily Mail article and a Reddit user's post), with the faces swapped using AI to create fake patient testimonials. The site also displayed AI-generated Ozempic box images with garbled text and listed real physicians who said they did not know they were on the site. The issue is not that Medvi used AI to make images. The issue is that the images were designed to look like real patient results when they were not. Medvi's site now carries disclaimers that some content may be AI-generated or AI-enhanced.

These flags matter because they show what can happen when a two-person front end scales to half a million customers faster than the trust, compliance, and support infrastructure behind it can keep up.

The definition game

Here is the part most coverage skips. $1.8 billion in revenue is not the same as a $1.8 billion company. Medvi is private. It has raised zero dollars in venture capital. It has no known valuation. Altman's prediction, as Evan Armstrong noted in Every, was almost certainly about valuation, not revenue.

That said, at 16.2% net margins on $1.8 billion, projected 2026 profit would sit around $291 million. A valuation above $1 billion is plausible under several frameworks. Hims, for reference, trades at roughly 5.5x revenue with $2.35 billion in 2025 sales and 2,442 employees.

Then there is the headcount question. "Two employees" excludes the contractors, the outsourced doctors, the pharmacists, the CareValidate team, the 201 to 500 people at OpenLoop, and the shipping logistics chain. The prediction was about one person doing a billion dollars. Not one person coordinating hundreds of other people doing a billion dollars.

The whole debate rests on a few definitional moves people make without noticing. Revenue or valuation? Payroll headcount or total embodied labor? AI-native product delivery or AI-enabled orchestration? One founder or one employee? Depending on which definitions you pick, Medvi either fulfills the prediction or misses it entirely.

The compounded GLP-1 market adds another layer of fragility. The FDA sent more than 50 warning letters to GLP-1 compounders in September 2025 and 30 more to telehealth companies in March 2026. Branded alternatives are getting cheaper: Novo Nordisk now offers some Wegovy telehealth subscriptions at $249 per month, and Lilly launched oral Foundayo at $149. The pricing gap that made compounded telehealth compelling is narrowing. Whether Medvi can sustain $1.8 billion into a tighter regulatory environment is an open question.

A longer trend than the headline suggests

The idea of a company that owns almost nothing and coordinates everything has been around longer than the AI hype cycle. Gerald Davis wrote about "web page enterprises" in 2016: firms that rent all inputs and keep only the coordination layer. Auren Hoffman argued in 2020 that fewer employees was becoming a status game. What AI changed is the cost of the coordination layer itself. Marketing, code, customer service, analytics: all of it got cheap enough for one person to run. Medvi is the most extreme version yet, but it is the latest chapter in a longer trend.

The real question

The answer to "are we at the one-person billion-dollar company?" may be: wrong question.

What AI enabled at Medvi is not a company where one person replaced the workforce. It is a company where one person orchestrates the workforce from a laptop. The doctors, pharmacists, compliance teams, and support staff are all still there. They moved off the cap table and off the payroll.

The real one-person billion-dollar company, where AI delivers the core value and not just the marketing wrapper, has not arrived yet. But the fact that this conversation is happening at all means something shifted. The question going forward is not whether it's here. It's what we're willing to count.