Sam Altman said a single person armed with AI would build a billion-dollar company. Most of us filed that under “2030 aspirational futurism.”

Matthew Gallagher didn’t get the memo. He did it in 2025.

From $20K to $1.8 Billion

Medvi is a telehealth startup selling GLP-1 weight-loss drugs online. Two employees — Gallagher and his brother Elliot. Everything else is AI. The numbers: $401 million in revenue last year, tracking toward $1.8 billion in 2026, with a reported 16.2% net margin. That’s roughly $65 million in profit for a two-person operation.

Gallagher isn’t a Stanford dropout or a YC alum. He’s a 41-year-old self-taught coder from Los Angeles who spotted the GLP-1 boom in September 2024 and built the entire platform in two months with $20,000.

Zero outside funding. No venture capital. Just AI tools and timing.

The AI Stack Running the Show

Here’s what replaces a traditional 200-person back office:

  • ChatGPT, Claude, and Grok for vibe-coding the website and backend
  • Midjourney and Runway for marketing creative and video ads
  • ElevenLabs for AI-powered customer communication
  • Custom AI agents handling order processing, performance tracking, and operations
  • AI chatbots fielding customer queries at scale

Gallagher even cloned his own avatar and voice to attend multiple customer meetings simultaneously. When your headcount is two, you improvise.

The growth curve tells the rest: 300 customers in month one. A thousand more in month two. By end of 2025, roughly 250,000 customers served.

Let’s Be Honest About What This Actually Is

Medvi is an extraordinary case study in AI-enabled efficiency. But calling it an “AI company” misses what’s actually driving revenue: GLP-1 drugs and massive consumer demand for affordable weight-loss medication.

The telehealth-to-prescription pipeline isn’t new. Companies have been running this playbook for years with hair loss, sexual health, and mental health meds. Medvi rode the same wave at the perfect moment — when GLP-1 demand was peaking and AI tools had matured enough to replace an entire workforce.

For context: Hims & Hers operates in similar territory and generated $2.35 billion with roughly 2,000 employees. Medvi is approaching similar revenue with 1,000x fewer people. The efficiency gap is staggering. The underlying business model is identical.

The Part Nobody Wants to Talk About

Venture capitalist Sheel Mohnot flagged serious concerns about Medvi’s marketing. He pointed to over 800 active ads allegedly featuring fake doctor images and AI-generated testimonials, plus 4,000 more ads from fabricated organizations. “This is highly illegal,” Mohnot wrote on X.

The regulatory picture is worse. The FDA has already sent 30 warning letters to telehealth sites marketing compounded GLP-1s. The DOJ has received enforcement referrals. Medvi’s entire revenue projection hinges on a regulatory window staying open — and there’s no guarantee it will.

This is the uncomfortable flip side of “AI democratizes entrepreneurship.” The same tools that let one person build a legitimate platform also let one person scale questionable practices to hundreds of millions of dollars before anyone catches up. AI doesn’t just compress timelines for building — it compresses timelines for everything, including the stuff regulators haven’t figured out yet.

What This Means for Everyone Else

Strip the controversy away and the signal is still loud: the traditional startup playbook — raise VC, hire aggressively, burn cash — just got an alternative ending.

For software-adjacent, service-oriented businesses that live primarily online, the Medvi model is now proven at scale. You don’t need permission, pedigree, or a hundred employees anymore.

The labor market implications are sobering. If a telehealth company can generate nearly $2 billion with two employees, what happens to the customer service reps, marketing specialists, web developers, and operations managers who would’ve staffed that company five years ago?

Gallagher himself acknowledged the limit: after scaling, he started assigning real human account managers to customers. But even those managers use AI to handle multiple customers simultaneously. The human touch isn’t gone — it’s been stretched incredibly thin.

Two Truths at Once

The Medvi story validates two uncomfortable realities simultaneously.

AI has genuinely lowered the barrier to building massive businesses. The tools work. The costs are low. The speed is real. Altman’s prediction wasn’t hype — it was early.

The businesses most likely to hit those numbers fastest are the ones operating in regulatory gray zones. Riding existing demand, using AI not just for efficiency but for scale that outpaces oversight. The most “successful” early examples of AI-native business might not be the ones worth celebrating.

Medvi is expanding into men’s health, meal delivery, women’s health, and skincare. If the regulatory environment holds, it could become a multi-vertical telehealth giant run by a skeleton crew.

But the real story isn’t Medvi. It’s the thousands of entrepreneurs watching this unfold and thinking: I could do that. Some will build incredible things. Some will build problematic things incredibly fast.

The age of the AI-native company isn’t coming. It’s here. And we’re going to need guardrails that can keep pace — especially when anyone with a laptop and a Claude subscription can scale to billions.