Home / Entrepreneurship Lab / Mini Case Studies
Entrepreneurship Lab

AI-Based
Case Studies.

Casos de Estudio
Basados en IA.

Deep dives into AI-based startups rewriting the rules of business — and what young leaders can learn from them. Real founders, real numbers, real lessons.

Featured Case

The billion-dollar
(almost) solo founder.

Most Revolutionary
Medvi · Matthew Gallagher
AI-powered telehealth · Founded 2024 · Los Angeles

In September 2024, a 41-year-old who grew up in a trailer park and dropped out of university launched Medvi, an online telehealth company selling GLP-1 weight-loss treatments. He had no co-founder, no venture capital, and no team — just $20,000, a laptop, and more than a dozen AI tools. He used AI to write the software, generate the website copy, produce the ad images and videos, handle customer service, and analyze the business. Everything regulated — doctors, prescriptions, shipping — he rented from existing platforms. Within two months he was live; within a year, Medvi had crossed $401 million in revenue. His only full-time hire was his younger brother.

$20K
Initial investment
2 mo
To launch
$401M
Year-1 revenue
1–2
People on the team

What young leaders can learn

→Own the customer, rent the rest. Gallagher didn't build hospitals or hire doctors — he built the brand and experience layer on top of infrastructure that already existed.
→AI is a force multiplier, not magic. His "builder instinct" and willingness to act mattered as much as the tools. AI amplified a sharp strategy — it didn't replace it.
→Background is not destiny. A trailer park and a dropout record didn't stop him. Access to AI tools is leveling who gets to build.

AI tools he used

ChatGPTClaudeGrok ReplitMidjourneyRunway ElevenLabsCustom AI agents

Code & software with ChatGPT, Claude, Grok and Replit; ad creatives with Midjourney and Runway; customer service with ElevenLabs and custom AI agents.

Inc. The no-employee billion-dollar startup ↗ CNBC Why AI founders are getting younger ↗ Crevio How Medvi did it: the playbook ↗
More Cases

Two more worth
studying.

Built by teenagers
Cal AI
Zach Yadegari & Henry Langmack · founded at 17–18

Two high school students built a photo-based calorie-tracking app: snap a picture of your meal, and AI estimates the calories and macros at ~90% accuracy. Built entirely in the age of large image models (using Anthropic and OpenAI models), it scaled to over 15 million downloads and $30M+ in annual revenue in under two years — then was acquired by MyFitnessPal in December 2025.

15M+
Downloads
$30M+
Annual revenue
<2 yr
To acquisition
Lesson: You don't need to wait. These founders built a category-leading product from their classrooms by pairing a simple, real need with the newest AI models.
OpenAIAnthropicRAG
→ TechCrunch The acquisition story → CNBC How a teen CEO built it
Built by students
Cursor
Anysphere · four MIT students

Four MIT students left school to rebuild the code editor around AI — an "AI-first" environment that writes, edits and reviews code alongside the developer. Launched in 2023, Cursor went from zero to $1 billion in annual recurring revenue faster than any business-software company in history (Slack took ~4 years; Cursor under 2). By early 2026 it had passed $2 billion ARR, used by over half the Fortune 500.

$2B
ARR by 2026
<2 yr
0 → $1B ARR
50%+
Of Fortune 500
Lesson: Architecture is destiny. They didn't bolt AI onto an old tool — they rebuilt the whole product around it. Solving your own frustration deeply can beat giants.
Frontier AI modelsVS Code
→ Fortune The rapid rise of Cursor → TechCrunch $2B ARR & enterprise growth
A note on the numbers: these figures come from the linked press reporting (NYT, TechCrunch, CNBC, Fortune and others) and reflect what founders and companies disclosed at the time. Some revenue claims are self-reported and hard to independently verify. Treat them as inspiring directional signals, not audited financials — and always read the original sources.
The Pattern

What these cases
have in common.

01

Small teams, huge output

One to four people did what once needed hundreds. AI collapsed the cost and time of building.

02

A real problem first

None started with "let's use AI." They started with a clear human need, then used AI to serve it.

03

Speed & judgment

They shipped fast and kept human judgment at the center. AI handled the work; they made the calls.

Your turn to build.

These founders started with the same tools you can access today. Explore the AI Toolkit to build the skills, then turn your idea into something real.

Explore the AI Toolkit ← Back to Lab