TL;DR: Mercor, founded in 2023 by three 21-year-old Thiel Fellows, has become the most capital-efficient AI startup ever measured. At $4.5 million in revenue per employee, they surpass Microsoft ($1.8M), Meta ($2.2M), and even Nvidia ($3.6M). The company hit a $10 billion valuation in October 2025 and is on track to reach $500M ARR faster than Cursor. Their secret? They pivoted from AI-powered hiring to becoming the essential human training layer for OpenAI, Anthropic, and Google DeepMind.
The Youngest Self-Made Billionaires in History
In October 2025, Forbes announced something unprecedented: three 22-year-olds had become the world’s youngest self-made billionaires. Not through crypto speculation or inheriting a tech empire, but by building a company that solves one of AI’s most critical bottlenecks.
Brendan Foody, Adarsh Hiremath, and Surya Midha—high school friends who competed together on the Bellarmine Speech and Debate Team in the Bay Area—had dropped out of college to pursue a simple idea. Just two years later, that idea became Mercor, a $10 billion company that’s fundamentally reshaping how AI models learn from human expertise.
The Numbers That Defy Traditional SaaS
Let’s start with the metric that makes Mercor extraordinary:
| Company | Revenue Per Employee | Notes |
|---|---|---|
| Mercor | $4.5M | Highest among AI startups |
| Cursor | $3.2M | Fastest to $1B ARR |
| Nvidia | $3.6M (FY 2025) | Chip monopoly |
| Meta | $2.2M (FY 2024) | 3.5 billion users |
| Microsoft | $1.8M (FY 2024) | Enterprise dominance |
| ElevenLabs | $825K | AI voice leader |
| Perplexity | $800K | AI search challenger |
Mercor’s efficiency isn’t just impressive—it’s 2.5x higher than Nvidia and more than double Meta’s. For a two-year-old startup, this defies everything we thought we knew about company scaling.
Funding Timeline
| Round | Date | Amount | Valuation | Investors |
|---|---|---|---|---|
| Seed | 2023 | Undisclosed | ~$30M | Initial investors |
| Series A | 2024 | $32M | ~$250M | General Catalyst, Benchmark |
| Series B | Feb 2025 | $100M | $2B | Felicis Ventures |
| Series C | Oct 2025 | $350M | $10B | Felicis, Benchmark, General Catalyst, Robinhood Ventures |
That’s a 5x valuation jump in just eight months—from $2 billion to $10 billion.
What Is Mercor, Actually?
Here’s where the story gets interesting. Mercor didn’t start as what it is today.
The Pivot That Changed Everything
Originally, Mercor was built to solve a straightforward problem: connecting freelance programmers in India with companies in the United States. They developed an AI platform that could interview programmers and match them with hiring companies.
It was a decent business. But then the founders saw a much bigger opportunity.
OpenAI, Anthropic, Google DeepMind, and every major AI lab shared a common bottleneck: they needed human experts to train their models. Not just any humans—they needed scientists, doctors, lawyers, engineers, and domain experts who could teach AI systems the subtleties that can’t be captured in code.
Mercor pivoted hard. They transformed from an AI-powered hiring platform into the essential human training layer for the entire AI industry.
The Business Model
Today, Mercor works like this:
- Recruitment: They identify and vet expert contractors—scientists, doctors, lawyers, bankers, journalists
- Matching: AI algorithms pair these experts with AI labs that need specific domain knowledge
- Training: Experts perform reinforcement learning tasks, verifying or disputing AI model decisions
- Infrastructure: Mercor provides the software layer for managing feedback loops at scale
The company charges an hourly finder’s fee and matching rate for each expert’s work. It’s essentially taking a cut of every hour spent training the world’s most advanced AI systems.
Current Scale
- 30,000+ expert contractors on their roster
- $1.5 million paid daily to contractors
- $85/hour average contractor compensation
- $500M ARR trajectory (faster than Cursor achieved it)
- Key clients: OpenAI, Anthropic, Google DeepMind
Why AI Labs Need Mercor
To understand Mercor’s explosive growth, you need to understand the RLHF bottleneck.
Reinforcement Learning from Human Feedback (RLHF)
Modern AI models aren’t just trained on data—they’re trained on human preferences. When ChatGPT gives a helpful response, it’s because thousands of hours of human feedback taught it what “helpful” means.
But here’s the problem: as AI models get smarter, you need smarter humans to train them.
A data labeling company can hire anyone to tag images of cats. But teaching an AI to reason like a lawyer? To diagnose like a doctor? To code like a senior engineer? That requires actual lawyers, doctors, and engineers.
The Scale AI Exodus
Mercor’s growth accelerated dramatically after a major industry shift. In mid-2025, Meta invested $14 billion in Scale AI and hired its CEO. This move prompted OpenAI and Google DeepMind to reportedly cut ties with Scale AI.
These labs needed a new source of expert training data. Mercor was perfectly positioned.
As the company wrote in a recent blog post:
“Since we founded Mercor almost three years ago, AI has advanced at an astonishing pace. But it still struggles with the subtleties that drive economically valuable work—balancing trade-offs, understanding intent, developing taste, and deciding what should be done, not just what can be done.”
Those subtleties require human experts. And Mercor has 30,000 of them.
The Founders’ Unconventional Path
Three Friends, One Vision
Brendan Foody, Adarsh Hiremath, and Surya Midha weren’t Stanford dropouts with family connections in Silicon Valley. They were high school debate partners from Bellarmine College Preparatory who shared a conviction that the traditional college path wasn’t necessary for success.
All three received Thiel Fellowships—Peter Thiel’s program that pays promising young people $100,000 to skip college and start companies. The irony isn’t lost on anyone that their company now employs PhD holders and medical doctors.
Leadership Evolution
The founders have shown impressive maturity in building their leadership team:
- Brendan Foody: CEO - handles strategy and fundraising
- Adarsh Hiremath: CTO - leads technical development
- Surya Midha: Board Chairman - focuses on governance
In May 2025, they hired Sundeep Jain, Uber’s former Chief Product Officer, as Mercor’s first President. Bringing in experienced executive talent while maintaining founder control shows the kind of strategic thinking that separates billion-dollar companies from failed startups.
All three founders made Forbes 30 Under 30 in 2025. By October, they were billionaires.
What Makes Mercor Different
1. Expert Quality Over Quantity
Unlike traditional data labeling companies that optimize for volume, Mercor optimizes for expertise. Their average contractor earns $85/hour—that’s not gig economy wages. They’re paying for genuine domain experts.
2. AI-Native Matching
Mercor uses AI to interview, evaluate, and match experts with clients. The same AI capabilities they’re helping to train are what powers their own operations. It’s a recursive business model.
3. Reinforcement Learning Infrastructure
Beyond matching, Mercor is building software infrastructure for reinforcement learning workflows. This makes them stickier with clients—they’re not just a talent marketplace, they’re a training platform.
4. First-Mover in Post-Scale AI Era
The Scale AI exodus created a vacuum. Mercor moved faster than anyone to fill it. Now they have relationships with every major AI lab.
The Revenue Per Employee Breakdown
Let’s examine why Mercor generates $4.5M per employee while other companies struggle to hit $500K:
Traditional SaaS Model
- Hire engineers to build product
- Hire salespeople to sell product
- Hire support staff to service customers
- Revenue scales with headcount
Mercor’s Model
- Small core team builds AI matching platform
- 30,000 contractors do the actual work
- Contractors aren’t employees (they’re “managed” talent)
- Revenue scales with contractor base, not core headcount
This is the key insight: Mercor’s contractors are their product, not their employees.
When OpenAI pays Mercor for 1,000 hours of expert training time, that revenue counts against Mercor’s ~50 core employees—not the hundreds of contractors who performed the work.
It’s a brilliant arbitrage of the revenue-per-employee metric.
The Broader Implications
For the AI Industry
Mercor’s success validates a crucial thesis: AI training is becoming its own industry. As models get more sophisticated, the demand for expert human feedback only increases.
For the Labor Market
Mercor is creating a new category of knowledge work. A molecular biologist can now earn $100/hour part-time teaching AI models about protein folding. A trial lawyer can monetize their expertise without leaving their practice.
For Startups
The 22-year-old billionaires story will inspire (and probably annoy) a generation of founders. But the real lesson is strategic: Mercor succeeded by pivoting quickly when they saw a better opportunity and moving faster than established competitors.
Challenges Ahead
1. Concentration Risk
Mercor’s client list is essentially the AI lab oligopoly. If OpenAI or Anthropic decides to build in-house training capabilities, Mercor’s revenue could evaporate quickly.
2. Margin Pressure
Paying contractors $1.5M daily while maintaining profitability requires careful management. As competition increases, contractor wages may rise faster than client rates.
3. Regulatory Uncertainty
How contractor classification rules evolve could impact Mercor’s business model. If those 30,000 contractors were reclassified as employees, the economics change dramatically.
4. AI Replacing Human Training
Ironically, Mercor could be training its way out of a job. If AI models become good enough to train themselves (or train each other), the need for human experts diminishes.
What’s Next
Mercor has outlined three priorities for their Series C capital:
- Expanding the talent network - recruiting more experts across more domains
- Improving matching systems - better AI for pairing experts with clients
- Building automation products - new tools to scale their operations
The company reportedly told investors they’re on track to hit $500M ARR faster than Cursor achieved it—which would make them the fastest-growing enterprise AI startup in history.
Key Takeaways for Founders
1. The Best Pivot is the Obvious One (In Hindsight)
Mercor’s pivot from “AI hiring tool” to “AI training infrastructure” seems obvious now. But recognizing that opportunity and executing the pivot in real-time required conviction and speed.
2. Marketplace Businesses Can Be Extremely Efficient
By keeping contractors off the payroll while building valuable infrastructure around them, Mercor achieves efficiency ratios that traditional SaaS can’t match.
3. Timing the Market Matters
Mercor’s growth accelerated when Scale AI’s relationships fractured. Being positioned to capture that moment wasn’t luck—it was preparation.
4. Young Founders Can Build Enterprise Companies
The conventional wisdom says enterprise sales requires gray hair and golf handicaps. Three 22-year-olds just closed deals with OpenAI and Anthropic.
The Bottom Line
Mercor represents something new in the AI ecosystem: a company that profits from making AI smarter while employing the humans who do the teaching.
At $4.5 million in revenue per employee, they’re more efficient than any tech giant. At $10 billion valuation, they’re more valuable than companies 10x their age. At 22 years old, their founders have achieved what most entrepreneurs dream about for decades.
Whether Mercor can maintain this trajectory depends on how the AI training market evolves. But for now, they’ve proven that the most valuable layer in the AI stack might not be compute or data—it might be the humans who teach machines how to think.
FAQ
What does Mercor actually do?
Mercor connects AI companies like OpenAI and Anthropic with expert contractors (scientists, doctors, lawyers, engineers) who train AI models through reinforcement learning and human feedback.
How does Mercor make money?
They charge an hourly finder’s fee and matching rate when clients hire experts through their platform. Think of it as a recruiting fee applied to ongoing work.
Why is Mercor’s revenue per employee so high?
Their 30,000+ contractors aren’t employees—they’re “managed talent.” Revenue from contractor hours counts against Mercor’s small core team of ~50 employees.
Who are Mercor’s customers?
OpenAI, Anthropic, and Google DeepMind are confirmed clients. They serve essentially every major AI lab.
How do I work for Mercor as a contractor?
Apply through mercor.com. They’re particularly interested in domain experts: scientists, doctors, lawyers, engineers, and other specialists.
Will AI training still need humans in the future?
As AI models tackle more complex tasks, they need higher-quality human feedback. Mercor bets that expert human training will remain essential even as AI improves.
Last updated: February 24, 2026