Mercor Raises $350 Million and Hits $10 Billion Valuation as Demand Grows for Human Experts to Train AI

The $10 Billion Gamble on Human Judgment: How Mercor Took Off in the Age of AI

A new kind of AI platform rockets to a $10 billion valuation by betting on human expertise over machine autonomy

In the sleek offices where AI pioneers dream up the next generation of machines, a quieter—and arguably more human—revolution is underway. This movement isn’t driven by fancy algorithms or brute computing power, but by people. Doctors, lawyers, and scientists are teaching machines how to think, not just compute.

Mercor, a three-year-old startup that links top experts with AI labs, has just raised $350 million in Series C funding at a $10 billion valuation. That’s five times its $2 billion valuation from earlier this year—a meteoric leap. The round was led by Felicis, with backing from Benchmark, General Catalyst, and newcomer Robinhood Ventures.

This surge isn’t just about hype. It signals a deeper shift in AI development. The bottleneck has moved from hardware and data to something far harder to replicate—human judgment. As models grow more powerful, they still need someone to teach them how to make good decisions in messy, real-world situations.

When Machines Need Mentors, Not Just Data

Mercor isn’t your typical data-labeling outfit. Instead of low-wage workers tagging images, the company relies on about 30,000 credentialed professionals—people who actually know their fields. Imagine a cardiologist fine-tuning how an AI detects heart disease or a lawyer teaching a model the nuances of legal precedent. That’s Mercor’s bread and butter.

Clients pay around $100 an hour for expert time. Mercor keeps about 30–35% of that, a healthy premium over traditional outsourcing rates. The difference is in the sophistication of the work. These experts aren’t clicking boxes; they’re shaping how machines reason.

“AI is unlocking new opportunities and boosting human capability,” CEO Brendan Foody said in the announcement. “Over the next decade, millions of people will train machines to understand judgment, nuance, and taste—the things only humans have.”

Interestingly, Mercor didn’t start this way. It began as a recruiting platform powered by AI. But as the market shifted in 2025, so did Mercor. Its pivot toward “expert-in-the-loop” training couldn’t have been better timed. The company landed squarely at the intersection of two exploding trends: massive AI models and the growing realization that quality—not just quantity—of training data makes all the difference.

Building the “Judgment Economy”

Mercor’s ambitions stretch beyond hiring experts. It’s developing the software backbone for what could become the “evaluation economy.” Think of it as the quality control layer for AI. The company builds tools and frameworks that help labs and enterprises test and refine their models before they go live.

This push toward “judgment ops”—structured systems for evaluating AI decisions—may prove to be Mercor’s strongest moat. In industries like finance, law, and healthcare, these evaluations could soon be mandatory for compliance. The rubrics, checklists, and standards Mercor creates might end up being just as valuable as the human experts themselves.

And the market? It’s massive. Banks rolling out AI-driven lending systems, hospitals deploying diagnostic assistants, or law firms using contract analyzers—all of them need to prove that their AI makes sound, explainable decisions. Regulators in the EU and beyond are already tightening the screws with acts like the EU AI Act, which could make human oversight a legal necessity.

Balancing Growth and Automation

Mercor’s biggest challenge may also be its irony. As AI gets smarter, the need for human oversight could shrink. Each new advance in self-learning and synthetic data means fewer hours of human input. The question is: can Mercor move fast enough up the value chain to stay indispensable?

Based on current data, the company’s marketplace volume could land anywhere between $216 million and $960 million per year. With its 30–35% take rate, that means net revenue somewhere between $70 million and $312 million. At a $10 billion valuation, investors are paying lofty multiples—anywhere from 30x to 140x revenue. That kind of math demands either blistering growth or a pivot toward higher-margin software income.

Legal Clouds and Fierce Competition

Not everything is rosy. Mercor is currently locked in a lawsuit with competitor Scale AI, which accuses it of stealing trade secrets. While details are scarce, such disputes often drag on, causing delays in big enterprise deals—especially in cautious sectors like finance and healthcare.

Competition is heating up too. Traditional outsourcing firms, new AI evaluation startups, and in-house teams from AI labs are all eyeing the same opportunity. Mercor’s edge lies in its credibility—its network of vetted experts and its robust software infrastructure. But staying ahead means constant reinvestment and innovation.

Big Names, Big Signals

The participation of Robinhood Ventures adds an interesting twist. By giving retail investors potential access to private startups like Mercor, Robinhood could help widen public awareness—and even boost Mercor’s recruiting power.

Meanwhile, Felicis, Benchmark, and General Catalyst doubling down sends a clear message: these heavyweight investors believe Mercor’s onto something big. Whether that’s rapid revenue growth or the creation of an entirely new layer of AI infrastructure, they’re betting that Mercor will define how humans supervise machines.

What It Means for Investors

For institutional investors, Mercor represents a high-stakes wager on human oversight becoming essential to AI deployment. Several tailwinds could drive further upside: more autonomous AI agents working in regulated industries, tighter compliance laws, and the growing need for standardized evaluation frameworks.

If Mercor successfully shifts its focus from service revenue to software subscriptions, margins could skyrocket. But make no mistake—the risks are real. Litigation could drag on, dependency on a few major AI labs could hurt stability, and international labor regulations add complexity. Worst of all, AI might simply outgrow the need for human tutors faster than Mercor can adapt.

Still, many in the market see Mercor as a strategic gem. For major software firms or cloud providers, acquiring Mercor could instantly strengthen their AI governance offerings. The company’s expert network and evaluation tools could easily slot into a larger ecosystem.

Disclaimer: This article is for informational purposes only. It’s not investment advice. Market conditions, competition, and execution can all change the game. Investors should do their homework and consult professional advisors before making any decisions.

At the end of the day, Mercor’s $10 billion valuation puts a price on something priceless—human judgment. It’s a bold bet that in the race to build smarter machines, teaching them how to think like us will matter just as much as teaching them to think fast. Whether that bet pays off will depend on how well Mercor—and the rest of the industry—can balance human wisdom with machine intelligence.