AI Training Startup Mercor Seeks $10B+ Valuation Amid Soaring Revenue

Mercor, a rapidly growing startup specializing in connecting large tech companies with domain experts for AI model training, is in advanced negotiations to raise a Series C funding round. The company is targeting a valuation of over $10 billion, a substantial leap from its previous $2 billion mark achieved just months ago.

Mercor’s Meteoric Growth

Founded in 2022, Mercor has quickly established itself as a major provider of expert talent to tech giants like OpenAI, Meta, Amazon, Google, Microsoft, Tesla, and Nvidia. The company's main revenue stream comes from supplying scientists, doctors, lawyers, and other specialists to help train and improve foundational AI models. For these services, Mercor charges an hourly finder’s fee and a matching rate, earning profits by effectively brokering high-skill labor.

Mercor’s annualized run-rate revenue is now approaching $450 million, compared to just $75 million reported in early 2025. Multiple offers from venture firms have reportedly come in, some valuing the startup even higher than its $10B target. Profitability has also set Mercor apart from similar fast-scaling AI startups—unlike some competitors, Mercor posted $6 million in profit during the first half of the year.

Investor Interest and Competitive Dynamics

Felicis Ventures, a significant backer from previous rounds, is expected to double down on its investment in Mercor. About two new capital partners are also said to be raising funds through special purpose vehicles (SPVs) to participate in the upcoming round. The demand for participation illustrates both excitement about the AI training sector and skepticism about its scalability, given emerging competition.

Mercor faces intense rivalry from firms like Surge AI and Scale AI, both of which are broadening their own portfolios to include advanced reinforcement learning (RL) and AI data labeling services. Notably, Surge AI is rumored to be raising at a $25 billion valuation, signaling a gold rush in the market for human expertise used in machine learning annotation and validation.

Expansion into Software and RL Solutions

To differentiate and diversify, Mercor has informed investors of its plans to increase investment in software infrastructure for reinforcement learning. RL is a sophisticated training method where AI systems are improved iteratively based on human feedback. Mercor’s ambition is to build a comprehensive AI-powered recruiting marketplace, tapping into automation to streamline how technical talent is discovered and deployed globally.

Legal and Market Pressures

The competitive atmosphere is not without friction. A recent lawsuit filed by Scale AI accuses Mercor (and a former Scale employee who jumped ship) of misusing proprietary documents—an allegation Mercor strongly denies. Even as legal battles loom, Mercor’s three co-founders—Brendan Foody (CEO), Adarsh Hiremath (CTO), and Surya Midha (COO), all in their early twenties and former Thiel Fellows or Harvard students—have attracted seasoned executive talent, recently bringing on Uber’s former Chief Product Officer Sundeep Jain as president.

Deep Founder Analysis

Why it matters

Mercor’s story demonstrates the explosive market demand for human-in-the-loop AI model training and the huge valuations now seen in this niche. The rapid growth implies that expertly curated human feedback remains a strategic bottleneck for quality AI development, particularly as generative AI progress accelerates. For founders or operators, this signals a shift: domain knowledge and access to subject-matter expertise are now as valuable as algorithms themselves.

Risks & opportunities

One risk is overreliance on large clients—Mercor’s biggest share of revenue reportedly comes from a handful of major AI labs, which could expose it to revenue shocks if buying patterns shift or if big players bring this function in-house. On the opportunity side, building sophisticated tools for reinforcement learning outsourcing and expert marketplaces could create defensible moats, making the company a platform rather than a basic staffing agency. Legal risks—such as trade secret litigation—are another factor to monitor, reminiscent of early talent wars in the cloud era.

Startup idea or application

Inspired by Mercor, a promising startup concept is a managed “expert auditing” marketplace for generative AI, where freelance specialists rigorously test and validate AI outputs for specific industries like health, law, or finance. Such a platform could use proprietary RL pipelines, offer audit trails for compliance, and even certify AI model reliability for enterprise buyers—bridging the gap between loose crowdsourcing and expert-driven evaluation.

Looking Forward: The Next Phase for AI Talent Marketplaces

Mercor’s trajectory is a harbinger for the next wave of AI-driven staffing and training platforms, particularly as more enterprise firms seek specialized oversight for AI deployments. As the competition heats up, leaders in the category will be those who successfully combine expert communities with robust, scalable, and secure software products.

AI Startups Human-in-the-loop Reinforcement Learning Expert Marketplaces Funding

Read related insights for founders: How OpenAI’s team changes will impact the human side of AI, Amazon-backed startups and content innovation, and Europe’s unicorn startups in 2025.

Visit Deep Founder to learn how to start your own startup, validate your idea, and build it from scratch.

📚 Read more articles in our Deep Founder blog.