Brendan Foody, CEO and cofounder of Mercor, has said an IPO is “potentially on the horizon,” his first public acknowledgment that the company could move beyond private fundraising. The remark comes as Mercor announced a $350 million Series C round that values the company at $10 billion. The financing was led by Felicis, with participation from Benchmark, General Catalyst, and Robinhood Ventures, and follows a $100 million Series B in February, also led by Felicis.
Mercor’s Early Pivot
Mercor was founded by three Thiel Fellows and began as a hiring startup, designed to assess candidates by analyzing résumés, interviews, and portfolios. In building that system, the company inadvertently amassed a network of specialized experts. That network became the foundation for recruiting highly skilled professionals to train AI systems.
This aligned with a growing need from AI labs for structured evaluation sets, reinforcement learning environments, and reliable human oversight. Today, Mercor manages more than 30,000 contractors who collectively earn over $1.5 million each day. Contractors are drawn from fields such as software engineering, finance, law, and medicine, with rates as high as $100 an hour. Work spans from lawyers tutoring legal AI systems to meme specialists helping xAI’s Grok capture internet culture.
Foody has described this as a new labor category. “Millions of people will spend the next decade teaching machines the judgment, nuance, and taste that only humans possess,” he wrote in a blog post. “Instead of doing predictable work repeatedly, they’ll teach agents how to do it once, so the agent can do it a million times.”
The Business Model
Mercor distinguishes itself by treating human expertise as infrastructure. “One of the most important things in building these environments to train agents is reliability,” Foody said on the TBPN show. “If there’s noise or errors, it’s very difficult for a model to learn effectively. You can’t just ask users to spend hours correcting mistakes. Instead, you need dedicated experts who can diligently analyze model performance, go through strict review processes, and build robust reinforcement learning environments.”
The company has grown an average of 51% month over month in revenue for the last six months, with a run rate above $75 million and continued profitability. Its business is organized around three priorities Foody has compared to Jeff Bezos’s Amazon playbook. Where Amazon emphasized more products, lower prices, and faster delivery, Mercor focuses on more candidates, better matching, and faster delivery of results.
Rivals in the Market
Mercor’s momentum has coincided with disruption among competitors. Scale AI, previously the market leader, sold a 49% stake to Meta for $14.3 billion earlier this year. The deal raised concerns about neutrality, and both Google and OpenAI reportedly cut ties. “It just doesn’t happen too often in startups where your biggest competitor gets torpedoed overnight,” Mercor cofounder Adarsh Hiremath told Forbes.
Other players remain active. Surge AI is reportedly targeting up to $1 billion in new funding, Turing AI reached a $2.2 billion valuation in March, and Invisible Technologies raised $100 million in September to pass a $2 billion valuation. Mercor, meanwhile, has remained profitable while scaling its contractor network and customer base.
Foody has acknowledged questions about defensibility, particularly against platforms like LinkedIn or Microsoft. His answer points to two advantages. The first is scale: a larger pool of experts creates stronger marketplace effects. The second is what he calls the “data flywheel,” where every customer interaction produces performance data that improves future matches across the platform.
He likened Mercor’s near-term market dynamics to Nvidia, which works with a handful of hyperscalers. In the long term, Foody believes every enterprise will require agent training to customize workflows, and Mercor will supply that expertise.
Asked about disintermediation risk if a dominant AI platform might bypass intermediaries like Mercor he pointed to the complexity of reliable training. “You need to ensure they are consistently accurate, and that takes real work,” he said.
Demand Across Industries
Mercor’s strongest categories are software engineering, finance, law, and medicine. Foody noted heavy investment in law firms in particular. He recounted a conversation with a lawyer who had been skeptical of AI tools a year ago but had since changed his view. “Harvey isn’t perfect, but it has better attention to detail and is more thoughtful than almost any junior at our firm. I’ve watched it do $100K of associate-level work in 10 minutes,” the lawyer told him.
Foody said the limiting factor in advancing such applications is not model capability but the availability of training data and evaluation sets. “That’s exactly what we’re helping build,” he added.
Foody frequently situates Mercor in the context of the broader knowledge economy. He estimates that businesses spend roughly $40 trillion annually on knowledge work, much of it repetitive. He argues that training agents to automate these workflows is structurally more efficient, and that Mercor is “only 1% of the way there.”
Contractors have trained agents to respond to customer service tickets, analyze data rooms in banking, or tutor chatbots in language and culture. “Instead of a customer support representative responding to hundreds of tickets, they can create an eval to train an agent once to handle those,” Foody explained.
Mercor’s comment about a potential IPO comes at a time when few AI infrastructure startups have publicly raised the possibility. Foody has not committed to a date but placed the idea on the record. As Foody put it: “Right now, businesses spend about $40 trillion a year on knowledge work and people doing largely monotonous tasks. All of that will transform into training agents to automate those workflows. Mercor is building the infrastructure to help humans teach models and fit into this AI economy. It’s an opportunity worth tens of trillions of dollars a year over the next decade and we’re only 1% of the way there, if that.”
