AI labs are increasingly turning to Mercor to access specialized industry knowledge that companies are unwilling to share directly. Instead of negotiating costly data-sharing agreements, these labs are tapping into a network of former senior employees from top firms in finance, law, and consulting—individuals who possess deep, firsthand expertise in complex workflows. Brendan Foody, the 22-year-old co-founder and CEO of Mercor, highlighted this shift during a panel at TechCrunch Disrupt 2025, describing the startup’s platform as a key bridge between former industry insiders and AI developers. Mercor’s model relies on paying highly skilled professionals—some earning up to $200 an hour—to complete tasks like filling out forms, writing reports, and describing internal processes for AI training. The company now employs tens of thousands of contractors and disburses over $1.5 million daily. Despite the high costs, Mercor remains profitable, as AI labs are willing to pay premium prices for this type of high-quality, domain-specific data. In just under three years since its founding, Mercor has grown its annualized recurring revenue to around $500 million and recently secured funding at a $10 billion valuation. Its clients include major AI players like OpenAI, Anthropic, and Meta. Foody noted that companies like Goldman Sachs are often reluctant to allow AI models to be trained on their data, as it could automate core parts of their business and disrupt their competitive edge. This resistance creates a gap that Mercor fills by connecting AI labs with people who once worked inside those organizations. While Mercor claims to prevent corporate espionage by instructing contractors not to upload documents from their former employers, the risk remains. Some contractors still work full-time jobs and submit data on the side, raising concerns about the potential leakage of sensitive information. Foody acknowledges that “there are things that happen” at scale, but maintains that the knowledge in a person’s mind belongs to them, not their company—a stance that contrasts with many enterprises’ views on intellectual property. Mercor’s approach has drawn attention from both supporters and skeptics. Some companies are embracing the shift, seeing it as part of a new era of work. Uber’s former chief product officer, Sundeep Jain, joined Mercor as president, signaling growing industry interest. Foody envisions a future where such platforms create a new kind of gig economy, similar to how Uber transformed transportation. The startup has also expanded into areas that blur the line between knowledge and data. For example, Mercor is currently seeking CTOs or co-founders of startups who can authorize access to production codebases for AI evaluation or training. While Mercor says some have accepted, it has not disclosed contract details. Mercor’s rise has coincided with challenges at Scale AI, which lost key partnerships after Meta’s major investment and the hiring of its CEO. This created an opening for Mercor, which has since quintupled in value. Still, it remains smaller than competitors like Surge and Scale AI, both valued at over $20 billion. Looking ahead, Foody believes AI will eventually surpass top-tier firms in consulting, finance, and law. He sees this transformation as a positive force that could drive economic abundance. “Over time, ChatGPT will be better than the best consulting firm, better than the best investment bank, and better than the best law firm,” he said. “That’s going to transform the economy radically.”