Venture Capital Videos

Mercor CEO on Why Application Layer Companies Have No Defensibility & Token Spend Exceeds Salaries

In a candid interview, Mercor CEO Brandon Foody discussed the growing challenges of building defensibility in the application layer of AI services, noting that token consumption for internal agents now outpaces traditional employee payroll. He highlighted the company’s rapid growth—adding $300 million in net new ARR within just 60 days after a security breach—and emphasized that the incident prompted Mercor to embed security as a seventh core company value. Foody explained that the model itself has become the product, making it difficult for downstream applications to create lasting moats. The firm’s token spend eclipses headcount costs, underscoring a shift in cost structures across AI firms. He also refuted rumors about losing major customers, confirming that relationships with OpenAI remain strong while Meta’s partnership is on pause. The hack was contained quickly, with external consultants engaged and communications transparent to customers. Notable moments included Foody’s admission, “We’re spending more on tokens for our internal agents than we are on employee headcount,” and his description of attackers using swarms of coding agents to exhaust codebases far faster than human teams. He warned that this new attack vector will fuel a surge in AI‑focused security engineering tools, positioning Mercur as both a target and a potential provider of defensive solutions. The interview signals that AI companies must prioritize security and rethink unit economics as token usage balloons. Investors and enterprise buyers will watch how firms like Mercur balance rapid growth with robust cyber defenses, while the emerging market for AI security solutions could become a major revenue driver in the next few years.