20VC: Mercor: From $1M to $500M in 17 Months: The Fastest Growing Company in the World | How to Think About Margins and Revenue Sustainability in AI | Why Evaluation Benchmarks in AI are BS Today with Brendan Foody

What It Covers

Brendan Foody, 22-year-old CEO of Mercor, details scaling from $1M to $500M revenue in 17 months by providing high-caliber human experts to train AI models, becoming the fastest-growing company in history.

Key Questions Answered

• •Revenue concentration strategy: Mercor's largest customer concentration mirrors NVIDIA's model, demonstrating that building for the best customers creates trillion-dollar businesses despite concentration risk concerns from traditional investors. The company quadrupled revenue after Scale AI's acquisition, validating this approach.
• •Data quality economics: Mercor pays marketplace workers $95 per hour versus competitors' $30 per hour, targeting Goldman bankers and McKinsey analysts instead of crowdsourced labor. This power law approach shows the top 10-20% of contributors drive majority of model improvement, creating defensible competitive advantages through quality.
• •Evaluation framework shift: Current AI benchmarks like Olympiad math and PhD-level reasoning disconnect from real enterprise needs. The market transitions toward aural environment evaluations that measure practical capabilities like drafting emails, scheduling meetings, and using multiple tools across 10-100 hour workflows rather than academic tests.
• •Synthetic data limitations: Total addressable market remains bounded by tasks humans perform better than models. Each time models improve at one complexity level, new tool combinations and longer task trajectories create fresh opportunities for human-generated training data, preventing synthetic data from fully replacing human contribution.
• •Capital efficiency paradox: Despite 50%+ month-over-month growth and ability to double overnight with capacity, Mercor remains profitable without trying. Foody questions whether burning hundreds of millions on supply/demand subsidies would accelerate market dominance, balancing aggression against building sustainable fundamentals for decade-long durability.