Mercor

AI Infrastructure · Data/Evals · Enterprise Software

Funding

Mercor raises $750M Series D at $20B valuation for AI evals

The AI training startup is pivoting its frontier-lab playbook of paying experts to grade models into a product every enterprise can buy.

• Mercor raised a $750 million Series D at a $20 billion valuation, with participation from Benchmark, Felicis, Fundomo, NVIDIA, Salesforce, Zoom, General Catalyst and funds managed by Apollo affiliates, among others.
• Mercor started as an AI hiring platform and now runs a network of more than five million experts who get paid to train and evaluate frontier AI models using their professional knowledge.
• The round includes an employee tender offer, though Mercor has not disclosed how much of the $750 million goes to buying out staff shares versus new company capital.
• Mercor builds custom evaluations from an enterprise's real workflows, then uses the results to post-train models starting from an open model rather than a frontier one.
• The $20 billion valuation is double the roughly $10 billion mark Mercor hit after its $350 million Series C just over a year earlier, in October 2025.
• Mercor's rise followed OpenAI and Google DeepMind reportedly shifting business away from rival Scale AI after Meta's roughly $14 billion investment in that company.
• The deal signals that the market for AI data and evaluation work is shifting from a few frontier labs to the entire enterprise economy, as companies look to own and customize their own models rather than rely solely on off-the-shelf ones.