Mercor, an AI training and evaluation startup that pays domain experts to help build and test models, raises $750 million in a Series D round at a $20 billion valuation. The company says it uses a large network of more than five million domain experts to provide human expertise used for training and evaluation, and it aims to take those capabilities to enterprise customers, including for AI agent evaluation.
Outlets frame the funding as a scaling moment for Mercor’s “human-in-the-loop” approach, with emphasis on how expert labor supports model performance. The Next Web highlights the round’s link to training and testing AI across a broad range of tasks and notes that the valuation is higher than in an earlier round. SiliconANGLE focuses on the enterprise push for AI agent evaluation. Channel NewsAsia reports the headline valuation and funding as the key development, indicating continued investor interest in AI training infrastructure.
Across coverage, the core facts—amount raised, valuation, and Mercor’s expert-driven evaluation model—remain consistent, with differences mainly in wording and the specific market or use-case angle each outlet foregrounds.
