Mercor, a company that pays experts to help train AI models, said on Nov. 10 that it had raised $750 million in a round that values it at $20 billion. That is twice the $10 billion it was valued at in October 2025, when it raised $350 million. Investors in the new round include Benchmark, Felicis, General Catalyst, NVIDIA, Salesforce, Zoom and funds managed by affiliates of Apollo.
Mercor's argument is that AI is only as useful as the human know-how behind it. It says more than five million experts in its network are paid to train the most advanced AI models with knowledge from their own fields. When it raised money in October 2025, it described a doctor teaching AI to spot early warning signs a person might miss, a banker training it to do financial analysis and a lawyer refining how it reasons about past court cases.
Mercor says it built that work for the leading AI labs, and that it is now bringing it to businesses across the economy. It already named businesses as well as labs among those it worked with in October 2025. Mercor says more companies want AI of their own. They can start from a capable model whose trained files are published, so anyone can download it and run it on their own computers. Then they give it extra training for their own work, which Mercor says takes a fraction of the resources needed to build a leading model.
Mercor says it works with experts to build tests from the way a company actually does its work. The tests show how well AI agents handle that work, and where extra training could help. Agents are AI that can carry out tasks on its own, such as using apps and websites, instead of only answering questions. To give agents somewhere to practice, Mercor bought Deeptune earlier in 2026. Mercor says Deeptune's team spent two years recreating hundreds of business apps where AI agents can practice work tasks.
Mercor says big tech companies, makers of downloadable AI models, app makers and Fortune 500 companies have started using the tools it built for the labs. In 2026 it opened offices in New York and London, to be closer to the businesses and labs it works for. Mercor says it also leans on AI in its own work. It says it spends three times as much on running trained AI models to answer requests as it pays its staff in salaries, and that this spending pays for itself by making its staff better at their jobs.
Part of the round was an employee tender, a sale in which staff can sell some of their shares, usually to investors. Mercor has not said how much of the $750 million went to buying staff shares, so it is not clear how much new money the company itself will have to spend. It also did not say which investor led the round, or give its sales or profits.