Mercor raises $750 million at a $20 billion valuation

Mercor raises $750 million at a $20 billion valuation

10 November 2026: Mercor, the San Francisco company that pays experts to train and evaluate AI models, has announced a $750 million Series D that values it at $20 billion, and says it will spend the money on infrastructure and research for evals and post-training. The round included an employee tender, so part of the total went to staff selling shares rather than to the company.

Not yet independently verified. Mercor names no lead investor, and says the round included an employee tender without stating how much of the $750 million it was. We have not confirmed the figures beyond the company's own post. We will update this when it can be confirmed, and remove this note.

Who put money in

In a post by co-founder and chief executive Brendan Foody, the company says the round came "along with participation from numerous investors," and names Benchmark, Felicis, Fundomo, funds managed by affiliates of Apollo, G Squared, General Catalyst, Glade Brook Capital, NVIDIA, Salesforce, Sapphire Ventures, StepStone Group, XTX Ventures and Zoom. It does not name a lead investor and does not say what share of the total was the employee tender.

What the money is for

Mercor says it will put the funds into evals, post-training and building more environments where agents can learn complex workflows. It acquired Deeptune earlier this year, whose team spent two years recreating hundreds of enterprise applications where AI agents can practise professional workflows.

The company says it has built the infrastructure to train and evaluate frontier models on real-world work, and is now bringing the same capabilities to enterprises. Its pitch is that a company can start from a capable open model, post-train it with a fraction of the resources needed at the frontier, and measure it against evals built from its own workflows.

The expert network

Mercor says more than five million experts in its network are paid to train frontier AI models on the strength of their own knowledge. It points to more than 800 distinct occupations in the United States, each with workflows that models have not been evaluated against. For enterprises it builds custom evals with experts in their fields, based on the company's actual workflows, to show how models perform against its requirements.

The post says the capabilities built for frontier labs are now adopted by big tech, developers of open models, application-layer companies and the Fortune 500.

Offices and a company statistic

The company says its team has grown beyond San Francisco, with new offices in New York and London opened this year to be closer to enterprises and labs and to recruit from strong talent pools. It also states that it already spends three times as much on inference as on employee salaries, and says that this pays for itself by making its team better at their jobs. Foody argues that if the ratio holds as AI spreads, it points to double-digit GDP growth over the next five years.

The post gives no founding year, revenue or customer numbers.

Company profile on TrustList: Mercor

Sources

• Mercor blog, Brendan Foody: Organizing human intelligence for the entire economy, 10 November 2026, 10 November 2026