The Pulse

Mercor raises $750 million to expand enterprise AI evaluations

Mercor raised a $750 million Series D at a $20 billion valuation, with the round including an employee tender. The company plans to invest in AI evaluations, post-training and environments where agents can practice professional workflows.

AI.info Team · · fact-checked · against 3 sources ↓ · Published 11 Nov 2026

Mercor has raised $750 million in a Series D round that values the company at $20 billion, as it expands its work evaluating AI agents for enterprises. The November 10 announcement said the round included an employee tender.

The round included Benchmark, Felicis, Fundomo and funds managed by affiliates of Apollo. Other participants included G Squared, General Catalyst, Glade Brook Capital, NVIDIA, Salesforce, Sapphire Ventures, StepStone Group, XTX Ventures and Zoom.

Funding the enterprise push

Mercor plans to increase investment in research and infrastructure for evaluations and post-training, as well as build more environments where agents can learn complex workflows. The company is extending capabilities it developed for frontier AI labs to businesses seeking to measure what agents can do, what they cost and how their performance could improve.

Its previous Series C was announced in October 2025 and raised $350 million in a round led by Felicis. It valued the company at $10 billion and also included Benchmark, General Catalyst and Robinhood Ventures.

Tests built around work

Co-founder and CEO Brendan Foody said Mercor spent the past two years building infrastructure that uses human expertise to train and evaluate frontier models on real-world work. The company is bringing those capabilities to enterprises across the economy, while also supplying data and evaluations for companies that want to adapt models to their own needs.

Mercor works with companies and specialists to create custom evaluations based on actual business workflows. Those tests measure models against a company's requirements and help identify where further training or practice in realistic environments could improve results.

Foody said professional work requires knowledge of both how a task is performed and how a particular company carries it out under its own systems, policies and quality standards. He also said the capabilities built for frontier labs are being adopted by large technology companies, open-model developers, application companies and the Fortune 500.

Experts and practice environments

Foody said Mercor has more than five million experts in its network who are paid to train frontier AI models using their specialized knowledge. Its APEX benchmark measures model performance on economically valuable tasks across thousands of real-world workflows, according to the company.

Mercor acquired Deeptune earlier this year to develop environments where agents can practice professional work. The Deeptune team recreated hundreds of enterprise applications over two years, including software ranging from spreadsheets to Salesforce, the company said.

In July, Mercor described each training environment as consisting of software, tasks for an agent to complete and verifiers that determine whether it succeeded. Mercor said it was already a Deeptune customer before the acquisition and that its own experts create tasks and verifiers for such systems.

Offices and spending

Mercor has opened offices in New York and London this year, adding to its presence in San Francisco. The company said the expansion puts it closer to the enterprises and labs it serves and helps it recruit.

Foody also said Mercor spends three times as much on inference as it does on employee salaries, a ratio he said pays for itself by helping employees work better. He argued that if that ratio holds as AI spreads through the economy, it points to something closer to double-digit GDP growth over the next five years, nearly three times the rate he said is usual.