Mercor buys Deeptune to expand AI training stack

Mercor said it will acquire Deeptune, a startup that builds simulated software environments for reinforcement learning and agent training. The companies said the deal is meant to connect Mercor’s network of more than five million experts with Deeptune’s platform for creating realistic training environments across many industries and workflows.

Financial terms were not disclosed. Reporting also noted that Deeptune had raised a $43 million Series A earlier in 2026, led by Andreessen Horowitz, and that Mercor chief executive Brendan Foody had personally invested in that round before the acquisition. Deeptune’s team will join Mercor in New York as part of the transaction.

The deal fits Mercor’s broader effort to own more of the AI training stack, from the tasks and evaluation work done by domain experts to the environments where models practice real-world actions. Deeptune had focused on recreating enterprise software such as spreadsheets and Salesforce so AI systems could train in settings that resemble production use, which makes the acquisition strategically relevant for labs and companies building agentic AI systems.