Mercor acquires Deeptune: AI agent simulation platform; $2B ARR, Mag Seven as customers

Mercor, the 23-year-old founder Brendan Foody's $10 billion AI unicorn, acquired Deeptune, an Andreessen Horowitz-backed startup that builds simulation environments where AI agents practice real-world tasks before production deployment. Deeptune's platform creates digital replicas of enterprise software (Excel, Salesforce, Slack, etc.) where frontier models like Anthropic and OpenAI's agents can learn, fail safely, and iterate. Financial terms were not disclosed; Deeptune's team is relocating to New York.

Foody was listed as an angel investor in Deeptune's $43 million Series A just three months before the acquisition, a strategic pre-move he confirmed: "the angel check was written with acquisition already in mind." The combination covers the full reinforcement learning stack: Mercor operates a network of over five million domain experts who build tasks and scoring rubrics that evaluate agent performance; Deeptune builds the applications those tasks run inside. Mercor claims all six of the Mag Seven except Tesla are customers for its reinforcement learning infrastructure.

Mercor hit $2 billion in annual recurring revenue in June 2026, up from $1 billion a year prior—a 24-month arc from $1M to $2B ARR described as "the fastest growth trajectory ever." The acquisition happens on the heels of a March 2026 data breach where hackers exploited LiteLLM (an open-source library) and exfiltrated four terabytes of Mercor's data including contractor biometrics and internal records on how Anthropic and Meta build RL pipelines. Despite the breach, Foody claims all frontier labs expanded relationships since March, signaling either remarkable customer loyalty or frontier labs having no alternative at scale.