Startup profile · funding coverage
Deeptune
RL environment vendor building high-fidelity computer-use and code training gyms for frontier AI labs.
Stage
acquired
Status
acquired
Coverage
full
Why Deeptune is interesting
a16z-led $43M Series A for agent-training infra; Mercor acquired Deeptune in July 2026 to combine expert eval networks with scalable RL environment software.
Key facts
- Acquired by Mercor (Jul 9, 2026) — team joined NYC to scale RL training environments (Mercor blog, Fortune)
- Series A (Mar 2026): $43M led by a16z; Inspired Capital and Abstract Ventures joined (Fortune, company blog)
Investors in our directory
Funds linked from Deeptune's profile — open a fund page for stage focus and related deal articles.
Industries
AI & Machine Learning Developer Tools
Industry hub pages are rolling out from our fund-derived baseline taxonomy.
Related funding articles
Venture Capital Tracker pieces that cover Deeptune's financing or category context.
FAQs about Deeptune
Funding context and why this company shows up in venture coverage.
Deeptune created 'training gyms' — sandbox replicas of enterprise software where AI agents learn tasks through reinforcement learning, used by frontier labs and Mercor as a customer.
Andreessen Horowitz (/fund/andreessen-horowitz) led the $43M Series A in March 2026. Inspired Capital (/fund/inspired-capital) participated alongside Abstract Ventures and angel investors.
Mercor acquired Deeptune in July 2026. The team joined Mercor in New York to scale RL environment production for frontier AI labs and enterprise agent deployments.
As RL scaled, the bottleneck shifted from models to environments — realistic software simulators where agents practice and get scored. Deeptune built that software layer.
Yes. Mercor announced the acquisition on July 9, 2026. Financial terms were not publicly disclosed.
Last updated: 2026-07-24