Deeptune Raises $42.2M for AI 'Training Gyms'
Deeptune, a startup that builds simulation environments or "training gyms" for AI agents, has closed a $42.2 million funding round. The company's tools are designed to improve agent reliability and test AI workflows. The investment highlights growing demand in the DevOps and engineering communities for reproducible, testable AI development.
The Series A funding round was led by venture capital firm Andreessen Horowitz and placed a valuation of $221.0 million on Deeptune. This new investment brings the total funding for the New York-based company to $45.3 million, following an earlier seed round. Deeptune, founded in 2022 by Lukas Schmit and Tim Lupo, has demonstrated significant early traction in the market. The company reportedly grew from zero to over $10 million in annual recurring revenue in just six months. The core of Deeptune's offering is its "training gyms," which are simulation environments that replicate popular workplace software like Slack and Salesforce. These virtual spaces allow for the reinforcement learning and benchmarking of autonomous AI agents in realistic scenarios. The investment from Andreessen Horowitz reflects a broader trend of significant venture capital interest in AI infrastructure. The firm has been actively funding companies that provide the foundational tools and platforms for building and deploying artificial intelligence. Deeptune's team includes engineers and operators with experience from prominent tech companies such as Glean, Uber, Retool, and Hebbia. The company is headquartered in New York City and is focused on partnering with frontier AI labs to advance the capabilities of next-generation models.
