Patronus AI raises $50M to build digital worlds for stress-testing AI agents

Patronus AI, founded in 2023 by former Meta AI researchers Anand Kannappan and Rebecca Qian, has raised $50 million in Series B funding. The round was led by Greenfield Partners, with participation from Notable Capital, Lightspeed, Datadog, and Samsung. The company's total funding now stands at $70 million.

The company uses what it calls "digital world models" to create replicas of websites and internal systems. In these environments, agents are stress-tested after training using reinforcement learning, which iteratively rewards successful task completion and penalizes errors. Patronus AI's approach is similar to how Waymo trained autonomous cars by first building synthetic worlds to test vehicles against rare hazards.

AI labs see great value in these simulations because they allow agents to try different, sometimes unpredictable scenarios. Unlike human-data firms like Mercor and Surge, Patronus operates by evaluating how agents behave without any human involvement.

According to co-founder Kannappan, the company is currently focused on verifiable problems in software engineering and finance, but plans to expand into harder-to-verify areas in the future. Patronus's main competitor is the internal teams that AI labs have already built to evaluate agent behavior.