Patronus AI secured a $50 million Series B round led by Greenfield Partners to expand its platform for testing AI agents in simulated digital environments. The San Francisco startup, founded in 2023 by former Meta researchers Anand Kannappan and Rebecca Qian, builds replicas of websites and internal systems where agents face stress-testing through reinforcement learning.

The company's approach mirrors how Waymo tested autonomous vehicles in synthetic worlds before real-world deployment. AI labs and startups are signing up rapidly, with Notable Capital's Glenn Solomon calling demand "nearly insatiable." Revenue has surged 15-fold over the past year, attracting investors including Lightspeed, Datadog, and Samsung.

Patronus focuses on software engineering and finance applications where outcomes can be verified, but plans to tackle harder-to-validate domains. Their simulations catch shortcut-taking behaviors that traditional benchmarks miss. The startup primarily competes against internal evaluation teams at major AI labs rather than other external vendors.

With $70 million total funding, Patronus aims to create environments where agents can operate for extended periods - 10 hours to 10 weeks - before deployment. This addresses a critical gap in AI safety testing as agents become capable of handling complex, multi-step tasks.