Patronus AI raises $50M to jailbreak your agents before hackers do
patronus ai raised $50m to build digital worlds that stress-test ai agents. the startup simulates environments to catch agent failures before deployment, serving frontier labs.
patronus ai secured $50m in series b funding to build simulated digital worlds for stress-testing ai agents, bringing total funding to $70m. founded by ex-meta researchers, the startup creates replicas of websites and internal systems to evaluate agent performance over long durations without human oversight. the technology addresses the gap between high benchmark scores and real-world reliability by catching agents that take shortcuts or fail complex tasks. with revenue growing 15-fold, patronus currently serves finance and software industries, using reinforcement learning to ensure agents perform correctly over 10-hour or 10-week cycles.