Patronus AI secures $50M Series B to scale digital world models

Patronus AI, a San Francisco-based AI infrastructure startup, has closed a $50 million Series B funding round to accelerate development of its Digital World Models, large-scale simulated environments designed to train, evaluate, and stress-test AI agents across complex digital workflows. The new capital brings the company’s total funding to $70 million and will be used to expand its research organization, grow its engineering team, and invest in the compute and infrastructure required to run these simulations at scale. The round was led by Greenfield Partners, with participation from existing backers Notable Capital, Lightspeed Venture Partners, Datadog, Samsung, Factorial Capital, Gokul Rajaram, and other AI and software executives.

Patronus AI builds what it describes as digital world models, effectively replicas of websites and internal enterprise systems where autonomous agents can practice and be evaluated using reinforcement learning and other techniques before being deployed in production environments. By providing a high-fidelity sandbox for long-horizon AI agents, the platform is intended to improve reliability and safety, allowing companies and leading AI labs to identify failure modes and edge cases without risking real systems or data. The Series B funding underscores growing demand for robust evaluation and simulation infrastructure as AI agents become more deeply integrated into business workflows, positioning Patronus AI as a key provider of tooling for stress-testing frontier AI systems.