Patronus AI, an AI infrastructure company focused on evaluation and simulation technologies for AI agents, has raised $50 million in Series B funding led by Greenfield Partners, with participation from Lightspeed Venture Partners, Notable Capital, Datadog, Samsung, Gokul Rajaram, Factorial Capital, and other AI investors. The company also introduced a preview of its first Digital World Model, designed to improve AI agent training through large-scale simulations.
Building Digital World Models
Patronus AI develops evaluation tools and simulation environments that help AI agents learn complex, long-horizon tasks. The newly unveiled Digital World Model simulates digital environments where AI agents can practice workflows involving coding, research, communication, and software interactions before deployment.
The company believes simulations will become a critical part of AI training, allowing agents to gain experience in realistic digital environments rather than relying only on static datasets.
Expanding AI Infrastructure
The Digital World Model uses language diffusion technology to predict environment behavior and guide AI agents across different tasks. The platform supports coding, dialogue, research, GUI interactions, and general tool usage, enabling more advanced autonomous workflows.
With the new funding, Patronus AI plans to expand its research organization, accelerate go-to-market efforts, and invest in the compute infrastructure required to develop and deploy Digital World Models at scale.