Patronus AI announced that it raised a $50 million Series B round and unveiled Digital World Models, a new class of large-scale simulation environments for AI agent training and evaluation.
The round was led by Greenfield Partners and included participation from existing investors Notable Capital, Lightspeed Venture Partners, Datadog, Samsung, Factorial Capital, Gokul Rajaram, and other AI and software executives.
The new funding brings Patronus AI’s total capital raised to $70 million. The company plans to use the funding to expand its research organization, grow its engineering team, and invest in the compute and infrastructure needed to train and run Digital World Models at scale.
Patronus AI is focused on AI evaluation, simulation infrastructure, and reliability testing for advanced AI systems. The company said it works with the majority of the world’s leading frontier AI labs and hyperscalers, and its revenue has grown more than 15x over the past year.
The company was founded by AI researchers and engineers with backgrounds at organizations including Meta AI, Amazon AGI, and Google. The team’s experience spans large language model evaluation, AI alignment, fairness, and embodied agents.
Patronus AI said the next phase of AI development requires moving beyond static benchmarks. As AI agents take on longer and more complex workflows, they need dynamic environments that resemble the digital systems they will operate in.
Digital World Models are designed to support training and evaluation across complex digital workflows. Patronus AI describes them as language diffusion world models that can scale the creation of simulation data for AI agent actions.
The company’s simulation infrastructure is intended to help AI systems train on realistic software, research, communication, and enterprise workflows. The goal is to create agents that can operate reliably across ambiguous, long-horizon tasks instead of simply performing well on narrow benchmark tests.
Patronus AI said its research focuses on generating environments where AI agents can encounter edge cases, recover from failures, and improve through repeated interaction. This includes simulation tooling, evaluation systems, and diffusion-based models that can generate increasingly sophisticated training environments over time.
The company also believes simulations will be critical for scalable oversight. As AI systems become more autonomous and operate across millions of workflows and decisions, Patronus AI said manual review alone will not be sufficient.
Patronus AI’s long-term vision is to build systems that can supervise, evaluate, and govern increasingly autonomous agents at scale before failures occur in production environments.
KEY QUOTES:
“Benchmarks were never the destination. Static evaluations tell you whether a model can answer a narrow question in a controlled setting. They do not tell you whether an agent can navigate ambiguity, recover from failure, or operate reliably across long, unpredictable workflows. That requires environments where systems can practice, adapt, and accumulate experience over time.”
“Manual review does not scale once AI systems begin operating across millions of workflows and decisions. That is why simulations matter. They create environments where AI systems can be tested, improved, and supervised before failures happen in production.”
Anand Kannappan, CEO and Co-Founder of Patronus AI
“Patronus AI is tackling one of the most important infrastructure problems in artificial intelligence. The future of AI will depend on systems that can learn and operate reliably in complex environments, and simulations are becoming essential to making that possible.”
Itay Inbar, Partner at Greenfield Partners
