Since launching less than three years ago, Patronus AI has become a leader in AI evaluation, simulation infrastructure, and reliability testing for frontier AI systems. Today, Patronus AI works with the majority of the world’s leading frontier AI labs and hyperscalers. The company’s revenue has grown more than 15x over the past year, reflecting growing demand for infrastructure that helps organizations train, evaluate, and deploy increasingly autonomous AI systems. The new funding brings Patronus AI’s total capital raised to $70 million. <\/p>\n
Patronus AI was founded by AI researchers and engineers with backgrounds at organizations including Meta AI, Amazon AGI, and Google. The team’s experience spans LLM evaluation, AI alignment, fairness, and embodied agents, providing the technical foundation for the company’s work in simulation and evaluation infrastructure. <\/p>\n
From Static Benchmarks to Simulated Digital Worlds <\/p>\n
The first phase of generative AI was built on static internet text and benchmark leaderboards. But as agents move into longer, more complex workflows, the limitations of that approach are becoming increasingly clear. <\/p>\n
An agent managing a customer escalation, navigating enterprise software, conducting research across thousands of documents, or debugging production infrastructure cannot be trained through benchmark memorization alone. These systems need dynamic environments that resemble the digital world they will actually operate inside. <\/p>\n
Patronus AI is building what it describes as Digital World Models \u2014 language diffusion world models that are designed to scale the creation of simulation data to train and evaluate AI agent actions across complex digital workflows. <\/p>\n
The company builds simulation infrastructure that allows AI systems to train on realistic software, research, communication, and enterprise workflows. Instead of optimizing for narrow benchmark performance, the goal is to produce agents that can operate reliably across ambiguous, long-horizon tasks. <\/p>\n
“Benchmarks were never the destination,” said Anand Kannappan, CEO and co-founder of Patronus AI. “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.” <\/p>\n
Introducing World’s First Digital World Models <\/p>\n
Patronus AI believes simulations will become one of the defining infrastructure layers of the AI era. <\/p>\n
The company’s research focuses on generating ecologically valid environments where agents can encounter edge cases, recover from failures, and improve through repeated interaction. This includes simulation tooling, evaluation systems, and diffusion-based Digital World Models that can generate increasingly sophisticated training environments over time. <\/p>\n
The approach is designed to address one of the largest unsolved problems in AI: scalable oversight. <\/p>\n
As AI systems become more capable, manual review becomes increasingly insufficient. Patronus AI’s long-term vision is to build systems capable of supervising, evaluating, and governing increasingly autonomous agents at scale. <\/p>\n
“Manual review does not scale once AI systems begin operating across millions of workflows and decisions,” said Kannappan. “That is why simulations matter. They create environments where AI systems can be tested, improved, and supervised before failures happen in production.” <\/p>\n
New Funding Fuels Research and Expansion <\/p>\n
With the new funding, Patronus AI plans to expand its research organization, grow its engineering team, and invest in the compute and infrastructure required to train and run Digital World Models at scale. <\/p>\n
“Patronus AI is tackling one of the most important infrastructure problems in artificial intelligence,” said Itay Inbar, Partner at Greenfield Partners. “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.” <\/p>\n
About Patronus AI <\/p>\n
Patronus AI is a simulation and evaluation infrastructure company building Digital World Models to accelerate the next generation of AI agents. Founded by former Meta AI researchers Anand Kannappan and Rebecca Qian, the company develops large-scale simulation environments, evaluation systems, and reliability infrastructure that help AI research and engineering teams to build and deploy trustworthy AI systems. <\/p>\n
For more information, visit\u00a0<\/a>https:\/\/www.patronus.ai<\/a>\u00a0 <\/p>\n