Laplace Intelligence brings AI-driven red-teaming to robotics and autonomous vehicles

At TechCrunch Disrupt in San Francisco on October 13, 2026, Howie Sun, co-founder and CEO of Laplace Intelligence, discussed how the Ann Arbor-based startup is bringing automated red-teaming to robotics and physical AI.

Red-teaming stress-tests a system by deliberately trying to make it fail. Sun said that as AI agents move from computers into the real world, ensuring physical AI does not fail in real-world settings will require AI to test AI. The company's goal is to help businesses discover and fix problems before deployment, and over time build a shared evidence base that customers, insurers, and regulators can trust.

Laplace was founded about a year ago out of the University of Michigan by Sun and chief scientist Henry Liu. Sun spent eight years in academia researching how to challenge autonomous driving AI in simulation; their 2023 paper appeared on the cover of Nature and drew interest from autonomous vehicle companies.

The startup combines a high-fidelity simulation reconstructed from real-world data with AI agents that continuously probe for weaknesses and corner cases. Its simulation environment is built on Nvidia's Cosmos platform, and developers can adjust scenes through language prompts. The AI agents can also design edge cases from sources such as news reports and police collision reports.

Laplace already has revenue from autonomous driving customers including Isuzu, autonomous trucking company Bot Auto, and Nvidia. It is also working with T-Mobile on a smart city pilot in San Jose, California. Sun described the AI-testing-AI approach as a new paradigm for physical AI testing and the company's biggest moat.