Laplace Intelligence uses AI to stress-test robots and self-driving cars in simulation

Laplace Intelligence, a startup founded out of the University of Michigan, builds automated red-teaming software that uses AI agents to find failure points in robots, autonomous vehicles and smart city systems before deployment.

CEO Howie Sun says its simulations are built from real-world data, such as Google Street View images and map data, and are based on a post-trained version of Nvidia's Cosmos.

Customers include Isuzu and Bot Auto, and Laplace is competing in TechCrunch Disrupt's Startup Battlefield.