Laplace Intelligence red-teams robots and self-driving cars before they hit the real world

Laplace Intelligence, a University of Michigan spinout based in Ann Arbor, is building automated red-teaming technology that stress-tests robots and physical AI systems by deliberately trying to make them fail in high-fidelity simulations. Co-founded by CEO Howie Sun and chief scientist Henry Liu, the startup grew out of academic research on challenging autonomous driving AI in simulation, including a 2023 paper that landed on the cover of Nature. Laplace's core product is a simulation environment built on a post-trained version of Nvidia's Cosmos world model, reconstructed from real-world data such as Google Street View imagery and real-time traffic information from providers like TomTom, in which AI agents continuously probe for weaknesses and design corner cases drawn from news reports, police collision records, and other sources. The company already generates a modest profit, with customers including Isuzu, Bot Auto, and Nvidia, and is working with T-Mobile on a free smart city pilot in San Jose, California. Its near-term goal, which it hopes to advance by competing in TechCrunch Disrupt's Startup Battlefield 200, is to raise funds and scale against larger rivals like Applied Intuition, betting that using AI to test AI will become a new paradigm and its biggest moat.

(Source：TechCrunch)