Laplace intelligence uses AI to safeguard physical robots and self-driving cars before deployment

Laplace Intelligence, a startup based in Ann Arbor, uses AI to stress-test robotics and autonomous vehicles in simulation before real-world deployment. Founded a year ago by Howie Sun and Henry Liu, former University of Michigan researchers, the company aims to ensure safety by identifying and fixing potential failures. Sun, with eight years of autonomous driving AI research, emphasized proactive testing as AI agents transition from digital to physical environments. Laplace’s high-fidelity simulations, based on Nvidia’s Cosmos platform, recreate real-world environments like Arizona highways and Paris pedestrian crossings. Developers can modify settings in real time using language prompts, incorporating data from sources like TomTom. The startup is competing in TechCrunch Disrupt 2026, with plans to raise funds and scale. Clients include Isuzu, Bot Auto, and Nvidia, with projects spanning autonomous trucks and smart city initiatives. Laplace’s AI agents automatically generate edge cases, challenging autonomous systems with rare scenarios. Competitors include Applied Intuition, and Sun believes the approach offers a competitive edge in the field.