Laplace Intelligence, a spinout from the University of Michigan, is employing AI agents and interactive simulation to stress-test robots and self-driving vehicles. The company, founded by Howie Sun and Henry Liu, aims to help companies discover and fix problems in physical AI before real-world deployment.
The Ann Arbor-based startup uses a high-fidelity simulation environment, a post-trained version of Nvidia’s Cosmos, which is grounded in real-world data. This allows for the reconstruction of environments ranging from highways to pedestrian crossings, with developers able to modify settings using language prompts.
A core aspect of Laplace's offering is its AI agents, which automatically design corner cases and challenging scenarios by drawing information from sources like news and police reports. This approach seeks to move beyond traditional methods of using recorded footage or human-designed edge cases for training autonomous systems.
Laplace Intelligence has generated revenue and achieved a modest profit, primarily from autonomous driving use cases. Its customers include Japanese trucking manufacturer Isuzu, Texas-based autonomous trucking firm Bot Auto, and Nvidia. The company is also involved in a smart city pilot for San Jose with Nvidia and T-Mobile, and is exploring humanoid robotics for warehouses.
Laplace is one of 20 startups competing in TechCrunch Disrupt’s Startup Battlefield, where it is presenting on Monday, October 13.