Analysis
Laplace Intelligence, a University of Michigan spinout based in Ann Arbor, builds automated red-teaming tools that stress-test robots and self-driving cars in simulation before they operate in the real world, according to TechCrunch. The company reconstructs real-world map data into simulations and runs AI agents through them to find the failure points that generalized robots -- with little training data of their own -- are most likely to hit.
Co-founder and CEO Howie Sun and chief scientist Henry Liu, who was Sun's PhD adviser at Michigan, founded the company about a year before the October 2026 profile. "No one knows what the edge cases will be in the real world," Sun said, describing the premise behind the product.
Already profitable, without a disclosed round
Laplace has not disclosed any funding, but the company says it is generating revenue and is modestly profitable, mostly through project-based contracts rather than recurring subscriptions. Named customers include Bot Auto, a Texas autonomous-trucking company, and an unnamed major Japanese automaker; it also ran a free smart-city pilot in San Jose with Nvidia and T-Mobile, and has early work underway with unnamed humanoid-robotics partners on warehouse use cases.
That profile puts Laplace in sharp contrast with its most direct comparison, Applied Intuition, which has raised $1.2 billion at a $15 billion post-money valuation building simulation and testing tools for the same autonomous-vehicle and robotics customers. Waabi, an autonomous-trucking company that also builds AI tools to generate edge cases internally, is a looser comparison -- it is a customer of its own simulation work rather than a vendor to the industry.
What to watch
Laplace's simulations run on a post-trained version of Nvidia's Cosmos platform and use traffic data licensed from TomTom, which ties its roadmap to decisions made by larger partners it doesn't control. The company presented at TechCrunch Disrupt's Startup Battlefield 200 on October 13, a showcase that has launched fundraising rounds for plenty of past participants. Whether Laplace follows that path with an outside round, or keeps growing on project revenue alone, will say a lot about whether safety-testing infrastructure can be a standalone venture-scale business or stays a niche service for a small set of robotics primes.
