Ann Arbor-based startup Laplace Intelligence is developing an automated AI testing platform to identify potential failures in robots, self-driving cars, and other physical AI systems before they are deployed. The platform uses realistic simulations and AI agents to generate challenging scenarios, allowing developers to test how autonomous systems respond to unexpected situations.
The company was among 20 startups selected to compete in Startup Battlefield 200 at TechCrunch Disrupt 2026, held in San Francisco from October 13 to 15. Its technology focuses on automated red-teaming, a testing method that deliberately challenges systems to identify weaknesses.
Founded in 2025 by CEO Howie Sun and chief scientist Henry Liu, Laplace originated from research at the University of Michigan. Sun spent eight years studying autonomous vehicle testing, including research published with Liu in Nature in 2023 that attracted interest from automotive companies.
Laplace Uses AI Simulations to Test Autonomous Systems
Laplace’s platform uses a post-trained version of Nvidia Cosmos, an AI world simulation platform, to reconstruct physical environments using real-world information. It combines map data, street imagery, and traffic information to create realistic simulations of locations such as highways and pedestrian crossings.
Developers can modify simulated environments using text prompts, including adding traffic congestion or changing road conditions. Laplace also incorporates traffic information from navigation providers such as TomTom to reflect real-world conditions.
Unlike simulations with static objects, Laplace models how vehicles and other participants respond to changing situations. During a demonstration, an AI-controlled vehicle swerved into oncoming traffic, prompting another simulated vehicle to take evasive action.
The platform also uses AI agents to automatically generate rare and difficult scenarios that could cause autonomous systems to fail. These scenarios can be developed using news reports, police collision records, and other sources rather than relying entirely on recorded driving footage or manually designed tests.
Sun told TechCrunch that testing general-purpose robots presents additional challenges because less real-world operational data is available. The company intends to use automated testing to identify potential failures before deployment.
Laplace Generates Revenue From Autonomous Driving Customers
Laplace has begun generating revenue and reported a modest profit, primarily from autonomous driving applications. Its customers include Texas-based autonomous trucking company Bot Auto and a major Japanese automotive manufacturer.
The startup is also collaborating with Nvidia and T-Mobile on a free smart city pilot in San Jose, California, focused on traffic operations. Separately, it is working with partners developing humanoid robots for warehouse applications.
Most customer agreements are currently project-based, although Laplace is working toward recurring revenue contracts. The startup also plans to raise funding to expand its operations and compete with larger companies.
One competitor is Applied Intuition, which has raised approximately $1.2 billion and reached a $15 billion post-money valuation. Sun said Laplace’s focus on using AI agents to continuously test other AI systems forms the basis of its approach to autonomous system validation.
Featured image credits: rawpixel.com
For more stories like it, click the +Follow button at the top of this page to follow us.