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

Key points

• Laplace Intelligence, an Ann Arbor startup, builds automated red-teaming for robotics and physical AI.
• CEO Howie Sun and chief scientist Henry Liu founded the company a year ago out of University of Michigan.
• The startup is competing in Startup Battlefield 200 at TechCrunch Disrupt in San Francisco on October 13-15.

Laplace Intelligence, an Ann Arbor-based startup, is developing automated red-teaming for robotics and physical AI. The company stress-tests systems by deliberately trying to make them fail, aiming to ensure safety before real-world deployment. It is one of 20 startups competing in Startup Battlefield 200 at TechCrunch Disrupt in San Francisco this week, October 13-15, and will present on the Disrupt Stage on Tuesday afternoon, October 13.

CEO and co-founder Howie Liu, along with co-founder and chief scientist Henry Liu, started Laplace a year ago out of the University of Michigan. Sun spent eight years in academia researching how to challenge autonomous driving AI in simulation. In 2023, he and Liu published a paper on the subject that appeared on the cover of Nature, leading to interest from autonomous vehicle companies.

How the technology works

Laplace uses a high-fidelity simulation reconstructed from the real world, paired with an AI agent that constantly probes for weaknesses. The simulation is a post-trained version of Nvidia's Cosmos, grounded in real-world data. It blends real-world map data to recreate environments true to scale, such as an Arizona highway or a Paris pedestrian crossing. Developers can use language prompts to change settings in real time, like adding afternoon traffic, and Sun says Laplace pulls real-time traffic information from providers like TomTom.

Within the simulations, the behavior of other actors is also simulated. In a demo, Sun directed the test vehicle to swerve into oncoming traffic, and the car that would have been hit moved out of the way. Sun called it 'truly an interactive simulation environment,' adding that everyone is reacting to everything. The AI agents automatically design corner cases, drawing from news and police reports to challenge autonomous vehicles and other physical AI.

Our goal is to help companies discover and fix problems before deployment — and, over time, build a shared foundation of evidence that customers, insurers, and regulators can trust

— Howie Sun

Customers and competition

Laplace is generating revenue and turning a modest profit, mostly from autonomous driving use cases. Customers include Japanese trucking manufacturer Isuzu and Texas-based autonomous trucking firm Bot Auto. Nvidia is also a customer, and the two companies, along with T-Mobile, are working on a free smart city pilot for San Jose, California. The startup is also in early stages with partners building humanoid robotics for warehouses.

Most contracts are project-based, and the company aims to convert them into recurring revenue. Its near-term goal is to raise funds and scale to compete with rivals, chiefly Applied Intuition, which has raised $1.2 billion at a $15 billion post-money valuation. Sun believes Laplace's core methodology of using AI to test AI is a 'new paradigm that changes the testing of physical AI entirely.'