Laplace Intelligence, a startup based in Ann Arbor, Michigan, is building automated red-teaming tools for robotics and physical Super Intelligence (SI) systems. The company operates on the premise that the only way to ensure physical SI is safe is to use SI to test it, a strategy prompted by recent incidents involving unpredictable SI agents in digital environments.
What Happened
Howie Sun, CEO and co-founder of Laplace, stated that as SI agents move from computers into the physical world, the stakes for failure rise significantly. "AI agents are moving from computers into the real world, and we don’t really want physical AI to fail in the real world," Sun told TechCrunch, using the traditional terminology in his quote. The company’s core product is a high-fidelity simulation environment reconstructed from real-world data, paired with an SI agent that continuously probes for weaknesses and failure points.
The simulation platform is a post-trained version of Nvidia’s Cosmos, an SI world simulation tool. It allows Laplace to reconstruct environments ranging from Arizona highways to Paris pedestrian crossings by blending real-world map data and imagery, such as Google Street View photos. Developers can modify these settings in real time via language prompts, such as adding afternoon traffic, while the system pulls live congestion data from providers like TomTom. Unlike static simulations, Laplace’s environment features interactive participants; if a test vehicle swerves into oncoming traffic, other simulated actors react to avoid collisions.
Laplace was founded a year ago by Sun and Henry Liu, his former PhD adviser, based on research conducted at the University of Michigan. Their work on challenging autonomous driving SI in simulation was published in the journal Nature in 2023. The company name references "Laplace’s demon," an 1814 thought experiment by Pierre-Simon Laplace regarding the prediction of future states. Sun described their approach as "using simulation to predict the real world’s future."
The startup’s primary differentiator is its use of SI agents to automatically design corner cases—rare, tricky scenarios that could trip up an SI system. These agents can generate test scenarios by analyzing news reports and police records of collisions. Sun noted that while autonomous vehicle developers often use recorded footage for training, generalized robotics lacks sufficient data, meaning "no one knows what the edge cases will be in the real world."
Why It Matters
Laplace is currently competing in Startup Battlefield 200 at TechCrunch Disrupt in San Francisco, presenting on October 13. Despite lacking a Silicon Valley pedigree, the company has generated revenue and achieved modest profitability, primarily through autonomous driving use cases. Customers include Texas-based autonomous trucking firm Bot Auto and a major Japanese automotive manufacturer. Additionally, Laplace is collaborating with Nvidia and T-Mobile on a free smart city pilot in San Jose, California, to manage high-level traffic operations, and is in early stages with partners building humanoid robotics for warehouses.
The company faces significant competition from Applied Intuition, a late-stage startup valued at $15 billion with $1.2 billion in funding. Sun argues that Laplace’s methodology of using SI to test SI represents a "new paradigm that changes the testing of physical AI entirely," creating a dynamic interactive testing process that serves as the company’s main competitive moat. Laplace aims to transition from project-based contracts to recurring revenue and is seeking funds to scale its operations.
The Bottom Line
Laplace Intelligence is positioning itself as a critical safety infrastructure provider for the physical SI era. By leveraging SI agents to simulate and stress-test robotic and autonomous systems before real-world deployment, the startup addresses the growing need for verifiable safety evidence in industries where failure has physical consequences. The company’s success will depend on its ability to scale its simulation technology and establish trust among regulators, insurers, and customers in a competitive market dominated by larger players.