Laplace Intelligence uses AI to red-team robots and self-driving cars

The core of the system involves AI agents that automatically design corner cases and failure scenarios to probe weaknesses in physical AI systems. This approach moves away from relying solely on recorded footage or human-designed edge cases, which is particularly important for generalized robotics where real-world data is scarce. Laplace claims its simulations are interactive, with simulated actors reacting realistically to test vehicles.

The startup is currently generating revenue from autonomous driving customers, including Isuzu, Bot Auto, and Nvidia. It is also working with partners on humanoid robotics for warehouses. Laplace is competing in the Startup Battlefield at TechCrunch Disrupt, aiming to raise funds to scale and compete with rivals like Applied Intuition, which has raised $1.2 billion at a $15 billion valuation.