NVIDIA outlines its robotaxi platform for fleet-scale autonomy

NVIDIA’s robotaxi platform is built around a three-computer solution covering AI model training, simulation and validation, and real-time in-vehicle computing. The stack includes NVIDIA DGX systems for training driving models, the NVIDIA Alpamayo portfolio of reasoning vision language action models and physical AI datasets, and reinforcement learning tools for adapting models to specific vehicles and deployment needs.

For testing and validation, NVIDIA Omniverse NuRec reconstructs real-world driving scenarios from sensor data, while NVIDIA Cosmos generates physically based variations across driving behavior, traffic, weather, lighting and sensor conditions. NVIDIA says adding meta-action and chain-of-thought reasoning data improved a VLA model’s trajectory prediction accuracy, reducing minimum average displacement error by 43%, from 2.08 to 1.18.

The in-vehicle layer centers on NVIDIA DRIVE Hyperion with DRIVE AGX. DRIVE Hyperion 10 combines dual NVIDIA DRIVE AGX Thor systems-on-a-chip with 14 high-definition cameras, nine radars, three lidars and 12 ultrasonics for real-time, 360-degree sensor fusion, supported by redundant compute and sensing for fail-operational driving.

NVIDIA’s ecosystem includes Uber, May Mobility, Bolt, Lyft, Grab, WeRide, Waymo, Wayve, Nissan, Autobrains, Zoox, Momenta, Pony.ai, Tensor, Waabi, TIER IV, Isuzu, Lenovo, DeepRoute.ai, Tesla, Mercedes-Benz, Stellantis, Lucid, Nuro, Hyundai Motor, Kia, Geely and Zeekr across robotaxi development, production programs and commercial deployment plans.