Robotaxi Leaders Build on NVIDIA Full-Stack Platform

NVIDIA has detailed how the world's leading robotaxi companies are building their autonomous fleets on its full-stack open platform. The timing is significant: the global robotaxi market is projected to reach $400 billion by 2035, with more than 6 million commercial vehicles expected to be in operation. Driverless fleets already carry passengers through some of the busiest and most complex streets in the world. Scaling from limited pilots to citywide and eventually global operations, however, demands enormous computing resources. Training perception and planning models, running large-scale simulation to validate edge cases, and performing onboard inference in real time all require specialized hardware and software working together. NVIDIA positions its platform as covering that entire pipeline, from data center training to in-vehicle compute. Robotaxi developers can use the same underlying stack across development and deployment, which simplifies engineering and helps ensure that what is validated in simulation behaves consistently on the road. The open nature of the platform also allows companies to customize models and integrate their own sensor suites and software. The post frames physical AI as the first commercial breakthrough of the coming decade, with autonomous vehicles leading the way. Robotaxis represent a demanding proving ground: they must operate safely in unpredictable urban environments, handle weather and construction, and do so without a human driver as backup. Meeting that bar at scale is a massive technical challenge, and NVIDIA is betting that its full-stack approach will be the foundation on which most major fleets are built.