NVIDIA has announced a major partnership with Uber to accelerate the global rollout of level 4 autonomous vehicle fleets, making the world robotaxi-ready. At the heart of the initiative is the new NVIDIA DRIVE AGX Hyperion 10 platform, a reference compute and sensor architecture that enables any vehicle to be level 4-ready. This platform integrates the NVIDIA DRIVE AGX Thor system-on-a-chip, a safety-certified operating system, and a comprehensive multimodal sensor suite including 14 high-definition cameras, nine radars, one lidar, and 12 ultrasonics. The DRIVE AGX Hyperion 10 is designed to be modular and customizable, allowing automakers and developers to build safe, scalable, and AI-defined fleets. With two DRIVE AGX Thor platforms based on the Blackwell architecture, each delivering over 2,000 FP4 teraflops of real-time compute, the system is optimized for advanced AI workloads such as transformer models, vision-language-action (VLA) systems, and generative AI — essential for handling complex urban driving scenarios. Uber will leverage this technology to scale its global autonomous ride-hailing network, targeting 100,000 robotaxis by 2027. The company will operate a unified network that seamlessly integrates both human-driven and autonomous vehicles. To support this, NVIDIA and Uber are building a joint AI data factory powered by the NVIDIA Cosmos world foundation model platform, designed to curate and process vast amounts of real-world and synthetic driving data. Leading automakers are already joining the ecosystem. Stellantis is developing AV-Ready platforms optimized for level 4 autonomy, integrating NVIDIA’s full-stack AI technology and collaborating with Foxconn on hardware. Lucid is advancing level 4 capabilities in its next-generation U.S. vehicles using the DRIVE Hyperion platform. Mercedes-Benz is testing future robotaxi integration with its MB.OS operating system, aiming to deliver a luxury, chauffeured level 4 experience in its new S-Class. In freight, Aurora, Volvo Autonomous Solutions, and Waabi are developing level 4 long-haul trucks powered by NVIDIA DRIVE AGX Thor, extending the reach of AI-driven autonomy beyond passenger mobility. NVIDIA is also introducing the Halos Certified Program, the industry’s first system to evaluate and certify physical AI safety for autonomous vehicles and robotics. The NVIDIA Halos AI Systems Inspection Lab, accredited by the ANSI Accreditation Board, is already working with companies like AUMOVIO, Bosch, Nuro, and Wayve to ensure robust safety and cybersecurity standards. To further advance AI in autonomy, NVIDIA is releasing the world’s largest multimodal AV dataset — 1,700 hours of real-world data from 25 countries, including camera, radar, and lidar inputs. This dataset will support the training and validation of foundation models and VLA systems. Jensen Huang, CEO of NVIDIA, emphasized that robotaxis are no longer science fiction but a tangible future, enabled by AI infrastructure. Dara Khosrowshahi, CEO of Uber, praised NVIDIA as the backbone of the AI era and affirmed the partnership’s role in transforming urban mobility. With growing collaboration across the ecosystem — including Avride, May Mobility, Momenta, Pony.ai, Wayve, and WeRide — NVIDIA is establishing a unified, scalable foundation for the future of autonomous transportation.