Toyota and NVIDIA Expand Partnership to Advance L2++ ADAS and AI Across Vehicles, Factories, and Smart Cities

On July 14, Toyota and NVIDIA announced an expansion of their collaboration, extending physics-based AI technologies from the automotive domain into robotics and intelligent infrastructure. Toyota is leveraging the NVIDIA DRIVE AGX platform and the safety-certified NVIDIA DriveOS operating system to develop its next-generation vehicles featuring L2++ advanced driver-assistance capabilities, enhancing the vehicle’s understanding of road environments and decision-making performance. The partnership spans multiple areas: In automotive development, Toyota utilizes NVIDIA’s accelerated computing, AI software, and simulation technologies to improve vehicle safety and intelligence. In software development, Toyota has trained and fine-tuned a MISRA-compliant Code Assistant AI model based on NVIDIA Megatron-LM, combined with datasets such as NVIDIA Nemotron, to efficiently generate, review, and validate safety-critical code—significantly shortening development cycles. In manufacturing, Toyota employs NVIDIA Omniverse libraries and the Isaac Sim open framework to build digital twin environments, simulating production processes and robotic movements to optimize workflows, reduce downtime, enhance efficiency, and lower costs. Additionally, Woven by Toyota, a subsidiary of Toyota, is developing a multimodal vision-language model—the Woven City AI Vision Engine—using NVIDIA H100 Tensor Core GPUs and Megatron-Core. This model analyzes urban traffic environments, predicts incident evolution, and enables real-time responses from transportation and infrastructure systems. The collaboration also extends to smart city development, advancing AI-driven urban mobility systems and traffic analytics technologies.