After Jensen Huang’s Tokyo Trip, Nvidia Deepens Its Bet on Japan’s Physical AI

Lead

Jensen Huang’s July 15–16 visit to Tokyo was more than a round of executive meetings. According to TechCrunch, Nvidia left Japan with a set of agreements that touch almost every part of the country’s technology base: government-backed AI infrastructure, robotics, automotive software, manufacturing simulation, and chip-material supply chains.

The message was consistent with Huang’s broader Asia tour: the next phase of AI will not be limited to chat interfaces or cloud services. It will move into factories, vehicles, robots, and machines that operate in the physical world. For Japan, that narrative is especially compelling because the country combines deep manufacturing expertise with a shrinking workforce and a policy drive for AI sovereignty.

Key points

• Noetra is Japan’s sovereign AI vehicle: Japan is bringing together roughly 44 domestic companies, with SoftBank, Sony, NEC, and Honda among the core players, to build AI for robots, vehicles, and factory floors. Tokyo is committing up to 1 trillion yen over five years to support homegrown physical AI foundation models.
• The compute layer still depends on Nvidia: Noetra is expected to oversee a Vera Rubin AI factory using Nvidia’s next-generation platform. The planned system, expected in 2028, includes 13,750 Vera CPUs and 27,500 Rubin GPUs, with a 140-megawatt footprint.
• Japan’s robotics leaders are aligning with Cosmos: Fanuc, Yaskawa, Kawasaki Heavy, Fujitsu, Hitachi, NEC, Sony, SoftBank, Kubota, and AIRoA plan to build on Nvidia’s Cosmos models. Nvidia also introduced Cosmos 3 Edge, designed to run on Jetson Thor chips inside machines, bringing model execution closer to the factory floor.
• Toyota broadens Nvidia’s role: Toyota has already committed next-generation vehicles to Nvidia’s Drive platform. The newer work extends Nvidia’s reach into production-line simulation, vehicle software, and traffic-reading systems, while Toyota continues to pursue a more driver-involved advanced assistance strategy than Waymo or Tesla.

Why it matters

The visit shows that Nvidia is not simply selling accelerators into another national market. It is embedding itself in Japan’s plan to reinvent industrial automation around AI. Japan wants 10 million AI-equipped robots across 18 sectors by 2040 and aims to capture more than 30% of the global AI robotics market by then. Those goals require models, data, deployment environments, and large-scale compute to move together.

Japan’s strengths are real: factory-floor data, industrial discipline, robotics know-how, and manufacturers that can test physical AI in complex production settings. Its vulnerability is also clear. The country wants its own data, compute, and AI stack, but the most advanced hardware for training models at this scale still comes from Nvidia.

For Nvidia, Japan could become a showcase for the post-chatbot phase of AI. If factories, robots, and cars are rebuilt around Nvidia’s software and silicon, the company’s role expands from chip supplier to foundational infrastructure provider for intelligent industry.

Source: TechCrunch AI