Huang’s Japan blueprint targets factory floors as Tokyo races for AI independence

Jensen Huang spent July 15 and 16 in Tokyo, leaving with commitments spanning Japan's industrial ecosystem: a national AI factory, partnerships with the country's robotics leaders, and agreements with chip-material suppliers for Nvidia's next-generation processors.

The visit marks a strategic turn toward what Huang calls "physical AI"—AI built to run robots, vehicles, and factory floors rather than data centers. "Japan invented modern manufacturing. Now, it has the opportunity to reinvent it for the age of intelligent industries," Huang said at a government event alongside Japan's trade minister.

Three projects anchor the effort.

Noetra: Japan's sovereign-AI bet. A consortium of 44 domestic firms, led by SoftBank, Sony, NEC, and Honda, is building foundation models designed for robots and factory automation. Japan's government is committing up to 1 trillion yen ($6.2 billion) over five years. Noetra will oversee construction of what Nvidia calls "the world's first national AI infrastructure"—a data center launching in 2028 with 13,750 Vera CPUs and 27,500 Rubin GPUs, delivering 140 megawatts. The effort runs in three phases: a reasoning model emphasizing Japanese-language skills by fiscal 2026, a multimodal version handling text, images, video, and audio by 2028, and "Real-world Native AI" for robots by 2030.

Robotics coalition on Cosmos. Fanuc, Yaskawa, Kawasaki Heavy, Fujitsu, Hitachi, NEC, Sony, SoftBank, Kubota, and AIRoA are committing to build on Nvidia's Cosmos models. Nvidia unveiled Cosmos 3 Edge—a version running on Jetson Thor chips inside machines themselves—giving partners a concrete tool to work with. Some companies are already testing shared control systems; Honda R&D and Omron are building applications now.

Toyota's dual commitment. Toyota, a major Nvidia customer across its stack, extended the partnership into manufacturing simulations for production-line design, vehicle software, and traffic-reading systems. At CES in January 2025, Toyota committed its next-generation vehicles to Nvidia's Drive platform. The company is taking a conservative approach to autonomous driving—advanced driver assistance that steers and brakes but requires a human operator—compared to Waymo and Tesla's less-dependent systems.

Why this matters. Japan faces a shrinking workforce and has set a target of 10 million AI-equipped robots across 18 sectors by 2040, backed by $65 billion in public and private physical-AI investment. The longer horizon is bolder: Japan's AI Robotics Strategy, released in March, aims to capture more than 30% of the global AI robotics market by 2040—a market valued at roughly ¥20 trillion ($133 billion).

Underlying the industrial case is a sovereignty play. As the U.S. and China extend their leads in large-scale AI, Tokyo wants its own data, its own compute infrastructure, and less reliance on systems it doesn't control. Prime Minister Sanae Takaichi's administration has positioned AI and semiconductors as the centerpiece of a growth plan targeting ¥370 trillion ($2.3 trillion) in combined public and private investment by 2040. Noetra's factory is the clearest concrete step.

The paradox is sharp: Japan's push for independence currently runs on American silicon. Huang's two-day visit—meetings with CEOs over lunch, dinners with supply-chain executives in a Kanda izakaya—followed a similar playbook to his recent Taiwan keynote and his 50,000-GPU deal in Seoul last fall. This time, the focus was robots, semiconductors, and the infrastructure underneath Japan's next manufacturing chapter.