Nvidia CEO Jensen Huang visited Tokyo for two days on July 15 and 16 to meet with Japan’s industrial and chip supply elite. It was a few weeks after my keynote speech in Taiwan and a few months after my visit to Korea. He signed deals that spanned Japan’s entire technology ecosystem: a national AI factory, partnerships with Japan’s top robotics companies, and chip material suppliers supporting Nvidia’s next-generation AI chips. His message was clear. Nvidia is targeting factory floors in Japan, and many of the country’s largest manufacturers are getting in on the action. Huang said the next chapter of AI belongs to the factory floor, robots and machines, and that Japan wants to build it.
Thirty years ago, Sega’s $5 million investment helped keep Nvidia afloat from near bankruptcy. Today, Nvidia and the Japanese industrial giant need each other again. This time, we need each other to build the era of physical AI, starting with these three projects:
Noetra — Sovereignty of Japan – AI Play. The country does not want its factories and robots to be run by American or Chinese AI. So the government gathered about 44 domestic companies, led by Softbank, Sony, NEC, and Honda, to build its own AI for robots, vehicles, and factory floors. Tokyo is investing up to 1 trillion yen ($6.2 billion) over five years, betting on homegrown “physical AI” – a foundational model built to run machines. Japan wants to own software brains. However, the hardware needed to build it still comes from Nvidia. The US chip giant is building the “Vera Rubin AI Factory”, a massive data center filled with next-generation chips scheduled for launch in 2028, featuring 13,750 Vera CPUs and 27,500 Rubin GPUs, delivering 140 MW. Noetra will oversee these efforts through data center construction plans. Noetra’s plan unfolds in three stages. Starting in fiscal year 2026, an inference model focused on Japanese language proficiency; An omnimodal version that handles text, images, video and audio by 2028; By 2030, the “real native AI” built to run robots will be rolled out in stages to external Noetra developers.
Robot Alliance – Japan’s industrial giants have lined up behind Cosmos. Nvidia is targeting factory floors in Japan, and many of the country’s top robotics and manufacturing companies are signing on. Fanuc, Yaskawa, Kawasaki Heavy, Fujitsu, Hitachi, NEC, Sony, SoftBank, Kubota and robotics group AIRoA said they plan to build on Nvidia’s Cosmos model, an open model effort that Nvidia launched in May with a handful of global AI labs. In Tokyo, Nvidia gave them a reason to commit when it unveiled the Cosmos 3 Edge, a version of the model that runs on the Jetson Thor chip inside the machine. Some are already testing shared control systems. Other companies, such as Honda R&D and Omron, are currently building on this tool. “The next frontier for AI is in the physical world, and this is a once-in-a-generation opportunity for Japan,” Hwang said in a company statement. “Japan invented modern manufacturing. Now we have the opportunity to reinvent it for the age of intelligent industry.”
Toyota — Automotive and physical AI. Toyota uses Nvidia chips in most of its stack. The next generation of vehicles will be launched on Nvidia’s Drive platform at CES in January 2025. The new work expands Nvidia into manufacturing, where simulations are used to design production lines, software that runs vehicles, and systems that read road traffic. Toyota’s cars will do the steering and braking, but will still run advanced driver assistance features that require a driver. This is a more conservative approach than Waymo and Tesla, which are developing systems that rely less on human drivers.
Prime Minister Hwang’s visit puts physical AI at the center of Japan’s industrial strategy, and Tokyo is spending money to support it. Japan, facing a shrinking workforce, wants 10 million AI-equipped robots across 18 sectors by 2040, backed by $65 billion in public and private physical AI investments.
The longer the game, the bigger it gets. Japan’s AI robotics strategy announced last March aims to occupy more than 30% of the global AI robotics market by 2040. This market has a market value in Japan of approximately 20 trillion yen, or approximately $133 billion. METI is funding home-based models to run the machines, and Noetra’s Nvidia-based factory is where models at scale equivalent to trillions of parameters are trained. The bet is that Japan’s factory floor data and manufacturing base is well-suited to physical AI.
Beneath the industrial case is the sovereign case. As the United States and China move ahead in large-scale AI, Tokyo wants its own data, its own computing, and less reliance on infrastructure beyond its control. Huang appeared with Trade Minister Ryosei Akazawa at the government’s physical AI launch event on July 16, and Prime Minister Sanae Takaichi also joined by video. Takaichi’s administration has made AI and semiconductors central to its growth plan, seeking 370 trillion yen ($2.3 trillion) in public and private investment by 2040. Noetra’s factory, which Nvidia bills as “the world’s first national AI infrastructure,” is the most obvious choice yet. At least for now, Japan’s push for independence hinges on American chips, which are based on American silicon.
In two days, Huang had lunch with the CEOs of Toyota, Fanuc, Yaskawa, Fujitsu and Kawasaki, and met almost every important name in Japanese technology, including dozens of supply chain executives, over skewers and whiskey at Kanda Izakaya.
This is the same playbook he ran a few weeks ago. It was Taiwan’s homecoming keynote, fried chicken, and a 50,000 GPU deal in Seoul last fall. This time it was Tokyo’s turn, with robots, supply chains and chips underneath.
If you purchase through links in our articles, we may receive a small commission. This does not affect our editorial independence.
