When Jensen Huang flew to Tokyo last month and signed seven Japanese industrial giants into Nvidia's (NVDA -0.85%) new physical AI coalition, he was locking in a massive, long-lived stream of demand for Nvidia's chips and software, and that is something I think investors should really pay attention to.
Nvidia CEO Jensen Huang. Image source: Nvidia.
"Physical AI" refers to AI that controls robots, factory lines, and machines in the real world instead of just chatbots on a screen. In mid-July, Nvidia announced that companies like Fujitsu, FANUC, Yaskawa Electric, and Kawasaki Heavy Industries, along with Hitachi, NEC, SoftBank, Sony, and Kubota, intend to build on its Cosmos, Isaac, Metropolis, and Jetson platforms as part of a "Cosmos Coalition" focused on physical AI. Put simply, these are some of Japan's biggest names in robotics, manufacturing, and communications agreeing to standardize on Nvidia's stack as they build the brains for next-generation industrial automation.
Behind the coalition sits an even larger national project. Nvidia is partnering with Noetra, a Japanese AI consortium backed by Sony, SoftBank, Honda, and dozens of other firms, to build what it calls the world's first national infrastructure for physical AI. The centerpiece of that effort will be a Vera Rubin AI factory that will feature 13,750 Vera CPUs and 27,500 Rubin GPUs, delivering about 140 megawatts of compute capacity to train and deploy physical AI models. Japan's industry ministry has framed this as the computing backbone for its FRONTia program, and expects that it will help the country hit its goal of capturing 30% of the global AI robotics market by 2040.
NASDAQ: NVDA
Key Data Points
What about Nvidia shareholders?
On top of that, Prime Minister Sanae Takaichi's government has laid out a plan to mobilize more than 370 trillion yen ($2.3 trillion) in combined public and private investment by 2040 across physical AI, semiconductors, and data centers. Noetra's own roadmap calls for roughly 1 trillion yen ($6.3 billion) of sovereign AI spending over five years to develop domestic foundation models for robots and industrial AI. When you add up the national AI factory, the robot makers building on Cosmos, and Japan's broader tech investment targets, you are talking about demand that lives comfortably in the trillion-dollar range over the coming decades.
For Nvidia shareholders, the important part is not just that these orders exist. It is that they represent multiyear infrastructure-level commitments that are hard to unwind. FANUC and Yaskawa are not going to rip out their control platforms every cycle. A sovereign AI factory is not built for a single experiment. By turning physical AI into a coalition and tying it to Japan's long-term industrial strategy, Huang is trying to make Nvidia's chips and tools the default choice for robots and factories in one of the world's most advanced manufacturing economies.
That is what confirmed demand really means here. It is not a one-off spike in GPU sales. It is governments and industrial giants literally planning their futures around Nvidia's hardware and software, which gives its shareholders much more visibility into where revenue and profit margins might come from years down the line.
