Nvidia built its empire on chips that power generative artificial intelligence, and that engine is still growing. The company posted record quarterly revenue of $81.6 billion, up 85 percent year over year. But with a valuation of 5 trillion dollars prices in boom years, CEO, Jensen Huang is under pressure to prove that expansion can extend beyond the US data center boom. His latest field focuses on Japan, where the government and major industrial players are preparing to deploy artificial intelligence in robots and factories.
During a two-day visit to Tokyo last week, Huang gathered Japan’s manufacturing giants, robotics innovators and software developers behind the “Physical AI,“Technology that allows robots and other machines to perceive their surroundings, make decisions and act in the real world. Japan, Huang argued, has a natural advantage for such a change. “Japan has historically been very good at precision manufacturing and very large-scale manufacturing, but now we have AI. You can combine the two technologies and create robotics 5 during a July event.”
Japan has strong incentives to embrace the partnership. It boasts world-class industrial facilities but faces a worsening labor shortage. premier Sanae Takaichiwhich has made semiconductors and artificial intelligence central to its growth agenda, is trying to marry that hardware strength with cutting-edge software. Japan accounts for about 70 percent of the global industrial robot market, but just over 10 percent of the service robot market, according to government data. Its updated strategy targets more than 30 percent of the emerging AI robotics market by 2040, representing $133 billion in business.
During his visit to Tokyo, he presented his vision to the country’s industrial elite. At a pub in Tokyo’s Kanda district, Huang gathered more than 30 executives from 16 major firms, including Tokyo Electron, Panasonic AND Mitsubishi Electricamong others – to discuss how Japan’s semiconductor supply chain can support an AI-led expansion. During the lunch, he also met with executives from Fujitsu, Kawasaki Heavy Industries, Fanuc and Yaskawa. According to Nvidia, all four companies are now developing physics-AI systems on its platform.
Tokyo’s reception contrasts with Nvidia’s experience in China. Huang visited the country in January AND returned in May as part of President Trump’s delegation, trying to rebuild Nvidia’s position in a market squeezed by US export controls and Beijing’s support for domestic chipmakers.
By comparison, Japan is putting government support behind infrastructure built around Nvidia’s technology. On the second day of his trip, Huang appeared alongside the Japanese Minister of Industry Ryosei Akazawa to launch the nation’s new government-backed physical artificial intelligence initiative. orAs part of that launch, Huang announced that Nvidia is partnering with Noetra Corpa Japanese AI consortium backed by Sony, SoftBank, Honda and roughly 40 other companies, to build what the chipmaker calls “the world’s first national AI infrastructure for physical AI”
“Japan must own, improve, secure and deploy AI Japan,” Huang said during the keynote. “After 15 years of work, physical artificial intelligence is here, the foundation of the next industrial revolution, and it must be made in Japan.”
The project centers on an “AI factory” powered by 13,750 Nvidia Vera CPUs and 27,500 next-generation Ruby GPUs. Expected to offer 140 megawatts of data center capacity, the facility will provide the computing needed to train complex physical AI models. Construction is scheduled to begin in April 2027, with operations coming online in June 2028.
The infrastructure will serve as the computing backbone for FRONTia, a government project supporting multimodal models for physical AI Noetra will lead the development of native models that run on the network. According to Noetra’s roadmap, the group will prioritize understanding and reasoning in Japanese before expanding to text, image, video and audio capabilities in 2028. By 2030, it aims to deploy “real-world native AI” for robotics and autonomous machinery.
This week, Nvidia brought the same message to SIGGRAPH, the ongoing annual computer graphics conference in Los Angeles. During Monday takeawayHuang tied Nvidia’s roots in computer graphics directly to its physics-based AI ambitions. “Whether it’s games, cinema, robotics or digital factory twins, the goal is the same: to create virtual worlds that behave with the fidelity and realism of the physical world,” he said in a pre-recorded video.
Before a robot can navigate its surroundings, it needs a model of how the physical world behaves. At SIGGRAPH, Nvidia released the Cosmos 3 Edge, a compact version of its world-class Cosmos model designed to run directly on hardware inside robots, vehicles and peripherals. By processing data locally rather than running queries to a remote data center, edge models enable real-time decision making. Companies currently evaluating the framework include Agile Robots, Doosan Robotics, Siemens and Skild AI.
However, not all training needs to be done in the physical space. During the keynote speech, Ming-Yu LiuNvidia’s vice president of Cosmos Lab unveiled Cosmos-Dreams, a set of virtual test environments for physical AI In a demonstration, the software generated an entire driving scene from a single video frame.
Cosmos was trained to act out a scene, predict outcomes, simulate consequences and choose an action, Liu said. These capabilities can be applied to machines ranging from humanoid robots to mechanical weapons and self-driving vehicles. “Each incarnation speaks a different language,” Liu said. “Our solution is to build a common vocabulary.”
