Nvidia built its empire on chips that power generative AI, and that engine is still going strong. The company posted a record $81.6 billion in quarterly revenue, up 85 percent year over year. But with a $5 trillion valuation in years of growth, CEO Jensen Huang is under pressure to ensure that expansion can extend beyond US data center growth. His latest centers on Japan, where the government and major industry players are gearing up to incorporate AI into robots and factories.
During a two-day visit to Tokyo last week, Huang brought together Japanese manufacturing giants, robotics inventors and software developers behind “virtual AI,” technology that allows robots and other machines to see their surroundings, make decisions and act in the real world. Japan, Huang argued, has a natural advantage for such change. “Japan has historically been very good at precision manufacturing and mass production, but now we have AI. You can combine the two technologies and create robots,” he told reporters during the July 15 event.
Japan has strong motivations for accepting partnerships. It boasts world-class industrial infrastructure but faces a dire labor shortage. Prime Minister Sanae Takaichi, who has made semiconductors and AI a centerpiece of his growth agenda, is pushing to combine that power with more advanced hardware and 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 revised strategy targets more than 30 percent of the emerging AI robotics market by 2040, representing a $133 billion business.
During his visit to Tokyo, he presented his vision to prominent people in the country’s industries. At a pub in Tokyo’s Kanda district, Huang gathered more than 30 executives from 16 major companies—including Tokyo Electron, Panasonic and Mitsubishi Electric, among others—to discuss how Japan’s semiconductor industry can support AI-led expansion. During lunch, he also met with executives from Fujitsu, Kawasaki Heavy Industries, Fanuc and Yaskawa. According to Nvidia, all four companies are now building wearable AI systems on their own.
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 team, trying to rebuild Nvidia’s position in a market squeezed by US export controls and Beijing’s support for domestic chip makers.
Japan, by contrast, is putting government support behind the infrastructure built around Nvidia’s technology. On the second day of his trip, Huang appeared alongside Japanese Industry Minister Ryosei Akazawa to launch a new government-backed AI initiative. AAs part of that initiative, Huang announced that Nvidia has partnered with Noetra Corpa Japanese AI consortium backed by Sony, SoftBank, Honda and nearly 40 other companies, to build what the chip maker calls “the first national AI infrastructure for virtual reality”
“Japan must own, develop, protect and use Japanese AI,” Huang said during his remarks. “After 15 years of work, virtual AI is here, the foundation of the next industrial revolution, and it must be made in Japan.”
The project focuses on an “AI factory” powered by 13,750 Nvidia Vera CPUs and 27,500 next-generation Rubin GPUs. Expected to deliver a capacity of 140 megawatts for the data center, the facility will provide the computing needed to train portable AI models. Construction is scheduled to begin in April 2027, and operations will come online in June 2028.
The infrastructure will serve as the computing backbone of FRONTia, a government project that supports multi-object models of virtual AI Noetra will lead the development of networked home models. According to Noetra’s roadmap, the group will prioritize the understanding and reasoning of the Japanese language before expanding into text, image, video and audio capabilities by 2028. By 2030, it aims to deploy “real-world native AI” to robots and autonomous machines.
This week, Nvidia brought that same message to SIGGRAPH, the annual conference on advanced computer graphics Los Angeles. On Monday key noteHuang linked Nvidia’s roots in computer graphics directly to its realistic AI ambitions. “Whether it’s games, cinema, robotics or industrial digital 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 environment, it needs a model of how the virtual world behaves. At SIGGRAPH, Nvidia released Cosmos 3 Edge, a compact version of its Cosmos world model designed to run directly on hardware inside robots, cars and peripherals. By processing data locally instead of sending queries to a remote data center, edge models enable real-time decision making. Companies currently testing the framework include Agile Robots, Doosan Robotics, Siemens and Skild AI.
However, not all training needs to take place in a virtual environment. During the keynote speech, Ming-Yu Liu, Nvidia’s vice president of Cosmos Lab, revealed Cosmos-Dreams, a series of virtual test areas for physical AI.
Cosmos was trained to interpret an event, predict outcomes, simulate outcomes and choose action, Liu said. These skills can be applied to machines ranging from human-made robots to mechanical arms and self-driving cars. “Every context speaks a different language,” Liu said. “Our solution is to create a common vocabulary.”
