Japan's Nvidia AI factory: what it means for NVDA stock
Japan's government, industrial leaders and Nvidia (NASDAQ: NVDA) have unveiled what Nvidia calls the world's first national AI infrastructure for physical AI. Noetra Corp. will build a Vera Rubin AI factory with 27,500 Rubin GPUs and 13,750 Vera CPUs, delivering 140 megawatts of capacity for METI's FRONTia Project—news closely watched by NVDA stock investors.
Key Takeaways
- Nvidia and Noetra will deploy a Vera Rubin AI factory with 27,500 Rubin GPUs, 13,750 Vera CPUs and 140 MW of data-center capacity.
- Japan's METI-backed FRONTia Project will use the site to train open multimodal foundation models for robotics and physical AI.
- Nvidia is also expanding Japan partnerships with Fanuc, Yaskawa, Fujitsu and others, plus a new Cosmos 3 Edge model.
- Noetra plans to start construction in April 2027 and begin operations in June 2028, per Reuters.
The announcement, detailed in an Nvidia investor relations release, frames the build as sovereign-scale compute for Japan's industrial AI push—and a demand signal that keeps NVDA stock in the spotlight.
What exactly did Japan and Nvidia announce?
Nvidia said it is working with Noetra Corp. to launch an NVIDIA Vera Rubin AI factory based on the NVIDIA DSX platform. The system will use Vera Rubin NVL72 racks, Spectrum-X Ethernet networking and BlueField DPUs.
Supported by Japan's Ministry of Economy, Trade and Industry (METI), the factory is positioned as the computing foundation for the FRONTia Project—formally, “Development of Multimodal Foundation Models with a View to AI Robotics and Physical AI.”
Nvidia describes the initiative as the world's first national AI infrastructure for physical AI, aimed at manufacturing, logistics, healthcare, telecommunications and related industries.
Why does this national AI factory matter?
Physical AI—models that help machines perceive and act in the real world—needs vast compute. The planned 140-megawatt facility is designed to train open multimodal foundation models that power AI agents, digital twins, robotics and other physical AI applications.
Pretrained model weights are to be shared broadly with domestic developers, alongside software such as NVIDIA Nemotron, Cosmos, Isaac GR00T open models and NeMo libraries.
Japan's AI Robotics Strategy, released in March, targets more than 30% of the global AI robotics market by 2040—an opportunity Nvidia estimates at $133 billion. As the factory expands, Nvidia says it will support training trillion-parameter-scale models.
“Japan invented modern manufacturing. Now, it is building the AI factories that will power the next industrial revolution,” Nvidia founder and CEO Jensen Huang said.
How is Nvidia expanding Japan's physical AI ecosystem?
Alongside the infrastructure news, CNBC reported that Nvidia unveiled Cosmos 3 Edge, a world model meant to help robots and vision AI agents perceive and navigate physical environments in real time.
Nvidia is also forming a Japan coalition that industrial firms including Fujitsu, Hitachi and Kawasaki Heavy Industries intend to join. Reuters reported partnerships with robotics leaders Fanuc and Yaskawa Electric, and said government-backed Noetra—whose investors include Sony—will buy 27,500 Nvidia Rubin chips.
According to Reuters, Noetra plans to begin construction in April 2027 and start operations in June of the following year. For more on similar breakthroughs, explore Future Tech & AI Wonders on BlasterPost.
METI Minister Ryosei Akazawa said FRONTia will serve as the core of Japan's physical AI ecosystem, while Noetra CEO Hironobu Tamba stressed that no single company can solve the compute, data and foundation-technology challenges alone.