NVIDIA and Japanese infrastructure startup Noetra Corp. announced a partnership to build what both companies call the world’s first national AI factory dedicated to physical AI. The facility will house 13,750 NVIDIA Vera CPUs and 27,500 NVIDIA Rubin GPUs, delivering 140 megawatts of data center capacity on the NVIDIA DSX platform with Spectrum-X Ethernet networking and BlueField DPUs.
Japan’s Ministry of Economy, Trade and Industry (METI) is backing the project, which will serve as the computing foundation for Japan’s FRONTia Project, a government initiative focused on developing multimodal foundation models for robotics and physical AI applications.
What the Factory Will Do
The AI factory is built around a specific mandate: train open multimodal foundation models that power AI agents, digital twins, robotics, and physical AI applications. Noetra plans to make pretrained model weights broadly available to Japanese developers and enterprises, alongside NVIDIA software including Nemotron, Cosmos, Isaac GR00T, and NeMo.
“Japan invented modern manufacturing. Now, it is building the AI factories that will power the next industrial revolution,” Jensen Huang, NVIDIA’s founder and CEO, said in the announcement.
Ryosei Akazawa, Japan’s Minister of Economy, Trade and Industry, framed the project as leveraging Japan’s manufacturing heritage: “By fostering collaboration between Japan and leading global innovators, and leveraging Japan’s strengths, such as its onsite expertise and manufacturing technology infrastructure, we will build highly reliable multimodal foundation models and contribute to solving global social challenges.”
The Robotics Strategy Behind the Hardware
The factory aligns with Japan’s AI Robotics Strategy, released in March, which sets a goal for the country to capture more than 30% of the global AI robotics market by 2040. According to NVIDIA’s press release, that target represents an estimated $133 billion opportunity.
As the facility expands, it will support training trillion-parameter-scale AI models. Noetra CEO Hironobu Tamba emphasized the collaborative approach: “Together with partners across Japan and around the world, Noetra will advance Japan-developed multimodal foundation models and accelerate the deployment of physical AI across Japanese industries by broadly sharing the results of our research.”
Sovereign Compute in a Fragmented Market
The announcement is the latest in a pattern of nation-state AI infrastructure investments. Japan joins a growing list of countries building dedicated compute capacity outside the US hyperscaler ecosystem. The factory’s architecture on NVIDIA Vera Rubin NVL72 racks represents NVIDIA’s newest generation hardware being deployed at national scale for the first time. According to NVIDIA’s Japan ecosystem overview, Japanese companies across healthcare, manufacturing, telecommunications, and logistics are already building on NVIDIA’s full AI stack, from BioNeMo for drug discovery to Cosmos for robotics simulation.
The explicit focus on agentic AI and robotics as primary use cases, rather than language model inference alone, signals a shift in how governments are thinking about AI compute investment. Japan is betting that the next generation of AI value creation will come from agents that interact with the physical world, not just chat interfaces.