Japan is moving ahead with what is being described as its first national AI factory, a large-scale computing project built around Nvidia technology. The system is planned to include 27,500 Rubin GPUs and 13,750 Vera CPUs spread across 382 Vera Rubin racks, creating a major domestic platform for advanced AI development.
According to the available details, Nvidia is working with Japan's newly formed Noetra consortium on the project. The facility is expected to draw 140 megawatts of power and serve as the computing backbone for the government's FRONTia program, underscoring the scale of the initiative.
Networking will rely on Nvidia's Spectrum-X Ethernet platform, which is intended to connect the massive number of processors inside the system. As capacity increases over time, the infrastructure is being designed to support the training of trillion-parameter AI models, a benchmark associated with some of the most demanding machine learning workloads.
The project highlights Japan's push to build national AI capacity with high-performance infrastructure rather than depending entirely on overseas resources. By combining thousands of GPUs, CPUs and high-speed networking in one installation, the AI factory is positioned as a foundation for the country's next phase of large-model computing.
