NVIDIA and RIKEN, Japan’s leading national research institute, have announced a bold expansion of the nation’s computational infrastructure with two next-generation supercomputers. These systems, powered by the NVIDIA GB200 NVL4 platform and linked via Quantum-X800 InfiniBand networking, are set to accelerate scientific inquiry across AI for science and quantum computing.
Driving Japan’s “Sovereign Innovation” Across Science and Industry
Together, NVIDIA and RIKEN are creating a powerful foundation for Japan’s sovereign innovation strategy. The first of the two new systems will host 1,600 Blackwell GPUs, focusing on AI-driven science research in fields like life sciences, materials science, climate and weather forecasting, manufacturing, and laboratory automation.
The second supercomputer, dedicated to quantum computing, will incorporate 540 Blackwell GPUs. It will enable advanced quantum research supporting hybrid simulation, quantum-classical computing, and novel quantum algorithm development.
In total, the two machines will bring 2,140 Blackwell GPUs on-site, strengthening Japan’s domestic infrastructure for high-performance computing.
Strategic Remarks: Building for Scale and Sovereignty
Ian Buck, Vice President of Hyperscale and High-Performance Computing at NVIDIA, emphasized the importance of the partnership:
“RIKEN has long been one of the world’s great scientific institutions, and today it stands at the forefront of a new era in computing. Together, we’re helping Japan build the foundation for sovereign innovation that will drive breakthroughs to solve the world’s most complex scientific and industrial challenges.”
Satoshi Matsuoka, Director of the RIKEN Center for Computational Science, underlined how this leap forward will transform Japan’s research capabilities: “Integrating the NVIDIA GB200 NVL4 accelerated computing platform with our next-generation supercomputers represents a pivotal advancement for Japan’s science infrastructure. Our partnership will create one of the world’s leading unified platforms for AI, quantum and high-performance computing, allowing researchers to unlock and accelerate discoveries in fields ranging from basic sciences to industrial applications for businesses and society.”
Anchoring Future Systems: FugakuNEXT and Co-Design
These supercomputers also act as proxy systems for the development of the upcoming FugakuNEXT supercomputer. In August, RIKEN, Fujitsu, and NVIDIA announced a joint initiative to design FugakuNEXT the next-generation flagship that succeeds the celebrated Fugaku system.
FugakuNEXT is being co-designed to leverage Fujitsu’s MONAKA-X CPUs tightly coupled with NVIDIA GPUs via NVLink Fusion. The projected performance gain is profound: up to 100× application performance over current CPU-based supercomputers, alongside future integration of quantum hardware.
Also Read: Red Hat Enhances AI Accelerator Experience on Red Hat Enterprise Linux
Software Ecosystem: Powering Science with CUDA-X and Emulation
To maximize GPU utilization, NVIDIA and RIKEN are developing floating-point emulation software that maps scientific workloads to Tensor Core GPU operations. This will enable established HPC applications to harness GPU acceleration efficiently.
RIKEN will also use the CUDA-X software stack. It includes over 400 optimized GPU-accelerated libraries, microservices, and development tools. This toolkit aims to spark research in AI, quantum, and high-performance computing.
Strategic Impact & National Significance
This partnership between NVIDIA and RIKEN is more than just a tech upgrade. It shows Japan’s strong commitment to building its own AI and quantum infrastructure. The deployment supports national goals for secure, advanced computing. It also aims to boost Japan’s competitiveness in global science and technology. Japan is merging AI, HPC, and quantum research into one platform. This move helps Japan excel in important fields like materials science, climate modeling, advanced manufacturing, and quantum algorithms.
