NVIDIA and RIKEN Advance Japan’s Scientific Frontiers With New Supercomputers for AI and Quantum Computing

NVIDIA announced that RIKEN, Japan’s premier national research institute, will integrate NVIDIA GB200 NVL4 systems into two new supercomputers — one focused on AI for scientific research and the other dedicated to quantum computing.

The first system will incorporate 1,600 NVIDIA Blackwell GPUs, powered by the GB200 NVL4 platform and connected with NVIDIA Quantum-X800 InfiniBand networking. This supercomputer will support RIKEN’s AI-for-science initiative, driving advancements in life sciences, materials development, climate and weather modeling, manufacturing, and laboratory automation.

The second system, designed for quantum computing research, will deploy 540 NVIDIA Blackwell GPUs using the same GB200 NVL4 architecture and NVIDIA Quantum-X800 InfiniBand networking. It will accelerate studies in quantum algorithms, hybrid simulation methods, and quantum-classical computing.

“RIKEN has long been one of the world’s most prestigious scientific institutions, and it now stands at the forefront of a new computing era,” said Ian Buck, NVIDIA’s vice president of hyperscale and high-performance computing. “Together, we’re helping Japan build a foundation for sovereign innovation that will drive breakthroughs across complex scientific and industrial challenges.”

“Integrating the NVIDIA GB200 NVL4 accelerated computing platform into our next-generation supercomputers marks a significant leap forward for Japan’s scientific infrastructure,” said Satoshi Matsuoka, director of the RIKEN Center for Computational Science. “This partnership will deliver one of the world’s most unified platforms for AI, quantum, and HPC, enabling researchers to speed up discoveries from basic science to industrial applications.”

Expanding Collaboration With RIKEN

The two new systems build on the August announcement of a joint effort between Fujitsu and NVIDIA to codesign FugakuNEXT — the successor to the iconic Fugaku supercomputer. These new GPU-accelerated machines will also act as proxy systems for developing hardware, software, and applications for FugakuNEXT.

FugakuNEXT is expected to use FUJITSU-MONAKA-X CPUs paired with NVIDIA technologies through NVIDIA NVLink™ Fusion, a new interconnect enabling high-bandwidth communication between Fujitsu’s CPUs and NVIDIA GPUs.

The system is projected to deliver up to 100× higher application performance compared with today’s CPU-based supercomputers and will eventually integrate production-grade quantum computing capabilities.

By combining MONAKA-X processors with NVIDIA’s newest GPUs, FugakuNEXT aims to redefine the future of scientific research through breakthroughs in HPC, AI, quantum computing, and hybrid approaches.

Overall, the growing partnership with RIKEN underscores Japan’s deep commitment to technological innovation — and NVIDIA’s role in strengthening the nation’s supercomputing and AI research infrastructure.

Advancing Science Through Supercomputing Software

NVIDIA is also working with RIKEN to develop floating-point emulation software that leverages NVIDIA Tensor Core GPUs, boosting the performance of traditional scientific computing. This approach enables applications at RIKEN and other supercomputing centers worldwide to fully utilize GPU power for both AI and HPC workloads.

RIKEN will additionally adopt NVIDIA CUDA-X™, a suite of over 400 optimized GPU-accelerated libraries and tools, to enhance its cutting-edge HPC applications and accelerate its AI-for-science and quantum computing programs.

The two new supercomputers are scheduled to begin operations in spring 2026, while FugakuNEXT is expected to go online by 2030.