RIKEN, Fujitsu, and NVIDIA Unveil FugakuNEXT: Japan’s Next-Gen AI-HPC Supercomputer for Scientific Innovation

Japan is advancing its scientific and technological leadership with the launch of FugakuNEXT, the next-generation supercomputer set to succeed the globally acclaimed Fugaku system. At the FugakuNEXT International Initiative Launch Ceremony in Tokyo on August 22, RIKEN, Japan’s premier research institution, announced a strategic collaboration with Fujitsu and NVIDIA to co-design the new system, marking a pivotal moment in the nation’s high-performance computing journey. This initiative goes beyond mere performance gains. FugakuNEXT is being built to address Japan’s most pressing scientific and societal challenges, including climate modeling, disaster resilience, drug discovery, and next-generation manufacturing. The project reflects a deliberate shift toward aligning computing power with real-world impact, integrating cutting-edge AI and simulation technologies to accelerate discovery. The collaboration brings together top minds from academia, industry, and government. RIKEN President Makoto Gonokami and Satoshi Matsuoka, director of the RIKEN Center for Computational Science and a leading figure in HPC architecture, delivered keynote remarks. Fujitsu Chief Technology Officer Vivek Mahajan underscored the company’s commitment to strengthening Japan’s technological sovereignty, while Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA, highlighted the importance of co-design in shaping a system that meets the demands of future research. The project responds to a call made by NVIDIA CEO Jensen Huang during his visit to Tokyo the previous year, urging Japan to build its own AI infrastructure using local talent and global technologies. FugakuNEXT embodies that vision, leveraging NVIDIA’s full software stack—including CUDA-X libraries like cuQuantum for quantum simulation, RAPIDS for accelerated data science, TensorRT for high-performance inference, and NeMo for large language model development—as well as specialized SDKs tailored for scientific and industrial applications. At its core, FugakuNEXT will be a hybrid AI-HPC system, uniquely designed to seamlessly integrate traditional scientific simulation with artificial intelligence workloads. It will feature Fujitsu’s MONAKA-X CPUs, engineered for high performance and efficiency, and will utilize NVLink Fusion technology to enable ultra-high-bandwidth communication between Fujitsu’s processors and NVIDIA’s accelerators. The system is expected to enable transformative capabilities such as automated hypothesis generation, intelligent code creation, and real-time simulation of complex experiments. RIKEN, Fujitsu, and NVIDIA will jointly develop software tools focused on mixed-precision computing, continuous benchmarking, and performance optimization to maximize the system’s effectiveness. Backed by Japan’s Ministry of Education, Culture, Sports, Science and Technology (MEXT), FugakuNEXT will serve as a national resource for universities, government agencies, and private sector partners. It represents more than a technological upgrade—it is a long-term investment in Japan’s scientific independence, innovation ecosystem, and global leadership. As Japan enters a new era of supercomputing, FugakuNEXT stands as a testament to the power of sovereign infrastructure, international collaboration, and a shared commitment to solving humanity’s most complex problems.