NVIDIA is expanding its partnership with Japan to build out the country's quantum and artificial intelligence infrastructure. This collaboration represents the first international segment of the Genesis Mission. The goal is straightforward. Both nations want to apply AI to scientific breakthroughs. According to an announcement on the official NVIDIA newsroom, the partnership spans multiple industries, research labs, and academic institutions to build a shared foundation for future research.
At RIKEN, Japan’s leading national research institute, 2 new supercomputers are beginning operations. Both systems run on NVIDIA Blackwell technology. The hardware is designed to accelerate research in fields ranging from life sciences to materials science.
The RIKYU system is focused directly on AI development. It deploys 1,600 Blackwell GPUs through the GB200 NVL4 platform. RIKEN will use this massive processing power to build open foundation models. The hope is to bring machine learning into everyday laboratory automation.
Then there is ROQUO, which functions as a hybrid quantum HPC system. This supercomputer integrates quantum processors with 540 Blackwell GPUs. It connects directly to local quantum computers at RIKEN facilities in Wako and Kobe, including the trapped ion Reimei system built by Quantinuum. Researchers have already started using the CUDA Q platform to generate quantum circuits for the Reimei hardware.
Connecting this hardware is only half the battle. AI is often the missing link needed to scale up quantum processors, but making it work smoothly is difficult. To solve this, NVIDIA is working with the National Institute of Advanced Industrial Science and Technology at their G QuAT center. They are using NVQLink to provide a low latency connection between the GPUs and quantum processors. They are also deploying Ising open models for automated processor calibration and error correction.
Simulating chemical systems is another major focus. Mitsubishi Chemical, Mizuho Bank, Keio University, and the University of Toronto worked with NVIDIA to build a GPU powered workflow for molecular spectral analysis. This is a vital tool for understanding molecular structures. During testing, NVIDIA GPUs achieved a 13.4x speedup over systems that rely on a CPU only. Researchers plan to use this speed to analyze extreme ultraviolet photoresist, a material needed for advanced semiconductor manufacturing.
Fujitsu is also working with NVIDIA to test CUDA Q for large scale molecular simulations. The tech company is trialing NVQLink to see if it can successfully manage hybrid quantum classical computing environments. This collective effort aims to establish a shared foundation for future scientific research.
