Tokyo, Japan – Hewlett Packard Enterprise (HPE) has been selected to build AIST’s next-generation supercomputer, ABCI 3.0, in Japan. This advanced system will leverage HPE Cray XD architecture, powered by NVIDIA H200 Tensor Core GPUs and NVIDIA Quantum-2 InfiniBand networking. Expected to be Japan’s fastest AI supercomputer, ABCI 3.0 will deliver approximately 6.2 exaflops of performance, significantly enhancing generative AI research and applications across various sectors, including healthcare, manufacturing, and energy.
ABCI 3.0: Enhancing Generative AI Research and Innovation
The new supercomputer, ABCI 3.0, is set to revolutionize the field of generative AI by providing unprecedented computational power and efficiency. Utilizing the latest NVIDIA technology, the system will facilitate complex AI model training and simulation, enabling breakthroughs in scientific research and technological innovation. AIST’s collaboration with HPE aims to accelerate advancements in AI, fostering innovation and addressing critical challenges in industries such as autonomous driving, drug discovery, and climate modeling.
Strategic Partnership to Drive AI Advancements in Japan
HPE’s partnership with AIST underscores a strategic commitment to advancing AI capabilities in Japan. By delivering the country’s most powerful AI supercomputer, HPE aims to support Japan’s leadership in AI research and development. The ABCI 3.0 will serve as a crucial infrastructure for academic and industrial stakeholders, promoting collaborative efforts and driving significant progress in AI-driven solutions. This initiative reflects HPE’s dedication to empowering global innovation through cutting-edge technology and strategic partnerships.
