In an effort to bolster its sovereignty in artificial intelligence (AI) and enhance its research and development capabilities, the National Institute of Advanced Industrial Science and Technology of Japan (AIST) will integrate thousands of NVIDIA H200 Tensor Core GPUs into its ABCI 3.0 supercomputing infrastructure. The HPE Cray XD system will feature the NVIDIA Quantum-2 InfiniBand network to deliver superior performance and scalability.
ABCI 3.0 is the latest iteration of Japan’s large open AI computing infrastructure, designed to advance research and development in artificial intelligence. This collaboration underscores Japan’s commitment to advancing its AI capabilities and strengthening its technological independence.
“In August 2018, we launched ABCI, the world’s first large-scale open AI computing infrastructure,” said AIST Executive Officer Yoshio Tanaka. “Building on our experience managing ABCI, we are now enhancing ABCI 3.0. In collaboration with NVIDIA and HPE, our goal is to develop ABCI 3.0 into an infrastructure that will advance generative AI research and development capabilities in Japan.”
Hirotaka Ogawa, AIST Solutions Solution Producer and ABCI Operations Chief, added: “As generative AI prepares to catalyze global change, it is crucial to rapidly cultivate research and development capabilities within Japan. I am confident that this major upgrade to ABCI, in collaboration with NVIDIA and HPE, will enhance ABCI’s leadership in domestic industry and academia, propelling Japan towards global competitiveness in AI development and serving as a foundation for future innovation.”
ABCI 3.0: A New Era for Japan’s AI Research and Development
ABCI 3.0, built and operated by AIST, its subsidiary AIST Solutions, and systems integrator Hewlett Packard Enterprise (HPE), is part of a broader effort by Japan’s Ministry of Economy, Trade, and Industry (METI) to strengthen computational resources through the Economic Security Fund.
Collaboration with NVIDIA is key in this project, following the visit of the company’s founder and CEO, Jensen Huang, last year, who discussed the future of AI with political and business leaders, including Japanese Prime Minister Fumio Kishida.
NVIDIA’s commitment to Japan’s future
Jensen Huang pledged to collaborate on research, especially in generative AI, robotics, and quantum computing, invest in AI startups, and provide product support, training, and education in AI.
During his visit, Huang emphasized that “AI factories”—next-generation data centers designed to handle the most computationally intensive AI tasks—are crucial for turning vast amounts of data into intelligence.
“The AI factory will become the foundation of modern economies around the world,” Huang said during a press meeting with Japanese media in December.
Unmatched Computing Performance and Efficiency
The ABCI 3.0 facility will offer:
6 exaflops of AI computing capacity.
410 petaflops of double precision, a measure of general computing capacity.
Each node will be connected via NVIDIA’s Quantum-2 InfiniBand platform with a bandwidth of 200GB/s.
NVIDIA technology will form the backbone of this initiative, with hundreds of nodes equipped with 8 H200 GPUs connected by NVLink, providing unprecedented computational performance and efficiency.
NVIDIA’s H200 GPUs are the first to offer over 140 gigabytes (GB) of HBM3e memory at 4.8 terabytes per second (TB/s). The H200’s larger and faster memory accelerates generative AI and language models, while advancing scientific computing for HPC workloads with improved energy efficiency and lower total cost of ownership.
The integration of advanced NVIDIA Quantum-2 InfiniBand with network computing—where network devices perform calculations on data, offloading work from the CPU—ensures efficient, high-speed, low-latency communication, crucial for handling intensive AI workloads and large datasets.
Japan’s Strategic Vision
METI’s substantial investment is a testament to Japan’s strategic vision to enhance AI development capabilities and accelerate the use of generative AI.
By subsidizing the development of AI supercomputers, Japan aims to reduce the time and costs of developing next-generation AI technologies, positioning itself as a leader in the global AI landscape.