English 
Sakura Internet Inc., a provider of digital infrastructure services, will begin offering “Sakura ONE,” a managed supercomputer that uses “NVIDIA Blackwell GPUs,” from Monday, October 20, 2025.
“Sakura ONE” is ranked in the international performance ranking “TOP500″ announced in June 2025 .”Sakura ONE” is a managed supercomputer rebuilt based on the system technology of the “NVIDIA H100 GPU” configuration, which ranked 49th in the world in the international performance ranking
In advance of September 2025, we will begin offering “Sakura ONE,” which uses the “NVIDIA H200 GPU.”
The newly offered “Sakura ONE” is designed specifically for the research and development of large-scale language model (LLM) learning and generative AI. It is comprised of servers equipped with eight of the latest GPUs, the “NVIDIA Blackwell GPU,” and can simultaneously utilize up to 48 servers, for a total of 384 GPU resources. This will enable us to provide high-performance computing resources from within Japan that can meet a wide range of computing needs.
こちらもお読みください： メシアス、Excelサポートを強化したSpreadJS V18.2Jを発表
さくら ONEは、コンピューティング環境の構築から日々の運用管理までを包括的にサポートするマネージドサービスで、最短30日間から日常的にご利用いただけるため、システム構築の負担なく、すぐにハイパフォーマンスコンピューティング環境をご利用いただけます。
The “NVIDIA H200 GPU” model, which is already available, is optimized for scientific and technical calculations, while the “NVIDIA Blackwell GPU” model is specialized for AI areas such as generative AI and LLM. By having both models, “Sakura ONE” can be used as a supercomputer that covers both the fields of science and technology and AI.
ソース さくら
日本語