日本和日本NVIDIA用Blackwell GPUs建造两台AI超级计算机,发射2026年,以促进科学和量子研究。 Japan and NVIDIA build two AI supercomputers with Blackwell GPUs, launching 2026, to boost science and quantum research.

NVIDIA和日本的RIKEN正在开发两台新的超级计算机,使用Blackwell GPUs和NVLink技术,定于2026年春季发射,以推动AI驱动的科学和量子计算研究。 NVIDIA and Japan’s RIKEN are developing two new supercomputers using Blackwell GPUs and NVLink technology, set for launch in spring 2026, to advance AI-driven research in science and quantum computing. 这些系统将作为FugakuNEXT的前体,FugakuNEXT是Fujitsu和NVIDIA共同开发的下一代超级计算机,预计到2030年将实现。 These systems will serve as precursors to FugakuNEXT, a next-generation supercomputer co-developed by Fujitsu and NVIDIA, expected by 2030. 与此同时,横跨亚洲、欧洲、中东和美国的十多个全球超级计算中心都采用了NVIDIA的NVQLink,这是一个开放的互连架构,能够使量子处理器与GPU平台的超低时间、高通量集成。 Meanwhile, over a dozen global supercomputing centers across Asia, Europe, the Middle East, and the U.S. have adopted NVIDIA’s NVQLink, an open interconnect architecture enabling ultra-low-latency, high-throughput integration of quantum processors with GPU platforms. 这一技术支持实时混合量子-古典工作流程,包括首次使用qLDPC代码对量子错误校正进行可缩放实时解码,这是向防错量子计算迈出的关键一步。 This technology supports real-time hybrid quantum-classical workflows, including the first scalable real-time decoding of quantum error correction using qLDPC codes, a key step toward fault-tolerant quantum computing.