印度正在建造土生土长的超级计算机,目标是到2030年实现完全独立,其国内含量达到70%。 India is building homegrown supercomputers, aiming for full independence by 2030 with 70% domestic content.

印度正在向2030年实现全国超级计算发展, 预计2032年将实现商业化, India is advancing toward fully indigenous supercomputing by 2030, with commercial systems expected by 2032, as domestic content in high-performance computing reaches 50% and aims to exceed 70% by decade’s end. 国家超级计算任务已通过自制Rudra服务器实现40千万亿次浮点次，目标到2026年达到90千万次浮点次。 The National Supercomputing Mission has already achieved 40 petaflops using homegrown Rudra servers, targeting 90 petaflops by 2026. 在 NSM 2. 0 下推出的快速计算新路线图着重于开发本地CPU、GPU和AI加速器。 A new roadmap for exascale computing under NSM 2.0 focuses on developing local CPUs, GPUs, and AI accelerators. 超过38 000个GPU已分发给研究人员和初创企业,而像MosChip这样的印度公司正在推进芯片设计。 Over 38,000 GPUs have been distributed to researchers and startups, while Indian firms like MosChip are advancing chip design. 政府的电子元件制造计划已批准了10个芯片制造和包装项目。 The government’s Electronics Component Manufacturing Scheme has approved ten projects in chip fabrication and packaging. 印度的电子制造业在十年中增长了六倍,移动生产增长了28倍,出口增长了127倍。 India’s electronics manufacturing has grown sixfold in ten years, with mobile production up 28 times and exports surging 127 times. 印度还在投资2公里和3毫米芯片设计,并通过2025年超级计算印度等活动促进全球合作。 The country is also investing in 2-nanometer and 3-millimeter chip design and promoting global collaboration through events like Supercomputing India 2025.