Nvidia said it was working to support the RISC-V chip architecture on its CUDA software platform, a move expected to boost the open-source movement that China is betting on as part of its tech self-sufficiency drive.
“We are porting CUDA to the RISC-V architecture,” Frans Sijstermans, vice-president of hardware engineering at Nvidia, said at the 2025 RISC-V Summit in Shanghai last week.
RISC-V central processing units (CPUs) could then be used as the main application processor in Nvidia systems, Sijstermans said, although he did not give a timeline for the plan.
RISC-V, the fifth generation of the open-sourced Reduced Instruction Set Computer architecture for CPUs, is free for anyone to use and modify. First developed in 2010 by researchers at the University of California, Berkeley, RISC-V is now managed by RISC-V International, a Zurich-based non-profit organisation.
The move represents a major step by Nvidia to boost the development of the open-source chip architecture in AI computing.