At the RISC-V Summit China held in China recently, Frans Sijsterman, vice president of hardware engineering at NVIDIA, said that CUDA will be added to support RISC-V architecture processors, which means that this key technology for deep learning and GPU acceleration applications will be able to run in a more diverse processor ecosystem, further expanding the possibilities of heterogeneous computing.
In the past, CUDA was mainly used to operate on x86 or Arm architecture processor platforms. The announcement of support for RISC-V architecture processors not only echoes NVIDIA's open strategy for CPU pairing in its GPU products, but also highlights NVIDIA's emphasis on the Chinese market. Especially in the current situation where Sino-US relations are becoming increasingly tense and the cost of Arm architecture licensing continues to increase, many Chinese companies have begun to replace previous x86 architecture or Arm architecture processor designs with open source RISC-V architecture processors. Expanding support for RISC-V architecture processors will also allow NVIDIA to embrace more development opportunities.
Exciting news from #RISCVSummitChina, as Frans Sijstermans from NVIDIA announces CUDA is coming to RISC-V! This port will enable a RISC-V CPU to be the main application processor in a CUDA-based AI system.#RISCV #RISCVEverywhere pic.twitter.com/08C2ghPHq9
— RISC-V International (@risc_v) July 18, 2025
NVIDIA has been working with IBM for a long time to integrate its GPU with POWER PC processors for accelerated computing. Later, it built the Grace CPU based on the Arm architecture design, and used NVLink technology to enhance the integrated computing efficiency between the GPU and CPU. In addition, it has been integrated with the x86 architecture for a long time in the past. The announcement that it will add support for RISC-V architecture processors will enable NVIDIA's GPU to have greater configuration and application flexibility, allowing users to build a more suitable accelerated computing platform according to their needs.
On the other hand, NVIDIA announced at Computex 2025 this yearNVLink Fusion Design, claiming to use this opportunity to cooperate with processor manufacturers such as Fujitsu and Qualcomm to promote the accelerated development of AI that is larger in scale and more in line with actual computing needs.
It is not yet confirmed when CUDA is expected to officially support RISC-V architecture processors, but considering NVIDIA's long-term strategy of emphasizing heterogeneous computing and platform openness, this move will not only strengthen its technological dominance in the global AI ecosystem, but is also expected to indirectly increase the visibility of RISC-V architecture in practical application areas.
