Top Scientist Calls for China's Development of Alternatives to Nvidia Software

In the world of artificial intelligence (AI) development, a key figure has urged China to develop alternatives to existing software platforms dominated by major players like Nvidia. Li Guojie, a respected computer scientist from the Chinese Academy of Sciences, highlighted two crucial issues that impact not only AI progress but also the ongoing technology rivalry between the United States and China.

Li emphasized that while Nvidia remains a giant in the field due to its advanced CUDA computing platform, there is a rising need for China to create its own solutions for long-term self-sufficiency in AI technology. The scientist pointed out that, despite the progress of Chinese AI accelerator chips—such as those from Huawei Technologies, Hygon Information Technology, and Cambricon Technologies—they still face challenges in matching Nvidia’s software ecosystem.

"Nvidia's strength is not just in its powerful chips but in the comprehensive CUDA ecosystem, which provides tools for engineers to create applications efficiently on their graphics processing units (GPUs)," Li explained. He stressed that for China to compete effectively, it needs to develop a robust alternative to this ecosystem.

Li acknowledged that initiatives like DeepSeek are making strides in this space; however, he noted that they have not fully transitioned away from CUDA. "There are still obstacles to bypassing CUDA completely," he commented. From his perspective, establishing a controllable AI software tool system that can outperform CUDA is essential for the country's technological independence.

The computer scientist drew parallels between China's journey to replace software and hardware systems, likening it to the efforts to replace Western platforms such as Microsoft Windows and Intel processors or Google's Android and Arm architectures. He stated, "Building an AI software ecosystem is an extremely challenging task that requires meticulous planning and persistent efforts over an extended period." He has been advocating for this approach to secure a strong position for China in the global tech arena.

AI, China, Nvidia