China’s Government Reportedly Aims to Ban NVIDIA AI Chips, Yet Weak Local Alternatives Prevent Full Implementation

Recent developments indicate that Beijing is increasingly concerned about its domestic industry’s reliance on American artificial intelligence (AI) technology and is actively seeking alternatives to mitigate this dependency.

Chinese Big Tech’s Shift Away from NVIDIA’s H20 AI Chip

China’s technology sector has heavily depended on AI chips and equipment from the United States. However, shifting geopolitical dynamics have prompted Beijing to address this vulnerability. One significant step has been the initiation of a regulatory investigation by China’s Cyberspace Administration, focused on uncovering potential security vulnerabilities—referred to as “backdoors”—in NVIDIA’s H20 AI accelerators. According to a Financial Times report, there are legislative discussions in China advocating for an outright ban of the H20 AI chip, although actual enforcement may prove challenging.

The backlash against NVIDIA’s AI chips appears to have intensified following comments from U. S.Commerce Secretary Howard Lutnick, who suggested that American technology should foster a dependency among Chinese developers. These remarks have been interpreted as insulting by China’s high-ranking officials, fueling a movement among local technology leaders to reassess their engagement with NVIDIA’s offerings.

Reports indicate that several Chinese companies have either significantly cut back their orders for the H20 chip or have abandoned them entirely, signaling a notable shift away from American AI technology. Concurrently, these firms are beginning to favor domestic alternatives, particularly from providers like Huawei and Cambricon, which are perceived to meet their inferencing workload requirements adequately.

China envisions a future where its AI computing demands can be met by homegrown solutions, with optimistic projections suggesting that by next year, increased production capabilities could realize this goal. However, the current landscape still heavily relies on NVIDIA. Recent insights revealed that DeepSeek faced delays in launching its next AI model, the R2, due to reliance on domestic chips before ultimately reverting to NVIDIA’s technology, underscoring the seriousness of China’s current dependence on American AI innovations.

For further reading, please refer to our full article and images.