Cold shoulder: Why Beijing is freezing Nvidia's access to the Chinese market

In a significant move, Beijing has reportedly suspended purchases of another AI chip from Nvidia, effectively cutting the company off from the Chinese market. Industry experts interpret this action as a sign of China's increasing self-reliance in chip manufacturing and a strategic maneuver to gain leverage in trade discussions. Just a few months ago, Nvidia's CEO Jensen Huang announced from China that the U.S. would permit the resumption of sales for its H20 graphics processing units tailored for the Chinese market. At that time, he also unveiled the RTX Pro GPU designed for AI applications in smart factories and logistics. However, the situation took a turn in August when Chinese regulators mandated tech firms to pause their purchases of Nvidia's H20s for a national security review. The regulatory scrutiny has since intensified, expanding to include Nvidia's RTX Pro 6000D chip, which now means the company is unable to sell any products to Chinese customers, according to a report by the Financial Times. This follows revelations that Nvidia has reportedly violated China's anti-monopoly laws, prompting an ongoing investigation. The motivations behind China's actions against Nvidia remain somewhat ambiguous. However, analysts in the tech and geopolitical fields suggest that these developments indicate a newfound confidence in China's domestic chip manufacturing capabilities, which the country appears to be leveraging against the U.S. Concerns over national security have been cited as rationale for the intervention, with the suggestion that Nvidia's chips could potentially be used for surveillance — an idea raised by American lawmakers. Experts believe this move is part of Beijing's broader strategy to encourage Chinese AI firms to explore local alternatives. Although many anticipate that exports will eventually resume due to strong demand from Chinese AI companies, some firms had already indicated plans to order substantial quantities of the RTX Pro 6000D. However, these efforts were halted when domestic regulators restricted access to Nvidia's products, after determining that local AI chip manufacturers had achieved performance levels comparable to Nvidia's offerings. Despite the progress, challenges remain for China's AI chip sector. Analysts point out that production capacity is a significant hurdle, as domestic firms still struggle to manufacture chips at the required scale. Reports suggest that local chipmakers are aiming to triple the country's AI processor output in the coming year, signaling increased faith in domestic supply capabilities. There are also signs of advancements within China's AI ecosystem. Huawei recently announced new AI compute infrastructure powered by its proprietary Ascend chips, claiming to have developed the world's most powerful AI solution. Research firm SemiAnalysis noted that Huawei's CloudMatrix system outperformed Nvidia's competing solutions in certain areas, despite the Ascend chips delivering only a fraction of Nvidia's performance. Meanwhile, Chinese startups like DeepSeek are working on AI models compatible with the next generation of homegrown chips, and major corporations such as Alibaba and Baidu are reportedly transitioning to using internally developed chips for their AI training needs. Despite these developments, experts remain cautious about China's ability to completely replace Nvidia's technology. Ray Wang, a semiconductor and technology analyst, expressed skepticism about the feasibility of advancing AI solely with domestic alternatives. Analysts suggest that China's actions may also be a tactic to gain leverage in broader trade negotiations with the U.S., as the country seeks to bolster its semiconductor independence while negotiating access to superior chip technologies. The Biden administration has indicated a willingness to allow even more advanced Nvidia chips into China, reversing previous export controls aimed at limiting China's access to cutting-edge technology. Observers note that Beijing's rejection of Nvidia's products could be an attempt to create negotiating space for future discussions on advanced GPU access, coinciding with other recent trade tensions, including an investigation into U.S. analog chip imports.