In a move that has sent ripples through global technology and political circles, the United States has decided to permit Nvidia, the world’s largest semiconductor company by market value, to export its powerful H200 artificial intelligence (AI) chips to China. The announcement, made by US President Donald Trump on December 8, 2025, via the social media platforms Truth Social and X, marks a significant shift in the ongoing tech rivalry between Washington and Beijing. Yet, the story is far from straightforward, as both countries have responded with complex measures that reflect the high stakes of AI supremacy and national security.
According to Financial Times, Chinese regulators are already discussing ways to limit access to Nvidia’s H200 chips, despite the US export green light. The chips, which belong to Nvidia’s previous Hopper generation, are considered the company’s second-best offering for AI applications, trailing only the latest Blackwell and Rubin series. Trump’s decision came after years of escalating restrictions, with previous US administrations—most notably under President Joe Biden—banning the sale of advanced AI processors to China over fears they could be used for military ends, including weapons development, cyberattacks, and surveillance.
Under the new US policy, Nvidia can sell the H200 chips to China, but only to customers approved by the US Department of Commerce and under stringent conditions designed to maintain strong national security. The department is currently finalizing the list of eligible buyers and the details of the export process. As part of the deal, the US government will receive a hefty 25 percent share of the sales revenue—a significant jump from the 15 percent previously demanded for the less powerful H20 chips. Trump’s announcement also clarified that the export of Nvidia’s latest Blackwell and forthcoming Rubin chips remains strictly off-limits.
"The approval applies only to 'approved customers' under conditions ensuring strong national security," Trump stated, emphasizing the cautious approach. Nvidia echoed this sentiment, telling tagesschau.de that the offer for approved commercial customers, vetted by the Commerce Department, "creates a well-considered balance. This is great for America." The same rules will also apply to other leading US chipmakers such as AMD and Intel, signaling a broader policy shift rather than a one-off concession to Nvidia.
The H200 chips are a substantial upgrade for Chinese buyers. As reported by SPIEGEL and the think tank Institute for Progress, the H200 is nearly six times more powerful than the previously permitted H20 model. This leap in performance could enable Chinese AI laboratories to build supercomputers approaching the capabilities of the best US AI supercomputers—though at higher costs and with certain limitations. Still, critics worry that even this partial opening could bolster China’s military and AI ambitions, the very concern that spurred the original US export restrictions.
Notably, Chinese President Xi Jinping reportedly responded positively to the development after being informed by Trump, according to SPIEGEL. Yet, the mood in the markets and among tech insiders has been more muted. While Nvidia’s shares initially jumped by as much as 2 percent in pre-market trading, the gains quickly faded, with the stock last up by just 0.6 percent, Reuters noted. Industry analysts attribute this lukewarm reaction to lingering uncertainty about how much access Chinese firms will truly have—and how much demand remains for US chips in China, given the country’s push for technological self-reliance.
Indeed, Beijing has not stood idly by. As Financial Times and Handelsblatt reported, Chinese regulators are actively considering their own restrictions, potentially allowing only limited access to the H200 chips. Officially, the Chinese government cites concerns about US backdoors and cyber-espionage, but the underlying motive is clear: to reduce dependence on American technology and promote domestic alternatives. In recent years, Chinese tech giants like Huawei, Biren Technology, and Cambricon have made significant strides in AI chips, eroding Nvidia’s once-dominant market share in China from around 90 percent before 2022 to just 10–20 percent today.
"China’s government has instructed domestic firms to favor homegrown products," Handelsblatt explained. Imports of US semiconductors are not outright banned, but they now require complex approvals, and in sensitive sectors such as the military and critical infrastructure, US technology is largely blocked. Even major Chinese tech companies—Baidu, Alibaba, Tencent, and ByteDance—have shifted their AI operations to local chips. Whether the new H200 access will reverse this trend remains to be seen.
The stakes for Nvidia are enormous. With a market capitalization of $4.51 trillion, the company is the world’s most valuable publicly traded firm. Its core business—graphic processors (GPUs) for AI training and inference—accounts for the lion’s share of its revenue. Nvidia’s global market share in AI chips stands at a commanding 80 to 90 percent, but competitors like AMD, Intel, and Google are catching up fast. In China, the competition is even fiercer, with local champions vying for dominance as US restrictions push innovation inward.
The US government’s balancing act is evident. The export approval is framed as a compromise between maintaining America’s technological edge and not pushing China to accelerate its own chip development. As Trump put it, "This approach is a middle ground between providing China with Nvidia’s newest chips and a complete export ban." Some US officials warn that a total ban would only strengthen Chinese firms like Huawei, while others point to America’s reliance on rare earths from China as a reason for caution. Still, critics—both in the US and abroad—continue to argue that any sale of advanced AI chips could inadvertently enhance China’s military capabilities.
Nvidia CEO Jensen Huang, who has maintained a close relationship with Trump, offered a pragmatic perspective: "You will not be able to stop China’s progress in artificial intelligence," he warned, as quoted by SPIEGEL. "Instead, strong domestic technology will be developed in China, which will later compete globally with US offerings, including those from Nvidia." Huang’s conclusion is clear: for the US to retain its leadership in AI, it must continue to sell chips to China, even as it navigates the risks.
As the details of the export process are hammered out by the US Department of Commerce—and as China weighs its own countermeasures—the world will be watching closely. The outcome could define not only the future of the global AI industry but also the broader contours of US-China relations in the years ahead.
