US imposes 25% tariff on AI chip exports to China

President Donald Trump has announced a 25% tariff on certain advanced AI chips from Nvidia and AMD, allowing their export to China while claiming a share of the sales revenue. The policy reverses a prior export ban on Nvidia's H200 chips but imposes the levy to fund US interests. Industry executives view it as a way to shield the arrangement from legal challenges.

On January 15, 2026, US President Donald Trump signed a proclamation imposing a 25% tariff on advanced computing chips imported into the United States and then re-exported abroad, particularly targeting sales to China. The measure affects Nvidia's H200 and AMD's MI325X processors, which are manufactured by Taiwan Semiconductor Manufacturing Company (TSMC) in Taiwan.

This policy builds on a December decision by the White House to permit Nvidia to ship H200 chips to China, reversing an earlier prohibition on advanced AI hardware exports. In exchange, the US government will take 25% of the sales value. Trump described the deal during an Oval Office statement: “We’re going to be making 25 percent on the sale of those chips, basically. So we’re allowing them to do it, but the United States is getting 25 percent of the chips in terms of the dollar value. And I think it’s a very good deal.”

The tariffs, enacted under Section 232 national security provisions, exempt chips imported for domestic US use, such as building AI infrastructure or data centers. They form part of Trump's broader trade strategy, which includes threats of up to 100% tariffs on semiconductors unless companies invest in US manufacturing. Nvidia has pledged $500 billion over four years for domestic production, while TSMC is investing $165 billion in Arizona facilities, where it began producing Nvidia's Blackwell chips in October 2025.

Nvidia welcomed the move, stating it “strikes a thoughtful balance that is great for America.” AMD affirmed compliance with US export laws. However, China's response remains uncertain; officials recently instructed logistics firms not to process H200 import requests, amid Beijing's push for domestic chip self-sufficiency. Tech firms like Alibaba, ByteDance, and Tencent favor the H200 for its performance.

The administration warned of potential broader tariffs on semiconductors following a ongoing national security probe, signaling continued tensions in global tech supply chains.

相关文章

U.S. officials and Nvidia executives shake hands over H200 chip at press conference approving exports to China, with flags and reporters.
AI 生成的图像

美国批准英伟达H200芯片对华出口

由 AI 报道 AI 生成的图像

美国商务部周二宣布,允许英伟达向中国客户恢复H200芯片的出货,此举是特朗普政府放松对华技术出口管制的最新举措。该芯片是英伟达第二先进的AI处理器,此前因担忧中国科技和军事实力而受限。

China has approved the import of high-end Nvidia AI chips following weeks of uncertainty. Over 400,000 H200 chips are set to reach major tech companies. This move reflects China's effort to meet technological demands while pursuing self-reliance.

由 AI 报道

Beijing has reportedly greenlit the sale of hundreds of thousands of Nvidia's H200 AI chips to Chinese companies, marking a key win after prolonged US export restrictions. This decision reflects a notable evolution in American technology policy toward China. Meanwhile, Nvidia CEO Jensen Huang has been enjoying casual outings in the country.

Three Supermicro employees face charges of conspiracy to smuggle restricted Nvidia H100, H200 and B200 chips to China. The alleged $2.5 billion scheme used dummy boxes, fake labels and a pass-through company. TechRadar describes it as the biggest heist of the US-China chip war.

由 AI 报道

在美国推动制造业回流之际,台湾积体电路制造公司(TSMC)在亚利桑那州凤凰城北部沙漠建造大型芯片工厂时,遭遇了意想不到的困难。不同于台湾的单一许可流程,美国复杂的监管体系要求公司从零建立约1.8万条规则,耗资3500万美元。公司主席魏哲家表示,这已不仅仅是商业投资,更像是一场昂贵的文化移植实验。

The US Supreme Court annulled most tariffs imposed by Donald Trump under the International Emergency Economic Powers Act (IEEPA) on Friday, in a 6-3 decision limiting its use for trade duties. Hours later, Trump signed an executive order for a 10% global tariff under Section 122 of the Trade Act of 1974, exempting T-MEC products. The measure takes effect on February 24.

由 AI 报道

Nvidia has announced plans to deploy up to 260,000 of its latest Blackwell GPUs in South Korea, partnering with the government and companies like Samsung Electronics, SK Group, Hyundai Motor Group, and Naver Cloud to build AI factories. The initiative will boost the country's AI computing capacity and offer significant opportunities for Samsung and SK hynix in supplying high-bandwidth memory chips. Nvidia CEO Jensen Huang praised South Korea's world-class memory technology during the APEC CEO Summit in Gyeongju.

 

 

 

此网站使用 cookie

我们使用 cookie 进行分析以改进我们的网站。阅读我们的 隐私政策 以获取更多信息。
拒绝