Nvidia, the Silicon Valley giant driving the AI boom, disclosed in a late-night regulatory filing on Tuesday that it will incur a $5.5 billion charge in the quarter ending April 27 due to new U.S. export controls targeting its H20 AI chips for China.
The U.S. Commerce Department confirmed the curbs, requiring a special license for Nvidia to sell its H20 chips—designed to comply with Biden-era restrictions on Chinese customers, marking a sharp escalation in Donald Trump’s trade war with Beijing.
Analysts estimate Nvidia’s sales to China could reach $17 billion this financial year, with the H20 chip alone accounting for about $12 billion of that revenue over the past year, though the licensing uncertainty threatens a potential “wipeout” of this product line.
Nvidia’s shares slid 7% in pre-market trading on Wednesday, while futures tracking the tech-heavy Nasdaq 100 index dropped over 2%, signaling broader investor concerns about the U.S.-China tech war’s impact on the semiconductor industry.
The fallout extended globally, with Dutch chip equipment maker ASML sinking 6% after disappointing orders, and stocks in Hong Kong declining. Alibaba fell nearly 4%, while Baidu and Tencent each dropped around 2%.
AMD, Nvidia’s closest rival in the AI data center chip market, saw its shares fall 7% in after-hours trading as the new controls also target its MI308 chip, while Broadcom and Intel declined 3% and 2%, respectively.
The H20 chip, less powerful than Nvidia’s top-tier GPUs used by Microsoft, OpenAI, Meta, and Amazon, was in high demand among Chinese tech giants like Alibaba, Baidu, and Tencent, despite China’s push for homegrown alternatives from Huawei.
The U.S. justified the crackdown by citing national security risks, with Nvidia noting concerns about H20 chips powering supercomputers in China for military applications like hypersonic weapons and nuclear modeling.
China accused the U.S. of weaponizing export controls to suppress its economic growth, while White House Press Secretary Karoline Leavitt urged Beijing to negotiate a new trade deal, stating, “the ball is in China’s court.”
Nvidia was informed of the new controls on April 9 and learned on Monday that the license requirement for H20 and similar chips would remain in effect indefinitely, compounding the company’s exposure to U.S.-China geopolitical tensions.
The Trump administration, which has raised tariffs on Chinese imports to 145% while sparing some consumer electronics temporarily, also launched a national security probe on Monday that could lead to new semiconductor tariffs.
In response, Nvidia announced plans on Monday to invest up to $500 billion in U.S. AI infrastructure over the next four years, partnering with Taiwan Semiconductor Manufacturing Company and Foxconn to bolster domestic capacity.
The Commerce Department’s latest measures build on Biden-era restrictions that barred Nvidia’s most advanced chips from China, prompting the company to introduce the H20 last year as a compliant alternative for the Chinese market.
Despite its reduced performance, the H20 faced potential exclusion from China due to Beijing’s energy-efficiency rules and tariffs on imports, even as local chip suppliers struggled to rival Nvidia’s technology.
An “AI diffusion” rule, set to take effect in May unless reversed by Trump, could impose even stricter export limits on powerful U.S. chips, using a tiered licensing system that favors only a select few countries.
The broader tech sector felt the strain, with Wall Street rattled by the global tech market’s vulnerability to U.S.-China trade policies, highlighting Nvidia’s pivotal yet precarious role in the AI and semiconductor industries.
The U.S.’s intensified export controls on AI chips underscore a critical moment in the U.S.-China tech rivalry, with Nvidia’s $5.5 billion hit signaling profound implications for innovation, supply chains, and the future of global AI development.
