Semiconductor giant Nvidia expects to take a charge of up to $5.5 billion in its current fiscal quarter after the U.S. government imposed new export restrictions on its H20 artificial intelligence chips, deepening a growing technology rift between Washington and Beijing.
In a regulatory filing on Tuesday, the company said that it was notified by the U.S. government on April 9 that a special export license is now required to ship its H20 integrated circuits to China, including Hong Kong and Macau, and all D:5 countries—a group that includes nations Washington deems as having national security sensitivities. The restriction, Nvidia said, will be in place for the indefinite future.
“First quarter results are expected to include up to approximately $5.5 billion of charges associated with H20 products for inventory, purchase commitments, and related reserves,” the company wrote, referring to its fiscal Q1, which ends April 27, 2025.
Washington tightens the screws on AI Tech
The restriction is part of a broader U.S. effort to prevent advanced semiconductors from aiding China’s military or supercomputing capabilities. The Commerce Department also applied similar licensing requirements to AI chips developed by AMD, including the MI308 processor. The measures follow a string of bipartisan calls in Congress to tighten controls after China’s rapid progress in AI development.
The U.S. government cited the H20’s memory and interconnect bandwidth as posing a risk of use in high-performance computing, including supercomputers that could be applied to military or surveillance functions.
Nvidia designed the H20 chip specifically to comply with earlier U.S. export controls introduced in late 2022, which banned sales of its more powerful H100 processor to China. The H20 features significantly lower compute performance and was marketed as a compliant alternative for Chinese customers.
Despite that effort, U.S. authorities determined the chip still met thresholds that raised concern. Analysts at Jefferies noted that no such export licenses have previously been granted, describing the requirement as an effective ban.
Strategic and Financial Blow
The financial impact reflects unsold inventory, supplier obligations, and reserves against future losses tied to the H20 line. Nvidia shares fell nearly 7% on Wednesday, erasing billions in market capitalization, while its suppliers also traded lower in Asia and Europe. Semiconductor equipment makers such as ASML, Samsung Electronics, TSMC, and Advantest all reported share price declines amid growing concerns over global trade volatility.
Though the $5.5 billion charge is a substantial one-off cost, some analysts view it as strategically more damaging than financially.
“This is a clear sign that Nvidia now has massive restrictions and hurdles in selling to China, as the Trump Administration knows there is one chip and company fueling the AI revolution—and it’s Nvidia,” wrote Dan Ives, Head of Tech Research at Wedbush Securities.
China’s AI boom meets US clampdown
The timing of the move is notable. Earlier this year, Chinese AI firm DeepSeek revealed its large language model R1, reportedly trained using Nvidia H20 chips at a fraction of the cost of U.S.-based training methods. The success of R1 stunned industry observers and triggered a surge in AI investment across China, pressuring U.S. lawmakers to act more aggressively to limit Chinese access to foundational AI infrastructure.
While Chinese firms like Huawei and Cambroon are developing domestic AI chips, analysts at Counterpoint Research note that their products still lag Nvidia’s in performance and software ecosystem maturity. However, there are growing concerns that repeated export curbs may accelerate China’s self-sufficiency and diminish the dominance of U.S. suppliers in the long term.
“Banning H20 performance levels makes little sense,” wrote Bernstein analysts in a note. “It simply hands the Chinese AI market to Huawei.”
Broader economic implications
The move also comes amid rising global uncertainty over trade policy. The World Trade Organization on Wednesday warned that global GDP growth is projected to decline by 0.6 percentage points due to new tariffs and restrictions. North America is expected to see a sharper contraction, with 1.6 percentage points shaved off expected growth.
The Biden administration—continuing Trump-era national security measures on semiconductors—has also signaled the possibility of future tariffs on chips, and investigations into semiconductor imports are ongoing.
At the same time, Nvidia has moved to align itself with U.S. industrial policy, announcing on Monday a $500 billion domestic investment plan over four years to build out AI supercomputers, data centers, and chip infrastructure. The White House praised the move, calling it “the Trump Effect in action.”
Semiconductors at the heart of tech geopolitics
Despite the projected charge, Nvidia remains a dominant player in the global AI chip market. The company’s H100 and upcoming Blackwell series remain central to the development of generative AI worldwide. But the latest regulatory shock underscores the growing risks for companies straddling the geopolitical divide between American innovation leadership and China’s digital ascent.
Ned Finkle, Nvidia’s vice president of government affairs, wrote in a company blog post that the adoption of AI around the world fuels growth and opportunity for industries at home and abroad. But the restrictions put that global progress “in jeopardy” and threatened to “derail innovation and economic growth worldwide,” he said.
As tensions rise, tech firms and investors alike are bracing for more policy-driven disruptions in the AI and semiconductor sectors—where commercial leadership increasingly intersects with national security.