Nvidia announced on Tuesday that it will incur $5.5 billion in charges due to US government restrictions on exporting its H20 artificial intelligence chip to China, a key market for one of its most popular products.
Nvidia will incur $5.5 billion in charges due to new U.S. export controls requiring licenses for its H20 AI chips destined for China and other countries. The move is part of efforts to curb the sale of advanced AI chips to preserve U.S. national security
Nvidia announced on Tuesday that it will incur $5.5 billion in charges due to US government restrictions on exporting its H20 artificial intelligence chip to China, a key market for one of its most popular products.
Nvidia’s AI chips have been a primary target of US export controls, as officials aim to block the sale of advanced chips to China to preserve the US lead in the AI race. In response to these regulations, Nvidia developed processors designed to comply with US restrictions while maximising performance within those limits.
Nvidia disclosed in a filing that the US government notified the company it would require a licence to export its H20 chips to China and several other countries. A US Commerce Department spokesperson announced late Tuesday that new licensing requirements would be implemented for exports of chips, including Nvidia’s H20, AMD’s MI308 and their equivalents.
"The Commerce Department is committed to acting on the President's directive to safeguard our national and economic security," said the spokesperson of the department that oversees US export controls.
The latest export restrictions represent a clear indication that Nvidia’s robust growth trajectory may be constrained by escalating geopolitical limitations. The United States has stated that these high‑performance CPUs could power military‑grade supercomputers. Nvidia is scheduled to announce its first‑quarter financial results on May 28.
The H20 chip was developed specifically for the Chinese market to comply with US export restrictions introduced in 2022 and revised in 2023. Despite its constraints, the chip is estimated to have generated between $12 billion and $15 billion in revenue last year.
Although the H20 chip is less efficient at training AI models compared to Nvidia’s processors sold outside China, it performs comparably during the inference stage, where AI models deliver responses to users. Inference is rapidly becoming the largest segment of the AI chip market.
Last month, Nvidia CEO Jensen Huang stated that the company is well‑positioned to lead this shift. However, on Tuesday, Nvidia announced that the US government has banned H20 sales to China, citing concerns that the chips could be used in supercomputers.