Semiconductor giant Nvidia is encountering new U.S. export restrictions on its H20 chips. In a filing released on Tuesday, Nvidia disclosed that the U.S. government informed the company that a license would be required to export its H20 AI chips to China indefinitely. The U.S. government justified the decision by citing the potential risk that the H20 chips might be utilized in a supercomputer in China.
Nvidia is projecting related charges amounting to $5.5 billion in its Q1 2026 fiscal year, ending on April 27. Following this announcement, Nvidia’s stock experienced a decline of approximately 6% in extended trading sessions.
The H20 chip represents the most advanced AI technology that Nvidia is permitted to export to China under both existing and previous U.S. export regulations. Last week, NPR reported that CEO Jensen Huang might have successfully negotiated to avoid new restrictions on the H20 chips during a meeting at President Donald Trump’s Mar-a-Lago estate. This negotiation reportedly involved a commitment to invest in AI data centers within the United States.
Coincidentally, Nvidia announced on Monday its plan to invest several hundred million dollars over the next four years to manufacture some AI chips domestically. Industry commentators noted that the company’s announcement lacked comprehensive details.
There have been calls from various government officials advocating for tighter export controls on the H20 chips due to their alleged use in training models for DeepSeek, a China-based AI startup. This includes the R1 “reasoning” model, which significantly impacted the U.S. AI market in January.
Nvidia did not provide any comments on the matter.