U.S. Shuts Down $160M Nvidia GPU Smuggling Ring Tied to China as Crackdown Intensifies
U.S. authorities announced Tuesday that they have dismantled another smuggling operation linked to China, one that illegally trafficked or attempted to funnel more than $160 million worth of export-controlled Nvidia AI chips overseas. Two individuals have been arrested, while a Houston firm and its owner have already pleaded guilty as part of the widening federal probe.
The case marks the latest escalation in Washington’s push to prevent China from obtaining advanced artificial intelligence hardware, including Nvidia’s high-performance GPUs. The crackdown, known as Operation Gatekeeper, uncovered a network working to redirect cutting-edge chips—often used in both civilian and military applications—to restricted destinations that could compromise U.S. national security, prosecutors said.
Texas Businessman Pleads Guilty in $160M Nvidia Chip Scheme
According to newly unsealed court filings, Alan Hao Hsu, 43, of Missouri City, Texas, and his company, Hao Global, pleaded guilty on October 10 to smuggling and unlawful export activity. Prosecutors say Hsu and his associates illegally exported or attempted to export at least $160 million worth of Nvidia H100 and H200 GPUs between October 2024 and May 2025.
Although the H200 and H100 are no longer Nvidia’s most advanced offerings, they are still restricted under U.S. export rules and require special licensing to ship to China. Hsu’s operation allegedly falsified shipping data, concealed the GPUs’ destinations, and routed hardware through China, Hong Kong, and other prohibited regions. Investigators traced more than $50 million in financing from China tied to the scheme.
Hsu remains free on bond and faces up to 10 years in prison when sentenced on February 18. His company could also face significant penalties, including fines up to twice the value of its illegal gains.
In a statement to CNBC, Nvidia emphasized that U.S. export controls are stringent, noting that “even sales of older generation products on the secondary market are subject to strict scrutiny.” The company said it continues working with regulators to curb the growing trade in illicit GPUs.
Additional Defendants Accused of Rebranding and Relabeling Nvidia Chips
Federal prosecutors also charged Fanyue Gong, a 43-year-old Chinese citizen living in New York, and Benlin Yuan, 58, a Canadian resident of Ontario. Both allegedly worked independently with Hong Kong logistics partners and a China-based AI company to evade U.S. chip restrictions.
According to the complaint, Gong used intermediaries and straw buyers to obtain controlled Nvidia GPUs by claiming they were destined for U.S. customers or for unrestricted countries. The hardware was then rebranded at U.S. warehouses under fake labels and exported as generic components to China and Hong Kong.
Yuan, who serves as CEO of the U.S. arm of a Beijing IT firm, is accused of recruiting inspectors to help conceal the true destinations, crafting cover stories for detained shipments, and managing storage for additional GPU exports.
If convicted, Yuan faces up to 20 years in prison for conspiracy to violate the Export Control Reform Act, while Gong could receive as much as 10 years for conspiracy to smuggle.
Growing Pressure to Close Export Loopholes
The investigation involved the Commerce Department’s Bureau of Industry and Security (BIS), which oversees export controls for advanced technologies. The bust is part of a recent wave of enforcement actions related to unauthorized Nvidia chip shipments, reflecting increasing pressure from lawmakers to close loopholes in existing regulations.
The case also arrives amid shifting U.S. policy. President Donald Trump signaled this week that he would allow Nvidia to sell its H200 model to “approved customers” in China and other markets—provided the U.S. government receives a 25% share of the sales.
While the H200 is not Nvidia’s flagship chip, it would become the most advanced GPU available to China, potentially easing the country’s ongoing shortage of AI computing power.
