The U.S. Attorney’s Office for the Southern District of New York has unsealed charges against associates of a U.S. server manufacturer, alleging a scheme to illegally divert billions of dollars worth of Nvidia-powered servers to China. This development underscores the ongoing geopolitical and technological battle for dominance in the artificial intelligence sector, as American companies grapple with preventing their advanced hardware from falling into the hands of Chinese competitors.
The indictment, unsealed on Thursday, names Yih-Shyan “Wally” Liaw, Ruei-Tsan “Steven” Chang, and Ting-Wei “Willy” Sun as defendants accused of violating the Export Control Reform Act. According to the U.S. government, these individuals conspired to circumvent stringent U.S. export controls that prohibit the sale of high-performance servers, equipped with Nvidia’s cutting-edge GPUs, to China without a specific license. These controls are explicitly designed to safeguard U.S. national security and foreign policy interests in an era of rapid technological advancement.
The core of the accusation centers on the alleged actions of Liaw, identified as a co-founder and board member of Super Micro Computer (SMCI), a prominent server manufacturer. Liaw reportedly controls a significant stake in Super Micro. Chang is described as a sales manager based in Taiwan, and Sun is identified as a contractor. While Super Micro is not named as a defendant, the company has taken swift action. A statement released by Super Micro confirms that Liaw holds the position of senior vice president of business development, Chang is a sales manager in Taiwan, and Sun is a contractor. The company has placed the two employees on administrative leave and terminated its engagement with the contractor, emphasizing that the alleged conduct is contrary to its corporate policies and its commitment to adhering to export control regulations. The fallout from the indictment was immediately apparent, with Super Micro’s shares dropping 8% in extended trading.
The alleged modus operandi involved a Southeast Asian intermediary acting as a shell company. This entity purportedly generated fraudulent documentation to create the illusion that the servers were destined for legitimate use within its jurisdiction. A separate logistics firm was allegedly employed to repackage the servers, effectively disguising their true destination before they were shipped to China. The indictment further details a sophisticated attempt to deceive the server maker’s compliance department. The defendants are accused of presenting “dummy” servers at the intermediary’s storage facilities to mislead the compliance team, while the actual high-performance servers had already been rerouted to China. This alleged deception was reportedly coupled with pressure tactics to secure approval for these illicit shipments.
The scale of this alleged operation is substantial. The indictment claims that since 2024, these illicit activities have generated approximately $2.5 billion in sales for the server maker. Specifically, between late April 2025 and mid-May 2025, an estimated $510 million worth of servers were diverted through the Southeast Asian company to China, without the requisite U.S. Commerce Department licenses for exporting servers equipped with Nvidia GPUs.
Nvidia’s graphics processing units (GPUs) have become the indispensable engine for training advanced generative AI models, driving unprecedented demand globally. The U.S. government has been actively working to restrict China’s access to these powerful chips, recognizing their potential strategic implications. While former President Donald Trump initially sought to curtail China’s acquisition of these processors, recent statements indicate a potential shift in policy. President Trump has reportedly communicated to China’s President Xi Jinping that the U.S. might permit Nvidia to supply its H200 GPUs to China under specific conditions designed to maintain strong national security. This is further corroborated by recent comments from Nvidia CEO Jensen Huang, who indicated that the company is resuming manufacturing to fulfill H200 purchase orders from China. Previously, Nvidia had received licenses to export its H20 chip to China, with CEO Jensen Huang agreeing to a stipulation that 15% of its China sales would be shared with the U.S. This complex interplay between technological innovation, corporate compliance, and national security concerns continues to shape the global AI landscape.
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/19934.html
