Federal prosecutors in Manhattan unsealed an indictment on March 20, 2026, charging three individuals associated with Super Micro Computer with conspiring to unlawfully divert more than $510 million worth of Nvidia-powered AI servers toward China in alleged violation of U.S. export controls. The case names a company co-founder, a regional sales manager, and a contractor, and it represents one of the largest alleged technology-diversion schemes since Washington tightened restrictions on advanced chip exports. The charges come as the U.S. government escalates enforcement against alleged networks that route restricted AI hardware to Chinese end users through front companies and third-country intermediaries.
Who Was Charged and What They Allegedly Did
The indictment, filed in the Southern District of New York, names Yih-Shyan “Wally” Liaw, Ruei-Tsang “Steven” Chang, and Ting-Wei “Willy” Sun. Liaw, 71, a Fremont, California, resident, served as senior vice president of business development and sat on Super Micro Computer’s board. Chang held the role of Taiwan sales manager at the company. Sun, a 44-year-old Taiwan citizen, worked as a contractor associated with the firm. All three face counts of conspiracy charges including violating the Export Control Reform Act, smuggling goods from the United States, and defrauding the U.S. government.
Liaw and Sun were arrested after the indictment was unsealed, according to published reports. Liaw was later released on bail, while Sun was held pending a bail hearing, according to Associated Press accounts. Chang, the Taiwan-based sales manager, had not been arrested at the time reporting was published. Prosecutors say the trio coordinated to disguise the ultimate destination of high-end servers built with Nvidia accelerators that fall under Commerce Department licensing requirements for shipments to China, allegedly relying on false documentation and misrepresented end users to move hardware that otherwise would have required government approval.
The $510 Million Diversion Allegation
The scale of the alleged operation sets it apart from prior export-control cases. Federal authorities described a $510 million network that prosecutors say funneled restricted AI technology, including systems referencing Nvidia H100 and H200 chips, toward Chinese customers. The Justice Department framed the enforcement action as the shutdown of a major China-linked AI tech smuggling network, signaling that the government views alleged diversion of restricted AI technology as a national security threat rather than a routine trade violation.
What makes this case distinctive is the alleged involvement of senior insiders at a publicly traded U.S. server manufacturer. Most export-control prosecutions target external brokers, freight forwarders, or shell companies operating at the margins of the supply chain. Here, prosecutors allege that a company co-founder with board-level access played a central role in the scheme. That distinction matters because it suggests that tightening rules at the border or screening third-party resellers may not be enough when the diversion originates within a firm that legitimately purchases and assembles restricted components.
According to descriptions in the charging documents, the defendants allegedly helped steer advanced servers toward Chinese entities by routing orders through intermediaries in other jurisdictions and by masking the presence of Nvidia accelerators subject to licensing rules. Authorities say the conduct spanned multiple years and involved repeated efforts to evade compliance systems, both inside the company and at customs checkpoints, in order to keep high-performance AI hardware flowing into a market where direct exports are tightly constrained.
Super Micro’s Response and Corporate Exposure
Super Micro Computer moved quickly to distance itself from the defendants. In a public statement, the company confirmed it had been informed that SDNY unsealed an indictment of three individuals associated with the company. It identified Liaw by his senior vice president and board titles, Chang as Taiwan sales manager, and Sun as a contractor. The company stressed that it is not named as a defendant in the indictment and said it is cooperating fully with authorities. Liaw and Chang were placed on administrative leave while the investigation proceeds.
Liaw also resigned from the company’s board, a step reported by Bloomberg and confirmed in coverage by the Wall Street Journal. The exit of a co-founder from a company he helped build underscores the severity of the legal exposure, even if Supermicro itself currently avoids criminal liability. Corporate governance experts note that boards in similar situations face pressure to demonstrate independence from any individuals under indictment, both to reassure regulators and to stabilize investor confidence.
The case also raises questions about the robustness of internal compliance systems at firms that work with restricted chips. Supermicro has emphasized its cooperation and the fact that prosecutors have not charged the company. But enforcement agencies frequently revisit corporate responsibility as investigations unfold, especially if evidence suggests that red flags were ignored or that compliance staff lacked authority to halt suspect deals. Competitors in the server and data-center market are likely to review their own export-control programs in light of the allegations.
Export Controls That Triggered the Case
The legal foundation for the charges rests on the Export Administration Regulations, which govern the sale and transfer of dual-use technologies. Advanced AI accelerators, including Nvidia’s H100 and H200 series, are classified under the Commerce Control List and require export licenses when destined for China and certain other countries. U.S. export rules governing advanced computing items have tightened in recent years, expanding licensing requirements for certain high-performance chips and related systems when destined for China and other restricted destinations.
The indictment alleges that the defendants circumvented these requirements by routing servers through intermediary destinations and using entities that concealed the true Chinese end users. This kind of transshipment fraud has become a primary enforcement challenge for the Commerce Department’s Bureau of Industry and Security and the FBI, which jointly investigate export-control violations. The Federal Register’s rule updates have steadily expanded the scope of covered semiconductors and computing systems, but enforcement ultimately depends on detecting deceptive paperwork, misdeclared shipments, and complex routing patterns that can obscure where sensitive hardware is actually headed.
Prosecutors in this case argue that insider knowledge of how export-screening processes work gave the defendants an advantage in structuring deals that appeared compliant on their face. By allegedly pairing that insight with overseas partners willing to act as nominal purchasers, they are accused of undermining a licensing regime designed to slow China’s access to the most capable AI training platforms. If the government proves its claims at trial, the case could become a template for future prosecutions involving employees at major technology suppliers rather than only outside intermediaries.
Broader Implications for AI Supply Chains
The indictment arrives as Washington and Beijing compete over who will control the most advanced AI infrastructure. U.S. officials have argued that restricting access to cutting-edge accelerators is necessary to prevent Chinese military and intelligence agencies from harnessing American-designed chips for surveillance, cyber operations, and weapons development. The alleged diversion of hundreds of millions of dollars in servers, if confirmed, would illustrate how lucrative the gray market for such hardware has become.
For global technology companies, the case highlights the tension between commercial demand and regulatory risk. China remains a major market for data-center equipment, but the rules governing what can be sold there are tightening and complex. Firms that assemble or integrate AI servers must verify not only where a shipment is going on paper, but also who will ultimately control and use the systems. That often requires deeper due diligence on customers and closer monitoring of channel partners to reduce the risk of diversion.
Regulators, meanwhile, are likely to treat the Super Micro case as evidence that compliance programs need to extend beyond paperwork checks and automated screening tools. Training for senior executives, clear escalation paths for employee concerns, and independent auditing of high-risk deals are all areas that may draw renewed scrutiny. If prosecutors secure convictions, they may also seek to use the outcome as leverage in negotiations with other firms, pressing them to adopt stricter internal controls in exchange for more predictable enforcement.
However the litigation unfolds, the indictment underscores that export controls on AI hardware are no longer an abstract policy debate but a frontline enforcement issue with personal consequences for executives and technologists. As the U.S. government continues to refine its rules and pursue high-profile cases, companies embedded in the AI supply chain will be pushed to choose between aggressive expansion into sensitive markets and the heightened legal and reputational risks that now accompany every shipment of advanced computing power.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
