TradingKey - The investigation into the flow of Nvidia (NVDA) AI chips to China is further spreading to its server partner Super Micro Computer (SMCI).
According to a November 17 report by The Information, Nvidia's former head of worldwide sales Jay Puri was one of the primary contacts for key company clients, including former Megaspeed CEO Alice Huang. Companies associated with the latter were alleged to have shipped at least $7.5 billion worth of export-restricted Nvidia AI chips to China through covert channels.
The report also noted that Puri has personal ties to server maker Super Micro Computer, which previously had multiple affiliated individuals criminally indicted for allegedly assisting in transferring servers equipped with Nvidia chips to China.
Meanwhile, short seller Muddy Waters also released a report questioning possible ties between Puri and former Super Micro Computer executive Cenly Chen, and disclosed that it holds a short position in Nvidia. However, these links have not been independently verified, and the relevant civil litigation allegations have not been judicially confirmed.
Super Micro Computer has already faced export compliance investigations, personnel indictments, and shareholder class-action lawsuits, and now new media reports have once again thrust its supply chain relationships into the spotlight.
Supermicro is a key partner in Nvidia's AI server ecosystem, responsible for integrating GPUs, networking equipment, storage, and liquid cooling systems into complete server racks, before selling them to cloud computing companies, data center operators, and large enterprises. As demand for AI infrastructure expands, GPU servers and rack-scale systems have become the primary drivers of the company's revenue growth.
This business model also means that Supermicro occupies a critical position in the U.S. export control enforcement chain. Nvidia typically sells chips to server manufacturers and system integrators, who are then responsible for vetting end customers, equipment applications, and delivery locations. Once servers are transshipped to China via third countries, regulators may trace the entire process from chip allocation and customer screening to logistical delivery.
An indictment released by the U.S. Department of Justice in March this year alleged that the individuals involved arranged approximately $2.5 billion in server transactions between 2024 and 2025, using falsified documents, substitute equipment, and complex transshipment routes to conceal the final destinations. The Justice Department charged three individuals associated with Supermicro, rather than the company itself.
Supermicro later launched an internal investigation overseen by independent directors and stated that it would continue to enhance its export compliance procedures. The company also cooperated with relevant authorities in Taiwan in an effort to strengthen supply chain tracking and prevent restricted servers from being illicitly transshipped.
Recent reports have broadened the focus of the investigation from isolated sales activities to the customer management relationship between Nvidia and Supermicro. If regulators determine that insiders at both companies had indications of irregular transactions but failed to take timely action, the scope of scrutiny could expand to GPU serial number tracking, sales approvals, end-user verification, and related-party disclosures.
The most critical potential risk for Super Micro Computer is that Nvidia may tighten GPU allocations and customer vetting procedures. The Information reported that Puri previously held the authority to allocate AI chips to server vendors such as Super Micro Computer. As his relationships with customers come under investigation, Nvidia's new management may re-examine existing sales processes and implement stricter end-user verification for orders from high-risk regions.
Nvidia Interim Chief Executive Officer Colette Kress has stated that the company will no longer deliver licensed Hopper data center chips to China. Meanwhile, the U.S. government has tightened restrictions on overseas data centers remotely providing advanced computing power to Chinese customers, signaling that regulatory focus is expanding from direct exports to third-country transshipments and offshore computing access.
This could increase order vetting times for Super Micro Computer and delay some server deliveries. Large AI racks typically require advance coordination of GPUs, networking components, memory, and liquid-cooling equipment. If Nvidia requires additional proof of customer identity, funding sources, or final deployment locations, Super Micro Computer's cycle from obtaining chips to recognizing revenue could be lengthened.
Stricter vetting could also affect the company's ability to win overseas customers. Even if orders themselves are entirely legitimate, customers from Southeast Asia, the Middle East, and other regions with higher transshipment risks may face more complex licensing and review procedures going forward. As a result, Super Micro Computer may forgo some high-risk transactions or bear higher legal, audit, and supply chain tracking costs.
If the scope of the investigation widens, the company may need to increase spending on legal counsel, audits, export licensing, and customer verification. Second is operational risk: a tighter GPU allocation system could lengthen delivery cycles and affect the timing of converting orders into revenue. Finally, there is a governance discount: if the market perceives that the company provides inadequate disclosures regarding related-party relationships, customer backgrounds, or end-user vetting, investors may demand a higher risk premium.
Super Micro Computer currently still retains distinct business advantages, including rapidly integrating Nvidia's next-generation GPUs, offering complete rack-scale solutions, and its early move into liquid-cooling systems. However, the company's gross margin itself is not high, standing at 10.8% for fiscal 2026, down from 11.1% in the prior fiscal year. Even if additional compliance costs do not significantly affect revenue, they could further squeeze profit margins.