Super Micro Co-Founder Charged: $2.5B AI Chip Smuggling
Three Super Micro insiders charged for diverting $2.5B in Nvidia AI chips to China. The DOJ's highest-profile AI chip export enforcement action yet.
Super Micro co-founder charged with AI chip smuggling: that is the headline the U.S. Justice Department produced Thursday, unsealing an indictment against three individuals who allegedly diverted at least $2.5 billion worth of Nvidia-powered AI hardware to China in direct violation of AI chip export controls in place since 2022. The case is the government’s most significant enforcement action against illegal AI compute flows to date, with immediate implications for how the AI infrastructure supply chain is audited and policed.
The charges represent the most significant enforcement action the U.S. government has taken to date against the illegal flow of advanced AI compute to restricted markets, and come as Washington continues to tighten its grip on the global supply of high-performance AI chips.
The Charges: A Two-Year Smuggling Operation
Prosecutors named Yih-Shyan Liaw (known as Wally), Ruei-Tsang Chang (known as Steven), and Ting-Wei Sun (known as Willy) in a federal indictment unsealed in Manhattan. Liaw co-founded Super Micro in 1993 and has served on its board of directors since 2023. Chang was a sales manager in the company’s Taiwan office. Sun was a contractor.
The alleged scheme was deliberately layered to evade detection. According to DOJ filings, the defendants shipped U.S.-made AI servers through Taiwan to Southeast Asian intermediary countries, where the servers were transferred into unmarked boxes before being forwarded to China. In a detail that illustrates the operation’s operational security, U.S. officials allege the group used hair dryers to remove serial number labels from real hardware and attached them to dummy machines left behind as decoys.
Super Micro was not named as a defendant and said it had cooperated with federal investigators. The company placed Liaw and Chang on administrative leave and severed ties with contractor Sun. Super Micro’s shares fell 8% in after-hours trading on the news.
Nvidia, which dominates the AI chip market and whose hardware is presumed to be at the center of the scheme, issued a statement emphasizing strict compliance obligations: “Unlawful diversion of controlled U.S. computers to China is a losing proposition across the board. NVIDIA does not provide any service or support for such systems.”
Why This Case Marks a Turning Point in AI Export Enforcement
The U.S. has restricted advanced AI chip exports to China since October 2022, with successive rounds of controls expanding the scope of restricted hardware. The infrastructure race driving demand for those chips continues at scale, as covered in our analysis of Colossus expansion plans. Despite those controls, intelligence assessments and investigative reporting have consistently pointed to significant volumes of controlled compute reaching China through third-party intermediaries.
What distinguishes this case is scale and the identity of the defendants. Charging a co-founder of a major U.S. AI infrastructure company signals that the government has moved beyond customs-level interdiction and is now pursuing enforcement inside the supply chain itself.
The $2.5 billion figure, if accurate, represents a meaningful fraction of the total AI server market. Analysts who cover AI infrastructure have noted that a single high-density GPU cluster can cost tens of millions of dollars, which means the alleged scheme involved hundreds of discrete server systems moving through a clandestine logistics chain over an extended period.
Enforcement actions of this type also have downstream effects on the legal supply chain. Companies that sell to Super Micro or adjacent server integrators are now likely to face heightened compliance requirements and third-party audit demands as the DOJ signals it will hold individuals, not just entities, accountable.
Google Releases Gemini 3.1 Pro as AI Model Competition Accelerates
Separately, Google this week released Gemini 3.1 Pro, an upgraded reasoning model that it says shows meaningful improvements on complex problem-solving benchmarks compared to its predecessor. The model is rolling out across the Gemini API in Google AI Studio, Vertex AI, the Gemini app, Android Studio, and the company’s agentic development platform Google Antigravity.
Google characterized 3.1 Pro as the “core intelligence” underlying its broader model family, including Gemini 3 Deep Think, which the company updated last week to target scientific and engineering reasoning tasks. The release of 3.1 Pro brings that reasoning capacity into everyday developer and enterprise workflows.
The timing is notable. Google is accelerating its model release cadence into Q2 2026 as OpenAI, Anthropic, and emerging competitors including xAI and Mistral all push new versions through the same quarter. Developer feedback since the 3.1 Pro preview opened has focused on improved instruction-following on multi-step agentic tasks, an area where the previous generation showed inconsistencies.
Perplexity Launches Health Feature, Joins AI Platform Race for Medical Data
Also announced this week, Perplexity launched Perplexity Health, a set of data connectors that allow the AI search platform to ingest and query personal health records. At launch, supported sources include Apple Health, electronic health records from more than 1.7 million care providers, and wearable platforms including Fitbit, Ultrahuman, and Withings.
The product positions Perplexity as the third major AI platform to integrate with Apple Health, following ChatGPT Health in January 2026 and Microsoft Copilot Health the week prior. The convergence of AI platforms on health data access raises both product and regulatory questions: what users gain in personalized health querying they trade against an expanding set of third-party data custodians handling sensitive medical records.
The HIPAA implications of AI platforms that aggregate health data remain an active area of legal interpretation. No regulatory guidance specific to AI health assistants has been issued by HHS as of this writing.
The Super Micro indictment in particular is expected to accelerate compliance audits across the AI server supply chain and may prompt Congressional inquiry into the adequacy of existing export control enforcement mechanisms. A court date has not yet been announced.