US and Taiwanese authorities separately indicted 11 people in two cases for illegally reselling US-manufactured semiconductor chips to China, according to the Institute for the Study of War's China-Taiwan update published October 9, 2026. The cases underscore how aggressively both governments are now enforcing export controls aimed at denying Beijing the advanced chips it seeks for military modernization and AI development.
The two cases
On October 6, Taiwanese prosecutors charged 10 individuals with reselling export-controlled semiconductor chips to China from 2017 through at least April 2026. The defendants allegedly forged end-user documents and falsely claimed to represent Taiwan's National Chung-shan Institute of Science and Technology (NCSIST) — the island's premier military research body — to purchase the chips. Taiwanese authorities stated that China's People's Liberation Army sought the advanced chips for use in missiles and radar systems.
Separately, US authorities arrested Greg Lui, a California resident, on October 1 for reselling servers containing roughly $300 million worth of export-controlled Nvidia chips designed to support large AI models. Lui allegedly first exported the servers to Singapore and Malaysia to circumvent US export restrictions before illegally selling them onward to China.
Background: the export-control regime and its weak points
US semiconductor export controls on China have tightened steadily since 2022, restricting sales of advanced AI chips, chipmaking equipment, and related technology. The controls rest on two pillars: licensing requirements for direct sales, and end-user/end-use certifications that are supposed to prevent diversion.
The smuggling cases reveal where the system leaks. Prosecutors have repeatedly found that traffickers falsify end-user documents and route exports through Southeast Asian countries — Singapore, Malaysia, Thailand, Vietnam — to conceal the ultimate destination. This "transshipment" pattern has become the signature enforcement problem of the chip war, and both cases announced this week follow that exact playbook.
Enforcement has also escalated in parallel with the controls themselves. What began as a Commerce Department licensing regime has become a criminal enforcement priority, with the Justice Department treating chip diversion as a national security offense — carrying far heavier penalties and investigative resources than trade violations typically attract.
Why it matters for founders and investors
For AI infrastructure startups, this is a compliance wake-up call. Companies buying, leasing, or reselling GPU capacity — especially through brokers in Southeast Asia or the Middle East — face genuine diligence obligations. Know-your-customer procedures that were optional in 2023 are now load-bearing: a startup that cannot document the provenance and end use of its compute supply is exposed.
For investors, the cases add a new risk factor to AI infrastructure deals. Data-center and neocloud investments should be underwritten with supply-chain diligence — where did the GPUs come from, who owned them before, and does the paperwork survive scrutiny? The $300 million figure in the Lui case shows the scale at which illicit diversion operates; legitimate buyers need to be certain they are not downstream of it.
There is also a macro signal: sustained enforcement keeps advanced compute artificially scarce in China, which shapes the competitive landscape for AI model development globally. Every smuggling indictment is, indirectly, a data point on how binding the export controls actually are.
What to watch
Watch for sentencing outcomes and whether prosecutors name additional co-conspirators — these cases often expand. Watch also for tighter end-use verification requirements from the Commerce Department, which would raise compliance costs for legitimate cross-border AI infrastructure deals. And founders should note the Taiwan angle: Taipei is prosecuting its own citizens for chip diversion, which means the enforcement net is genuinely bilateral, not just American.
Source: Institute for the Study of War (https://understandingwar.org/research/china-taiwan/china-taiwan-update-october-9-2026/)
Read Next
OpenAI Releases Codex Security as Open-Source CLI and SDK
Databricks Unveils Always-On Genie, an AI Agent That Stops Waiting to Be Asked
DeepMind Launches AlphaGenome 2, a DNA Model That Reads Genomes Cell by Cell
ElevenLabs Unveils 13 New Products and Upgrades at Its New York Summit
Google Launches Gemini Decision 1, a Model That Answers Instead of Writing
Runway Open-Sources Praxis-1, Its Robot Model That Learned by Watching Video



