Taiwanese prosecutors have indicted nine people, including employees of Nvidia and Super Micro Computer, over an alleged scheme to illegally export AI servers equipped with restricted Nvidia chips to China. Prosecutors say the defendants knew about export-control requirements and Nvidia and Super Micro's compliance procedures but allegedly falsified documents and coordinated shipments for profit.
The investigation reportedly involves 130 Super Micro B300 servers that were declared for use in Taiwan. Prosecutors allege that 74 were ultimately exported to China, either directly or through intermediaries in countries including Indonesia, Japan and Hong Kong, while Taiwanese authorities intercepted the remaining 56. Eight defendants face charges including breach of trust and document forgery, while another faces related financial-misconduct charges.
The case highlights how difficult it is to enforce U.S.-led AI chip restrictions across a global supply chain. The United States has restricted exports of advanced AI technology to China since 2022, while Taiwan plays a critical role in manufacturing and packaging Nvidia's chips. Authorities have already uncovered other alleged schemes involving Southeast Asian intermediaries, showing that restricted AI hardware can be rerouted through third countries rather than moving directly from the United States or Taiwan to China.
The broader takeaway is that AI export controls are becoming a supply-chain enforcement problem as much as a policy problem. As demand for advanced computing remains enormous in China, companies and intermediaries have strong financial incentives to circumvent restrictions. Taiwan's prosecution therefore demonstrates that Washington's attempt to limit China's access to frontier AI hardware increasingly depends on cooperation from semiconductor hubs, server manufacturers, logistics companies and governments around the world.
