US moves to block Nvidia and AMD AI chip shipments to Chinese firms operating overseas – Firstpost

The United States has moved to close a potential loophole that may have allowed some of the world’s most advanced artificial intelligence chips from Nvidia and AMD to reach Chinese companies through overseas subsidiaries, marking a fresh escalation in Washington’s efforts to curb Beijing’s AI ambitions.

In new guidance issued over the weekend, the US Department of Commerce said it would enforce licensing requirements on advanced AI chip exports to entities headquartered in China, even if those entities are located outside mainland China.

The move effectively targets Chinese companies operating through subsidiaries in countries such as Malaysia and other technology hubs, where they may have been able to procure cutting-edge processors despite broader US export restrictions.
The guidance affects high-performance AI chips, including Nvidia’s latest Blackwell and Rubin processors as well as AMD’s MI350X accelerators, which are considered critical for training and deploying advanced AI models.

The development highlights concerns within Washington that Chinese firms may have been accessing restricted technology through overseas operations since May 2025, when the Trump administration decided not to enforce the AI Diffusion Rule introduced during the final days of the Biden administration.

Industry estimates suggest the scale of exports may have been significant. One semiconductor supply-chain expert told Reuters that “hundreds of thousands” of advanced chips could have reached Chinese-linked entities during the period.

Technology policy experts described the gap as a major vulnerability in the US export-control framework. Chris McGuire, a former US State Department official, said the loophole allowed overseas subsidiaries of Chinese companies to acquire Nvidia’s advanced Blackwell chips without obtaining a licence.

However, the latest guidance stops short of requiring data centres to discontinue the use of affected chips already deployed in their facilities. It also does not mandate service providers to halt maintenance or support for existing advanced computing systems.

The action comes as the US intensifies efforts to maintain its lead in artificial intelligence and prevent China’s access to cutting-edge semiconductor technologies viewed as strategically important for both commercial and military applications.