[News] U.S. Reportedly Mulls AI Chip Curbs on Malaysia, Thailand—Taiwan Server Firms at Risk

According to Bloomberg, sources say the U.S. Commerce Department is drafting new rules to restrict shipments of advanced AI chips to Malaysia and Thailand, as part of a broader effort to crack down on suspected chip smuggling into China.

As noted by Tom’s Hardware, citing Bloomberg, a preliminary version of the rule would require companies to obtain a U.S. export license before shipping AI GPUs to the two Southeast Asian nations. The plan is still under review and subject to change, the report adds.

Meanwhile, Bloomberg, citing sources, indicates that the restrictions may include a temporary provision allowing firms based in the U.S. and allied countries to continue shipping AI chips to Malaysia and Thailand without a license for a limited time after the rule is published.

Bloomberg notes that Southeast Asia has become a focal point of U.S. concerns over the global spread of AI chips. According to the report, Companies like Oracle are investing heavily in Malaysian data centers, and trade data cited in the report shows a recent surge in chip shipments to the region.

Uncertainty Grows for Malaysia and Thailand’s AI Server Hubs

Economic Daily News notes that the proposed U.S. measures could disrupt AI server production in Southeast Asia, affecting Taiwanese firms such as Wiwynn and Quanta. Quanta’s Thailand facility mainly produces AI server motherboards, while Wiwynn has the most extensive Malaysian operations among Taiwanese peers. Its AI server rack assembly plant reportedly began mass production in 2023, with motherboard production starting in Q4 2024, as the report notes.

The Economic Daily News report also highlights Malaysia’s rise as a global server manufacturing hub, with Johor serving as the country’s largest data center investment zone and ranking ninth in the Asia-Pacific data center market. The area has drawn multinational players, including Australian data center giant AirTrunk, as the report highlights.

Additionally, Reuters notes that Microsoft plans to launch its first Malaysian cloud region—with three data centers—by mid-2025, nearly a year after announcing a $2.2 billion investment in the country.

Potential Factors Behind U.S. Scrutiny of Southeast Asia

As for why these Southeast Asian nations are under U.S. scrutiny, Tom’s Hardware notes that while Malaysia is not listed as one of NVIDIA’s primary sources of revenue, it has recently become a significant destination for computing equipment and components sourced from Taiwan in recent quarters.

Meanwhile, Bloomberg reports that in Singapore, prosecutors have charged three individuals with deceiving customers about the true destination of AI servers. These servers, originally shipped from Singapore to Malaysia, may have contained advanced NVIDIA chips.

Read more

• [News] Malaysia’s April NVIDIA GPU Imports Hit Record USD 2.74B, Fueling China Rerouting Fears
• [News] Malaysia Reportedly to Monitor NVIDIA AI Chip Shipments Amid U.S. Concerns Over Transfers to China

(Photo credit: NVIDIA)