The U.S. government is intensifying its effort to keep advanced Nvidia chips from reaching China through indirect routes. Malaysia is now in the spotlight, as officials suspect that high-end semiconductors may still be moving toward China despite export controls.
At the center of the issue is a fast-growing data center market, strict U.S. rules on AI chip exports, and concern that shipments can be redirected after reaching third countries. The result is a sharper focus on monitoring, licensing, and the companies that operate global AI computing infrastructure.
Malaysia Comes Under Closer Scrutiny
U.S. officials are pushing for stricter oversight of high-end Nvidia chip shipments to Malaysia. Malaysian Trade Minister Zafrul Aziz told the Financial Times that U.S. officials are demanding strict monitoring of all such shipments.
The concern is straightforward: officials suspect advanced semiconductors are still reaching China even though export controls are meant to block that flow. Malaysia matters because it has become one of the fastest-growing markets for data centers.
That growth has been large enough to attract over $25 billion in investments during the past 18 months. For U.S. regulators, a booming data center industry can also create a larger surface for possible circumvention of export restrictions.
Malaysia is now creating a task force to tighten regulations around the sector. The move suggests that the country is trying to respond to external pressure while also managing the rapid expansion of a strategically important industry.
Why Nvidia Chips Are The Focus
Nvidia chips sit at the center of the current AI computing boom. The source article focuses on high-end Nvidia chips and servers equipped with Nvidia chips, which are important because advanced AI systems require significant computing power.
The U.S. government’s concern is not only about direct exports. It is also about whether restricted technology can move through other markets before reaching China. In this case, officials are examining whether Malaysia could become part of that route.
Recent events in Singapore appear to have strengthened those concerns. Authorities there arrested nine people allegedly involved in fraudulent sales of Nvidia chip-equipped servers.
That case does not prove every shipment to nearby markets is suspicious. It does, however, show why regulators are treating the movement of AI hardware as a compliance issue that extends beyond a single border or a single buyer.
The Wider Export-Control System
The Malaysia pressure fits into a broader U.S. effort to tighten control over AI chip exports to China. The U.S. government has increasingly restricted chip exports through a new three-tier licensing system for AI chips.
Under that system, 18 close allies including Germany and Japan receive preferential treatment. Roughly 120 countries face strict quantity limits. China faces a complete export ban alongside Russia, Iran, and North Korea.
The Trump administration is also planning additional restrictions. These include limits on maintenance work by Tokyo Electron and ASML at Chinese chip factories, as well as stronger sanctions against companies like ChangXin Memory Technologies and SMIC.
The structure of the rules shows how broad the policy has become. It is not limited to chip sales alone. It also reaches maintenance work, cloud computing capacity, company sanctions, and the geographic placement of AI infrastructure.
Cloud Providers Face Capacity Limits
The export-control effort also affects major cloud providers. Companies such as Microsoft, Google, and Amazon now face special regulations limiting their AI computing power outside the U.S. to 50% of their total capacity.
In countries outside the closest allies, only 7% of computing power is allowed per country. Companies can apply for special permits, but the rule still places a clear ceiling on how much AI compute can be deployed in many markets.
These limits matter because cloud providers can offer access to computing power without physically selling chips to every customer. Controlling chips alone may not be enough if the same capabilities can be accessed through cloud infrastructure.
For regulators, that makes cloud capacity part of the same export-control problem. For companies, it means AI infrastructure decisions are increasingly tied to government rules about where computing power can be placed and how it can be used.
Nvidia Warns Against The Measures
Nvidia has strongly opposed these restrictions. The chip manufacturer calls them a "far-reaching intervention" and warns that they could threaten U.S. leadership in AI development.
That objection points to the central tension in the policy. The U.S. government is trying to limit China’s access to advanced AI hardware, while Nvidia argues that broad controls could weaken the position of U.S. companies in AI.
Malaysia’s task force and the demand for shipment monitoring show how that tension is spreading across the global technology supply chain. Data centers, cloud providers, chipmakers, and governments are all being pulled into the same enforcement challenge.
The immediate issue is whether high-end Nvidia chips can be kept from reaching China through Malaysia. The larger issue is whether export controls can keep pace with a global AI market built on hardware, servers, maintenance networks, and cloud computing capacity that crosses borders.