U.S. blocks China’s remote access to AI chips after Nvidia scandal; China slams move
The U.S. announced an interim final rule aimed at stopping Chinese firms from reaching advanced American AI chips through data centres in third countries. The concern is that controls on chip sales could be bypassed when customers use computing facilities abroad. The rule adds a licensing requirement for exporting, reexporting or transferring advanced semiconductors to Indonesia, Malaysia, Singapore, Thailand and Brazil, which the U.S. calls higher-than-normal diversion risks. It also addresses remote access to chips and servers, targeting computing arrangements as well as direct shipments. The Commerce Department says the rule is meant to prevent restrictions on sensitive U.S. technology from being circumvented. Beijing called it an abuse of export controls and warned of harm to ordinary trade and global semiconductor supply chains. The measure was announced just before President Donald Trump’s planned meetings with President Xi Jinping.
What new restriction did the U.S. announce on advanced AI chips and data centres?
The U.S. announced an interim final rule aimed at stopping Chinese firms from reaching advanced American AI chips through data centres in third countries. The concern is that controls on chip sales could be bypassed when customers use computing facilities abroad.
The rule adds a licensing requirement for exporting, reexporting or transferring advanced semiconductors to Indonesia, Malaysia, Singapore, Thailand and Brazil, which the U.S. calls higher-than-normal diversion risks. It also addresses remote access to chips and servers, targeting computing arrangements as well as direct shipments.
The Commerce Department says the rule is meant to prevent restrictions on sensitive U.S. technology from being circumvented. Beijing called it an abuse of export controls and warned of harm to ordinary trade and global semiconductor supply chains. The measure was announced just before President Donald Trump’s planned meetings with President Xi Jinping.
What does “remote access” to AI chips mean?
The article does not give a technical definition of “remote access.” In ordinary cloud computing, it means a customer uses computing resources over a network while the hardware remains somewhere else. The user sends tasks to a remote system and receives the results, without physically possessing the chips.
For example, a data centre in Malaysia could house advanced chips while a customer in another country connects to its computing services. The chips stay in the data centre, but the customer can use their processing capacity remotely. The article says Chinese firms have sought access to Nvidia technology through data centres in countries such as Malaysia.
That distinction matters because export controls often focus on where technology is shipped or transferred. The new U.S. rule is expressly intended to address remote computing arrangements involving data centres in third countries. It seeks to prevent existing limits on access to sensitive U.S. technology from being circumvented through those arrangements.
How many countries does the rule name as higher-than-normal risks for diversion, and which are they?
The rule names five countries as presenting a “higher-than-normal risk of diversion.” They are Indonesia, Malaysia, Singapore, Thailand and Brazil. The U.S. says the list matters because advanced chips could be redirected or made available in ways that undermine existing restrictions.
The rule requires a license to export, reexport or transfer advanced semiconductors to any of those countries. This requirement applies in addition to current licensing rules for advanced semiconductors. The article also describes a separate, related focus: remote access to advanced chips and servers through data centres in third countries.
The announcement puts Singapore in particular under scrutiny because cloud computing firm Megaspeed is based there. The article says the company is suspected of helping Chinese firms get around U.S. export controls. China criticized the new rule, arguing that it interferes with normal trade and could affect the stability of the global semiconductor industry and supply chain.
Why have Chinese companies sought access to Nvidia chips through data centres in other countries?
The article gives a direct reason: Nvidia is barred from selling its most advanced chips in China. Chinese companies have instead sought access to the technology through data centres located in other countries. This route makes the location of the hardware different from the location of the company using its computing power.
The article points to Malaysia as one example of a country where Chinese firms have sought access through data centres. It also describes scrutiny of Nvidia’s chip provision to companies operating data centres in Southeast Asia. Singapore-based cloud computing firm Megaspeed is suspected of helping Chinese firms get around U.S. export controls.
The new U.S. rule is meant to close off or control these kinds of routes, including remote computing arrangements. It requires licenses for advanced-chip exports, reexports and transfers to five named countries. China objects that the restrictions interfere with normal trade and technology cooperation and could disrupt global semiconductor supply chains.
What license will companies need to export, reexport, or transfer advanced chips to the named countries—and how could that affect access?
The rule requires a license to “export, reexport, or transfer (in-country)” advanced semiconductors to Indonesia, Malaysia, Singapore, Thailand or Brazil. The U.S. identifies these countries as having a higher-than-normal risk of diversion. The license requirement is in addition to current licensing requirements for advanced semiconductors.
This can affect access because companies must seek approval for the specified transactions rather than moving the chips without that license. The rule also targets remote access to advanced chips and servers, aiming to stop third-country data centres from providing a route around existing restrictions. The article does not specify how applications will be decided or whether any particular transfer will be approved.
The measure comes as U.S. officials scrutinize chip access through Southeast Asian data centres, including the Singapore-based firm Megaspeed. China says the rule harms legitimate business interests and could destabilize global semiconductor supply chains. The new requirements will shape how firms arrange chip transfers and computing access in the named countries.
Why are advanced chips important for developing and running artificial intelligence?
The article does not explain how AI chips work. As established background, advanced chips perform large numbers of calculations that AI systems need during development and use. They provide computing capacity for processing data and carrying out the operations that produce an AI system’s outputs.
That capacity can be concentrated in data centres. A customer may use the computing power of chips housed in a facility without owning or physically receiving those chips. The article describes Chinese companies seeking access to Nvidia technology through data centres in other countries after Nvidia was barred from selling its most advanced chips in China.
This helps explain why the U.S. rule covers both chip transfers and remote access to chips and servers. The Commerce Department says it is intended to stop restrictions on sensitive U.S. technology from being circumvented through third-country data centres. China argues the measures interfere with normal trade and threaten semiconductor supply-chain stability.
How can a data centre let customers in another country use its chips, and why does that make export controls harder to enforce?
The article does not describe the technical process in detail. In ordinary cloud computing, a data centre houses servers and chips, while customers connect over a network to send tasks and receive results. The hardware stays in the facility; the customer uses its computing capacity remotely rather than taking possession of the chips.
For example, the article says Chinese companies have sought Nvidia technology through data centres in countries such as Malaysia. If customers can use chips from abroad, the location of a shipment may not show who ultimately benefits from the computing power. That is why controls aimed only at direct chip sales may not cover every route to access.
The U.S. Commerce Department says its interim final rule is designed to address remote access through data centres in third countries. It also requires licenses for specified chip transfers to five countries. The article does not detail the enforcement process, but says the rule is intended to prevent circumvention of existing restrictions.
This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.
Read more in the JupiteX app
Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.
Or read more news on the web