China Signals Retaliation Over US Curbs on Remote Access to Advanced Chips
Key Facts
China vowed to take countermeasures after a new US rule took effect targeting remote access to advanced computing chips and servers, expanding the scope of restrictions from physical shipments to the use of computing capacity through overseas data centers. China’s Ministry of Commerce said Washington was “generalizing the concept of national security and abusing export controls,” and was interfering in normal trade and technical cooperation between Chinese companies and companies in other countries. It added that the measure harms the stability of global semiconductor supply chains, violates important understandings reached by the two countries’ presidents and undermines market rules and the international trading system. The ministry said Beijing “will take countermeasures” if Washington proceeds along this path, expressing hope that the differences can be addressed through consultation and cooperation. The tensions come as Nvidia closed at $156.14 on November 16, according to EL7 data.
The US Bureau of Industry and Security issued the interim final rule on November 16, saying it was intended to prevent existing restrictions from being circumvented through remote computing arrangements at data centers in third countries. Under the rule, overseas data centers serving companies in countries including China, Russia, Iran and Cuba must identify their end users. This extends oversight to the party using the computing capacity, even when the servers are outside the country targeted by the restrictions. The bureau said the rule was designed to ensure that restrictions on access to sensitive US technology cannot be circumvented through remote computing services. The mechanism is intended to prevent the benefits of advanced chips from reaching unauthorized users through service providers and infrastructure in other countries.
The rule adds licensing requirements for chip shipments to Indonesia, Malaysia, Singapore, Thailand and Brazil, countries that it said pose higher-than-usual risks of diversion to unauthorized destinations. These requirements are not limited to data centers in those countries; they cover chip shipments to them, while end-user identification requirements apply to overseas data centers serving companies in countries including China. The distinction between the rule’s two tracks reflects the broader geographic and regulatory scope of its application, from tracking the beneficiary of computing to scrutinizing shipment destinations. China’s Ministry of Commerce said the US measure interfered with normal commercial and technical cooperation between companies and destabilized semiconductor supply chains. Beijing also said the measure conflicted with understandings between the two countries’ presidents and undermined the international trading system.
Estimates cited by Reuters point to the scale of investment exposed to changes in the terms for expansion in the region. Malaysia has approved data center-related investments worth 385.7 billion ringgit, equivalent to $95 billion, since 2021. In Thailand, there were 42 data centers with a technical load of 3,400 megawatts and investments of about 750 billion baht as of February 2026. Growing data center projects helped lift Malaysia’s construction sector 6% in the third quarter of 2026, Reuters reported. An official at the Thai Data Center Association said the equivalent of about 300 megawatts of GPU capacity was expected to enter the country over the following 12 to 18 months, destined for Chinese end users. Capital expenditure linked to that capacity was estimated at about $12 billion to $18 billion, putting planned computing projects at the center of the restrictions’ potential impact on sector expansion.
A Saxo analyst warned that overseas data center operators relying on Chinese customers could face lower utilization, pricing pressure and delays to expansion. He said Chinese AI developers could bear higher costs and face slower development timelines. Bloomberg quoted Nomura analysts Jialong Shi and Rachel Guo as saying that tighter access to advanced computing resources could constrain AI model training and slow the growth of cloud computing revenue in China. An analyst at Counterpoint Research said Chinese companies would no longer be able to bypass the ban on physical hardware by renting advanced US computing capacity from overseas data centers. He added that the regulatory burden had shifted substantially from hardware manufacturers to cloud service providers and companies. These estimates place model developers, infrastructure operators and cloud service providers among the parties that could all be affected by the rule’s implementation.
Singaporean authorities said they were studying the details of the measure and engaging with affected companies and US authorities to discuss its scope and implementation. Singapore said it takes its international reputation seriously and does not accept companies using their association with the country to circumvent or violate other countries’ laws, including export controls. Nvidia, for its part, announced on November 17 the creation of a standing board committee to oversee national security policies, export control compliance and global trade risks. The company said the committee would meet at least once each quarter. On November 12, the company had announced that Jensen Huang would resign, effective immediately, as chief executive, company president and a member of the board, and that Colette Kress would be appointed interim chief executive. Nvidia was scheduled to announce its third-quarter fiscal 2027 results on November 17 at about 1:20 p.m. Pacific time, followed by an earnings call at 2 p.m.