Trump Administration Working on Rule to Limit China’s Remote Access to Advanced Chips
Key Facts
The Trump administration is working on a rule aimed at limiting China’s remote access to advanced chips, according to the headline of a report published by The Information, addressing a gap in U.S. technology controls. The headline did not specify the rule’s stage of development or details, and the publicly available body of the report did not provide that information. The report says the issue concerns the ability of Chinese artificial intelligence companies to obtain computing capacity at data centers outside China through cloud services. That means the regulatory debate extends to the use of chips after they are located outside Chinese territory, and not just their direct export. But the initiative remains uncertain: the report said the remote-access gap had not been resolved.
According to the report, a small team at the U.S. Commerce Department was preparing a scaled-back version of an earlier AI diffusion rule. It said the department could present the proposed rule to AI companies and industry trade groups for feedback as early as September 2026. That puts data center operators and cloud service providers at the center of the debate because they allow customers to use computing capacity remotely. Measures cited in the report include imposing “know your customer” requirements to verify customers’ identities and how they use the services provided. These requirements link regulatory compliance with monitoring the ultimate user of computing capacity, according to the proposal described in the report.
The discussion of cloud access comes amid existing export rules for advanced chips. In a directive dated 31 May 2026, the Commerce Department said an export license is required for advanced computing chips destined for entities headquartered in China or Macau, even when the entity itself is located outside those destinations. The department said the licensing requirement continues to apply to those entities. But it also said data center operators are not required, because of that directive, to stop the ongoing use, storage or servicing of chips. The distinction highlights the scope of the existing directive: it sets licensing requirements while leaving operators’ ongoing activities outside the shutdown requirement addressed in the directive.
The AI diffusion rule that preceded this effort was issued by the Biden administration on 15 January 2025, with compliance requirements scheduled to begin on 15 May of that year. The Commerce Department said in May 2025 that it would rescind the rule, while a report by The Information said the department had not been enforcing it, although it had not formally rescinded it at the time the report was published. That background helps explain the administration’s effort to prepare a lighter alternative, according to the report. The potential new rule remains separate from the decision to rescind the earlier one, and its details did not appear in the published headline. The publicly available report does not establish that the proposal reached the approval or implementation stage.
A proposed chip deal between Oracle and Tencent provides context for the potential scale of demand for computing outside China, although the account of its cancellation has not been independently confirmed. On 4 November, Reuters reported, citing the Financial Times, that Oracle had canceled a multibillion-dollar chip lease with Tencent after pressure from the Trump administration, adding that Reuters could not verify the report. The Financial Times said the agreement would have given Tencent access to about 100,000 advanced AI chips through Oracle data centers in Southeast Asia under a five-year lease. The deal report does not establish that its cancellation resulted from the cloud rule under consideration, and the available source does not link the two events. As described in the report cited by Reuters, the deal illustrates how chip capacity could reach a Chinese customer through data centers in another region.