- China's Commerce Ministry said on Tuesday it "strongly opposes" the new U.S. interim rule restricting remote access to advanced AI chips, servers, and computing infrastructure.
- The U.S. measure, developed under the direction of the Treasury Department and National Security Council, aims to prevent countries subject to U.S. export controls from bypassing restrictions through data centers in third countries and could affect global cloud providers, AI infrastructure operators, and chipmakers.
- Nvidia (NASDAQ:NVDA), the central corporate stakeholder, reported blowout quarterly results last month, but faces leadership uncertainty after CEO Jensen Huang's sudden resignation.
Closing the Backdoor
The U.S. Commerce Department has issued a new interim rule restricting remote access to advanced AI chips, servers, and computing infrastructure. The measure, developed under the direction of the Treasury Department and National Security Council, aims to prevent countries subject to U.S. export controls from bypassing restrictions through data centers in third countries. The rule extends existing technology safeguards to remote computing access, potentially affecting global cloud providers, AI infrastructure operators, and chipmakers.
The initiative aims to address a workaround that has frustrated national security officials: Chinese firms barred from buying advanced chips directly can instead access them via cloud services hosted in third countries. Commerce is considering additional controls involving Singapore, Malaysia, Thailand, Indonesia, Brazil, South Africa, and Norway, though some administration officials are pushing for worldwide coverage.
Reporting published November 15 described internal divisions, with Commerce favoring a narrower approach that would cover future shipments, exempt already-equipped facilities, and require notification—but not automatically a license—when Chinese customers train models. Other accounts suggested shipments of Nvidia Blackwell-class or more advanced chips to these countries would require licenses, with conditional exemptions for US-operated data centers that restrict access from arms-embargoed countries, including China and Russia.
A Commerce spokesperson did not respond to a request for comment.
The U.S. measures announced on November 16 cover overseas data centers supplying "significant volumes" of computing power to companies headquartered in China. Affected operators have a 90-day window to obtain licenses to continue those services or risk future export restrictions. Shipments of covered advanced chips to Singapore, Thailand, Malaysia, Indonesia and Brazil face additional licensing requirements because Washington considers them higher-risk diversion destinations, with exceptions for U.S.-headquartered companies operating data centers in those countries and for certain retail and social-media recommendation systems. On November 17, Foreign Ministry spokesperson Guo Jiakun called the restrictions economic coercion and "long-arm jurisdiction," arguing that they interfere with legitimate technological cooperation and disrupt supply chains; China's statement did not announce a specific retaliatory measure.
Nvidia in the Crosshairs
While the rule is a government-policy story, Nvidia is the central corporate stakeholder. The company supplies the advanced processors and computing infrastructure used to train and run AI models, and its Blackwell systems are specifically identified in reports about the proposed controls. Nvidia argues that broad access to American infrastructure sustains US technology leadership, warning that restricting it risks ceding opportunities to foreign competitors such as Huawei.
The stakes are enormous. Nvidia's latest quarterly results, for the period ended July 26, showed revenue of $96.2 billion, up 106% year over year, with data center revenue surging 117% to $89 billion. GAAP net income jumped 126% to $59.7 billion. Those results, however, predate the current policy uncertainty.
Investors are also digesting a leadership shakeup. On November 12, Nvidia announced that Jensen Huang had resigned immediately as president, CEO, and director. CFO Colette Kress became interim CEO while retaining her finance role; Harvey C. Jones became board chair, and the company has begun a search for a permanent CEO. Separately, independent directors said they had retained former acting US Attorney Joon H. Kim and law firm Wachtell, Lipton, Rosen & Katz for an independent investigation. These governance developments are separate from the export-control debate, though they add to the uncertainty surrounding the company.
Nvidia is scheduled to release its next quarterly results on November 17, which will give management an opportunity to address exposure, compliance, and leadership questions.
Economic Ripple Effects
The central industry shift is from controlling where chips are shipped to controlling who can use their computing power after installation. Cloud services make those two questions very different: hardware can remain in Malaysia while a customer in China uses it over a network.
Data-center host economies are particularly exposed. Research cited by The Wall Street Journal estimates that Malaysia has attracted more than $34 billion in data-center investment, largely from American and Chinese businesses. Changes to customer eligibility could affect project financing and expected utilization. On November 13, Malaysia said it would not allow itself to be used to circumvent export controls and was investigating Megaspeed and its Malaysian affiliate over possible chip re-exports. Megaspeed says its business complies with applicable rules.
Cloud competition could also shift. If the reported conditional exemptions survive, US-operated facilities could face a different regulatory path from local or Chinese-linked operators. The actual competitive effect will depend on licensing conditions and customer-screening requirements.
Investment relocation is already underway. Providers are looking beyond Southeast Asia—including Japan, Brazil, Central Asia, and Europe—for capacity serving Chinese customers, suggesting that narrow geographic controls may redirect activity rather than eliminate it.
Political and International Context
The new rule is the latest step in a multi-year effort to limit Chinese access to advanced American technology. Existing export controls, affirmed by BIS guidance in May 2026, require licenses for specified advanced computing exports to entities headquartered in designated restricted countries, even when the recipient is located elsewhere. The new rule extends those restrictions to remote access.
Congress is also moving. The Remote Access Security Act passed the House on January 12, 2026, seeking explicit authority over remote access to controlled technology. The Senate has not yet acted, and House passage alone does not make it law.
The debate pits national-security advocates seeking to deny China frontier computing capabilities against officials and businesses concerned about American sales, technological influence, and trade relations. Countries hosting the infrastructure face pressure to preserve investment while preventing their facilities from becoming export-control workarounds.
A related November 15 New York Times investigation examined Commerce Secretary Howard Lutnick's former firm Cantor Fitzgerald and its business relationships with potentially affected data-center companies. The report explicitly said there was no evidence those investments influenced his deliberations; Commerce said he complied with ethics agreements.
What to Watch
The decisive questions are whether the interim rule covers existing facilities, how it defines restricted customers and computing thresholds, whether access requires a license or merely notification, and which operators qualify for exemptions. November 15 reporting indicates these choices remained contested.
Expert assessments emphasize the risk of displacement. Council on Foreign Relations senior fellow Chris McGuire told The New York Times (NYSE:NYT) that limiting access only in selected countries would allow Chinese customers to rent capacity elsewhere. Foundation for American Innovation researcher Ryan Fedasiuk described tighter oversight of these arrangements as inevitable.
A parallel trend is the development of nationally controlled, or "sovereign," AI infrastructure. On November 9, Nvidia announced a collaboration with Korea's Rebellions and SK Telecom (NYSE:SKM) combining Nvidia infrastructure with Korean-designed AI processors and local data-center capabilities. That is not itself an export-control response, but it illustrates the increasingly country-specific organization of AI computing.
The most consequential shift is that export policy is moving beyond the chip's physical destination toward the identity and activities of its remote user. Whether the new controls meaningfully restrict Chinese access—or mainly move that access elsewhere—will depend on the interim rule's coverage and enforceability.