U.S. Restricts Remote Computing Access, Targeting Data Centers Exceeding 12 Million Performance Points
Key Facts
On November 16, the U.S. Department of Commerce’s Bureau of Industry and Security issued an interim final rule restricting remote access to advanced computing chips and servers, targeting data centers outside the United States that provide large amounts of computing capacity to companies based in China or Macau. The rule requires licenses to export, reexport or transfer advanced chips to those centers, and also requires supply-chain participants to conduct customer due diligence. The department said the aim is to prevent foreign parties from using advanced computing capacity based on U.S. technology in ways that threaten national security or foreign policy interests. Its scope also extends to data centers serving companies headquartered, or whose parent companies are headquartered, in destinations within Country Group D:5. This makes cross-border chip use an explicit part of export controls, even when capacity reaches the customer through a hosted computing service outside the customer’s country.
The rule sets a quantitative threshold for centers serving the companies concerned: it covers use of more than 12,000,000 total processing performance, or TPP, at a single data center or across interconnected centers. It also adds a licensing requirement for advanced chips destined for 5 countries: Indonesia, Malaysia, Singapore, Thailand and Brazil. Shipments intended for companies headquartered, or whose ultimate parent companies are headquartered, in the United States are exempt from this conditional requirement. The exemption ties treatment to the headquarters of the beneficiary company or its parent, as well as the shipment’s destination. The licensing requirement for data-center access does not apply to commercial recommendation systems, including those used by online retailers and social media companies.
The rule gives data centers already providing significant computing capacity to entities headquartered in Macau or in Country Group D:5 time to obtain a license allowing service to continue. Those centers must secure the license within 90 days of the rule’s effective date. This means the measure applies not only to new export applications, but also to the continuation of some existing services after the provisions take effect. For center operators, continued provision of computing capacity is tied to customer reviews and meeting licensing requirements within the specified period. The deadline sets a practical timetable for applying the restrictions to computing services that were in place before the rule was published.
The announcement came on a day of broad pressure on U.S. stocks: by 10:15 a.m. Eastern time on November 16, the S&P 500 was down 1.3% and headed for a third consecutive daily loss. The Nasdaq fell 2.1% at the same time, while the Dow Jones lost 332 points, or 0.6%. The Associated Press said semiconductor stocks were affected by broader concerns about the AI sector and a leadership crisis at Nvidia following Jensen Huang’s resignation. AMD shares fell 4.7% in trading on November 16. Nvidia’s close at $162.85, meanwhile, was from the November 13 session and does not by itself establish the cause of the decline or a connection to the rule issued on November 16.
Nvidia said in a filing with the U.S. Securities and Exchange Commission that Jensen Huang resigned immediately as chief executive and a member of the board, and that Colette Kress would become interim chief executive while continuing as chief financial officer. The chairman said the board expected operations, customer commitments and the product roadmap to continue without disruption. Amid the leadership turmoil, Reuters quoted an Ameriprise strategist as saying, "Nvidia will go through a period of volatility," adding that this would keep trading in the technology and AI sectors volatile for the rest of the year. Reuters also reported that Chinese AI companies were trying to rent computing capacity powered by Nvidia chips outside China, amid concerns about tighter remote-access restrictions. These commercial concerns intersect with the new rule’s scope at data centers providing computing capacity to specified customers outside China and Macau.
Attention now turns to Nvidia’s quarterly results, scheduled for Tuesday, November 17, as Kress takes temporary leadership of the company. The report was scheduled for after the U.S. market close, with Kress set to speak to Wall Street. According to the company’s filing, fiscal 2027 third-quarter results are due at about 1:20 p.m. Pacific time, followed by the earnings call at 2 p.m. The announcement comes as the company faces regulatory scrutiny and concerns about AI company valuations, alongside investor questions about leadership continuity. The company set the results date in line with its previously announced schedule, while implementation of the remote-access rule and the company’s earnings path have become two concurrent areas of focus.