US makes data centers worldwide license-bound if they may serve Chinese AI customers
The Commerce Department's interim rule also adds licensing for chip shipments to Malaysia, Singapore, Indonesia, Thailand and Brazil, TechCrunch and Al Jazeera reported.
Add FinanceInTime as a preferred source on Google
The US Commerce Department's Bureau of Industry and Security (BIS) issued an interim final rule on Monday, November 16, that requires licenses to send AI chips and servers to data centers anywhere in the world if they may provide computing power to companies in China or other restricted countries, according to TechCrunch and Al Jazeera.
Who needs a license, and for what
Under the rule, companies must get licenses to export, reexport or transfer AI chips and servers to data centers worldwide if they may provide compute to companies in China or other restricted countries, TechCrunch reported. The remote-access rules target companies based in the D:5 group of embargoed countries, which includes China, Russia, Iran, North Korea and 18 others.
Before the rule, Chinese firms were barred from importing the most advanced graphics processing units but could legally use top-of-the-line computing power at data centers in countries such as Singapore, Malaysia and Thailand, Al Jazeera reported. BIS called the change "a logical extension of existing controls," according to TechCrunch's reading of the rule text.
A computing-power threshold and five added countries
The rule applies to data centers or networks that can supply more than 12 million total processing performance (TPP), an export-control measure of computing power, TechCrunch reported. It estimated that 12 million TPP is roughly the maximum computing power of about 750 Nvidia H100 or H200 chips. The rule also covers indirect access through go-betweens or brokers, and exporters and data center operators must monitor their customers' ownership and activities.
It adds license requirements for exports of advanced chips to Malaysia, Singapore, Indonesia, Thailand and Brazil. The Commerce Department said these countries "present a higher-than-normal risk of diversion," per TechCrunch. The requirement does not apply to chips destined for companies based in the United States. The rule also does not apply to commercial recommendation systems such as those used in e-commerce or social media, and it rescinds the Biden administration's AI Diffusion Rule from January 2025.
Ninety days for data centers already serving Chinese customers
Existing data center operators providing services covered by the rule must obtain licenses within 90 days of the rule coming into effect, TechCrunch reported. Fortune reported that the rule takes effect immediately once it is included in the Federal Register, and noted that the Commerce Department's press release says the rules were "written by and at the direction of the Department of Treasury and the National Security Council."
Reactions and the investigation behind the timing
Singapore's Ministry of Energy, Trade and Industry and Ministry of Digital Development and Information said they were "studying the details of this new measure carefully" and "engaging affected companies to understand the potential impact on their businesses," Al Jazeera reported.
The rule followed the resignation last week of Jensen Huang as Nvidia's chief executive. FinanceInTime covered the board's position before the resignation. Al Jazeera reported that the Commerce and Justice departments publicly confirmed late on Monday a criminal investigation into Megaspeed, a Singapore-based data center company. Beijing has denied the spying allegations made in media reports, according to the same report.
In Congress, the Remote Access Security Act, which also targets remote access, passed the House in January and is awaiting consideration in the Senate, according to Fortune.
Capacity abroad now carries a license risk that spending plans did not list
We think the rule moves a cost that used to sit with the buyer onto the builder. Operators of large data centers outside the United States now have to know who owns their customers, and the sources show they have 90 days to apply for a license if they already serve Chinese companies. None of the reports say how much existing capacity that covers.
Our line weighs capital spending against revenue. The threat is that capacity built on the assumption of Chinese demand turns out to need licenses it does not get. The opportunity is that operators with documented, non-Chinese customers may find the rule narrows competition. We would watch whether license decisions arrive inside the 90 days, and our view would change if the final text or later guidance narrows the 12 million TPP threshold.
The view in this final section is FinanceInTime's editorial opinion, for information only. It is not investment, tax, legal or other advice, and it does not take your circumstances into account. Forecasts are uncertain and past results do not guarantee future ones.
Sources
- US tightens AI chip export controls to deter Chinese companies from accessing compute remotely — TechCrunchSecondary source
- US moves to block China's remote access to AI chips amid Nvidia scandal — Al JazeeraSecondary source
- The Nvidia scandal fuels tougher China chip rules as Commerce targets Singapore and Malaysia — FortuneSecondary source
More in AI & Tech
View All AI & Tech


