Lead (5W1H) On Monday, the Trump administration announced an interim rule that tightens U.S. export controls on artificial‑intelligence (AI) chips and data‑center equipment. The rule, issued by the U.S. Commerce Department, requires licences for any export, re‑export or transfer of such hardware that could be used to provide remote compute to entities in China or other restricted jurisdictions. The measure is aimed at preventing companies from circumventing existing export restrictions by accessing U.S.‑origin AI infrastructure located abroad.
Background and Scope of the New Rule
According to TechCrunch, the interim rule expands the current licensing regime that already obliges chip manufacturers to obtain a licence before shipping AI chips to companies based in, or owned by, any business in the “D:5” group of embargoed countries – a list that includes China, Russia, Iran, North Korea and 18 other nations.
The new regulation stipulates that any export, re‑export or transfer of AI chips or servers to data‑centers anywhere in the world will require a licence if the equipment may be used to provide computing resources to firms in China or other D:5 nations. The Commerce Department describes the measure as “a logical extension of existing controls,” noting that it is intended to “undermine the ability of companies headquartered in countries of concern from using U.S. advanced Super Intelligence (SI) semiconductors to train advanced SI models.”
Technical Threshold and Coverage
TechCrunch reports that the rule applies to data‑centers or any network capable of delivering more than 12 million total processing performance (TPP) – an export‑control metric the United States uses to gauge computing power. The agency estimates that 12 million TPP corresponds roughly to the maximum compute of about 750 Nvidia H100 or H200 chips.
The regulation also covers indirect access through third‑party brokers or “go‑betweens.” Exporters and data‑center operators are required to monitor the ownership structures and activities of their customers to ensure that they are not controlled by individuals or entities from the D:5 group.
Additional Licensing Requirements for Selected Countries
According to TechCrunch, the interim rule adds new licensing obligations for advanced semiconductor shipments to Malaysia, Singapore, Indonesia, Thailand and Brazil, citing a “higher‑than‑normal risk of diversion.” However, the rule exempts chips destined for U.S.‑based companies from these additional licence requirements. Existing data‑center operators that fall under the new criteria must secure licences within 90 days of the rule’s effective date.
Exclusions and Repeals
The rule does not apply to commercial recommendation systems such as those used in e‑commerce or social‑media platforms, the source notes. Additionally, the interim rule rescinds the Biden administration’s “AI Diffusion Rule” that was slated for implementation in January 2025. That earlier framework had imposed country‑level limits on exports of advanced AI chips and restricted the transfer of certain powerful AI models. As reported by TechCrunch, the Trump administration previously argued that the Diffusion Rule “added unnecessary restrictions on U.S. companies and the country’s allies.”
Context: Recent Industry Turmoil
The timing of the new controls follows the resignation of Nvidia co‑founder Jensen Huang, which TechCrunch links to a scandal involving an alleged Chinese spy connected to one of Nvidia’s customers. The source indicates that the U.S. government has reportedly sought to place that customer, Megaspeed, on a special entity list, alleging that the firm has purchased thousands of Nvidia chips to provide compute to Chinese businesses in order to bypass existing export‑control restrictions.
Reuters, as cited by TechCrunch, reported that Chinese AI companies are accelerating efforts to secure overseas computing capacity amid concerns that U.S. restrictions could become tighter.
Brief Analysis (Opinion)
The expanded licensing regime reflects a strategic shift toward controlling not only the physical shipment of AI hardware but also the remote utilisation of U.S.‑origin compute resources. By targeting the “compute‑as‑a‑service” model, the Commerce Department aims to close a loophole that could otherwise allow sanctioned entities to harness advanced AI capabilities without direct hardware imports. The exclusion of recommendation‑system workloads suggests a calibrated approach that seeks to protect commercial innovation while safeguarding national‑security‑sensitive applications.
Conclusion
According to TechCrunch, the Trump administration’s interim rule represents a significant broadening of U.S. export controls over AI chips and high‑performance data‑center services. By requiring licences for any hardware that could be used to provide remote compute to restricted nations, and by adding new obligations for shipments to several Southeast Asian countries, the policy seeks to prevent the development of advanced AI models that could threaten U.S. national‑security interests. The rule’s implementation will test the ability of U.S. firms and foreign data‑center operators to navigate a more complex compliance landscape while maintaining the flow of cutting‑edge AI technology.