The Trump administration has formulated draft regulations that could grant the United States unprecedented authority over the global distribution of artificial intelligence processors. By requiring companies to secure federal approval before finalizing AI chip exports from industry leaders like Nvidia and Advanced Micro Devices (AMD), these proposed rules aim to position Washington as the gatekeeper of global digital infrastructure.
If implemented, these new regulations would expand existing restrictions—currently affecting about 40 nations—to cover hardware shipments anywhere in the world. Rather than acting as a blanket ban, the framework ensures the international community utilizes American AI technology under carefully monitored conditions.
A Tiered Approval Process for Computing Power
The drafted framework establishes a tiered licensing system based on the volume of computing power requested. For smaller-scale shipments involving up to 1,000 of Nvidia’s latest GB300 graphics processing units (GPUs), the review process would remain relatively straightforward and include potential exemption opportunities.
As the scale of the data center increases, so do the regulatory hurdles. Companies seeking to build larger clusters must obtain pre-clearance before applying for export licenses. To secure approval, these organizations could be forced to disclose their underlying business models or permit on-site inspections by US government representatives.
For massive deployments exceeding 200,000 Nvidia GB300 GPUs owned by a single company in one country, the requirements become significantly more demanding. Such monumental projects necessitate direct involvement from the host nation’s government. The United States would only authorize these exports to allied nations willing to make rigorous security commitments and provide matching financial investments into the American AI sector. Although the draft does not mandate a specific investment ratio, a previous chip export agreement required the United Arab Emirates to invest one dollar in the US for every dollar invested domestically before licenses were granted.
Replacing Previous Regulatory Frameworks
This proposed strategy represents the administration’s most significant move toward a comprehensive global semiconductor export policy since it discarded the previous administration’s approach earlier in the year. Officials have criticized the prior AI diffusion rule, which imposed strict sales caps and controlled exports to most countries, calling it excessively burdensome.
The new approach is designed to facilitate American dominance while ensuring strict oversight. However, the exact impact will depend heavily on execution. Rapid license approvals would allow the global AI infrastructure buildout to continue with merely added paperwork, whereas bureaucratic bottlenecks could severely disrupt international project planning.
Diplomatic Leverage and Foreign Investments
The global mandate leaves several key questions unanswered, particularly regarding the financial expectations placed on nations with massive computing ambitions. Countries like India and France are planning large-scale data centers requiring one gigawatt or more of power, but it remains unclear how much matching investment the US might demand.
Furthermore, the regulations could become a powerful tool in broader diplomatic negotiations. The administration has previously threatened semiconductor export controls as a retaliatory measure against digital services taxes imposed by entities like the European Union.
Another area of uncertainty surrounds the regulation of model weights—the numerical parameters crucial for processing data and decision-making in AI software. While previous frameworks utilized sweeping restrictions on where frontier model weights could be hosted, the new proposal addresses these highly valuable intellectual properties on a case-by-case basis through individual licensing.
Countering International Competition
Foreign leaders have expressed discomfort at having their technological development tied to Washington’s approval, yet they face a market with limited alternatives. While Nvidia remains the undisputed market leader, the primary alternative is China’s Huawei Technologies Co. To discourage the adoption of alternative technologies, Washington has warned that deploying Huawei AI accelerators anywhere globally could violate existing American trade restrictions.
China’s technological ambitions remain a driving force behind these policy shifts. The United States continues to limit China’s domestic AI chip production by restricting semiconductor manufacturing equipment exports. Concurrently, the administration allowed Nvidia to re-enter the Chinese market to directly compete with Huawei.
The global framework would enhance Washington’s visibility into international chip movements to detect semiconductor smuggling. This oversight could prevent chips from being diverted to China. Additionally, the United States could use licenses to restrict Chinese companies from accessing AI computing power abroad. Recent shipments to the UAE were conditioned on the buyer refusing computing services to Chinese AI firms. Similar terms might apply in Southeast Asia, where companies like Alibaba Group rent computing time on Nvidia chips they cannot purchase directly.
Current Status of the Draft
The drafted framework is not yet finalized. Federal agencies are actively providing input, meaning the proposed rules could undergo substantial revisions or be shelved for other priorities. Neither the Commerce Department’s Bureau of Industry and Security nor chipmakers Nvidia and AMD have commented on the ongoing policy discussions.
