The United States is reportedly drafting a new set of export control rules that could dramatically expand Washington’s authority over the global sale of advanced AI chips from tech firms like Nvidia and AMD. The proposal would shift US export controls from a country-specific framework – largely focused on China – to a broader, global licensing system covering high-performance AI accelerators. Under this draft, firms like Nvidia and AMD would be required to obtain government permits to ship high-performance AI accelerators to customers worldwide, reports Bloomberg. Notably, these chips, which power large language models, generative AI systems, and other compute-intensive applications, can cost tens of thousands of dollars each and are essential for modern AI development.
The proposed measures clearly indicate growing concern in Washington over the national security implications of advanced AI technologies. Currently, US export controls target specific countries, including China, Russia, and a handful of other nations considered strategic rivals. But under the new draft framework, the US government would gain the authority to oversee sales of these chips globally, not just to traditional enemies. According to the report, officials are reportedly evaluating mechanisms that would allow the government to review transactions, impose per-customer caps, and ensure that these powerful processors are not used in sensitive military, intelligence, and dual-use applications abroad.
At the same time, the proposed rules also appear to consider quantitative controls. Reports suggest that regulators are weighing imposing limits on the number of high-end chips any single foreign customer can purchase, a step designed to prevent large-scale acquisitions that could bypass existing restrictions. Meanwhile, the timeline for implementing these rules is still uncertain.
The potential rule will mainly seem to target the distribution of the most advanced AI processors, especially Nvidia’s H200 series GPUs and AMD’s MI300 family accelerators, which are engineered for large-scale AI workloads. These chips are considered critical infrastructure for generative AI systems, cloud computing platforms, autonomous systems research, and advanced defense applications. The move becomes even more significant as Nvidia currently dominates the AI accelerator market with an estimated global market share of 80% or more in high-end data-center GPUs.
However, such rules could push other countries to accelerate their domestic AI chip development programs. For example, China has been investing heavily in semiconductor self-sufficiency and the design of advanced AI hardware, aiming to reduce reliance on US suppliers. Chinese technology firms are reportedly exploring alternatives to Nvidia and AMD chips, while domestic companies backed by giants like Huawei and Alibaba are racing to close the performance gap in high-end AI accelerators. Beyond China, nations across Europe and Asia are also boosting their AI hardware capabilities, not only to encourage innovation but also as a safeguard against potential supply restrictions from the United States.
The Tech Portal is published by Blue Box Media Private Limited. Our investors have no influence over our reporting. Read our full Ownership and Funding Disclosure →
