⚡ Quick Summary
- Trump administration circulating 129-page draft requiring country-by-country export licensing for AI chips
- Nvidia and AMD would need government approval before selling GPUs to any foreign country
- White House distanced itself from leak but Commerce Department did not directly deny the plans
- Separate proposal may require large GPU buyers to invest in US-based AI infrastructure
What Happened
The Trump administration is reportedly circulating a 129-page draft regulation that would fundamentally reshape how American AI chips reach the global market. According to reports from Axios and the Financial Times, the Department of Commerce is preparing rules that would require chipmakers like Nvidia and AMD to obtain explicit government approval before exporting AI-capable processors to any country — effectively placing Washington at the centre of every significant GPU transaction worldwide.
The proposed framework would mandate country-by-country export licensing, giving federal regulators granular control over where America's most advanced silicon ends up. This represents a significant escalation from existing controls and has drawn immediate comparisons to the Biden-era AI Diffusion rules, which the Trump administration previously rescinded as "burdensome, overreaching, and disastrous."
A White House official pushed back on the leaked draft, stating it "does not reflect what President Trump has said on export controls nor does it reflect the direction of the Trump administration on encouraging export of the American AI stack." However, the Commerce Department's public response notably avoided directly denying the existence of country-by-country licensing plans.
Background and Context
The United States has been tightening its grip on AI chip exports for years, but the policy landscape has been anything but consistent. The Biden administration introduced the AI Diffusion framework in January 2025, which proposed capping AI chip sales to most nations while exempting trusted allies and imposing total bans on Russia, China, and Iran. That framework never took effect — the incoming Trump team killed it before implementation.
Since then, the administration has pursued a patchwork approach. In January 2026, it announced that exports of Nvidia and AMD GPUs to China would only proceed if domestic supply was deemed sufficient. Meanwhile, large-scale GPU deals with Middle Eastern nations — particularly the UAE and Saudi Arabia — were greenlit as part of broader diplomatic and investment agreements.
The semiconductor industry sits at the intersection of national security and economic competitiveness. Nvidia alone generated over $120 billion in profit recently, even as its China revenue dropped to effectively zero. The company's dominance in AI training hardware makes its products de facto strategic assets, and whoever controls their distribution wields enormous geopolitical leverage. For businesses relying on enterprise productivity software and cloud-based AI tools, the downstream effects of chip supply restrictions could be significant.
Why This Matters
These proposed rules would give the US government unprecedented control over the global AI compute supply chain. Every major AI training cluster, every cloud hyperscaler expansion, and every national AI programme outside the United States would effectively need Washington's blessing to proceed. This is not merely a trade regulation — it is a mechanism for maintaining American primacy in artificial intelligence.
The strategic logic is straightforward: if you control the chips, you control the pace of AI development worldwide. Nations that cannot access cutting-edge GPUs cannot train frontier models. Companies that cannot procure hardware cannot compete. The proposed rules would ensure that the United States remains the gatekeeper of global AI capability, with all the diplomatic leverage that entails.
However, the risks are equally significant. Heavy-handed export controls could accelerate China's domestic chip development programme, which is already making progress with alternatives to Nvidia hardware. Beijing has reportedly ordered its industries to cancel Nvidia orders, suggesting it is already preparing for a future without American silicon. If the controls push too hard, they may achieve the opposite of their intent — fostering a parallel AI ecosystem that the US cannot influence at all.
Industry Impact
For Nvidia and AMD, the implications are enormous. Country-by-country licensing would create massive administrative overhead and uncertainty for customers planning multi-billion-dollar infrastructure investments. Cloud providers like Microsoft Azure, Google Cloud, and Amazon Web Services — whose global expansion depends on predictable GPU supply — would face new planning challenges. Organisations that depend on these platforms for everything from affordable Microsoft Office licence deployments to enterprise AI workloads could see ripple effects.
The Financial Times separately reports that the Commerce Department may require corporations purchasing large volumes of AI chips to invest in infrastructure on American soil. This "build here to buy here" approach aligns with the administration's broader strategy of using trade policy to reshore technology investment, but it adds yet another layer of complexity for multinational technology companies.
Semiconductor stocks are likely to experience volatility as investors parse the difference between leaked drafts, official denials, and eventual policy. The pattern of the past year — announce, deny, modify, implement — has made planning exceptionally difficult for companies operating at the frontier of AI hardware.
Expert Perspective
Industry analysts note that the draft rules reflect a fundamental tension within the administration between security hawks who want maximum control over AI capabilities and commerce advocates who fear that over-regulation will push customers toward Chinese alternatives. The Commerce Department's statement that it "will not" return to the AI Diffusion rule, while simultaneously preparing what appears to be an even more comprehensive licensing regime, illustrates this internal contradiction.
Trade policy experts suggest that the most likely outcome is a hybrid approach: broad licensing requirements coupled with expedited approval for allied nations and pre-negotiated deals like those already in place with Gulf states. This would give the administration both control and flexibility, though it would still represent a significant new burden on chipmakers and their customers.
What This Means for Businesses
Businesses that rely on cloud computing and AI services should monitor these developments closely. If export controls tighten significantly, cloud providers may face constraints on expanding capacity in certain regions, potentially affecting service availability and pricing. Companies planning AI infrastructure investments should factor regulatory risk into their timelines.
For small and medium businesses, the most immediate impact may be indirect — through pricing changes for cloud AI services or delays in the rollout of AI-powered productivity tools. Ensuring your core productivity stack is solid, including securing a genuine Windows 11 key and keeping software licences current, remains a practical step regardless of how the geopolitical landscape shifts.
Key Takeaways
- The Trump administration is drafting rules requiring country-by-country export licensing for AI chips from Nvidia and AMD
- A 129-page draft is circulating through government, though the White House has distanced itself from the leak
- The rules could give Washington unprecedented control over global AI compute supply
- Separate proposals may require large GPU buyers to invest in US-based infrastructure
- The policy creates tension between national security goals and commercial competitiveness
- China is already moving to reduce dependence on American chips
Looking Ahead
The coming weeks will determine whether these draft rules evolve into formal policy or join the long list of trial balloons that the administration has floated and abandoned. Regardless of the specific outcome, the direction of travel is clear: the US government intends to maintain firm control over who gets access to the world's most powerful AI hardware. The question is not whether controls will tighten, but how much friction the government is willing to introduce into a market that drives hundreds of billions in American exports.
Frequently Asked Questions
What are the proposed AI chip export rules?
The Trump administration is reportedly drafting regulations that would require American chipmakers like Nvidia and AMD to obtain government approval before exporting AI-capable processors to any country, on a country-by-country basis.
How would this affect cloud computing services?
Cloud providers expanding globally may face GPU supply constraints in certain regions, potentially affecting service availability, pricing, and the rollout of new AI-powered features.
Is China developing its own AI chips in response?
Yes, China has been accelerating domestic chip development and has reportedly ordered its industries to cancel orders for Nvidia hardware, signalling a strategic shift toward self-sufficiency in AI silicon.