In August 2025, reports said Nvidia and AMD would give the U.S. government 15% of revenue from certain China-bound AI-chip sales in exchange for export authorization or eased restrictions. The chips named were Nvidia’s H20 and AMD’s MI308. The arrangement was unusual because permission to sell sensitive technology appeared to be tied to a percentage of sales to a specific country—not because it was clearly a conventional tax. The White House confirmed the basic revenue-sharing arrangement, but public reporting did not establish a final contract or settle its legal mechanics.
What the reported arrangement involved
Reports published August 10–12, 2025 described a 15% revenue share on certain China-bound sales of Nvidia H20 and AMD MI308 AI accelerators. The apparent exchange was a payment to the U.S. government for permission to export chips that had been restricted or subject to licensing. Axios reported the arrangement; the Associated Press also reported the 15% figure.
That is a useful description of the reported framework, not proof of identical signed agreements. The public accounts did not specify whether the 15% applied to gross sales, net revenue, profit, or another defined base; which indirect transactions counted; how payments would be collected; or whether Nvidia and AMD had the same terms. “Revenue-sharing deal” is therefore shorthand for a reported export-licensing condition, rather than a description of a publicly available contract.
What was confirmed—and what remained unclear
- Reported and acknowledged: The White House confirmed the basic revenue-sharing arrangement to news organizations, and coverage identified a 15% share tied to certain sales of H20 and MI308 chips.
- Not publicly established: A final legal instrument, the precise revenue calculation, collection mechanism, duration, statutory authority, accounting treatment, and whether the two companies had matching obligations.
- Not implied: Unrestricted permission to export any AI chip to China. Licensing conditions, product scope, eligible customers, and Chinese rules still mattered.
In reports from August 2025, the administration was still working through the arrangement’s legality and mechanics. The Los Angeles Times covered the legal questions; the reported uncertainty is a reason not to treat “15% export tax” as a settled legal classification.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
Why Nvidia and AMD needed U.S. authorization
The arrangement followed a shift in U.S. export policy. The Trump administration restricted sales of the H20 and MI308 to China in April 2025, citing national-security concerns. Later, the companies sought or received authorization for some exports under licenses. The exact status could vary by product, customer, and transaction. The Congressional Research Service’s overview of U.S. export controls and advanced semiconductors provides context for the restrictions and the H20’s place in the product landscape.
The H20 was designed for the China market and was less capable than Nvidia’s most advanced products available elsewhere. In July 2025 reporting, Commerce Secretary Howard Lutnick described it as Nvidia’s “fourth-best” AI chip. Its lower performance did not make it an unrestricted consumer product: export controls and case-by-case licensing still applied. AMD’s MI308 was also named in coverage of the reported arrangement, though the public details about AMD’s terms were less complete.
Why analysts called it “unique”
Governments routinely use tariffs, export controls, licensing requirements, fees, and penalties. The reported arrangement stood out because it appeared to combine five features: a particular company, a particular product category, a particular destination—China—a payment calculated as a share of sales, and permission to export sensitive AI hardware.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
TechInsights analyst Dan Hutcheson told EE Times he could not identify a comparable product- and company-specific charge on sales to one foreign country. “Unique,” “rare,” and “unprecedented” were analysts’ and commentators’ characterizations, not settled legal findings. EE Times also described the structure as an “inverted tariff” or “tariff at source”: a useful analogy for a charge connected to exports, but not necessarily its formal legal name.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Was it a tariff, tax, fee, or licensing condition?
Calling the arrangement a 15% export tax is common shorthand, but the legal form was not publicly settled. A conventional tariff generally applies to imports entering a country through a statutory or trade-policy mechanism. The reported payment, by contrast, was linked to sales revenue and to permission to export specified goods. Nor did the reporting establish that it was a routine fixed administrative licensing fee.
The closest cautious description is a reported revenue-sharing condition associated with export authorization. Whether the payment would legally count as a tax, fee, penalty, voluntary contribution, or another kind of charge depended on terms and authority that had not been publicly laid out. That distinction matters: the label affects questions about who can impose the payment, how it is collected, and how companies would account for it.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
What Nvidia and AMD said
Nvidia
Nvidia said it follows U.S. government rules governing its participation in global markets and expressed hope that export controls would allow the United States to remain competitive in China and worldwide. The cited public coverage did not provide a detailed company explanation of the reported revenue-sharing mechanics. EE Times summarized Nvidia’s position.
AMD
AMD reportedly said U.S. authorities had approved applications to export some AI processors to China and emphasized that it complies with U.S. export controls. The cited reporting does not establish that AMD publicly described a complete 15% payment agreement with terms identical to Nvidia’s. Reuters analysis carried by Investing.com discussed the arrangement and its risks.
Recommended Free Tools
Why the arrangement mattered to the companies and Washington
Potential gains and costs for Nvidia and AMD
Authorization could reopen access to Chinese customers, help companies sell products designed for that market, and preserve a place for U.S. hardware in Chinese AI systems. Those gains would come with a direct reduction in revenue on covered sales, as well as additional compliance and accounting work. A policy that can be changed or withdrawn also makes planning less predictable. Nvidia had warned that tighter restrictions could cost it billions in additional revenue, according to the AP’s reporting.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
There were strategic costs beyond the payment itself. If Chinese buyers see access to U.S. chips as conditional or politically fragile, they may favor domestic alternatives. A company-specific bargain could also leave competitors uncertain about whether comparable terms are available to them.
Potential gains and risks for the U.S. government
The arrangement could generate a financial return while allowing limited exports, maintain Chinese reliance on U.S.-designed hardware, or provide leverage in trade negotiations. But monetizing access to technology that the government had restricted on security grounds could weaken the clarity of the controls’ national-security rationale. It could also encourage companies to seek individual deals with the executive branch rather than rely on generally applicable rules, while inviting litigation, congressional resistance, or similar charges by other governments.
The wider U.S.–China trade and security context
The reported arrangement sat between two competing objectives: limiting China’s access to computing that could support AI and military applications, and preserving U.S. companies’ market share in China. President Trump had spoken about seeking a share of Nvidia’s China sales and possible future chip tariffs. Commerce Secretary Lutnick also linked resumed Nvidia sales to a broader trade understanding involving rare-earth magnets, according to the AP’s account. That context does not establish that the chip revenue-sharing terms were part of a comprehensive treaty.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
U.S. approval did not guarantee Chinese sales
Export permission from Washington was only one side of the transaction. Reporting based on people familiar with the matter said Chinese authorities urged or discouraged some domestic companies from using Nvidia H20 chips, especially in government-related or security-sensitive settings. It was not clear whether the same guidance explicitly covered AMD’s MI308, and the reports do not establish a universal Chinese ban. Investing.com reported on the H20 guidance.
That creates a practical failure point: a U.S. license may permit a sale without ensuring that a Chinese buyer can, will, or is authorized to use the chip. Supply also could not necessarily restart immediately. Nvidia said it had not shipped H20 to China for months in the cited coverage, so resuming sales could require production, testing, and customer approvals. EE Times reported that context.
What legal questions the arrangement raised
The central issue was not simply whether the executive branch can restrict exports; U.S. export controls already operate through government licensing authority. It was whether a license could be conditioned on a percentage of private-company sales, and under what legal authority. The answer could depend on the actual terms and how the payment was structured. The public reporting did not establish that the arrangement was illegal, nor did it resolve the legal questions.
- Would the payment be a tax, tariff, fee, penalty, or negotiated licensing condition?
- Would Congress need to authorize a revenue-based charge, and how would export-control statutes apply?
- Would the condition be applied consistently to similarly situated companies and transactions?
- Who would collect the money, into which account, and under what accounting treatment?
Members of Congress sought an explanation. An August 15, 2025 letter from the House Select Committee on the CCP and a Senate letter dated August 15, 2025 provide evidence of lawmakers’ concern about the basis and details. Critics also raised questions about taxation, delegation of authority, and whether the payment sat comfortably with the purpose of export controls; those concerns were not judicial rulings.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteCould it become a precedent?
There were at least three plausible readings of the reported arrangement:
- A one-off exception: A temporary political compromise tied to a particular trade negotiation and a narrow set of chips.
- A new tool of economic statecraft: A model in which export access is exchanged for revenue or other concessions.
- An unstable hybrid: A politically announced framework that might not become a durable, standardized legal program.
The administration could argue that export licenses are discretionary and may carry conditions intended to protect U.S. interests. Critics could answer that charging a percentage of private sales to a designated market is different in kind from ordinary licensing oversight. Without public terms and a clear legal mechanism, it was not possible to know which interpretation would prevail or whether the approach would be repeated.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
