The United States has not established a simple, codified 25% tax on Nvidia’s China chip sales. President Donald Trump said in December 2025 that Nvidia could export limited quantities of its H200 artificial-intelligence chips to approved Chinese customers while the U.S. received 25% of future sales revenue. But Nvidia’s subsequent SEC filing described a 25% tariff applied when H200 chips entered the United States and said the company had generated no H200 revenue under the licensing program at the time of filing.
That distinction matters. A revenue share, licensing fee, export tax and import tariff are different mechanisms. The policy is therefore real as a reported administration arrangement, but its legal form, implementation and financial consequences remain more complicated than the headline suggests.
The short version
- Trump said the U.S. would receive 25% of future revenue connected with approved Nvidia H200 exports to China.
- Nvidia’s filing did not report a completed 25% government revenue-sharing payment. It referred instead to a 25% tariff upon importation into the United States.
- The earlier H20 arrangement involved an approximately 15% revenue expectation, not the later 25% H200 figure.
- Nvidia reported about $60 million in H20 revenue under licenses, but no H200 revenue under the newer program in the filing cited here.
- Critics say the approach may be legally vulnerable, inconsistent with national-security logic and a dangerous precedent for export controls.
What happened, and which chips are involved?
The dispute began with the Nvidia H20, a lower-performance artificial-intelligence accelerator designed for the Chinese market. On April 9, 2025, the U.S. government told Nvidia that H20 exports to China, Hong Kong, Macau and certain other destinations required licenses. Nvidia recorded a $4.5 billion charge in fiscal 2026’s first quarter related to excess H20 inventory and purchase obligations.
In August 2025, Nvidia received licenses for some H20 sales. The Associated Press reported that Nvidia and AMD had agreed to provide approximately 15% of revenue from certain licensed China chip sales to the U.S. government. Trump said he had initially sought 20% from Nvidia’s H20 sales and later agreed to 15% after discussions with Nvidia CEO Jensen Huang. Nvidia did not publicly confirm the precise quid pro quo, but said it would comply with U.S. export rules.
Crashes, 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 minuteWindows 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 reinstall#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
Nvidia later disclosed approximately $60 million in H20 revenue under those licenses. Its filing said U.S. officials had expressed an expectation that the government would receive 15% or more of licensed H20 revenue, while also noting that no regulation codifying that requirement had been published.
The later controversy concerns the H200, a more capable successor-generation accelerator. In December 2025, Axios reported that Trump would allow limited H200 exports to specified Chinese customers while seeking 25% of future sales revenue for the United States. Nvidia’s filing said the H200 license covered small amounts for specific China-based customers and required inspection by U.S. officials before shipment.
The authorization was not a blanket reopening of the Chinese market. More advanced Blackwell products remained restricted in the reporting cited here. Nvidia’s next-generation Rubin architecture was expected to supersede Blackwell in the second half of 2026, according to Axios.
Timeline of the policy
| Date | What happened |
|---|---|
| April 9, 2025 | H20 exports to China and specified destinations became subject to licensing requirements. |
| Fiscal 2026 first quarter | Nvidia recorded a $4.5 billion charge tied to H20 inventory and purchase obligations. |
| August 2025 | Licenses were granted for some H20 sales; the reported U.S. revenue expectation was approximately 15%. |
| December 2025 | Trump reportedly announced conditional approval for limited H200 exports and a 25% share of future sales revenue. |
| Later Nvidia filing | Nvidia reported approximately $60 million in licensed H20 revenue and no H200 revenue under the new program. |
| August 16, 2026 | The latest verified status in the available record still did not establish a straightforward, codified 25% revenue-sharing tax. |
What could “25% cut” mean?
The phrase sounds clear but can describe several legally different arrangements:
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
| Mechanism | Meaning | Why it matters here |
|---|---|---|
| Revenue share | The government receives a percentage of the seller’s sales proceeds. | This is how Trump’s reported H200 announcement was described. |
| Licensing fee | A payment connected to receiving government authorization. | The administration could characterize a payment this way rather than as a tax. |
| Export tax | A levy imposed on goods leaving the country. | Critics say a percentage of export revenue could function like one. |
| Import tariff | A charge imposed when goods enter a country. | Nvidia’s filing specifically described a 25% tariff upon H200 importation into the United States. |
Those descriptions should not be merged. Trump’s political statement about receiving 25% of future sales revenue and Nvidia’s corporate disclosure about a 25% tariff may be related parts of the same policy, but the public record cited here does not prove that every China sale creates a direct 25% payment to the U.S. Treasury.
Why Nvidia wants to keep selling in China
Nvidia’s commercial argument is that a total ban could sacrifice the Chinese market without eliminating Chinese demand for advanced computing. Chinese customers might instead buy from Huawei or other domestic suppliers, giving those companies more revenue, data-center experience and software development momentum.
Nvidia also has an interest in maintaining its broader hardware-and-software ecosystem. Customers that continue using Nvidia accelerators may continue developing with its CUDA software platform, retaining skills, applications and supplier relationships that can support future sales if restrictions change.
From that perspective, accepting a 15% or 25% economic reduction could be preferable to losing the market entirely. Nvidia might also pass some costs to customers through higher prices, although its filing warned that it might not be able to pass the full cost through. The H200 is also older than Nvidia’s Blackwell and upcoming Rubin generations, which may make limited China sales less costly than permitting access to the company’s newest products.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →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
That is a commercial strategy, not proof that Chinese buyers will choose Nvidia. China could reject or restrict the chips for political and security reasons, favor domestic alternatives, or use Nvidia hardware while accelerating work on substitutes.
Why critics say the arrangement makes no sense
1. The constitutional question
Article I, Section 9 of the U.S. Constitution says that “No Tax or Duty shall be laid on Articles exported from any State.” Derek Scissors of the American Enterprise Institute argued, as reported by the AP, that taking a percentage of export sales looks functionally similar to an export tax.
That does not automatically prove the arrangement unconstitutional. The central issue is how the charge is structured. A government payment described as a licensing condition or fee for an authorization may be defended differently from a tax imposed on exported goods. The unresolved questions include whether the executive branch has authority to impose the payment without Congress, what service the fee pays for and whether the amount is reasonably related to that service.
2. It can turn national security into a price list
The basic contradiction is straightforward: if an AI chip is too risky to export, why should payment make it safe? Critics argue that national-security controls should be based on the product’s capabilities, destination, customer and intended use—not on how much a company is willing to surrender.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →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
Members of both parties on the House Select Committee on China objected that the arrangement could signal to China and U.S. allies that export restrictions are negotiable. A company could appear to buy permission for a product that the government has otherwise considered strategically sensitive.
3. It creates a pay-to-play precedent
Analysts cited by TIME raised the possibility that other companies could seek similar deals. If the approach spreads, future export-control decisions might become commercial negotiations: a restricted product could be approved if the seller accepts a sufficiently large payment.
That could make policy less predictable for manufacturers, customers and allies. It could also weaken the credibility of restrictions on more advanced chips. A company planning factories, inventories and customer support needs to know whether rules will be determined by stable technical criteria or renegotiated case by case.
4. It may damage U.S. companies even when sales are allowed
Nvidia’s filing warned that export controls could cause customers to switch to suppliers based in China, Europe or Israel. Compliance obligations, inspections, end-use monitoring, customer vetting, tariffs and the risk of future rule changes may make Nvidia products less attractive even when a license exists.
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
Chinese customers may also view U.S. hardware as politically unreliable. They could decide that investing in domestic alternatives is safer over the long term, reducing the value of Nvidia’s temporary access.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What Nvidia’s filing actually shows
Nvidia’s disclosures provide a more cautious picture than the headline “25% cut” suggests:
- H20 sales were disrupted. The April 2025 licensing requirement contributed to a $4.5 billion charge tied to H20 inventory and purchase obligations.
- Licensed H20 revenue was limited. Nvidia reported approximately $60 million in H20 revenue under licenses granted in August 2025.
- No H200 revenue had been reported. Under the H200 licensing program, Nvidia said it had generated no revenue at the time of the cited filing.
- The H200 authorization was narrow. It applied to small amounts for specified China-based customers and required inspection before shipment.
- The tariff could reduce competitiveness. Nvidia warned that it might not be able to pass the full cost to customers and that the rules could lead to litigation or additional expenses.
- The rules could change again. Nvidia warned that U.S. export controls might be expanded, modified or replaced.
These details are important for investors and other readers interpreting the policy’s financial impact. A license is not the same as a completed sale, and an announced arrangement is not the same as a final regulation or settled accounting treatment.
The key strategic trade-off
The policy comes down to a difficult question:
Does allowing a controlled sale create more U.S. strategic value than banning the sale entirely?
Free tools Windows power users keep installed
One-click scans. No signup required.
Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Supporters can argue that a limited sale:
- keeps Nvidia connected to Chinese developers and customers;
- preserves U.S. influence over part of the AI hardware and software stack;
- generates revenue for a U.S. company;
- reduces the chance that Huawei and other Chinese suppliers take the entire market; and
- limits access to older or less capable products rather than Nvidia’s newest systems.
Opponents can argue that a sale:
- gives Chinese companies useful AI computing capacity;
- could support military, surveillance or other sensitive applications;
- encourages China to use the chips while developing replacements;
- signals that restrictions can be negotiated for money; and
- undermines confidence in U.S. suppliers among customers elsewhere.
Neither side can resolve that debate simply by pointing to the 25% number. The relevant comparison is not “25% of revenue versus 0%.” It is the long-term value of controlled access versus the security and competitive risks of supplying the market at all.
What remains unresolved
As of August 16, 2026, the public record summarized here did not answer several practical and legal questions:
- Has the 25% payment been formally codified in a regulation, final license condition or other enforceable instrument?
- Who legally owes the payment—the exporter, importer, customer or another party?
- Is the percentage calculated from gross sales revenue, profit, a customs value or another amount?
- Does it apply to every H200 transaction or only specified licenses, customers and quantities?
- Can Chinese authorities block or limit imports even after the United States grants a license?
- Could the arrangement be challenged in court as an unauthorized fee or unconstitutional export tax?
- Will a similar approach apply to AMD, Intel or other semiconductor companies?
Those questions determine whether the arrangement is a workable policy or mainly a political announcement layered onto an already complex licensing system.
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.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
