On April 16, 2025, semiconductor shares slid after Nvidia and AMD disclosed that new U.S. export-license requirements could make China-bound AI chips harder to sell. Nvidia initially estimated up to $5.5 billion in charges tied to its H20 accelerator; AMD warned of up to approximately $800 million for its MI308. Those estimates were accounting charges tied to inventory and commitments—not equivalent amounts of lost sales. The longer-lasting concern was whether the companies could keep serving China’s AI market at all.
What changed in April 2025
The immediate catalyst was a new U.S. licensing requirement. Nvidia said the U.S. government informed it on April 9, 2025, that exports of its H20 integrated circuits and certain other circuits meeting specified memory-bandwidth or interconnect-bandwidth characteristics would require a license for destinations including China, Hong Kong, Macau and designated D:5 countries. On April 14, Nvidia said it was told the requirement would remain in effect indefinitely. Nvidia’s April 2025 filing describes the notification and its initial estimate of the financial impact.
AMD disclosed on April 15 that a new license requirement applied to its Instinct MI308 products. Its filing covered exports to China, including Hong Kong and Macau, and D:5 countries, as well as certain companies headquartered in or ultimately owned by entities headquartered in those jurisdictions. AMD’s filing set out the affected product and its potential charge.
The companies described a license requirement, not a blanket legal ban on every chip they make or every shipment to China. A license requirement can have a similar commercial effect when approvals are unavailable, but the distinction matters: licenses can be granted selectively, and later developments showed that some shipments did resume.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#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 these two chips were affected
Nvidia’s H20 was a China-focused AI accelerator designed to meet earlier U.S. export-control limits. The new rule covered H20 and other circuits with specified technical characteristics. AMD’s MI308 is an Instinct data-center GPU. The disclosed restrictions were product- and destination-specific; they did not, by themselves, establish that all Nvidia or AMD products were barred from China.
What the announced charges meant
The figures were not government fees or direct forecasts of sales lost. They reflected the accounting consequences of inventory, supplier purchase commitments and related reserves that became more difficult to use or sell once China-bound shipments required licenses.
- An accounting charge reduces the recorded value of inventory or recognizes obligations associated with commitments and reserves.
- Lost revenue is sales that may not happen if a product cannot be shipped or a customer switches suppliers. It is a separate commercial risk, not the same figure as an inventory charge.
- Cash impact can differ from the accounting charge. The timing and amount depend on supplier commitments, production, reuse or redirection of materials, and whether products are eventually sold.
Nvidia initially estimated up to $5.5 billion in H20-related charges. It later reported a $4.5 billion charge in fiscal Q1 2026, saying some materials could be reused. Its filing also reported $4.6 billion in H20 sales in that quarter before the new licensing requirements. The quarterly sales and the charge measure different things and should not be treated as interchangeable. Nvidia’s fiscal Q1 2026 filing gives both figures and the explanation for the lower charge.
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
AMD initially warned of up to approximately $800 million in charges and later recorded approximately $800 million in MI308-related inventory and related charges in the second quarter of 2025. These amounts signaled exposure to a sudden change in what could be shipped, not a dollar-for-dollar measure of AMD’s sales loss.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Nvidia and AMD: initial estimates and later outcomes
| Company | Affected product | Initial April 2025 disclosure | Later reported outcome |
|---|---|---|---|
| Nvidia | H20 and certain comparable circuits | Up to $5.5 billion in inventory, purchase-commitment and related reserve charges | $4.5 billion H20-related charge in fiscal Q1 2026; some materials could be reused |
| AMD | Instinct MI308 | Up to approximately $800 million in inventory, purchase-commitment and related reserve charges | Approximately $800 million recorded in Q2 2025; approximately $360 million later reversed after some licenses and shipments |
The figures come from the companies’ filings: Nvidia’s April disclosure, AMD’s April disclosure, Nvidia’s fiscal Q1 filing and AMD’s later annual filing.
Why semiconductor shares beyond the two companies fell
Contemporaneous coverage on April 16 reported Nvidia and AMD down roughly 6% or more during the selloff and the VanEck Semiconductor ETF down more than 4% in the reported session. These are reported market moves, not verified closing-price returns. The contemporaneous report also described declines among other semiconductor names.
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
The spillover reflected uncertainty across the industry, not proof that every falling company had direct H20 or MI308 exposure. Investors were weighing the chance of more export controls, difficulty obtaining licenses, disruption to supplier commitments, weaker Chinese AI-infrastructure spending and the possibility that Chinese accelerator suppliers would win customers. Broader trade-policy uncertainty also contributed to the market backdrop; it should not be confused with the specific chip-license requirements.
Why the risk went beyond a one-time write-down
The charge was the visible near-term cost. The strategic question was whether U.S. chipmakers could reliably sell advanced accelerators into China, a market where customers were building AI infrastructure. Nvidia’s previously restricted China opportunity was already smaller than before the October 2023 controls, but the H20 had offered a route to serve some demand within the earlier limits.
If customers cannot count on access to a product, they may redesign systems around alternatives. That can help domestic Chinese accelerator vendors develop customer relationships and software ecosystems. Nvidia warned that restrictions could strengthen competitors’ ecosystems, a risk that can extend beyond the immediate revenue from any one product. At the same time, China exposure is only one part of either company’s global business; the disclosures do not establish that worldwide demand for AI accelerators, networking or data-center systems disappeared.
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
AMD’s smaller potential charge still mattered to investors because it showed the policy could affect more than one U.S. supplier. It also illustrated the fragility of designing a product for a restricted market: compliance with one set of limits did not guarantee lasting access when rules changed.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What happened after the April selloff
Nvidia: selective H20 licenses and limited H200 authorization
Nvidia later said the U.S. granted licenses in August 2025 allowing certain H20 shipments to certain China-based customers. It reported approximately $60 million in H20 revenue under those licenses. In February 2026, the U.S. granted a license allowing small amounts of H200 shipments to specified China-based customers, subject to inspection and other conditions. These were limited authorizations, not a general reopening of the market.
Nvidia also said U.S. officials had expressed an expectation that the government receive 15% or more of revenue from licensed sales. The company said no regulation codifying that expectation had been published. It should therefore not be described as an enacted tax, fee or formal royalty. These developments are detailed in Nvidia’s 2026 filing.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteBest 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
AMD: some MI308 licenses and a partial charge reversal
AMD said it obtained some licenses to ship MI308 products to certain China-based customers and began shipments in the fourth quarter of 2025. It reversed approximately $360 million of its earlier charges. That reversal indicates some value became recoverable; it does not mean the original exposure disappeared or that all planned shipments were approved. AMD also identified Chinese import controls, customer demand and U.S. licensing as factors affecting future sales. AMD’s later annual filing describes the licenses, shipments and reversal.
How investors can assess the next export-control headline
A single charge does not answer whether the long-term business impact is manageable. Investors can separate the issues by asking:
- How large is the charge? Compare it with the company’s quarterly revenue, gross profit and cash flow, while keeping the accounting period clear.
- Is it recurring? A one-time inventory adjustment is different from repeated restrictions that continually limit market access.
- Can value be recovered? Look for evidence of component reuse, redirected inventory, licenses and completed shipments rather than assuming inventory is either fully lost or fully salvageable.
- What is the strategic cost? Consider whether customers are moving to alternatives and whether competitors are building durable software and developer ecosystems.
Licenses are only one condition for a sale: Chinese import rules and customer demand can also matter. A product may remain technically exportable yet lose commercial relevance if approvals are uncertain or customers commit to alternatives. The April 2025 shock therefore cannot be judged by the charge alone; it also exposed the risk of treating China access as dependable for advanced U.S. AI chips.
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.
