How Export Controls Can Propagate Upstream Through a Supply Chain
August 6, 2026
Altsets
Research by Altsets Research
Trace regulatory demand shocks from a restricted customer product into economically exposed suppliers while keeping product relevance and purchase commitments separate from company-wide exposure.
Data used:Altsets Supply Chain Intelligence: 90k+ entities, 400k+ relationships, 20+ years of history.
Key findings
- Nvidia recorded a 4.5B USD H20 excess-inventory and purchase-obligation charge in fiscal 2026 and another 0.4B USD H200-related charge in the first half of fiscal 2027 as export restrictions affected demand.
- Nvidia represents 27.88% of SK Hynix revenue and 17.62% of Micron revenue in the displayed relationship data, making both suppliers logical follow-up targets while product-specific impact remains a separate question.
Trace an export-control shock upstream by defining the restricted product and market, identifying the affected company, ranking its mapped suppliers by economic exposure, and verifying which suppliers actually serve that product. The regulation may target one customer-facing product while the demand and inventory effects reach selected suppliers rather than the entire upstream network.
The regulation starts at Nvidia, but the research does not stop there
Altsets maps Nvidia as a customer of both SK Hynix and Micron. The displayed supplier-revenue exposure is Nvidia at 27.88% of SK Hynix revenue and Nvidia at 17.62% of Micron revenue. Those percentages are large enough to make Nvidia-specific demand relevant to both suppliers.
They do not prove that H20 or H200 restrictions affect the entirety of either relationship. That product match has to be established separately.
Export controls are not the same as tariffs
A tariff increases cost on an allowed trade flow. An export control can prevent or restrict the transaction itself. That difference matters. Nvidia's filings describe products that could not be sold into intended markets, leading to excess inventory and purchase-obligation charges. The economic mechanism is therefore:
- product access is restricted;
- expected demand falls or shifts;
- inventory becomes harder to sell;
- purchase commitments may become excessive;
- upstream order requirements can change.
That is a different propagation path from a tariff pass-through model.
The inventory charge reveals committed supply risk
Nvidia's fiscal 2027 filing reported excess inventory purchase obligations of 2.138B USD at July 26, 2026. The company also disclosed inventory and excess-purchase-obligation provisions and explained that the H20 and H200 restrictions contributed to large charges. That matters for supplier research because customer commitments can remain economically relevant even after downstream demand changes.
A supplier can therefore experience a delayed rather than immediate impact. Existing purchase commitments may cushion the first period. Future orders can still weaken.
Relationship percentages identify where to look next
The proprietary exposure does not tell us which supplier provided which H20 or H200 component. Its value is prioritization. If Nvidia-specific revenue represents 27.88% of SK Hynix revenue and 17.62% of Micron revenue in the displayed data, then a large Nvidia demand-policy shock belongs in the research process for both companies.
The next questions are Is the affected Nvidia product relevant to this supplier? Are the components fungible across other Nvidia products? Can the supplier redirect output? Are there long-term purchase commitments? Does demand outside the restricted geography offset the loss? Is the product transition already moving to another architecture?
Product substitution can weaken the propagation
A regulatory shock at one Nvidia product does not necessarily reduce total Nvidia demand. Customers can shift to another approved product. Nvidia can redesign products.
Demand can migrate geographically. Other Data Center products can grow. The supplier may serve multiple Nvidia architectures. That is why upstream exposure should be treated as a conditional pathway rather than a mechanical revenue loss.
Export controls can also create competitor effects
Nvidia's filing warns that restrictions can encourage investment in foreign competitors that are less constrained by US rules. That creates a second-order research question. A regulation can simultaneously reduce sales of the restricted product, shift demand to a substitute architecture, alter supplier mix, accelerate local alternatives, change inventory commitments, and change geographic production decisions. A relationship graph can be expanded as those new counterparties emerge.
A repeatable export-control workflow
- Define the restricted product and market.
- Quantify the direct financial effect at the regulated company.
- Map economically important upstream suppliers.
- Rank them by supplier-revenue exposure to the regulated customer.
- Validate whether their products participate in the affected architecture.
- Review purchase commitments and inventory.
- Check substitution and redirection possibilities.
- Monitor competitor and geographic shifts.
- Keep product-specific conclusions separate from company-wide relationship percentages.
- Re-run the network after the policy response changes commercial relationships.
The tariff stress-test guide models cost changes when trade still occurs. This article addresses a different mechanism: what happens when regulation changes whether or where the product can be sold at all. The segment-exposure guide explains why company-level supplier percentages should not be assigned directly to one restricted product. For relationship methodology, read the Altsets supply-chain data methodology.