When trade policy rewrites the business case – and why tech strategy can’t ignore geopolitics

The difference between an attractive business case and a multibillion-dollar write-down can be a government decision.

NVIDIA discovered that in April 2025, when the US government introduced new licensing requirements for exports of its H20 AI chips to China. Suddenly, a product with an established customer base and commercial rationale faced a fundamentally different route to market.

As a result, NVIDIA recorded a $4.5 billion charge associated with excess inventory and purchase commitments, and said it was unable to ship a further $2.5 billion of H20 revenue during the quarter.

In effect, a product NVIDIA had specifically designed for the Chinese market following earlier export restrictions had its commercial premise rewritten overnight by another policy decision.

“Trade policy determines which business models are executable.”

Policy changes your route to market

The NVIDIA story didn’t end with the original restriction.

Licences later allowed limited H20 sales into China, generating approximately $60 million. Then, in early 2026, the US granted licences permitting small quantities of H200 chips to specific Chinese customers – but required those chips to undergo inspection in the United States before shipment.

For NVIDIA, that routing matters. The firm states that H200 chips entering the US under the licensing programme become subject to a 25% tariff, which the company may not be able to pass fully to customers.

Policy isn’t sitting outside the business model here. It’s actively shaping the route to market and economics of the product.

For leadership teams, this creates an entirely different perspective. Does an opportunity still make commercial sense once… 

• Compliance changes how the product must move;

• What it costs;
• And whether customers can actually receive it?

Addressable markets are politicised

“The size of a market matters less if policy determines how much of it you can actually serve.”

ASML demonstrates the same issue at market level.

The Dutch semiconductor-equipment manufacturer is a prominent player in the global chip ecosystem. Chinese customers represented 29.1% of its 2025 net sales, down from 36.1% in 2024. 

ASML explicitly identifies “macroeconomic uncertainty – including technological sovereignty and export controls” as risks capable of changing sales volumes, product mix, timing, cost and competitiveness.

That’s a company at the centre of the semiconductor ecosystem effectively acknowledging that geopolitics has become a variable in its commercial model.

For businesses operating in fields like semiconductors, AI, quantum, sensors or advanced computing, market opportunity can’t be evaluated through demand alone. A market of enormous theoretical value is not fully addressable if licensing restrictions narrow the products available to it.

A customer pipeline is less valuable if political intervention can constrain delivery, and a supply chain optimised for cost may be economically fragile if access to critical technology depends on government permission.

Industrial policy changes where investment makes sense

Trade policy can also reach into capital allocation.

In January 2026, the United States and Taiwan announced a trade and investment agreement. Taiwanese semiconductor and technology businesses committed to at least $250 billion in new US investment, supported by a further $250 billion in credit guarantees. The agreement explicitly links trade treatment with expanded US semiconductor, AI, and advanced-manufacturing capacity.

At almost the same time, the US introduced a 25% tariff on certain advanced computing chips (including NVIDIA’s H200 and AMD’s MI325X), while providing exemptions for specified uses that support US technology capacity and domestic manufacturing.

In this case, trade policy is influencing not simply what crosses a border, but the architecture of the industry itself:

1. Where factories are built
2. Where capital flows
3. Which supply chains make sense

“Leadership teams need policy-adjusted business cases.”

A factory with the lowest nominal production cost isn’t necessarily economically superior if tariffs, export controls or industrial incentives move against it.

A technically brilliant product roadmap isn’t commercially robust if its target customer base can be curtailed by government policy.

Trade policy belongs upstream

The takeaway here is this: any international policy assumptions embedded in major commercial proposals should be visible, tested, and defensible. Before approving a market entry, manufacturing location, acquisition, supplier relationship or product roadmap, leadership should understand:

• What regulatory permissions does this business case assume?

• Which markets or customers could become restricted?

• How exposed are the economics to tariff or industrial policy change?

• What happens to expected returns if those assumptions move?

• And, critically, how reversible is the decision?

“Geopolitics need to be built into the investment thesis. Trade policy analysis belongs upstream of major decisions.”

Strong technology businesses build enough visibility to understand where geopolitics touch their commercial model while alternatives are still available. For strategically sensitive technologies, policy is one of the variables from which business strategy should be built.