Export controls are a growth constraint. Treat them like one.

Every growth plan contains assumptions: how much demand exists; what customers will pay; which markets offer the strongest opportunity; and what investment will be required.

For tech firms, when those assumptions collide with export controls, the question is: can we supply our technology on the terms our growth plan assumes?

“Addressable markets have regulatory boundaries.”

Because export controls can influence:

• Which version of a product reaches a market
• Which customers receive it
• How long a sale takes
• What technology can be transferred
• Whether the economics still make sense

Export controls affect the product roadmap…

Following earlier US restrictions on advanced AI chips, NVIDIA developed its H20 processor specifically for the Chinese market within the applicable export control thresholds. Then, in April 2025, the US government announced that exports of H20 chips to China would require a licence.

NVIDIA subsequently recorded a $4.5 billion charge associated with excess H20 inventory and purchase commitments, while saying it had been unable to ship an additional $2.5 billion of H20 revenue.

NVIDIA later said the prevailing rules and geopolitical environment left it unable to create and deliver a competitive data-centre product for China that could satisfy both US and Chinese government requirements.

AMD faced a similar problem with its MI308 processors. The same licensing requirements introduced in April 2025 led to approximately $800 million in inventory and related charges, although the company later secured licences allowing some shipments to Chinese customers.

“When regulations determine which version of a product reaches a market, export control is already a growth decision.”

…and the wider growth plan

Exposure reaches beyond the final shipment. For instance, leadership must consider:

• Market entry: can we serve this market with the technology we intend to sell?
• Product roadmap: could specifications change classification or licensing requirements?
• Sales pipeline: which customers, destinations, or end uses introduce licence dependency?
• R&D and partnerships: what technology or know-how will collaborators need to access?
• Forecasting: how much revenue or inventory depends on an authorisation remaining available?

Seeing these dependencies early gives you more options.

A product specification may be changed. A sales team can qualify an opportunity differently. Licensing lead time can be reflected in the pipeline. Strategy and horizon scanning can help leadership anticipate regulatory shifts earlier, and investment can be directed towards markets with a more durable route to revenue.

But once the product is built, inventory committed, and customer expectations set, the available choices narrow quickly.

Licensing sits upstream of revenue

Software companies encounter similar commercial pressures. On 29 May 2025, Synopsys received notice of new US export restrictions covering certain electronic-design-automation software and technology supplied to China.

In response, Synopsys suspended financial guidance for both its third quarter and the full financial year while it assessed the potential impact on the business, its operating results, and financial condition. The restrictions were subsequently rescinded in July.

No container had to stop at a port for export controls to reach the revenue forecast.

ASML shows another dimension. Following expanded restrictions on advanced semiconductor manufacturing technology in late 2024, the company said it expected its China business to account for around 20% of total 2025 net sales, while confirming that export controls were affecting the expected geographical mix of its business.

“Export controls reach into product strategy, forecasting, and geographic growth.”

Bring the constraint forward

The practical response is to treat export control assumptions like pricing, production capacity, financing, or component availability: as inputs into the commercial model.

Before committing heavily to a market, customer, or product, firms should understand:

• How is our technology classified?
• Which destinations, customers, or end uses introduce licensing requirements?
• What lead time should Sales build into the deal?
• Could technical specifications change the control position?
• Which partnerships require controlled technology or knowledge to be shared?
• How much of our forecast revenue depends on an authorisation remaining available?

The earlier these questions enter the growth process, the more commercial options you retain.

A constraint discovered during product development may still be designed around; one discovered after inventory has been committed and customers are waiting is much harder to absorb.