US chip export controls are rules restricting the sale of advanced semiconductors, chipmaking equipment and related software to designated countries, above all China. Administered mainly through the Commerce Department’s Entity List and licensing regime, they aim to slow a rival’s progress in artificial intelligence and advanced computing, and they have turned product specifications into instruments of foreign policy.
How the regime works
The controls set technical thresholds, on computing performance, interconnect speed and manufacturing capability, above which exports need licences that are often denied. Chipmakers respond by engineering products just under the lines, regulators respond by moving the lines, and licences can be granted selectively, making revenue from an entire market a policy variable. Allies matter too: controls on lithography and tooling only bind because Dutch and Japanese suppliers apply parallel rules.
Market consequences
For investors the controls create a revenue cliff and a licence lottery in the world’s second-largest chip market, force duplicate product lines, and accelerate China’s drive for domestic substitutes, seeding future competitors. They also create headline risk in both directions, because loosenings can add billions of addressable market as quickly as tightenings remove it. Every AI-supply-chain earnings call now carries an export-policy paragraph.
Reading the headlines
Three questions organise any development: which products and thresholds are affected, whether licences will actually be granted, and how the target responds, through stockpiling, substitution or retaliation. The stock reaction usually keys off the first; the strategic consequence usually hides in the third.
Where you’ll meet this in our coverage
Nvidia H200 China Exports Begin: Trivial Volumes, a Major Policy Turn