Chip controls tighten around the thinnest wafers

On 14 April the United States extended its export controls to cover the extreme-ultraviolet lithography systems used for 2-nanometre chips, and added a dozen fabrication facilities in three countries to its entity list. The rule itself is short, a tightening of definitions and a widening of jurisdiction. The consequences are anything but short, because they land on an industry that had already rearranged itself around the previous round of controls, and the rearrangement is now incomplete.

The background is a pattern that has repeated since 2022. Each round of restrictions closes the routes that Chinese chipmakers used to obtain advanced equipment, and each round is followed by a quieter contest over the loopholes: third-country transfers, front companies, second-hand machines, and the slow accumulation of domestic tooling that makes each control less decisive than the last. What makes this round different is the object itself. At 2 nanometres, the lithography is no longer just expensive, it is nearly monopolized, with a single supplier in the Netherlands and a handful of buyers worldwide, which means the control is less about stopping one country than about governing the entire market structure of the most advanced node.

The Chinese response has been a familiar combination: formal protest, followed by a renewed push for domestic alternatives, followed by the quiet purchase of every legal machine that still qualifies. The domestic lithography programme is further along than most outsiders assume and further behind than its own press releases suggest, and the gap between those two realities is the space in which the whole sanctions game is played. The controls do not stop Chinese progress so much as price it, forcing the domestic industry to spend years and billions closing a gap that would have narrowed anyway, but more slowly.

There is a cost to this policy that American debates rarely price in. Every round of controls teaches the rest of the world that supply chains built on one country’s equipment are leverage that can be pulled at any time, and the lesson is being learned in Seoul, in Tokyo, in Taipei and in every European capital that hosts a fab. The long-term effect is not that China becomes self-sufficient, which remains doubtful, but that the entire world diversifies away from the tools that the United States controls, which is a different and more corrosive outcome. Sanctions that work perfectly create their own replacement industry.

None of this is an argument that the controls are wrong. There are capabilities that any government would be reckless to sell, and the security case here is real. It is an argument about expectations: the most advanced node in the world is not a faucet that can be turned off cleanly, it is a system that will find its own equilibrium, and the equilibrium after a decade of tightening looks like two separate supply chains with different speeds, different prices and different customers. That world is now being built, one export licence at a time, and the paragraph published on 14 April is just the latest brick in it.