The semiconductor chokepoint — Taiwan, TSMC, and great power risk

Modern economies don't just need chips. They need the latest chips — the 3-nanometer and 5-nanometer logic devices that power AI accelerators, advanced smartphones, automotive computing, military signal processing. Those chips are made at scale in one place: Taiwan's TSMC.

Samsung and Intel are years behind on leading-edge yield. The Netherlands' ASML makes the EUV lithography machines required — and ships them slowly. Even with unlimited capital and political will, replicating TSMC's ecosystem outside Taiwan takes a decade.

This is why every major power has been racing to onshore chip production — the US CHIPS Act, the EU Chips Act, Japan's Rapidus consortium. None of these will replace TSMC. They aim to ensure that if TSMC went offline, domestic critical needs (defense, infrastructure) could still be met.

From Beijing's vantage, Taiwan's chip dominance is leverage that grows less useful by the year. The PRC's own SMIC has crossed 7nm; restrictions have slowed but not stopped progress. From Washington's vantage, the vulnerability cuts the other way: the longer Taiwan is the world's fab, the higher the cost of letting it change hands.

The economic incentive on both sides is to avoid catastrophe. The political incentives are less stable.

­The leading-edge concentration is even sharper than the headline share suggests once the foundry-design split is decomposed. The most advanced process nodes — currently 3-nanometer and 2-nanometer — are produced exclusively by TSMC in Taiwan and, with a lag of one to two technology generations, by Samsung Foundry in South Korea. Intel Foundry Services, which Pat Gelsinger committed to bringing back to leading-edge competitiveness, has slipped its 18A schedule and is years from contesting TSMC's volume at the most advanced nodes. The legacy-node share is more diversified — China's SMIC and YMTC, US-based GlobalFoundries, European players like ST and Infineon — but legacy-node chips are not where the AI accelerator or smartphone application processor demand sits.

The upstream equipment dependency is the second binding chokepoint. ASML in Veldhoven holds an effective monopoly on extreme-ultraviolet lithography, the technique required for everything at and below 7-nanometer. ASML's machines cost roughly $200 million each for EUV and over $400 million for the next-generation high-NA systems, and the company ships them at the rate of dozens per year rather than hundreds. The Dutch government's export-control alignment with the United States, formalised through 2023 and 2024 amendments, now prevents the most advanced EUV machines from being shipped to Chinese customers — a constraint that translates directly into the Chinese mainland foundry industry's ability to push past the 7-nm process barrier.

Chris Miller's *Chip War* (Scribner, 2022) provides the canonical history of how this geography emerged. The book's argument is that the present concentration is the result of forty years of outsourcing, capital-cost compounding, and process-knowledge accumulation that cannot be replicated through subsidy alone, no matter how large the subsidy. The US CHIPS and Science Act of 2022 appropriated $52 billion in direct funding for domestic fab construction; the EU Chips Act allocated roughly €43 billion; Japan has committed comparable sums via the Rapidus consortium and TSMC's Kumamoto site. These programmes are large in absolute terms and small relative to TSMC's annual capital expenditure of $30-40 billion, all of which is already deployed against process roadmaps the subsidy recipients cannot match.

The strategic implication is asymmetric. If a Taiwan Strait contingency interrupted TSMC's output, the immediate global effect would not be a clean substitution from Samsung or Intel — it would be a multi-quarter rationing event in which the available leading-edge capacity is allocated to the highest-priority defense and infrastructure uses, with consumer-electronics, automotive, and general-purpose computing demand left to absorb most of the shock. The economic estimate associated with Rhodium Group's 2022 work on this scenario put global GDP impact in the range of 5-10% over a two-year window, with the largest absolute losses falling on the Taiwan economy itself and the largest GDP-share losses falling on advanced-economy electronics-exporters.

Beijing's calculation runs the other direction. The cost to China of a contingency that disrupts TSMC includes Chinese semiconductor-dependent industries — automotive, telecommunications equipment, consumer electronics — being cut off from the chips that currently constitute their critical inputs. SMIC's progress past 7-nm using DUV lithography is the most-watched indicator of how rapidly that constraint is loosening; the consensus among industry analysts (Doug Fuller at Carnegie Endowment, Lennart Heim at RAND) is that the upstream-equipment constraint binds harder than the process-knowledge constraint, and binds for a long time. The geopolitical implication is that the chokepoint cuts both ways and that the disincentive to disrupt the existing arrangement is substantial on both sides — which is exactly the deterrence logic that the present equilibrium rests on.

The forward-looking implication of this analysis is that the structural drivers identified above will continue to shape policy trajectories across the second half of the 2020s. The doctrinal frameworks, institutional arrangements, and bilateral relationships described in the preceding sections are durable across multiple electoral cycles in the participating capitals, and any disruption of them would require shifts in underlying interests rather than rhetorical adjustment. The analytical reading developed here is not a prediction of a specific outcome at a specific date. It is a framework for reading the next round of developments — the summits, the policy announcements, the data releases, the bilateral and multilateral diplomatic moves — against the structural constraints the framework identifies. Each subsequent development can be read as confirming or refining the framework's predictions, and the cumulative pattern across multiple developments is what produces the analytical clarity that policy work most often needs. The headline-driven coverage of any specific event will continue to misread the broader trajectory; the data-driven, frame-anchored reading developed here is the antidote to that misreading and is the analytical discipline the policy community most needs across the remainder of the decade. The arithmetic of the underlying interests does not change quickly. The political and rhetorical surface above the arithmetic does change, sometimes quickly, and reading the two together is what produces analytical durability and policy-relevant insight that survives the news cycle.

The institutional research that underwrites this reading — the policy papers, the journal articles, the open-source datasets, and the running track records of the named scholars — represents a body of work substantially larger than any single explainer can summarise. Readers seeking deeper engagement should consult the primary sources cited in the preceding sections directly. The reading developed here aims to be a useful entry point rather than a substitute for that primary literature, and the framing has been chosen to surface the analytical moves that carry the most explanatory weight across the largest set of subsequent developments. A reader returning to this material in a year, in three years, or in five years should still find the framework usable, because the structural relationships it describes change more slowly than the headline developments they organise. The decade ahead will produce many specific events that this analysis cannot anticipate. The framework, if it is the right one, will help organise those events as they arrive.