Written for geopolitical-risk and economic-exposure work: it traces a mechanism, not a market view.
Advanced logic manufacturing is concentrated in Taiwan to a degree with no parallel, and that concentration is a present structural exposure rather than a scenario. What a Taiwan-specific interruption would do is not documented and is graded as an unconfirmed scenario throughout.
How to read the grades
- ConfirmedDocumented as having occurred, with sources.
- Plausible exposureA mechanism Vigil assesses as likely; not documented as having occurred.
- Unconfirmed scenarioNamed because it is worth watching. Not asserted.
A step can never be graded more firmly than the step it depends on: a consequence cannot be better established than its cause. That rule is enforced when this site is built, not applied by hand — a chain that broke it would fail the build rather than publish.
- Trigger
The present concentration of advanced semiconductor manufacturing in a single geography, and the dependencies that concentration creates
- Affected asset, route or regionConfirmed
The large majority of the world's most advanced logic chips are made in a cluster on Taiwan's western plain. The concentration is a documented present fact, though the precise geographic share is not a published figure: what can be stated is that TSMC held about 73 per cent of the pure-play foundry market by revenue in Q2 2026, and that processes at 7nm and below were roughly 74 per cent of its own wafer revenue in Q1 2026 — a company share of a market segment and a share of one firm's revenue respectively, neither of which converts into a share of world capacity by location. The concentration exists alongside a second fact: the cluster depends on imported specialty materials, imported energy, continuous high-quality power and open air and sea freight. Capacity is being built in the United States, Japan and Europe — TSMC's Arizona programme is $265bn across six logic fabs, of which one has run high-volume N4 production since late 2024 and the next targets N3 volume production in the second half of 2027 — while Taiwan-based advanced capacity expands in parallel. Diversification is real, measured in years, and is not the same as the concentration falling.
Documented as having occurred, with sources.
- Operational disruptionConfirmed
The cluster has already been stressed by causes unrelated to the strait, and those episodes are the available evidence. A severe drought in 2021 constrained water supply to fabrication and required trucked delivery and priority allocation. Export controls from 2022 reorganised the industry's trade geography and made the cluster a subject of policy in several capitals simultaneously. Neither was a security event, and both demonstrated how sensitive the system is to inputs that are not chips.
Documented as having occurred, with sources.
- Exposed sector or commodityConfirmed
The exposed parties are industries with no connection to the dispute. Automotive, consumer electronics, industrial equipment, medical devices and defence manufacturing all stop for a missing component regardless of what the component costs, which is the mechanism that makes semiconductor supply propagate faster than its share of trade suggests. The 2020–2023 global shortage — caused by pandemic demand shifts, not by anything in the strait — is the empirical record of how that propagation behaves, and it reached vehicle production lines on other continents within months.
Documented as having occurred, with sources.
- Broader economic significanceUnconfirmed scenario
What a Taiwan-specific interruption would do is not documented and Vigil does not assert it. Named as worth watching rather than asserted: the 2020–2023 shortage arose from demand shifts against continuing supply, whereas an interruption at the source would be a different mechanism, and no published analysis establishes how the two compare. The observations that can be made are narrower and firmer — diversification reduces exposure at trailing nodes faster than at leading ones, inventory buffers in dependent industries are measured in weeks rather than quarters, and the affected parties would include states with no position on the political question. Vigil publishes no estimate of magnitude, duration or cost, and their absence is deliberate.
Named because it is worth watching. Not asserted.
Sectors and commodities exposed
Named as plain labels rather than a controlled vocabulary, so this list cannot drift from the commodity names the module's economy section already uses.
What remains unknown
- No defensible public figure was found for the share of world sub-7nm capacity physically located on Taiwan. Company market share and company revenue mix are published and are different measures; this chain states them as such and does not convert them into a geographic share.
- No published analysis establishes how a source interruption would compare to the demand-driven 2020–2023 shortage, which is why the final step is graded as an unconfirmed scenario.
- Site-level capacity, yields, customer allocation and inventory positions are commercially confidential and are not published.
- Estimates of the timeline for meaningful leading-edge diversification vary widely between published assessments and are contested.
- Vigil publishes no magnitude, duration or cost estimate for any interruption scenario, and the absence is deliberate rather than a gap awaiting research.
Readings the evidence also supports
- Diversification investment can be read as reducing global exposure to Taiwan, or as reducing everyone else's dependence while leaving Taiwan's own vulnerability unchanged and diminishing whatever deterrent value the concentration carries. Both are argued seriously and they imply opposite conclusions about whether diversification improves stability.
- The 2020–2023 shortage can be read as evidence of how badly an interruption would propagate, or as a poor analogue because its cause was demand rather than supply. This chain records it as the available evidence while grading the inference from it as unconfirmed.
Listed because the record's own assessment is not the only one its sources permit — not as a hedge on the assessment above.
How much leading-edge capacity is actually PRODUCING outside Taiwan?
Announced, under construction and in volume production are three different figures separated by years, and only the third changes the concentration this chain describes. The checkable milestones are TSMC Arizona fab two on N3, targeted for volume production in the second half of 2027, and fab three on N2 and A16 by the end of the decade — against continued expansion of advanced capacity on Taiwan itself, which is the comparison that matters.
Inventory practice in dependent industries
Buffers grew after the 2020–2023 shortage and have since been trimmed. The trend is the clearest measure of how exposed customers actually are.
Export-control changes in any jurisdiction
They reorganise the trade geography faster than any factory can be built, and they now originate in several capitals.
Water and power allocation to the cluster
The 2021 drought showed these are live constraints in ordinary years.
Open in Explore Interactive: select this record and follow what connects to it.
Assessed as of2026-08Last reviewed2026-08-28