Taiwan makes roughly nine of every ten advanced chips in the world. This extreme concentration is, so it is said, its best protection: it is the "silicon shield". No one would dare destroy the factory on which the entire planet depends, and everyone has an interest in defending it. But to reassure Washington and guard against China, the giant TSMC is shifting part of its production off the island. And each exported factory wears the shield a little thinner: in seeking to protect itself, Taiwan weakens what made it indispensable.
1 The sacred mountain
An industrial dominance without equal in recent history.
護國神山
The mountain that protects the nation
In Taiwan, the company TSMC is nicknamed "護國神山", the "sacred mountain that protects the nation". The name is no exaggeration: the island produces about 90% of the planet's most advanced semiconductors, the ones that power artificial intelligence, smartphones and modern weapons, and TSMC alone accounts for nearly 70% of the world foundry market. The concept of the "silicon shield", coined by the journalist Craig Addison in 2001 and taken up by President Tsai Ing-wen in 2021, captures this idea: Taiwan's centrality in the global digital economy protects it better than an army.
Advanced chips
~90%
of world production at the finest nodes (below 7 nm).
TSMC, foundry
~70%
of the world foundry market in 2025 (TrendForce).
2 Deterrence through dependence
The logic of the shield: no one destroys what they depend on.
The weapon of indispensability
Too precious to be attacked
The reasoning is elegant. Because the whole world, China included, depends on Taiwanese chips, an invasion or a blockade of the island would inflict a devastating shock on the global economy, estimated by some at around $10 trillion in the first year. Beijing would lose its access to the most advanced chips; Washington and its allies, having an economic as much as a strategic interest in the stability of the strait, would be prompted to defend the island. Mutual dependence would serve as deterrence: you do not break the machine you need.
A deterrence, not a guarantee
The argument has its force, but also its limits, which we will revisit. The shield has never stopped a missile: it alters a cost-benefit calculation. And it rests on a single condition: that the concentration of chips remains in Taiwan. Anything that dilutes that concentration mechanically weakens the island's deterrent power.
3 The hollow prize
A conquered factory would be worth nothing.
The "broken nest"
You do not capture a fab the way you capture a mine
The shield has a second, more subtle mechanism: even seized by force, a Taiwanese factory would be a "hollow prize". A cutting-edge foundry does not run on its own: it depends on foreign inputs, in particular the EUV lithography machines of the Dutch firm ASML, of which Taiwan makes none, and on materials supplied for half by Japan; it also depends on engineers and technicians who would not stay under occupation. American strategists have even theorized a "broken nest" policy: making the island not only useless but costly to conquer, so as to deny Beijing any industrial gain.
Upstream dependence, another lock
This point reveals a truth about the shield: it rests not only on TSMC, but on an entire chain in which Taiwan is only one link, central though it is. ASML holds 100% of EUV, Japan about half of the materials. It is this interdependence that makes the fab "uncapturable": the wealth is not in the walls, it is in the network that feeds them.
4 The great relocation
To diversify the risk and reassure the ally, TSMC settles elsewhere.
Off the island
Arizona, Kumamoto, Dresden
Under the twin pressure of the Chinese threat and American demands, TSMC is expanding beyond Taiwan. In Arizona, its commitment rose from $65 billion to $165 billion in March 2025, presented as the largest foreign direct investment in U.S. history: three factories in service or planned, already producing 4 nm and aiming for 3 nm and then 2 nm. To these are added a factory in Japan (Kumamoto) and one in Germany (Dresden). An agreement signed in January 2026 sealed the exchange: the United States caps its tariffs on Taiwanese goods in return for at least $250 billion of investments.
TSMC in Arizona
$165 billion
of commitment (from $65 billion to $165 billion in March 2025).
US-Taiwan agreement
Jan. 2026
capped tariff against ≥ $250 billion of investments.
5 The paradox
In protecting itself, the island weakens its protection.
The leverage that erodes
Reassuring Washington means handing it the shield
Here is the heart of the matter. The silicon shield rested on one thing: that production be concentrated in Taiwan, and therefore irreplaceable. By moving it to the United States, one dilutes precisely that irreplaceability. The more America makes its own advanced chips, the less it needs Taiwan; and the less it needs Taiwan, the less pressing its interest in defending the island. The relocation, intended to reassure Washington and spread the risk, therefore transfers part of the leverage off the island. It is the shield that one wears down by moving it.
The shield turned into a bargaining chip
The U.S. Secretary of Commerce, Howard Lutnick, put it bluntly in the autumn of 2025: "if you have 95% [chip production], how am I going to get it to protect you? You're going to put it on a plane? You're going to put it on a boat?". In other words: concentration is no longer presented as a protection, but as an argument for demanding a share. The shield becomes a negotiating card, and the implicit guarantee of defence becomes a quid pro quo.
6 The hollowing-out
In Taiwan, the fear of being drained of its substance.
掏空
"Protection money"
On the island, the debate is heated. The opposition speaks of "hollowing-out" (掏空): by exporting its most precious factories, Taiwan would drain itself of its substance and lose the protection it derived from them. Former president Ma Ying-jeou called the giant Arizona investment "protection money" and a "major national security crisis". The underlying fear: that the island ends up paying twice, by ceding its industry and by losing the shield it constituted, with no firm guarantee in exchange.
The government's reply
Taipei seeks to reassure: the most advanced technology, it says, stays home. The CEO of TSMC has insisted that "the most advanced research centre and the most advanced factories are in Taiwan", where thousands of engineers develop the next generation. The Vice Premier was categorical about a massive relocation: "I told the U.S. side very clearly: it's impossible." The shift would therefore be controlled, not an abandonment.
7 A capped erosion
In fairness: the erosion is real, but slow and bounded by design.
The safeguards
The law locks in the lag, the leading-edge node stays in Taiwan
Honesty calls for qualifying the paradox. First, a rule: Taiwanese fabs abroad may produce only with at least one generation of lag behind the most advanced node made on the island (the so-called "N-1 rule"), written into law. Next, a fact: volume production of 2 nm began in Taiwan first, in late 2025, and 80-90% of TSMC's capacity remains on the island, which is still building a dozen new factories. Finally, an economics: manufacturing in the United States costs, according to founder Morris Chang, from 50% to nearly double the Taiwanese cost, and the "one-hour ecosystem" of the island's technology parks remains irreplaceable in the short term.
An insurance, not a move
The erosion of the shield is therefore real, but spread out and capped by design. Taiwan exports part of its capacity, not its technological core: the leading edge, the research and the ecosystem stay on the island. What is being moved is not the shield itself, it is a margin; and that shift looks less like an abandonment than like a costly insurance, taken out for fear of China and demanded by the American ally.
8 Target or asset?
What if dependence, far from deterring, were an incentive?
The underlying debate
A shield, or a target painted on the roof?
The deepest debate is not about the solidity of the shield, but about its meaning. Several analysts argue that the world's dependence on Taiwanese chips could incite Beijing to seize the island rather than deter it: controlling the most strategic resource of the century is a temptation, not only a risk. Others recall that the Chinese motive is above all political, that it predates the chip industry, and that Beijing could accept enormous economic costs in the name of reunification. On this reading, the concentration would make Taiwan a target as much as a protected asset, and the "shield" would be in part a reassuring myth.
Neither talisman nor illusion
The truth doubtless lies between the two: the silicon shield is neither an infallible talisman nor a pure illusion. It raises the cost of an aggression without making it prohibitive; it draws attention as much as it protects. What is certain is that its logic rests on scarcity: as long as Taiwan remains irreplaceable, the calculation holds; as production disperses, it weakens.
9 The insurance that is spent
It remains to be seen how long the shield will hold.
The paradox owned
Protecting itself by wearing itself down
Taiwan is caught in a dilemma with no easy way out. Concentrating everything keeps the shield up, but exposes the island to losing everything in the event of conflict, and irritates an ally that wants its share. Dispersing reassures the ally and spreads the risk, but wears down the leverage that made the island indispensable. The relocation is therefore an insurance that is spent: each exported factory buys a little immediate security at the price of a little future deterrent power. The shield does not break; it thins, slowly, with each transfer.
The compass
①
The shield rests on scarcity. Taiwan's protection comes from the irreplaceability of its chips: anything that dilutes it weakens the leverage.
②
The paradox is real but capped. The leading edge, the research and the ecosystem stay on the island; the law locks in the lag of foreign fabs.
③
Deterrence or incentive? The debate is not settled: concentration can protect the island or turn it into a target. This sheet sheds light on a debate; it does not offer a forecast.
Morris Chang's word
The founder of TSMC, Morris Chang, summed up the era in a single phrase: "Globalization is almost dead". The silicon shield was a child of globalization, that interdependence which made war too costly for all. As each power wants its own factories, this cement cracks. And the day Taiwan is no longer indispensable, one must hope that other reasons still make it worth defending.
Read alongside: The battle for copper, the other face of sovereignty through strategic resources; and The obscure metal without which weapons fall silent, another chokepoint in a global chain. Reference: abbreviations & acronyms (TSMC, EUV, nm, FDI).