There is only one place on Earth where the most advanced chips in existence can be made. It is not the United States. It is not China. It is not South Korea, nor Japan, nor Germany. It is an island of thirty-six thousand square kilometres with twenty-three million inhabitants, located a hundred miles off the Chinese coast, which China considers part of its territory and whose de facto independence has been sustained, since 1949, exclusively by the willingness of the United States to defend it.
Taiwan produces approximately ninety-two per cent of the world’s most advanced logic chips, those below five nanometres. A single Taiwanese company, TSMC, controls around seventy per cent of global semiconductor foundry revenue. In advanced chips, its share is closer to ninety per cent. There is no second supplier capable of doing what TSMC does, at the scale TSMC does it, with the quality TSMC achieves. Samsung fabricates advanced chips but with lower yields and a declining market share. Intel has spent a decade trying to recover ground and just needed a capital injection from the American government to avoid disintegration.
This is not an economic statistic. It is a civilisational vulnerability. If tomorrow China blockades or invades Taiwan, the lights of the twenty-first century go out.
The Bottleneck
To understand the magnitude of the problem you have to understand what TSMC does and why nobody else can do it.
TSMC does not design chips. It does not sell products to the end consumer. What it does is manufacture the chips others design: Apple, Nvidia, AMD, Qualcomm, Broadcom, and dozens more. When Apple unveils a new iPhone with a faster processor, that processor was designed by Apple but manufactured by TSMC. When Nvidia sells the graphics cards that train the artificial intelligence models transforming the global economy, those chips were manufactured by TSMC. When the Pentagon needs advanced semiconductors for its latest-generation weapons systems, those semiconductors come, directly or indirectly, from TSMC.
Advanced semiconductor manufacturing is probably the most complex industrial process humanity has ever developed. A modern three-nanometre chip contains billions of transistors arranged in structures whose dimensions are comparable to a handful of atoms. Fabrication requires extreme ultraviolet lithography machines costing upwards of three hundred million dollars each and made by only one company in the world, the Dutch firm ASML. It requires cleanrooms where the concentration of airborne particles is millions of times lower than in a normal room. It requires thousands of engineers with decades of accumulated process experience that cannot be learned from a book. It requires a supply chain of hundreds of ultra-pure materials and components sourced from around the world.
TSMC mastered this process over thirty years of obsessive investment, talent accumulation, and incremental optimisation. Its founder, Morris Chang, an MIT- and Stanford-trained engineer who worked at Texas Instruments, created the company in 1987 with an idea that at the time seemed modest: instead of designing and manufacturing its own chips, TSMC would manufacture the chips others designed. The “pure-play foundry” model freed designers from the need to own their own factories and enabled a specialisation that proved devastatingly efficient. Today the global semiconductor ecosystem divides into two: those who design and those who fabricate. And the one who fabricates, in the advanced category, is essentially one.
Why Reshoring Will Not Work in Time
The United States and Europe recognised the problem. The American CHIPS and Science Act mobilised incentives that catalysed more than four hundred and fifty billion dollars in private investment. TSMC is building factories in Arizona. Intel received massive federal support. Samsung is expanding its presence in Texas. The European Union launched its own European Chips Act. Japan is subsidising the construction of TSMC factories in Kumamoto.
All of this is necessary and welcome. And all of it is insufficient.
The fundamental problem is that advanced semiconductor manufacturing does not relocate the way a car assembly line does. It is not a matter of building a facility, installing machines, and hiring workers. It is a matter of replicating an ecosystem that took three decades to develop, that depends on tacit knowledge accumulated in thousands of engineers, that requires a supply chain of overwhelming complexity, and that operates with margins of error measured in atoms.
The TSMC factory in Arizona illustrates the difficulties. It was announced in 2020 with a projected production start in 2024. It was delayed. Costs rose. TSMC had to bring in hundreds of Taiwanese engineers because it could not find local staff with the necessary experience, which generated tensions with American unions. When the factory is fully operational, it will produce chips at four- and five-nanometre processes, a generation behind what TSMC manufactures in Taiwan. Cutting-edge production, the three-nanometre and two-nanometre nodes entering mass production in 2025, will remain on the island.
Even in the most optimistic scenario, where every planned factory in the United States, Europe, and Japan is built on time and reaches competitive yields, total capacity outside Taiwan will not amount to more than a fraction of global advanced capacity before 2030 or 2032. An analysis by the Institute for Economics and Peace estimated that American fab capacity could increase by two hundred and three per cent by 2032, but that Taiwan would remain the critical node for the most advanced production. In other words, the window of vulnerability is open and will remain open for at least a decade.
There is an additional argument that is rarely mentioned. TSMC has a limited incentive to transfer its most advanced technology off the island. The geographic concentration of cutting-edge production in Taiwan is, from the Taiwanese perspective, an insurance policy. As long as the world depends on Taiwan for its chips, the world has an existential interest in Taiwan not being invaded. It is what some analysts call the “silicon shield”: the idea that TSMC’s importance to the global economy deters Chinese aggression because a war would destroy the semiconductor production on which China itself depends.
The Real Deterrence: American Power
The silicon shield is an elegant concept, but it needs handling with care. It is not chip dependency that prevents China from invading Taiwan. It is the military power of the United States.
China receives more than half of Taiwan’s chip exports. A blockade or invasion of the island would indeed be an act of monumental economic self-harm for Beijing. But history is full of countries that made strategically irrational decisions for reasons of national prestige, ideology, or miscalculation. Japan attacked Pearl Harbor knowing it could not win a prolonged war against the United States. Argentina invaded the Falklands knowing it could not sustain a war against the United Kingdom. Russia invaded Ukraine underestimating Ukrainian resistance and the Western response. Economic rationality is not insurance against military aggression. If it were, there would be no wars.
What does deter military aggression is the capability and willingness to respond. And in the case of Taiwan, the adversary is the United States. The Seventh Fleet operates in the Western Pacific. American bases in Japan, South Korea, and Guam surround the theatre of operations. American naval and air superiority in the region, though it has diminished relative to Chinese military growth, remains real. And more important than capability is the signal: the United States has maintained for decades a policy of “strategic ambiguity” regarding Taiwan that, in practice, has been steadily clarifying. Biden said explicitly that the United States would defend Taiwan in the event of an invasion. Trump, in his second term, has not contradicted that position.
China knows this. The People’s Liberation Army has been building amphibious capacity, modernising its fleet, and developing anti-ship missiles designed to keep American aircraft carriers away from the Taiwan Strait. But an amphibious invasion across a hundred miles of open sea, against a mountainous island with a hostile population and a reasonably equipped military, under the threat of intervention by the world’s foremost military power, is an operation of a complexity and a risk that even a bold Chinese strategist would think twice about. Three times, actually, because if the invasion fails, the Chinese Communist Party does not survive the humiliation.
This does not mean invasion is impossible. It means it is unlikely as long as the United States maintains the capability and the will to intervene. And that, in turn, means the security of the global semiconductor supply chain depends, in the final analysis, on the same factor nearly everything in the contemporary international order depends on: American military power.
What the Dependency Means
The strategic implications of semiconductor concentration in Taiwan are difficult to overstate.
First, dependence on TSMC conditions the decision-making of every major power. The United States cannot afford to lose Taiwan not only for reasons of strategic credibility in the Indo-Pacific but because losing access to TSMC would mean losing the technological edge that sustains its military supremacy. The chips on which American weapons systems depend, the artificial intelligence that feeds targeting systems like Palantir, the processors that run satellites and command-and-control networks, all come from Taiwan. A war in the Taiwan Strait is not a regional conflict. It is a systemic event that would reconfigure the global economy and the balance of power in ways no model can anticipate.
Second, China faces a dilemma with no military solution. If it invades Taiwan and destroys TSMC’s factories in the process, it loses the chips its own economy depends on. If it invades and captures them intact, something technically near-impossible given the extreme fragility of the facilities, it would need decades to operate them without the Taiwanese engineers who would presumably flee or be evacuated. And if it blockades the island without invading, it triggers the very global economic crisis it wants to avoid, because every affected country would align with the United States against China. There is no scenario in which Chinese military force solves the semiconductor problem.
Third, China’s attempt to develop domestic capability, through SMIC and massive state investment in semiconductors, has produced real but limited results. SMIC can fabricate chips at mature twenty-eight-nanometre processes and, according to some sources, has managed to produce seven-nanometre chips. But the gap with TSMC at advanced three- and two-nanometre nodes is enormous, and American export restrictions preventing China from accessing ASML’s EUV lithography machines make closing that gap extraordinarily difficult. China can make chips for its own basic needs. It cannot make the chips it needs to compete in artificial intelligence, high-performance computing, and cutting-edge military systems. That dependency is Beijing’s greatest strategic vulnerability, and there is no solution in sight.
Silicon as the Oil of the Twenty-First Century
There is an analogy circulating in strategic circles: semiconductors are the oil of the twenty-first century. Like oil in the twentieth century, chips are an essential resource on which all advanced economic and military activity depends. Like oil, their production is geographically concentrated in ways that create strategic vulnerabilities. And like oil, control of the sources of production is a first-order factor in geopolitical power.
But there is a crucial difference. Oil is a natural resource extracted from the ground. If a country loses access to one source, it can look for another. Saudi oil can be substituted, at a cost, by American shale or Norwegian crude. Advanced semiconductors are not a natural resource. They are a product of unprecedented technological complexity, manufactured by a single company with three decades of accumulated advantage, in one specific spot on the planet. There is no available substitute. There is no alternative well to drill.
That is the reality the world lives with today. The phone you are reading this on, the server processing this page, the artificial intelligence system assisting medical research, the processor guiding a precision missile, all depend on an island in the Western Pacific that an authoritarian government claims as its own and that a single company makes run.
Morris Chang, who is ninety-three and retired from TSMC in 2018, built what may be the most strategically important company in history. Not the largest, not the most profitable, but the most important, in the sense that its absence would be irreplaceable within any relevant timeframe. If TSMC stopped functioning tomorrow, the artificial intelligence industry would halt, smartphone production would collapse, Western defence supply chains would enter crisis, and the global economy would suffer a shock it would take years to recover from. A single point of failure for the entire technological civilisation.
Reshoring is necessary. Diversification is urgent. But we should be honest: the window of vulnerability will last at least a decade. And during that decade, peace in the Taiwan Strait, and with it the stability of the global economy and the supply chain on which everything depends, will rest on the same thing it has rested on since 1949: the Seventh Fleet of the United States Navy continuing to sail the Western Pacific, and Beijing believing that using it is not a bluff.
That is all. That simple. That fragile.