1. The Secret of Veldhoven
Southern Netherlands, ten kilometers southwest of Eindhoven. A small town of 45,000 people called Veldhoven — a place you can barely find on a map. The home of the most precise machine humanity has ever built sits in this town.
The ASML headquarters campus employs more than 20,000 people. At the center of the campus, in the De Run 1000 building, 4,000 engineers from 90 countries gather to build a single machine. The price of one machine shipped from this campus is $220 million — roughly ₩300 billion. The newest High-NA unit runs $380 million. The Hyper-NA, scheduled for release in 2030, is expected to exceed $700 million.
Without this machine, TSMC cannot make chips. Samsung cannot produce semiconductors. NVIDIA's AI accelerators cannot exist. Apple's iPhone chips cannot be made. One hundred percent of the world's advanced semiconductors pass through this machine.
At the end of Chapter 6 we saw the ceiling of Singapore. A country that is a hub without a core. The Netherlands is the opposite. In a small town that barely shows up on a map, it builds the core of the world's semiconductor industry. In March 2026, the Eindhoven city council approved the construction of a second ASML campus — targeting another 20,000 hires. This quiet small town is the heart of the world's semiconductor geopolitics.
2. Painting with Light
Making a semiconductor is, in essence, painting a picture with light. The process of projecting light onto a silicon wafer to etch circuit patterns is called lithography. To draw finer circuits, you need light of a shorter wavelength — the way a finer painting needs a thinner brush.
ASML's EUV (extreme ultraviolet) machine uses light with a wavelength of 13.5 nanometers. More than fourteen times shorter than the existing DUV (deep ultraviolet). This light makes it possible to draw circuits finer than 2 nanometers, and today's smartphone application processors, AI accelerators, and high-performance server chips are all made with this technology.
The way this light is produced is astonishing. Inside the machine, the following sequence repeats 50,000 times a second. A droplet of liquid tin, a third the thickness of a human hair, is fired at 70 meters per second. A low-power laser flattens the droplet into a disk. A high-power laser instantly vaporizes the disk into plasma. The plasma emits EUV light. Roughly a hundred layers of molybdenum-silicon mirrors guide that light onto the wafer with precision. All of it happens in a vacuum — EUV light is absorbed by air.
More than 100,000 parts make up a single machine. The most precise machine in human history. And there is exactly one company on earth that can build it.
3. A Thirty-Year Moat
ASML did not become a monopolist overnight. It is the product of thirty years of investment, failure, and global collaboration.
Founded in 1984 as a joint venture between Philips and ASMI. In 1986 it entered an optical partnership with Germany's Carl Zeiss. That partnership was decisive. Without the ultra-precise multi-layer reflectors Zeiss developed over fifteen years — mirror surfaces with a tolerance below 0.1 nanometer, refusing an error the size of a single atom — there is no way to focus EUV light onto the desired position. There is exactly one company in the world that can make these mirrors: Zeiss. In 2016, ASML bought 25 percent of Zeiss for €1 billion, elevating the relationship from a supply contract to a capital-level bond.
Germany's TRUMPF spent ten years developing the world's only ultra-high-power carbon-dioxide laser. Output 20 millijoules, frequency 50 kilohertz — the most powerful industrial pulsed laser in the world. Without it, you cannot vaporize tin 50,000 times a second. Without TRUMPF, ASML's EUV machine does not work.
In 2013, ASML acquired Cymer, a San Diego-based light-source technology firm. No one at the time expected that acquisition to later carry geopolitical implications.
In 2012, the customers became investors. TSMC $1 billion. Samsung $1 billion. Intel $4 billion. Total $6 billion. This was not a simple equipment purchase. It was a case of technology that could only exist if built together. Each of these companies watched its own competitors invest in the same firm — and had to invest anyway, because without ASML nobody could make the next generation of chips.
Total R&D investment: roughly $9 billion. About ₩12 trillion. Thirty years.
But this moat contains a paradox. The parts ASML manufactures in-house come to only about 15 percent of the total. The other 85 percent are supplied by more than 800 vendors around the world. Zeiss's mirrors, TRUMPF's lasers, Cymer's light-source systems — if any one of these three is cut off, ASML production as a whole stops. If ASML stops, the world's production of advanced semiconductors stops. The most precise machine in human history is simultaneously the most fragile machine in human history. Three separate single points of failure.
The numbers show the scale of that fragility. In 2025, ASML shipped 48 EUV units. Including advanced DUV machines, the total was 179. Every smartphone chip, every AI accelerator, every data-center server in the world depends on a supply of 48 EUV machines a year. That number is the true meaning of bottleneck. These are not mass-produced goods. Each one is closer to a handcrafted work of art than to a manufactured product.
This is the moat. The kind of moat that money alone cannot cross. A moat built by time.
4. The Forced Choice in the Prime Minister's Office
January 2023. The Hague. The office of the Dutch prime minister.
The American request was clear. Extend export restrictions on ASML's advanced DUV equipment to China. In the form of a U.S.–Netherlands–Japan trilateral arrangement.
There had been warnings. In 2020, the Dutch government had quietly revoked ASML's export license for EUV shipments to China. ASML was unable to deliver already-built EUV machines to Chinese customers. On DUV equipment, however, an informal "gentleman's agreement" with Washington was in effect — an unspoken understanding that a certain volume would be permitted. One report has it that ASML sold volumes above that line to China. Later came a report that ASML had offered to serve as America's "eyes and ears" in China. ASML denied it. But one thing is clear. The January 2023 meeting was not sudden. It was the final decision reached after three years of push-and-pull.
Now one more step — stop the export not only of EUV but also of DUV ArF immersion systems.
The Dutch dilemma was not simple. ASML is the heart of the Dutch economy. Market capitalization around $553 billion. Europe's largest tech company. 2025 revenue €32.7 billion (about ₩48 trillion). Net profit €9.6 billion (about ₩14 trillion). Annual spending on suppliers alone: €16 billion (about ₩22 trillion). 44,000 employees. The single largest private-sector R&D investor in the Netherlands. Tens of thousands of jobs sit on this company's supply chain.
China was a core customer. In 2023, the China share of revenue was 29 percent. In the first quarter of 2024 it surged to 49 percent — a wave of pre-emptive orders placed before the restrictions took effect. The Dutch government was being asked to give that revenue up.
But the Netherlands chose the United States. The reasons were layered. NATO security dependence. The EU-wide direction of semiconductor supply-chain restructuring. According to one report, the United States brought Ukraine aid to the negotiating table. And one paradoxical logic — export controls also prevent ASML technology from being transferred to Chinese competitors.
From September 2023, the export of advanced DUV equipment to China was banned. In September 2024, the scope was widened. Selling ASML's 1970i and 1980i immersion DUV systems to China now required an export license. In effect, the export of every advanced lithography machine to China was cut off.
In Chapter 5 we saw Israel choose the American side decisively over the Haifa port question. The Netherlands made the same structural choice. But there is a difference. Israel chose voluntarily, on the logic of a security alliance. In the Dutch case, the United States made the Netherlands flip its own switch.
5. Who Holds the Switch
ASML is a Dutch company. But the switch is in American hands.
The core of it is Cymer. Acquired in 2013, the firm is headquartered in San Diego. It holds the core technology of the EUV light-source system. The moment American technology is involved, the U.S. Export Administration Regulations (EAR) apply. Even though ASML is a Dutch legal entity, it cannot export any machine containing American-origin parts without American permission.
This structure is operating directly on Korean companies.
Samsung produces 35 to 40 percent of its total NAND flash at its Xi'an fab in China. SK hynix produces 40 percent of its total DRAM at its Wuxi fab and 20 percent of its NAND at its Dalian fab. The China share of production for both companies is the lifeline of the Korean semiconductor industry.
The shock of August 2025: the U.S. Commerce Department revoked the Validated End User (VEU) status of Samsung and SK hynix. With that status, their Chinese fabs could use American equipment without separate licenses. Once the status was gone, every single piece of equipment required individual approval from the U.S. government. In December 2025 came a partial easing — allowed for existing facilities only, no capacity expansion and no technology upgrades.
Lay the structure out and it looks like this. Samsung and SK hynix cannot make leading-edge semiconductors without ASML's equipment. ASML's equipment depends on American technology. Therefore the United States can exercise an indirect veto, through ASML, over the Chinese production of Korean companies. This is how the geopolitics of the technology bottleneck actually works. Whoever holds the bottleneck holds the switch. And that switch is not in Veldhoven. It is in Washington.
6. Four Hundred Years of Recurrence
In Chapter 1 we saw the triple monopoly of the Hanseatic League bypassed by Dutch merchants. The result of that bypass was the VOC — the Dutch East India Company.
Founded in 1602, the VOC was the first true multinational corporation in human history and the first joint-stock company. Its strategy was simple. Seize the chokepoints of the spice trade routes. Cloves from the Maluku Islands. Nutmeg from the Banda Islands. Pepper from the Malabar Coast of India. The cloves the VOC bought in Indonesia sold in Europe at fourteen to seventeen times the purchase price. In 1670, the VOC fleet accounted for roughly half of the total tonnage of the entire European merchant marine.
Four hundred years have passed. This small country of 18 million is once again in the same position.
Seventeenth century — the bottleneck of the global spice trade. Nineteenth and twentieth centuries — the port of Rotterdam, the gateway of European cargo. Twentieth century — Royal Dutch Shell, the oil supply chain. Twenty-first century — ASML, the bottleneck of semiconductor manufacturing. Where the VOC held the maritime trade routes, ASML holds the technical bottleneck of semiconductor manufacturing. The structure is identical. Sole supplier. Every competitor has to pass through this one firm. When a state controls the bottleneck, geopolitical leverage is generated. No substitute.
This country's strategic cast of mind is not to become the strongest but to become the one nothing can do without. In Chapter 3 we saw the indispensability Lee Kuan Yew deliberately designed. The Dutch indispensability is of a different character. It is a historical DNA built over four hundred years of repeatedly finding and occupying bottlenecks. Not design, but accumulation.
China is trying to break this monopoly. Shanghai's SMEE has begun producing DUV equipment at the 28-nanometer level, but the gap to ASML's EUV is at least ten years. There are reports of a secret EUV project involving former ASML engineers. The Chinese government's goal is 2028; the realistic projection is after 2030.
But there is a fundamental barrier. You can copy the blueprint of a machine. You cannot copy, in a short period, the optical system Zeiss spent fifteen years developing and the laser TRUMPF spent ten years developing. And there is something more important. The feedback data accumulated over decades of actual mass production — failure patterns, optimization know-how, the fine-grained balance of interactions between parts — cannot be bought and cannot be stolen. In Chapter 5 we saw the cases of Israeli cyber-technology leakage. But ASML's moat is not a few lines of code. It is the total sum of thirty years of experience. A moat built by time cannot be crossed with money.
7. When a Monopoly Makes a Country
The case of ASML is the most complete contemporary realization of this book's core proposition — the in-between strategy.
The Netherlands is neither the United States nor China. But America's Intel, Taiwan's TSMC, Korea's Samsung, and China's SMIC all cannot exist without ASML.
In Chapter 4 we saw Taiwan's TSMC function as the silicon shield. ASML is the only supplier of the tools that make that shield. The technology that makes the shield carries stronger leverage than the shield itself — because shields can be made in multiples, but the tools that make the shields come from exactly one place.
This is the structural difference between monopoly leverage and competitive leverage. TSMC is the world's best foundry, but it has Samsung as a competitor. Samsung is the world's largest memory maker, but SK hynix and Micron exist. Being the best at making something delivers market advantage. Being the only entity that can make something creates the necessity of existence itself. ASML is the latter. No competitors. No substitutes. Pricing authority sits with ASML. The customers have no option.
Apply the core formula and ASML is a perfect case. Technological innovation (thirty years of EUV development) produced concentration of capital ($553 billion market capitalization, Europe's largest tech company). That concentration became social tension (the central variable in the U.S.–China technology-hegemony race), which in turn triggered institutional redesign (the birth of the export-control regime).
But the limits of ASML are also clear. The switch sits in Washington. The Netherlands has not been able to convert the benefits of monopoly fully into its own strategic asset. The structure ends up using the monopoly as a geopolitical instrument for the United States. The country has become the one that cannot be done without, but the authority to use that cannot be done without belongs to the stronger party.
This structural tension is also being picked up at the European level. Italy's Bocconi research institute (IEP) has named ASML the most important bargaining card of the EU. The reading is that when the United States pressures Europe on security or trade, ASML's monopoly position raises the negotiating power of Europe as a whole. But it is not clear who this card actually belongs to — the Netherlands, Brussels (the EU), or Washington.
There is another worry on the Dutch government's mind. The concern that ASML could leave the Netherlands. There have been reports that France actively proposed attracting ASML, and ASML itself has at times hinted that if export controls restrict its growth it may consider relocating headquarters. In response, in 2026 the Dutch government approved the second campus and pledged investment in surrounding infrastructure — a retention strategy. The paradox of a country that holds a monopoly firm. It is not the firm that needs the country. It is the country that needs the firm.
This is what standing in the between means. A position essential to both sides, yet not free of the pressure of either. A monopoly can make a country. But monopoly alone does not become sovereignty.
Korea in the Mirror
The face of Korea that ASML reflects is the bottleneck of the bottleneck.
The structure of ASML dependence. Samsung and SK hynix are semiconductor firms at the world's highest level. But neither of them can make leading-edge chips without ASML's equipment. As of 2025, ASML holds 100 percent of the EUV market and 94 percent of the lithography market overall. One EUV machine: ₩300 billion. A High-NA: ₩500 billion. The entire value chain of the Korean semiconductor industry depends on a single firm in Veldhoven.
The bottleneck of the bottleneck. A more serious structure sits on top of this. Because ASML's equipment contains American technology, the United States can exercise an indirect veto, through ASML, over the Chinese production of Korean companies. Samsung's Xi'an fab (35 to 40 percent of NAND) and SK hynix's Wuxi fab (40 percent of DRAM) now hinge on U.S. export licenses. Korea depends on a bottleneck (ASML), and on top of that is bound to the bottleneck that controls the bottleneck (American technology). A bottleneck of the bottleneck.
The possibility called HBM. The closest breakthrough candidate is SK hynix's HBM (high-bandwidth memory). As of 2025, a market share of 62 percent. Effectively the sole supplier of the HBM3E that goes into NVIDIA's AI accelerators. 2026 volume is already sold out. On the next-generation Rubin platform, SK hynix's share is projected at roughly 70 percent. But HBM is not yet ASML. Micron (21 percent) and Samsung (17 percent) exist as competitors. HBM has not yet reached the level of technologically irreplaceable.
The ASML benchmark. Four things are required to reach true monopoly leverage. The position of sole supplier. A technology gap maintained for more than ten years. A core process on which every competitor depends. And the position from which state control of exports can affect the entire industry of a competitor country. SK hynix's HBM meets half of these criteria. The other half can only be filled by the accumulation of time and investment.
The lesson of ASML is clear. Being the only entity that can make something creates stronger indispensability than being the best at making something. As with the moat the Netherlands built with thirty years and $9 billion — only the accumulation of time and money and failure creates a true monopoly. Can Korea make an ASML moment? The answer depends on how far Korea can widen the HBM technology gap and how long it can sustain it.
The next mirror is a large country holding the card of natural resources — Indonesia's nickel ladder.