How Dutch start-up ASML rose to rule the world

ASML (Amsterdam: ASML) is the only commercial supplier of EUV lithography systems – machines that sit at the heart of the modern semiconductor industry. An extreme ultraviolet (EUV) lithography system is roughly the size of a double-decker bus, weighs roughly 180 tonnes and costs more than €180 million. It fires lasers at tens of thousands of microscopic droplets of molten tin every second, creating a plasma that emits light at a wavelength of 13.5 nanometres. That light is used to print the extraordinarily fine patterns from which the most advanced computer chips are made.

In 2025, ASML recognised revenue from 48 EUV systems and generated €32.7 billion of sales overall, including €8.2 billion from servicing and upgrading its installed base. The Dutch company does not manufacture every component itself. Instead, it coordinates a network of specialist suppliers that has taken decades to assemble. German company Carl Zeiss makes the mirrors at the heart of the optical system. Trumpf supplies the lasers used to generate the EUV light, while other specialists provide components ranging from wafer stages to sensors and software.

Its customers include TSMC, Samsung and Intel, companies with the scale and expertise to develop much of their own technology. Yet when it comes to leading-edge lithography, they have little choice but to buy from ASML. The machines are expensive because the economic cost of falling behind in semiconductor manufacturing is greater still. So how did a struggling Dutch start-up build a position that has become so difficult to challenge? And can that advantage survive the next stage of the industry’s development?

Try 6 free issues of MoneyWeek today

Get unparalleled financial insight, analysis and expert opinion you can profit from.

Start your trial

Sign up for MoneyWeek’s free twice-daily newsletter.

Join more than 165,000 subscribers and keep yourself informed with latest financial news, insights and expert analysis to help you understand what really matters when it comes to your finances.

ASML was established in 1984 as a joint venture between Philips and ASM International. The two companies brought different capabilities to the new business. Philips had already developed the PAS 2000 lithography system, while ASM International specialised in semiconductor manufacturing equipment. The timing was difficult. The company entered a competitive market dominated by established American and Japanese companies and initially had few customers. It also struggled financially. ASM International eventually withdrew after struggling to justify continued investment in a business that was consuming cash, while Philips was cutting costs. ASML survived because its management and shareholders continued to back the technology and because Philips provided further support.

ASML spent its first decade building the business and developing its technology. Rather than trying to make every critical component itself, ASML worked with specialist suppliers and concentrated on integrating their technologies into a complete lithography system. The breakthrough came in the 1990s with the development of the PAS 5500. Its productivity and resolution helped ASML win the customers it needed to become profitable. The company went public in 1995, giving it access to the capital needed to expand its research and manufacturing operations.

The approach mattered because lithography was becoming too complex for one company to master every discipline itself. ASML could concentrate on integrating the system while drawing on specialists in optics, lasers, metrology and other fields. The strategy eventually produced a decisive lead over its two big rivals, Nikon and Canon. But the biggest test was still to come. EUV had been discussed for years as the technology that might allow chipmakers to keep shrinking transistors, yet turning the idea into a machine capable of high-volume production proved difficult. ASML and its customers spent years developing the technology, with Intel, Samsung and TSMC helping to fund the effort.

In 2012, those customers committed €1.38 billion of research and development funding to ASML’s next-generation lithography programme and took minority stakes in the company. Intel alone committed €829 million of R&D funding alongside an investment of up to 15% in ASML. The arrangement illustrates how dependent ASML’s customers had become on its technology. They were helping finance development because the machines would determine how far their own manufacturing could advance. The result was to bring EUV into commercial production, but by then ASML had spent decades building the supplier network, engineering expertise and relationships needed to make it work.

competitive moat – the technology, expertise and relationships that make the business difficult for rivals to challenge. Its position in EUV was therefore built long before the first commercial machines were sold. To understand why ASML can charge hundreds of millions of euros for a single machine, one must consider the economics of modern semiconductor manufacturing. A cutting-edge fabrication plant, or “fab”, can cost more than $20 billion to build and equip. Much of that investment goes on the machinery inside it, with lithography scanners among the most expensive tools.

For a company like TSMC, the economics of a fab come down to how many useful chips it can produce, how quickly, and for how much. Smaller transistors allow more computing power to be packed onto each wafer – a thin disc of silicon on which chips are made – while higher throughput and better yield improve the economics of the plant. This is why lithography matters. The machine prints the patterns that become the transistors and other structures on the chip. If those patterns cannot be printed accurately enough, the economics of the entire fab suffer.

That gives ASML an unusual position. Its customers include some of the world’s largest companies and most buyers of that size would have considerable bargaining power over suppliers. When it comes to leading-edge lithography, their options are much narrower. TSMC, Samsung and Intel cannot simply switch to another supplier offering an equivalent EUV system. A more capable machine can also allow a chipmaker to produce more wafers from an expensive fab, while improvements in accuracy and yield can increase the proportion of those wafers that become saleable chips. The return from a lithography investment therefore depends on the additional output and yield it enables, rather than simply on the purchase price. This helps explain why ASML’s customers have been prepared to commit capital years before they receive their machines.

The company now sells its low-NA EUV systems for roughly €180 million and high-NA ones for more than €350 million. NA means numerical aperture, a measure of the optical system’s ability to resolve very small features. The higher the NA, the more detail the machine can print. Customers continue to buy the machines because the cost has to be weighed against the price of falling behind. EUV system sales reached €11.6 billion in 2025, while ASML’s gross margin was 52.8%. Those figures suggest that customers have little choice but to accept the price of staying at the leading edge.

Customers cannot treat an ASML machine as a one-off purchase. Once a fab is built around its equipment, maintaining and upgrading that equipment becomes part of the production process. ASML continues to service its machines throughout their working lives, helping customers maintain performance and install upgrades. Its installed-base business therefore grows as the number of machines in operation increases, creating a source of revenue alongside new system sales.

Why can’t another company simply build an EUV machine and compete with ASML? The answer lies in how much has to work together for a lithography system to operate in a semiconductor fab. ASML acts as the system architect, drawing on a network of specialist suppliers that has developed alongside the company. ASML then has to integrate all of the high-precision parts into a machine that can operate reliably in production. Precision matters because an error that would be invisible to the human eye could make a machine unusable for leading-edge production. This is very hard for a competitor to replicate. Roughly 80% of ASML’s bill of materials is sourced from its global supplier network. A rival would have to recreate that network, integrate thousands of precise components and persuade the world’s leading chipmakers to install an unproven machine in their most valuable factories.

ASML also has an advantage that is harder to see. Every machine it installs adds to its practical knowledge of how its technology behaves in a real production environment. ASML has about 10,000 customer-support employees working around the world to keep its systems running. That operation is a continuing relationship with customers after the initial sale; engineers work with chipmakers to maintain performance, solve problems and install improvements. A machine can be designed on a drawing board, but making it run reliably, at high throughput and with the precision required for commercial chip production, is a different problem. Each generation provides information that can be fed into the next one.

ASML therefore remains the company best placed to develop whatever lithography technology chipmakers require next. But that position does not make its earnings immune to disruption. Geopolitical restrictions can limit where it sells its machines. The semiconductor cycle can cause large swings in orders. Changes in chip design could reduce the industry’s reliance on successive generations of lithography.

China accounted for 29.1% of ASML’s sales in 2025, although much of this was mainstream semiconductor equipment rather than the EUV systems at the centre of its monopoly. Export controls have already restricted the sale of some of ASML’s systems and the provision of certain services to Chinese customers. Losing access to part of the Chinese market reduces revenue, but avoids giving Chinese chipmakers access to an alternative source of EUV machines.

The nearer-term risk is more familiar. ASML sells relatively few machines at very high prices, making the timing of customer capital expenditure critical. It recognised revenue from 327 new and refurbished lithography systems in 2025, compared with 418 in 2024 and 449 in 2023. Its largest customer accounted for 23.9% of sales and its two largest customers for 38%. A major customer delaying a fab or moving its spending into a different year can therefore produce a significant swing in ASML’s results.

The more fundamental threat is technological. ASML’s business depends on chipmakers continuing to invest in new generations of process technology. The company itself acknowledges that customers could delay adoption of new technologies if the economics do not justify the cost, or shift towards architectures that rely less on lithography. The price of each technological advance is rising. High-NA EUV machines cost hundreds of millions of euros, while chipmakers can also improve computing performance through chiplets, advanced packaging and architectural changes. If these approaches deliver better returns than simply shrinking transistors, the industry could become less willing to pay for each successive generation of lithography.

For now, there is little evidence that this is displacing leading-edge lithography. AI is increasing demand for advanced logic and memory, and ASML says growth in both markets helped drive demand for its EUV systems in 2025. This remains the most important long-term question hanging over the investment case. Export controls can reduce sales, a semiconductor downturn can defer orders and customers can become more selective about expensive new technologies. So far, none has emerged at commercial scale as a way of displacing ASML’s leading-edge lithography technology.

For shareholders, ASML offers a combination of growth and recurring revenue. In 2025 it generated €32.7 billion of sales, a 52.8% gross margin and €9.6 billion of net income. Installed-base management revenue, mainly servicing and field options for machines already in customers’ fabs, reached €8.2 billion, up 26% on the previous year. The outlook has strengthened since then. ASML reported €9.3 billion of sales in the second quarter of 2026, with a 54% gross margin, and raised its full-year revenue guidance to €43 billion-€45 billion, up from €34 billion-€39 billion at the start of the year. It says AI-related investment is driving demand for advanced chips and customers are accelerating expansion plans.

AI should still be viewed as an accelerant rather than the foundation of the investment case. The more durable thesis is that computing demand continues to rise and chipmakers need increasingly capable manufacturing equipment. ASML does not need every future semiconductor application to be an AI application for its long-term case to work. ASML remains a capital-equipment company and its earnings are therefore exposed to the spending decisions of its customers. Even a customer convinced of the long-term need for additional capacity can delay an order for a year. The result is an earnings profile that is more cyclical than the competitive advantage might suggest.

ASML shares have generally commanded a premium valuation because investors recognise the quality of the business. At roughly €1,700 a share, the stock trades on around 35 times consensus 2026 earnings and 25 times projected 2027 earnings. Those multiples can look reasonable if earnings grow rapidly. ASML is now aiming for €43 billion–€45 billion of revenue in 2026, up from €32.7 billion in 2025, while analysts expect further earnings growth in 2027. The valuation leaves less room for disappointment. A weaker semiconductor cycle, slower high-NA adoption or further restrictions on China could cause investors to reassess the multiple even if ASML’s competitive position remains intact.

The investment question is whether the future earnings from ASML’s competitive position justify the price of the shares. A business can continue to compound while its shares fall if investors have already priced in too much of that growth. ASML has spent four decades building one of the most formidable competitive positions in the global economy. The investment case now depends on whether it can turn that position into enough future earnings to justify the premium investors are being asked to pay today.