Discover more from Bismarck Brief
The Tradition of Knowledge Behind ASML
Europe’s most valuable tech company is the only manufacturer of the machines needed to make advanced computer chips. This market monopoly rests on a Dutch tradition of knowledge and industrial policy.
ASML is a publicly-traded Dutch technology company with a market capitalization of $236 billion as of October 2023, making it Europe’s largest tech company.1 Headquartered in a suburb of the small city of Eindhoven, the semiconductor company is one of the largest global suppliers of photolithography machines, which are used by semiconductor manufacturers to create the required patterns on silicon wafers used in computer chips. ASML also holds a monopoly on the development and manufacturing of extreme ultraviolet lithography (EUV) machines, which are necessary to produce the most advanced chips available today, such as the H100 Tensor Core graphics processing unit (GPU) designed by Nvidia and manufactured in Taiwan by TSMC.
You can listen to this Brief in full with the audio player below:
Entrepreneurs, investors, technologists, and even government officials in the developed world have come to expect that increasing the total amount of available computing power will result in economically transformative artificial intelligence that will offer technological solutions for a wide range of economic and political challenges arising over the last twenty years from the repeatedly disappointing performance of dominant institutions. This “scaling hypothesis,” as opposed to building artificial intelligence through novel architecture or data sets for training, necessarily relies on vastly increasing manufacturing of the most advanced chips, which so far can only be done by expanding the operations of a handful of companies like TSMC, Intel, or Samsung.
These companies, in turn, all rely on ASML alone to supply the EUV machines necessary for expansion. Through its EUV monopoly, ASML is an irreplaceable supplier for all of the largest chip manufacturers. The company is arguably the single hardest layer to recreate in the chip manufacturing stack, not only because of the vast capital costs that would be necessary to build the complex machines, but also because its success rests on a unique Dutch tradition of knowledge in photolithography that goes back to the 1960s, if not earlier.
As a result, no competitor to ASML is likely to emerge. ASML’s close relationships with the other major semiconductor manufacturers and its key component suppliers, several of whom are themselves monopolies on various other parts of the supply chain, are further obstacles. Likely only extensive subsidies and targeted government policy could end ASML’s monopoly on EUV machines, but the only industrialized country with both the incentive and ability to follow through on this is, for now, China.
ASML is a Monopoly Through Technical Superiority
ASML has about a 90% market share in photolithography machines and a 100% market share in the most advanced form of photolithography, extreme ultraviolet lithography.2 This effective monopoly is straightforwardly the result of technical superiority. Photolithography is the process of forming patterns on a silicon wafer using light, in this case at an extremely small scale. To build machines capable of reliably doing this is an extremely difficult feat of engineering. Light is passed through a photomask that contains the desired circuit pattern. The wafer that the light is aimed at is coated with chemically sensitive photoresist, which reacts to the light to form the required patterns. Modern photolithographic techniques use light beyond the visible spectrum, since shorter-wavelength light allows for higher transistor density on the wafer, which means more computing power per chip.
The process is conceptually similar to using a sharper pencil to write smaller letters on paper. “Deep ultraviolet” (DUV) techniques use wavelengths as short as 193 nanometers, while extreme ultraviolet (EUV) techniques use even shorter wavelengths of 13.5 nanometers. To manufacture and control these extreme ultraviolet wavelengths requires the heating of tin to a temperature of 220,000 degrees Celsius, forty times hotter than the surface of the Sun, in order to transform the tin into plasma. In the machines, the tin is heated by a laser that travels through vacuum, striking each speck of tin twice at frequencies of 50,000 impacts per second. The resulting light is then directed towards the photomask by mirrors composed of over one hundred layers, each layer several nanometers thick. The correct positioning of the wafer is critical to achieving maximum yield. These machines keep the wafers positioned with a quarter of a nanometer’s precision.
There are no how-to guides or written instructions for reliably reproducing such a difficult technical feat. Achieving it consistently relies on intellectual dark matter: the tacit knowledge, idiosyncratic skill, and unwritten rules that are necessary for success but cannot be made legible to outsiders or non-experts except through training with the current expert practitioners, in whose heads the knowledge actually lives.3 ASML staff are deployed along with its EUV machines to oversee their operation and maintenance: they remain on site throughout the lifetime of a given machine. The chip manufacturers are therefore dependent on the company for tacit technical knowledge as well the machines themselves.4 Much of this tacit knowledge relates to how light refracts in different contexts. The maintenance of this knowledge in an organized manner by the company means it carries a unique tradition of knowledge.5
Advancing this tradition is very expensive. ASML employs over 14,000 people in its R&D departments, out of over 39,000 employees worldwide, and spends almost $1 billion on R&D every quarter.6 To build a working EUV machine took thirteen years from the first prototype shipped in 2006 to the first EUV-enabled products released in 2019, and cost roughly $10 billion in R&D and capital expenditure costs, according to some external estimates.7 The company itself estimates that it spent $6 billion in R&D alone over 17 years.8 Including the costs borne by component suppliers, the total EUV development costs likely ran into the tens of billions.
Given the costs ASML paid to develop EUV, its prices are unsurprisingly high. The machines cost around $200 million each. In 2022, the company shipped just 54 of them, though it also profits from maintenance contracts on installed machines.9 Next-generation “High-NA” EUV machines are currently in production; the “high numerical aperture” refers to a greater ability to collect and focus light. Over five prototypes have been ordered by Intel and other customers. These are selling for $340 million per unit and forthcoming production models will be priced still higher.10
ASML generated almost $22 billion of revenue in 2022 at a 35% profit margin before interest and taxes (EBIT) and is growing revenue at around 20% annually.11 The company forecasts that by 2030 it will be generating between $45-60 billion of revenue annually.12 The company usually trades on the stock market at a high multiple of its current earnings, indicating that investors have high expectations of the company’s future growth and profitability. While its margins are substantially lower than those of TSMC, at around 65%, and Intel, they are comparable to many of the largest companies using the chips it helps to make.
In 2022, ASML operated at a higher gross profit margin than Apple, despite the latter’s pricing power due to its luxury branding. Smartphone manufacturers and cutting-edge AI companies alike are ultimately dependent on the semiconductor industry and its key suppliers. This means that the semiconductor firms and their key suppliers are able to capture a large portion of the windfall from the growing demand for high-performance computing, new models of smartphones and laptops, and more exotic use cases like augmented reality or virtual reality devices. This windfall goes in large part to the semiconductor design and fabrication companies, who are both dependent on ASML.
In turn, ASML inevitably shares some of its part of the windfall with its own suppliers. By one count, only 15% of the components of an EUV machine are manufactured by the company.13 The other 85% are manufactured by external suppliers. ASML assembles its machines itself at a single factory in Veldhoven, a suburb of the Dutch city of Eindhoven.14 ASML maintains close relationships with its most important suppliers, especially the privately-owned German optics manufacturing firm Carl Zeiss. Zeiss supplies the layered mirrors that reflect extreme ultraviolet light through the photomask onto silicon wafers. It is the world’s leading manufacturer of lithography optics equipment and has been a key supplier to ASML since the latter’s founding. ASML has described the relationship as “two companies, one business,” although throughout the 1990s the relationship was less close.15
In 1996, Zeiss attempted to sell its new optics technology to ASML’s then-competitor Silicon Valley Group (SVG), which caused a crisis in relations between the firms. SVG had grown out of the U.S. optics firm Perkin-Elmer and become one of Intel’s main suppliers. ASML management made it clear to Zeiss that they regarded this as unacceptable, in part because the optics had been developed due to ASML’s investments. Then-Zeiss CEO Peter Grassmann was invited to join ASML’s board of directors and the two companies signed an exclusivity deal the year after. In 2016, ASML paid €1 billion for a 25% stake in Zeiss, with its money going towards Zeiss’ R&D for optics that will constitute the core of the next generation of EUV equipment.16
Other key suppliers have been bought outright, partly in order to ensure quality control and align the supplier more closely with the company’s key priorities. Current CEO Peter Wennink reportedly warned one supplier that “if you don’t behave, we’re going to buy you.”17 Light source manufacturer Cymer was acquired in 2012 for $2.5 billion, explicitly in order to speed up the development of EUV and keep control of the EUV light source technology.18 ASML has also purchased Hermes Microvision, a manufacturer of inspection equipment, acquired in 2016 for $3.1 billion. Going back further, the company even bought out key competitors, including SVG, which it acquired in 2001. ASML’s business model and its pattern of acquisitions show a company that is entirely focused on advancing the cutting edge of technical superiority in a very narrow domain, rather than trying to increase margins for profitability or expand into new domains to build a conglomerate. This is unsurprising given that the company is built around a tradition of knowledge passed down by Dutch scientists.
The Dutch Tradition of Knowledge That Birthed ASML
ASML’s research and product development efforts have been led largely by 66-year-old Martin van den Brink, who today serves as co-president and chief technology officer of the company, roles he has held since 2013. He is paid equally to the company’s CEO and other co-president, Peter Wennink, who first joined the company in 1999 as chief financial officer, having previously worked at Deloitte as an accounting partner.19 Wennink and van den Brink are two of the six members of the ASML management board, but van den Brink is the only one whose history at the company dates back to its founding in 1984. The other management board members, as well as the supervisory board members, are relatively short-term appointees typically drawn either from the company’s executives or European industry broadly.
As a young man, van den Brink earned degrees in electrical engineering and physics from the HAN University of Applied Sciences in Arnhem and the University of Twente respectively. These are both Dutch universities in small cities. He joined ASML as an engineer at the age of 27 years old and has worked there ever since. Van den Brink led the technical development of ASML’s first successful lithography machine, the PAS 2500, which secured the company its first sales to TSMC and AMD and allowed it to turn its first profit in 1989.20
Van den Brink was also closely involved in the engineering of the PAS 5500, which secured ASML its first sales to IBM. Refurbished PAS 5500s remain in operation today and ASML continues to offer support for the model.21 By 1994, at 37 years of age, van den Brink was promoted to lead the company’s R&D department, although he was first sent on a six-week management course to learn how to handle the egos of the technical managers he had bruised in previous years.22 Subsequently, van den Brink also led the development of ASML’s first “step-and-scan” model, introduced as the deep ultraviolet lens PAS 5500/500, the machine that powered the company to market leadership when released in 1997. This model was a technical design advance on previous machines called “steppers.” Steppers, among other things, combined optical alignment and projection systems i.e. the wafer-positioning laser was sent through the same lens that projected the light wave onto the wafer.
Van den Brink was also pivotal to the decision to take a seemingly less ambitious path to product development after 1986. ASML’s main competitors at the time, the Japanese companies Nikon and Canon, sought to jump directly from 436 nm “g-line” steppers to 248 nm deep ultraviolet steppers, then considered the cutting edge. Van den Brink was aware of the technological challenges that the shift to DUV would bring and ensured that ASML instead first put its main efforts behind a reliable and fast intermediate 365 nm “i-line” stepper.23
This product came to market in 1991 and its sales allowed the company to take market share from its Japanese rivals throughout the 1990s. ASML did not introduce a deep ultraviolet product until 1997. Under van den Brink, the company was also the first to commercialize immersion lithography from 2003 onwards, a generational shift in production techniques that replaced the traditional air gap between lens and wafer with water. These are both examples of van den Brink’s deep technical expertise in photolithography guiding business decision-making and demonstrate the centrality of this expertise to the company’s commercial success.
Van den Brink subsequently oversaw the lengthy development of commercially-viable extreme ultraviolet lithography machines. In 2006 he gave a speech at the International Symposium on Extreme Ultraviolet Lithography that outlined his belief that EUV was the only cost-effective option for increasing the number of chips per silicon wafer.24 Although commercial development overran by a decade longer than planned per van den Brink, the trust ASML built with its customers had attracted large-scale investments to finance the final stages of EUV’s development, including from TSMC, Samsung, and Intel, who collectively bought a 21% stake in ASML in 2012, with Intel buying the largest piece, up to 15% for $4 billion.25
Van den Brink has had long-term allies within ASML, most notably Frits van Hout, a fellow physicist who studied at Oxford and in Switzerland then also joined the company in 1984. He served as project manager for the PAS 2500’s development alongside van den Brink, who led the systems engineering. Van Hout left the company in 1992 but rejoined in 2001, two years after van den Brink’s promotion to the company board, and was himself added to the company’s board in 2009. From 2013 to 2021, he oversaw the last stages of the development of EUV as Chief Program Officer and managed relationships with ASML’s key customers and suppliers as Chief Strategy Officer. He retired in 2021.26
Van den Brink himself is expected to retire in 2024, having persuaded the board and Carl Zeiss to make the required investments in High-NA EUV against, according to him, significant opposition.27 It is unclear, however, if he will actually do so, especially given that he has already explicitly set the company’s main strategic priorities for the next decade and may be tempted to stay and drive through his plans to maximize the efficiency of the company’s current products and push EUV to the limit of its potential.28
Van den Brink did not found the tradition of knowledge that has underpinned the company’s success, but inherited it and extended it. His first manager at ASML was the Dutch engineer Herman van Heek, co-inventor of the prototype of the world's first wafer stepper. Van Heek co-developed it with his colleague Gijs Bouwhuis at the physics and materials research laboratory of the Dutch electronics conglomerate Philips, called NatLab, which was comparable in size and scope to, for example, AT&T’s famous Bell Labs. Throughout the 1970s it employed over 2000 people. NatLab’s greatest successes were technologies key to the invention of the cassette tape and the CD, both consumer electronics innovations that Philips successfully commercialized.
In 1971, van Heek and Bouwhuis developed the world’s first design for a “stepper,” which was called internally the Silicon Repeater. This technology was patented and the intellectual property later transferred to ASML. While a working prototype was built by 1973, the NatLab scientists failed to persuade Philips’ Science and Industry Division to commercialize it, although development of the stepper continued throughout the 1970s inside NatLab. The Repeater’s eventual commercialization at Philips was secured by the intervention of Wim Troost, then director of a unit of the Science and Industry division, but later ASML’s second CEO. Progress towards making a working model ready for commercial use was slow. Philips lacked the scientific expertise to develop a commercially viable model and the NatLab scientists were eventually called back in to develop a working control system. This shows how the tradition of knowledge was essential to commercial viability before ASML even existed.
By 1982, Philips had sold its first stepper, known as the PAS 2000, to IBM, its first customer beyond Philips’ own subsidiaries. But, at the time, the conglomerate’s senior management were looking to trim its size. An unprofitable cutting-edge semiconductor equipment business was an obvious target for divestment. Wim Troost, however, believed in the Repeater’s potential and so reached out to another Dutch semiconductor firm, ASM, to form a joint venture, after three other potential joint ventures fell through. ASM contributed $2.1 million in cash to the joint venture and Philips contributed seventeen PAS 2000 machines and 47 staff. The nascent company began operations in 1984 as “ASM Lithography,” now officially just ASML.
Van den Brink was one of the early external hires of the new company, but was discouraged by the initially lethargic corporate atmosphere and announced his intention to quit. In response, van Heek offered him the chance to lead the design of the lens projection and alignment systems of the PAS 2500. Van den Brink also insisted on spending a day a week at NatLab itself, learning from the men who had designed the stepper and continued improving it throughout the 1970s, including Gijs Bouwhuis and Joseph Braat, the Repeater’s optical designer.
ASML is thus a success of Dutch scientists and engineers working on photolithography at Philips and later ASML passing down this unique knowledge from one generation to the next. It is unclear whether van den Brink has a successor in mind who can replace him as the technical head of the company. The other management board members are not technical. From 2018 to 2021, the head of research at ASML was the 56-year-old Dutchman Hein Otto Folkerts, who is now the head of development and engineering for sensors and mechatronics.29 The vice president of R&D since 2022 has been Rafael Howell, a 47-year-old American.30 Both men have spent all or almost all of their careers at Philips or ASML. NatLab was disbanded in 2001 and its facilities became part of the so-called High Tech Campus Eindhoven, a corporate industrial park.
Through NatLab, ASML’s tradition of knowledge originated under the patronage of the Philips industrial empire. NatLab was founded in 1914 by the founders of Philips themselves, the brothers Gerard and Anton Philips, who had founded the company with their father in the 1890s to manufacture lightbulbs. The Philips brothers came from an illustrious family: their father was a first cousin of Karl Marx and their grandfather Lion Philips, a wealthy tobacco merchant, was Marx’s main financial sponsor.
The company grew rapidly and was given royal Dutch patronage as early as 1916; the company officially renamed itself to Royal Philips in 1998.31 The Philips family were major patrons of the city of Eindhoven and family members served as company CEO until 1977. In European social democracies like Sweden and the Netherlands, governments formed broad alliances with industrial family dynasties to support their business empires in exchange for financing generous welfare states. ASML is ultimately a product of this system and the low social mobility of social democracies likely still helps the company keep much of its talented workforce in the Netherlands rather than Silicon Valley.
Favored by Dutch Industrial Policy and U.S. National Security
Throughout its history, ASML has benefited from European industrial policy. In its early years, the company survived on grants from the Dutch Ministry of Economic Affairs, which wanted to preserve and develop the domestic semiconductor industry. It also won subsidies from the European Economic Community (EEC), the predecessor of the European Union.32 Nederlandsche Middenstandsbank and ABN Amro, two of the largest Dutch banks, also served as patrons, repeatedly providing ASML with cheap loans. Prior to its initial public offering in 1995, the company still owed large debts to Philips and the banks, which it used to argue that the Dutch tax authorities should estimate the company’s value at zero and consequently levy no tax on the 5% shareholding awarded to a group of forty company insiders. The tax authorities agreed on a $5 million valuation, considerably below the almost $600 million valuation the company later listed at.33
In 1988, when ASM wanted to sell its stake in ASML, Philips assumed its stake, including the debts ASML owed to ASM. At the time, ASML was losing money, but Philips kept the company alive and continued to lend it money. This was partly due to the intervention of German board member Gerd Lorenz, who viewed chip production and lithography as strategic assets for Europe. Against a background of growing concern over the economic rise of Asian countries and Japan in particular, both European and U.S. elites wanted to retain control of cutting-edge technologies, much as anxieties over the rise of China have driven sanctions and restrictions on Chinese companies today.34
ASML further benefited from U.S. technology policy. Much of the early scientific work on EUV was, in addition to Intel, funded by the U.S. Department of Energy. While the relevant scientific discoveries took place in Japan in the early 1980s, the wider Japanese research community was at the time extremely skeptical of its commercial viability. American researchers were similarly unconvinced, but Bill Brinkman of Bell Labs was able to persuade the Department of Energy to fund further research at Livermore Labs.35 Brinkman had a long career in the U.S. scientific bureaucracy; he subsequently served in the Obama administration as head of the Department of Energy’s Office of Science.36
From the U.S. government’s perspective, the technical leap forward to EUV offered an opportunity to break the dependence of the U.S. semiconductor industry on Nikon and Canon, and help to secure the domestic industry’s long-term future by reasserting its technical superiority over the main Japanese firms.37 By 1997 it was sufficiently clear to leading U.S. semiconductor manufacturers that EUV was worth their investment, despite the potential of alternative techniques like electron beam or X-ray lithography. Intel, AMD, and Motorola created a joint venture, “EUV LLC” to fund development. They in turn brought Silicon Valley Group on board as a partner to develop the products stemming from their collective research efforts. ASML, which had already been pursuing its own EUV efforts in partnership with Zeiss and Oxford Instruments, began working with EUV LLC in 1999 and became a fully-fledged partner upon its 2001 acquisition of SVG.38
While the Japanese companies Nikon and Canon were not part of the coalition, some information from the program was shared with them, and the Japanese government funded their own EUV development efforts. Nikon built a working prototype in 2008 but delayed and eventually abandoned its EUV program after the 2008 global financial crisis and Intel’s 2012 decision to invest in ASML, signaling its commitment to buying their machines.39 Though U.S. policymakers could have blocked ASML’s takeover of SVG on antitrust grounds, they chose to let it go through, requiring only that ASML spin out SVG subsidiary Tinsley, a key supplier of lens-polishing technology for U.S. satellites. Intel’s own advocacy for the merger appears to have been a key factor in assuaging the Bush administration’s concerns. The U.S. Semiconductor Industry Association also backed ASML’s takeover of SVG.
Intel preferred SVG survive under foreign ownership—albeit the ownership of another Western firm—rather than fold due to its inability to compete at scale with Nikon and Canon. Moreover, Intel already viewed EUV as vital for its long-term future, and while SVG held key optics technology for EUV’s development, ASML was thought of as a better platform manufacturer.40 The rising capital intensity of the lithographic industry made corporate consolidation inevitable and ASML became the de facto Western champion by the late 1990s. The fact a Dutch rather than U.S. company became the champion suggests U.S. policy was not driven by a live player interested in building up U.S. manufacturing, but reactive to narrow concerns like preventing Japanese monopolies.
For the Netherlands, ASML is today a national economic champion. Both manufacturing and R&D largely take place at the company’s facilities in and around Eindhoven, which has become a growing technology hub largely thanks to the company.41 ASML has invested heavily in the local Eindhoven University of Technology, recently announcing a new on-campus research facility and the funding of 40 PhD positions a year in plasma physics, AI, and lithography.42 Across Europe, ASML sustains a network of suppliers who depend on its success, such as Zeiss and German laser manufacturer Trumpf. For the U.S., ASML has become an effective means of restricting China’s access to the most cutting-edge chips, fulfilling a national security objective through “friendshoring.” From an economic perspective, access to EUV has helped Intel’s fabs stay competitive with TSMC and Samsung. This together means the U.S. government is unlikely to expend substantial effort to try and unseat ASML.
The Future of ASML
Demand for computer chips has been rising for decades and this is unlikely to change even if the current sharp upswing in demand due to tech companies pursuing artificial intelligence development fades away. This is because the global proliferation of information technology and the internet over the last thirty years, in everything from the workplace to the military, has rested on the development of ever more powerful computer chips that can do more useful work on smaller devices at lower costs. In the short-to-medium term, demand for GPUs driven by artificial intelligence is a major tailwind for ASML. All the largest tech companies in both the U.S. and China are seeking to rapidly increase their AI capabilities primarily through adding more computing power, rather than through improvements in model architecture or novel sources of training data.
OpenAI’s GPT-4, for example, was trained using 66 times the compute that the company deployed to train GPT-3. Google has reportedly trained their Gemini model, due for release later in 2023, on five times the compute used to train GPT-4. Meta, Amazon, Google, Microsoft, and Tesla are all buying up as many Nvidia GPUs as TSMC can make, while also developing their own chips designed in-house specifically for AI model training and inference, such as Microsoft’s Athena or Amazon’s Trainium. The chips are intended for their own in-house model development, but surplus capacity will also be rented to other players in the ecosystem, including well-funded startups such as OpenAI, Anthropic, Inflection AI, and xAI.
ASML’s financial future seems assured so long as it doesn’t face a competitor on the cutting edge of photolithography, which itself seems highly unlikely due to the capital costs that would be necessary and the difficulty of reproducing such a unique tradition of knowledge. TSMC, for example, was incubated by the Taiwanese government in the 1980s and led by Morris Chang, by then a veteran of the semiconductor industry with a long career in the U.S. at Texas Instruments, then the incumbent. On the face of it, any competitor to ASML would most likely need to both find a willing government backer and poach high-level members of ASML’s technical team. The only country that plausibly has the ability and incentive to attempt this, or anything similar, is China.
Since 2019, ASML’s expansion in China has been restricted by U.S. export controls aimed at limiting Chinese technological advances in AI. The company is already banned from selling EUV machines to China, although for now there are no sufficiently advanced Chinese fabs that could actually use them. Until recently, ASML generated around 14% of its revenue from China and at least some of its own suppliers manufacture their equipment there.43 In June 2023, further export controls were announced as part of a mutual agreement between the U.S., the Netherlands, and Japan. These restricted the export of ASML’s most advanced DUV machines.44
ASML’s senior executives have repeatedly argued against the global trend towards regionalization and reshoring of semiconductor supply chains, arguing that in the long run this will simply force China to develop its own end-to-end domestic semiconductor supply chain and that, in the words of CEO Peter Wennink, “the laws of physics in China are the same as here.”45 If the Chinese government pursues such a policy, ASML is likely to be the target of intensive Chinese industrial espionage, especially given its close links to its local university and its decision to open up a R&D facility in Taiwan.46 In all likelihood this is already happening, with, for example, at least one report of an ex-ASML employee stealing IP and now being employed by Huawei.47 Industrial espionage will not, however, be able to recreate the whole tradition of knowledge underpinning ASML’s technical superiority. If China succeeds in matching ASML in the future, it will be due to the exceptional achievement of recreating its own tradition of knowledge in “the laws of physics.”
Despite its arguments for keeping the semiconductor industry globalized, ASML and its domestic position as a Dutch national champion mean the company will ultimately comply with U.S. efforts to choke off Chinese access to cutting-edge chip manufacturing technology. The Netherlands are a close U.S. ally and the company’s compliance at the behest of Dutch authorities illustrates Europe’s fundamental strategic dependence on the U.S. and inability to pursue an independent foreign policy.
ASML’s technological future is less assured than its financials or politics. Martin van den Brink has “mapped out” the course of the next ten years of ASML’s strategy, focused on deploying the High-NA EUV machines and reducing the costs of both EUV and High-NA patterning.48 He believes there is still room for improvement in making EUV machines more sophisticated and economical. Van den Brink has reiterated his suspicion that there will not be another generational breakthrough in EUV after High-NA, the so-called “Hyper-NA,” because he believes the costs of manufacturing still smaller chip structures might be insurmountably high.49
The company has a research group working on “Hyper-NA,” however, just in case. So long as van den Brink is leading ASML, there is good reason to think the company will continue making incremental improvements in photolithography and maintain its monopoly, perhaps even making one more big breakthrough. But the key juncture for the company will be the day when van den Brink inevitably retires or dies, when technical leadership will fall to an as-yet unknown successor. While the company would likely stay immensely profitable for a long time, similarly to Apple, that might ultimately leave an opening for technological disruption from a new competitor. Such a competitor would have to bet on a new technological paradigm of lithography and is today more likely to emerge in China or even South Korea rather than the U.S., let alone anywhere in Europe.
“Largest Companies in the EU by Market Capitalization,” Companies Market Cap, 2023, https://companiesmarketcap.com/european-union/largest-companies-in-the-eu-by-market-cap.
“Fitch Affirms ASML at 'A'; Outlook Stable,” Fitch Ratings, April 5, 2023, https://www.fitchratings.com/research/corporate-finance/fitch-affirms-asml-at-a-outlook-stable-05-04-2023.
Samo Burja, “Intellectual Dark Matter,” Medium, July 16, 2019, https://medium.com/the-long-now-foundation/intellectual-dark-matter-2e5890aa8d8f.
Sam Shead, “ASML: The Biggest Company in Europe You've Probably Never Heard Of,” CNBC, November 24, 2021, https://www.cnbc.com/2021/11/24/asml-the-biggest-company-in-europe-youve-probably-never-heard-of.html.
Samo Burja, “On the Loss and Preservation of Knowledge,” Samo Burja, March 8, 2018, https://samoburja.com/on-the-loss-and-preservation-of-knowledge.
Cheng Ting, “ASML Says Decoupling Chip Supply Chain is Practically Impossible,” Financial Times, June 27, 2023, https://www.ft.com/content/317be8b3-48d9-411e-b763-261a179c9d0d.
E.g Kalani Scarrott at “50 | ASML & EUV Lithography Deep Dive with Asianometry,” Compounding Curiosity, January 17, 2023, https://compoundingpodcast.com/ep50.
“EUV Lithography Systems – Products,” ASML, https://www.asml.com/en/products/euv-lithography-systems.
Peter Wennink and Roger Dassen, “Transcript Investor Call,” ASML, January 25, 2023, https://www.asml.com/-/media/asml/files/investors/financial-results/q-results/2022/q4/investor-call-prepared-remarks.pdf?rev=b05d17e3b37d437dac4f72de2b6f5829.
Toby Sterling, “Intel Orders ASML System for Well Over $340 Mln in Quest for Chipmaking Edge.” Reuters, January 19, 2022, https://www.reuters.com/technology/intel-orders-asml-machine-still-drawing-board-chipmakers-look-an-edge-2022-01-19.
Data from Koyfin.
“ASML reports €21.2 billion net sales and €5.6 billion net income in 2022,” ASML, January 25, 2023, https://www.asml.com/en/news/press-releases/2023/q4-2022-financial-results.
Chris Miller, Chip War: The Fight for the World’s Most Critical Technology, Simon & Schuster, 2022, p.228.
“Inside ASML, the Company Advanced Chipmakers Use for EUV Lithography,” CNBC, March 22, 2022, https://www.cnbc.com/2022/03/23/inside-asml-the-company-advanced-chipmakers-use-for-euv-lithography.html.
ASML, “Post on X: "Today, our CTO Martin van den Brink spoke at our strategic partner ZEISS in Oberkochen, Southern Germany to mark their 50 year anniversary in semiconductor optics...” X, September 14, 2018,
“ZEISS and ASML Strengthen Partnership for Next Generation of EUV Lithography,” ASML, November 3, 2016, https://www.asml.com/en/news/press-releases/2016/zeiss-and-asml-strengthen-partnership-for-next-generation-of-euv-lithography.
Chris Miller, Chip War: The Fight for the World's Most Critical Technology, Simon & Schuster, 2022. pg. 229.
Roberta Cowan, “Chip Gear Maker ASML Buys Cymer for $2.5 Billion,” Reuters, October 16, 2012, https://www.reuters.com/article/us-asml-cymer-idUSBRE89G08R20121017; Daan van Monsjou, “Hoe ASML de macht greep met euv - Deel 2,” Tweakers, April 8, 2023, https://tweakers.net/reviews/10832/2/cto-martin-van-den-brink-over-de-obstakels-naar-asmls-euv-alleenheerschappij-lichtbronproblemen-en-overname-cymer.html.
“ASML Annual Report 2022,” ASML, 2022, https://www.asml.com/en/investors/annual-report/2022.
Rene Raaijmakers, “ASML’s Architects”, Techwatch Books, 2020, p.417 (Kindle edition)
Sander Hofman, “Three decades of PAS 5500 – Stories,” ASML, May 7, 2021, https://www.asml.com/en/news/stories/2021/three-decades-of-pas-5500.
Rene Raaijmakers, “ASML’s Architects”, Techwatch Books, 2020, p.504 (Kindle edition)
Rene Raaijmakers, “ASML’s Architects”, Techwatch Books, 2020, pp. 409-13 (Kindle edition)
Martin van der Brink, “The Only Cost Effective Extendable Lithograph Option: EUV,” International Symposium on Extreme Ultraviolet Lithography, October 17, 2006, https://euvlsymposium.lbl.gov/pdf/2006/pres/D2%20Keynote_van%20den%20Brink.pdf.
Edwin Chan, “Intel Funds Next-Gen Chipmaking, Buys Into ASML for $4.1 Billion,” Reuters, July 9, 2012, https://www.reuters.com/article/ctech-us-intel-asml-idCABRE86819B20120709; Peter Clarke, “TSMC to Invest in ASML,” EE Times, August 6, 2012, https://www.eetimes.com/tsmc-to-invest-in-asml.
Peter van der Hurk, “Farewell to a “Big Family of Top-Class People:” Frits van Hout retires from ASML,” ASML, April 23, 2021, https://www.asml.com/en/news/stories/2021/frits-van-hout-retires-from-asml.
Paul van Gerven, “Hyper-NA After High-NA? ASML CTO Van Den Brink Isn't Convinced,” Bits&Chips, September 22, 2022, http://bits-chips.nl/artikel/hyper-na-after-high-na-asml-cto-van-den-brink-isnt-convinced.
“Hein Otto Folkerts - Cluster Head of Development & Engineering for Sensors and Mechatronics - ASML,” LinkedIn, https://nl.linkedin.com/in/hein-otto-folkerts-77521812.
The Associated Press, “Philips Name to Go Royal,” The New York Times, February 20, 1998, https://www.nytimes.com/1998/02/20/business/philips-name-to-go-royal.html.
Rene Raaijmakers, “ASML’s Architects”, Techwatch Books, 2020, p. 375 (Kindle edition)
“Asml Market Cap 1995,” StatMuse, https://www.statmuse.com/money/ask/asml-market-cap-1995.
Rene Raaijmakers, “ASML’s Architects”, Techwatch Books, 2020, p. 401 (Kindle edition)
Sander Hofman, “Making EUV: From Lab to Fab – Stories,” ASML, March 30, 2022, https://www.asml.com/en/news/stories/2022/making-euv-lab-to-fab.
Robert Service and David Malakoff, “DOE Science Chief to Step Down,” Science, March 20, 2013, https://www.science.org/content/article/doe-science-chief-step-down.
Michael Borrus, “Foreign Participation in US-Funded R&D: the EUV Project as a New Model for a New Reality.” eScholarship, 1998, https://escholarship.org/content/qt2b8155q0/qt2b8155q0_noSplash_65e4d3c7b90dafb87401616eb50de282.pdf.
Nelson M. Felix and David T. Attwood Jr. "EUV Lithography Perspective: from the beginning to HVM (Conference Presentation)", Proc. SPIE 11323, Extreme Ultraviolet (EUV) Lithography XI, 113232O (28 April 2020); https://doi.org/10.1117/12.2572271. Available at https://play.cadmore.media/Player/x6152710307001?userstatus=0
John Walko, “Nikon said to be delaying EUV lithography development,” EDN, May 15, 2009, https://www.edn.com/nikon-said-to-be-delaying-euv-lithography-development/; See here: https://www.fitchratings.com/research/corporate-finance/intel-deal-positive-for-asml-semiconductor-industry-17-07-2012
Mark LaPedus, “Who loses if Bush blocks ASML-SVG merger? ASML, SVG, and Intel,” EE Times, May 2, 2001, https://www.eetimes.com/who-loses-if-bush-blocks-asml-svg-merger-asml-svg-and-intel.
Andy Bounds, “Made in Eindhoven: the Small Dutch City that Became a Tech Powerhouse,” Financial Times, February 13, 2023, https://www.ft.com/content/c702c9f0-7f45-47d2-a0b3-6a6525e4b583.
“ASML and Eindhoven University of Technology Strengthen Longstanding Collaboration,” ASML, April 24, 2023, https://www.asml.com/en/news/press-releases/2023/asml-and-tue-strengthen-longstanding-collaboration.
Jim Pollard, “ASML Suppliers Plan to Cut China Exposure As Chip War Heats Up,” Asia Financial, March 11, 2023, https://www.asiafinancial.com/asml-suppliers-plan-to-cut-china-exposure-as-chip-war-heats-up.
Pieter Haeck, “The Netherlands to Block Export Of Advanced Chips Printers to China,” Politico.eu, March 8, 2023, https://www.politico.eu/article/netherlands-impose-restrictions-chips-export-to-china-asml/; Cagan Koc and Diederik Baazil, “ASML to Face New Export Restrictions to China as Soon as June 30,” Bloomberg, June 22, 2023, https://www.bloomberg.com/news/articles/2023-06-22/dutch-to-publish-new-chip-export-controls-as-soon-as-next-week.
Cagan Koc. “ASML Says Chip Controls Will Push China to Create Own Technology,” Bloomberg, January 25, 2023, https://www.bloomberg.com/news/articles/2023-01-25/asml-says-chip-controls-will-push-china-to-create-own-technology#xj4y7vzkg; Cheng Ting, “ASML Says Decoupling Chip Supply Chain Is Practically Impossible,” Financial Times, June 27, 2023, https://www.ft.com/content/317be8b3-48d9-411e-b763-261a179c9d0d.
“ASML Announces Plans for New Taipei City Plant,” Taipei Times, November 18, 2022, https://www.taipeitimes.com/News/biz/archives/2022/11/18/2003789129.
Paul van Gerven, “Hyper-NA After High-NA? ASML CTO Van Den Brink Isn't Convinced,” Bits&Chips, September 22, 2022, http://bits-chips.nl/artikel/hyper-na-after-high-na-asml-cto-van-den-brink-isnt-convinced.