The Lithography Wars: The Story of How Nikon Lost to ASML

0:00 Inside this clean room, under conditions more sterile than an operating theater, is a machine that costs $400 million. It's the size of a double-decker bus and is shipped in three separate Boeing 747s. It is, without exaggeration, the most complex and precise machine ever built by human hands. But it doesn't build rockets or cure diseases. It prints. It prints patterns of light onto silicon, a process so fundamental that without it, the device you're watching this on, the internet that delivered it, and the entire digital civilization we've built would simply not exist. This is a photoiththography machine, and for decades, the undisputed masters of this universe were Japanese, a name

0:46 synonymous with precision, optics, and quality. Nikon. In the 1990s, Nikon along with its domestic rival Canon was an unshakable titan controlling over 75% of the global market for these critical machines. They were the gatekeepers to the future. So, how did this happen? How did a small Dutch company born in a leaky shed next to a Philips office in Eintoven not only challenge the giant, but completely dethrone it? This isn't just a story about a corporate rivalry. It's a tale of two profoundly different philosophies. It's about the clash between a culture of perfection and a culture of collaboration. It's the story of how one company's

1:31 fatal caution costed everything and how another's audacious world spanning gambles won at a monopoly on the most important manufacturing tool on Earth. This is the story of the lithography wars. To understand Nikon's power, you have to go back to the 1980s. This was the era of Japan incorporated. The Japanese economy was a miracle, and its semiconductor industry was the crown jewel. By the late 1980s, Japan's share of the global semiconductor market had skyrocketed to over 50%, overtaking the United States. By 1989, six of the top 10 semiconductor companies in the world were Japanese. This dominance was built

2:19 on a unique model. Massive vertically integrated conglomerates or ketssu working in concert with a supportive government. Companies like Hitachi, Toshiba, and NEC weren't just making chips. They were building entire ecosystems where every part of the process from materials to final product was controlled inhouse or by a close partner. And at the heart of this ecosystem was the machinery to make the chips. The key technology of the era was the stepper. Think of it as a hyper advanced slide projector. It takes a master blueprint of a chip circuit called a reticle and projects a miniaturized image onto a section of a silicon wafer. Then it steps to the next

3:04 section and repeats the process. The company that could build the most precise lenses and the most stable stepping mechanism would win. And nobody was better at this than Nikon. Building on decades of worldclass camera and lens manufacturing, Nikon's mastery of optics was legendary. The Japanese government through its Ministry of International Trade and Industry, MITI, had specifically directed Nikon to enter the lithography business in the late 1970s, viewing it as a strategic national technology essential for Japan's economic ambitions. Their relationship with the powerful Japanese chip makers was symbiotic. As a key member of the Mitsubishi Ketsu,

3:49 Nikon had a built-in loyal customer base. These deep insular relationships provided a constant feedback loop, allowing them to perfect their machines in a closed circle of innovation. By the mid 1990s, Nikon and its rival Cannon together held a commanding 3/4 of the entire lithography market. They seemed invincible. This fortress-like structure, however, contained the seeds of its own demise. The very self-sufficiency that made Nikon a titan of the 1980s also made it culturally and structurally resistant to the kind of disruptive external ideas and collaborative risk sharing that would become essential in the coming decades. While Nikon was perfecting every

4:35 component in house, a different model was emerging elsewhere. The vertically integrated Japanese system was about to collide with a new horizontal and globalized world. a world that Nikon was illequipped to navigate. But while Nikon was perfecting its steppers, a new idea was emerging. Instead of exposing a whole chip pattern at once and then stepping, what if you scanned the pattern across the wafer like a modern office scanner? This is step and scan approach had two huge advantages. First, it allowed for much larger chip designs without needing impossibly large, flawless, and expensive lenses.

5:22 Second, by sweeping the image across the wafer, it averaged out any tiny imperfections in the lens system, leading to better uniformity and fewer defects across the chip. This wasn't just an improvement, it was a fundamentally superior architecture. And the company that bet everything on it was the Dutch underdog, ASML. ASML's origins were anything but glamorous. Spun out of the Dutch electronics giant Phillips in 1984, their first office was a leaky shed in Heinhovven. For years, they struggled, burning through cash with few customers and facing fierce competition. But they had a powerful idea and crucially a key partner from the very beginning, the German optics specialist Carl Zeiss, who

6:09 would design and build their complex lens systems. This reliance on an external partner born of necessity would become a defining feature of ASML's corporate DNA. Their breakthrough came in the early 1990s with the pass 5,500 platform, a step and scan system that was more productive and offered better resolution than the competition. It was this machine that finally got them in the door with major chip makers outside of the Japanese ecosystem, including a pivotal early order from a fledgling Taiwanese foundry called TSMC. Nikon wasn't oblivious to this new technology. They eventually developed their own scanner, the NSR-S20A,

6:55 which they released in 1995. But their corporate focus and the bulk of their R&D remained on their established and highly profitable stepper technology. They were the masters of a mature technology while ASML was a hungry innovator in a new one. You can see the difference in their very design philosophy. ASML's systems were modular, built with components from a global network of specialized suppliers. Nikon's machines were a product of their vertically integrated culture, highly precise but monolithic with most critical components developed in house. This was the first hint of the strategic divergence that would define the war to come. One company was building a fortress, the other was

7:41 building a network. This situation perfectly illustrates a classic business trap. As the market leader with profitable established products and loyal customers who relied on them, Nikon was rationally disincentivized from aggressively pursuing the new scanner architecture that would cannibalize its existing business. Their customers wanted better steppers, not a risky new machine. ASML, with almost nothing to lose and no legacy business to protect, could go allin on the future. Nikon's failure to dominate the scanner transition wasn't necessarily a result of incompetence, but a predictable outcome of a market leader being trapped by its own success. A pattern that would repeat itself with devastating consequences.

8:32 By the early 2000s, the entire semiconductor industry was hitting a physical wall. To continue shrinking circuits according to Moore's law, chipmakers needed to use light with a shorter and shorter wavelength. But building reliable lasers for deep ultraviolet DUV light was incredibly difficult and expensive. The solution was deceptively simple and it was championed by a brilliant Taiwanese engineer at TSMC named Burn J. Linn. What if you put a drop of ultra pure water between the final lens and the silicon wafer? Light bends when it passes from one medium to another. By replacing the air with water, which has a higher refractive index, you can bend the light more sharply. This effectively

9:18 increases the lens's power, its numerical aperture, or NA, and allows you to print much finer, more detailed patterns using the same 1 193 nanometer light source that everyone was already using. This was immersion lithography. This is the moment the war was truly won and lost. ASML saw immersion as the future and went allin. They deepened their collaborative partnership with a key customer who was desperate for this technology, TSMC. In 2003, TSMC led by Dr. Lynn. Lynn placed the industry's first ever order for a productionready immersion tool, the ASML Twin Scan XT1250i to be delivered in 2004.

10:04 It was a massive company-defining gamble for both of them. Nikon on the other hand was cautious. As the established leader, they viewed the new technique as risky, worrying about practical engineering problems like water bubbles, contamination, and defects caused by leftover water marks. Instead of fully committing to this unproven method, they hedged their bets. They poured resources into extending dry lithography, developing a proprietary technology called Pano, which used polarized light to improve image contrast without the need for water. They believed they could push the existing paradigm just a little further. A conservative incremental strategy that had always worked for them in the past. It was a fatal

10:50 miscalculation. dry lithography hit its physical limits just as ASML and TSMC were collaboratively solving the engineering challenges of immersion. Here, ASML's unique machine architecture gave them a critical advantage. Their twin scan dual stage platform measured the wafer on a completely separate dry stage and then moved it to a second stage for the wet exposure process. This brilliant design isolated the complex immersion process, minimized contamination risk, and dramatically boosted throughput compared to single stage designs. Nikon's competing tandem stage platform was a different concept. Combining an exposure stage with a calibration stage, which proved to be less productive and

11:37 ultimately less successful. The market's verdict was swift and brutal. Chipmakers desperate for the next node flocked to ASML for their proven high throughput immersion tools. The numbers tell the story. In 2006, the first full year of the immersion market, ASML captured a staggering 72.4% share. By 2011, that had grown to an overwhelming 82%. While Nikon's share had collapsed to just 18%. The war for the cutting edge of lithography was effectively over. ASML's victory wasn't just about choosing the right technology. It was about how they developed it. Their deep partnership with TSMC represented a fundamental

12:25 shift from a supplier ccentric innovation model where Nikon the expert would build a perfect machine and deliver it to a customercentric collaborative one. The idea for immersion was championed from within TSMC, the customer facing the manufacturing wall. ASML didn't just sell a machine. They co-developed a solution to their customers most critical problem. In an industry of skyrocketing complexity, this collaborative model proved decisively superior. Even with immersion, Moore's law demanded the next leap. The industry needed to move to an even shorter wavelength of light, extreme ultraviolet or EUV.

13:11 But this was a challenge of a completely different magnitude. EUV light with a wavelength of just 13.5 nanome is so energetic that it's absorbed by everything including air and even the highest quality glass lenses. This meant an EUV machine had to operate in a perfect vacuum. It couldn't use lenses, only a series of impossibly perfect, defect-free multi-layer mirrors. And the light source itself was a physics nightmare. The eventual solution was mind-bogglingly complex, firing a high-powered industrial laser at a microscopic droplet of molten tin 50,000 times a second to create a plasma that emits just enough EUV light to be captured. The technical hurdles and the

13:57 projected costs were so immense that both Nikon and Canon after initial research officially abandoned their EUV programs deeming them commercially and technically unfeasible. ASML pressed on but by 2012 even they were on the brink. The R&D costs were astronomical running into the billions of dollars with no clear finish line in sight. So their leadership did something unprecedented in the history of manufacturing. They went to their biggest customers, the very companies whose future existence depended on EUV, and made them an offer they couldn't refuse. In a deal that would permanently reshape the semiconductor industry, three of the world's largest chip makers, Intel, Samsung, and TSMC,

14:44 collectively invested over€5 billion directly into ASML. They provided€1.4 €4 billion in non-refundable R&D funding and bought a combined 23% equity stake in the company. Think about that. The customers became co-owners and co-funders of the machine they desperately needed to buy. It was a brilliant symbiotic solution to a problem that no single company, no matter how large, could solve alone. It was the ultimate expression of ASML's collaborative philosophy. For Nikon, this was the end of the road. Trapped in a more traditional, conservative corporate structure and lacking the global, open, and trusting relationships that ASML had cultivated, they simply

15:30 could not match this level of shared investment and risk. They had no choice but to formally concede the leading edge market entirely, focusing instead on older, less profitable lithography technologies. The war was over. ASML had won a total victory. Today, ASML holds a 100% monopoly on EUV lithography machines, the only tools capable of producing the world's most advanced chips. The price of a single EUV machine has soared past $200 million. And the newest high NA models, which enable even smaller features, cost nearly $400 million each. And every advanced chip maker in the world, Intel, Samsung, TSMC, has to buy

16:16 them from the very company they helped create. But was Nikon's loss entirely its own fault? Or was it a symptom of a larger national malaise? The 1990s and early 2000s were Japan's infamous lost decades, a period of prolonged economic stagnation following the spectacular collapse of its asset bubble. This environment fostered a deep-seated corporate culture of risk aversion. It's fair to ask, could any Japanese company, no matter how well-run, have realistically funded the EUV moonshot in that economic climate? Some analysts argue that the problem was more structural. The entire Japanese semiconductor industry failed to adapt

17:03 to the new horizontal fableless foundry model that was reshaping the globe. They clung to their vertically integrated past while the world moved on. Nikon's intense focus on its domestic market and its tight-knit ketsu partners may have blinded it to the global collaborative approach that was becoming necessary for survival at the cutting edge. Then there's the other side of Nikon's business. cameras. Throughout the 2000s, as the lithography wars raged, Nikon was enjoying a golden age. The transition to digital photography was a massive boom, and its DSLR cameras were hugely popular and immensely profitable. Did the steady, reliable profits from the camera division make the company as a whole

17:50 complacent? It's possible that this success starved the far riskier, more capital-intensive lithography business of the bold investment and high-risisk appetite it desperately needed to compete with a single-minded all or nothing competitor like ASML. When one part of your business is printing money, it becomes much harder to justify betting the entire company on a high-risk long-term R&D project. There's also a geopolitical shadow to this story. In the 1990s, the US government, wary of Japan's technological dominance in semiconductors, actively funded domestic efforts to create an American lithography champion through consorcia like Seitech. When that failed, some

18:36 argue that US policy, which restricted Japanese access to foundational EUV research emerging from US national labs, inadvertently favored the European underdog ASML, who was allowed to participate? Was the deck stacked against Nikon from the start? Ultimately, Nikon's defeat was not the result of a single bad decision, but a cascade of interconnected failures rooted in its corporate DNA. Its cultural conservatism led to technological caution, choosing the safety of dry lithography over the gamble of immersion. Its financial conservatism driven by the success of its camera division and Japan's broader economic stagnation prevented the necessary scale of investment in EUV and

19:23 its insular self-reliant structure made the kind of collaborative funding model that saved ASML conceptually and practically impossible. Each failure compounded the last, creating a perfect storm where the company's historical strengths became its fatal flaws. So, how did Nikon lose? They lost because they were busy perfecting the last war while ASML was inventing the next one. They lost because they valued self-reliant internal perfection over open global collaboration. In the end, the most complex machine in the world had become too complex for any one company or even any one country to build alone. ASML didn't just build a better machine.

20:09 They built a better business model. They created an open global risk sharing network that treated its suppliers like Carl Zeiss and its customers like TSMC as essential partners in innovation. The result is the world we live in today. A world where a single company in the Netherlands holds a terrifying and absolute monopoly on the technology required to build the future of computing. This has turned ASML into one of the most critical geopolitical choke points on the planet. The ongoing tech war between the United States and China is in many ways a war for control over access to ASML's machines. US export controls are designed specifically to prevent China from acquiring EUV

20:55 technology, a move that puts immense pressure on ASML and the Dutch government, who are caught between their largest trading partners. The fate of global technological progress now rests in part on the strategic decisions made in this one small town in the Netherlands. The lithography wars may be over, but their lessons echo. They serve as a stark reminder that in the relentless march of technology, no empire is forever. And that sometimes the future isn't won by the strongest or the most precise, but by those brave enough to share the risk and humble enough to ask for