ASML: from a leaky shed to the chip industry’s key player

ASML: from a leaky shed to the chip industry’s key player

ASML top executives Peter Wennink (CEO) and Martin van den Brink (CTO) will retire on Thursday. Frenchman Christophe Fouquet is set to replace Wennink, while Van den Brink’s role will vanish. The ASML left behind by the pair is incomparable to the miniscule operation it was 40 years ago. As a linchpin in the chip industry, it has endured great challenges during that period, from economic vagaries, a huge partner network, Japanese competition and (frequently U.S.) regulations.

In 1984, ASM Lithography was founded as a joint venture of ASM International and Philips. The fledgling company operated out of a leaky shed, as ASML itself still likes to remind us. In the few years immediately after its founding, it found itself mostly in troubled waters with small margins, high development costs and uncertain economic conditions. Martin van den Brink has steered ASML’s R&D direction since the beginning. That direction was ambitious, but an absolute necessity to secure a place in the highly competitive chip industry, even then. It cost lenders Philips, ASM and the Dutch government millions, with financial results solidly in the red for years. ASML did emerge as the clear No. 3 in the chip machine sector behind Japanese powerhouses Nikon and Canon during that time. It mainly served the “B-segment” among chip manufacturers. For example: it wasn’t able to sell to market leader Intel but did do business with its challenger, AMD.

The early days of this company, first marketed with the ill-advised acronym ALS, were characterized by difficulty in finding customers. By the mid-1980s, it was already widely known that the future was to be defined by microchips, but all the production steps were pricey. In the long run, the vertically integrated Philips proved not a suitable home, partly because of an undesirable political corporate top that was slow to make decisions. It was also far more concerned about selling electronic commodities than continually pushing the bleeding edge of the hardware inside. In addition, its partner ASM International proved unable to make the necessary investments in time.

Turnaround

Crucial to ASML’s change in fortunes was the 1991 PAS (Philips Automatic Stepper) 5500, which marked the beginning of the company’s march toward market leadership. Compared to competing machines, it significantly reduced production time for customers. Its performance and yield rate (the percentage of successful chips from a layer of silicon) impressed IBM, which was the first to buy a 5500. Gradually, ASML reached the runner-up position behind Nikon due to orders from other parties. An advantage for the new customer base was the modular nature of the PAS 5500: this design allowed the roughly ten components that made up the machine to be upgraded independently. This brought down production costs dramatically for ASML and its partners, a benefit it passed on to customers.

The resulting extra financial clout led to ever-higher demands on suppliers of the required components. Alongside several top executives, Van den Brink was regularly involved in such discussions. These partners themselves had to invest heavily to make chip innovations possible, something that has led some of them to bankruptcy over the years. Nevertheless, German lens specialist Zeiss is the leading example of a successful, long-term partnership with ASML. It now spends astronomical R&D sums on new lithography lenses, but it is allowed to charge top prices for them in turn.

Fundamental to the ASML success proved to be the respect it had with regard to external expertise. French IT services provider and longtime partner Capgemini emphasizes this hallmark. ASML, according to the Capgemini representatives, will always tell you if you’re doing a good job. If you aren’t, for that matter, they’ll bluntly tell you the opposite as well.

IPO and expansion drive

ASML’s IPO took place in 1995, with Deloitte as its accounting partner. Peter Wennink, working at Deloitte, became involved with the Veldhoven-based company as a result. He joined ASML in 1999 and served as CFO from that year until 2013.

Read more: Imec: Strong cooperation between EU and US needed

In the late 1990s, Taiwanese chip giant TSMC was already the most prominent customer, as it is today. A good relationship with TSMC chief Morris Chang and investments in a Taiwanese ASML site led to an intimate relationship between the two companies. Intel, which produced the most advanced chips at the turn of the century and served (and serves) most of the PC market, joined later.

Intel actually only became involved once its production partner Silicon Valley Group (SVG) was acquired by ASML. This move completed the market share catch-up with the Japanese competition. Since then, ASML has been the largest player in the lithography market, and by a country mile at that.

EUV investment

The PAS 5500 and its successors are all DUV (deep ultra-violet) machines, a term that refers to the wavelength of light the equipment uses to turn chip designs into a reality. The more precise these designs, the more advanced the chips. DUV has remained the standard in this field since the early 1990s, but EUV (extreme ultra-violet) has proven to be the next step. ASML invested in this technology back in the early ’00s, despite the fact that it knew it’d be a long-term bet. In fact, it would take a decade and a half for EUV to become production-ready.

In the meantime, ASML survived serious economic challenges after 9/11 and the subsequent 2008 financial crash. It did cost thousands of employees their jobs in both cases. But, by the late ’00s, ASML’s market share was already two-thirds of the entire chip machine industry. That leadership role offered ASML no guarantees, but because no other party could keep up with its technological lead, the company’s position was unchallenged.

In 2015, EUV technology debuted in an ASML machine. Intel refrained from buying such a tool years earlier because of doubts about the technology’s maturity. The American chipmaker had been at the cradle of EUV and contributed to its development for years, but thus did not initially reap the benefits. TSMC was convinced as early as 2010 when it appeared that Apple would be willing to take orders once EUV entered the fray. This fact shows more clearly than ever what ASML’s power means: Intel was soon surpassed technologically in a relatively short time by TSMC and Samsung, which did buy EUV machines. Their customers, such as AMD and Nvidia, either made enormous inroads in their markets as a result or came to utterly dominate them.

The acquisition of US-based Cymer for nearly $2 billion in 2012/13 allowed EUV development at ASML to make a final sprint to the finish. Cymer’s expertise was in the area of the EUV light source: it knew how to optimize this component so that chip designs could be drawn as precisely as possible. Starting in 2015, ASML sold numerous EUV machines, each worth hundreds of millions of euros, to various customers. Around 2018, Apple and TSMC decided it was finally time to bring an EUV product to market, with future iPhones being built by them.

Later, Intel also knocked on the door for an EUV machine, when it had clearly missed the boat. This error in judgment seems to have finally been corrected: the first High-NA EUV machine, the successor to EUV, now sits in an Intel factory.

Tip: Newest ASML machine at Intel is ready to go, with plenty of R&D ahead

Profit explosion and geopolitical plaything

Now that we can take stock of the EUV era, there is only one conclusion to draw. The success of EUV has taken ASML to new heights. Whereas an ASML share would cost you about 100 euros in early 2015, it’s worth about 900 euros in 2024. At the time of writing, the company value is 340 billion euros. While this is less impressive than the approximately 2 trillion euros Nvidia is worth, this AI chip maker could never have topped that unimaginable market cap without ASML equipment. Any tech giant with a chip design has ASML to thank for it turning into a physical product.

Also read: ASML sends High-NA EUV machine to second customer

This, incidentally, has also caused the great geopolitical tension ASML faces. To the annoyance of outgoing CEO Wennink, EUV machines are not allowed to be supplied to Chinese companies. According to him, it’s a counterproductive move: eventually Beijing will innovate to make similar technology itself. But regardless of Wennink’s wishes, he simply has to deal with the demands of the US, the EU and the Netherlands, all of which have a say in where ASML machines end up.

News organizations like Bloomberg, CNBC and others tend to talk about a “chip war” in this area. In the process, many a journalist suddenly discovered that a Veldhoven company, obscure to them, is pulling the strings behind the scenes. At the hands of Washington, Brussels and The Hague, ASML has now become a geopolitical plaything, with numerous countries craving more autonomy in the chip sector. Whether there is a “war” at hand or not, ASML gets all the pros and cons of being at its core. The demand for chip machines is high; so are the restrictions.

In turn, the Netherlands wishes to keep ASML in the Netherlands. Operation “Beethoven” served to keep ASML a Dutch entity. Still headquartered in Veldhoven, it has been running into growth limits for some time. France, Germany and other countries have already knocked on the door to be the next main stage, but the Dutch government recently managed to satisfy ASML in the nick of time. It is Wennink’s final success story as CEO, while for Van den Brink (like Wennink, born in 1957) it is also time to retire.

New chapter

The new ASML chapter is heralded by the appointment of Christophe Fouquet. The Frenchman has been with ASML for 15 years, but emphasizes a significant change. Globalization now characterizes the company rather than its homegrown Philips roots. It has long since grown far beyond the borders of the Netherlands, with tens of thousands of employees spread across North America, Europe and Asia. The fact that every government is itching to claim a meaningful piece of the company emphasizes that under Peter Wennink and Van den Brink, ASML has become an indispensable part of the chip sector.

Reading tip: ASML is allowed to expand in Dutch city of Eindhoven

At the same time, an entirely different factor will change ASML’s role: the laws of nature. Although a successor to the very young High-NA tech is already in the works with “Hyper-NA,” the dimensions deployed within chip production are already almost as small as currently seems physically possible. Van den Brink is not even convinced that another step beyond High-NA will follow. After forty years of pushing boundaries, there suddenly appears to be a technological limit, hardcoded into reality itself. He has believed this for years; the main concern is that Hyper-NA will never be economically viable. The margins for error are probably simply too small, Van den Brink notes.

The focus, as assessed by Bits & Chips as well as others, will be on efficiency and not “firefighting” through tumultuous innovation led by Van den Brink. As things stand in the here and now, ASML is the established player, and it cannot count on the same performance leaps the chip world has made so far.