Can any manufacturer challenge ASML?

The advancement of Moore's Law is no longer as fast as before, below 2 nanometers and even into the angstrom level, and competitors are trying to catch up. Can Dutch semiconductor equipment manufacturer ASML maintain its monopoly leadership in the extreme ultraviolet (EUV) lithography equipment market?

Will the world's most advanced high numerical aperture EUV machine become the ultimate weapon for customers to achieve technological leaps? How will geopolitical uncertainty change ASML’s strategy going forward?

G. Dan Hutcheson, Vice Chairman of TechInsights, who has been covering ASML since the 1970s when it was a subsidiary of Philips, and Marc Hijink, a journalist for the Dutch newspaper NRC and author of the book Focus – The ASML Road, address these questions in an insightful presentation.

While ASML and other semiconductor equipment suppliers have benefited from demand driven by U.S. export controls on China, geopolitics could have a long-term impact on the industry as additional capacity and ecosystems are established.

Hutcheson and Hijink see ASML following in the footsteps of its customers and diversifying its production away from the Western Pacific Corridor to the United States and Europe. Still, there may be concerns about the efficiency of its customers' fab operations.

“We’re in a new world of lower tool utilization,” Hutcheson said. “The problem is, if you can’t find workers to run the tools or repair them to keep them running, your fab is useless. "

Hijink observed that despite the localization efforts of various countries and regions, there is a talent shortage problem as South Korea and Taiwan remain important components of ASML's expansion, and Asia will remain an important chip production center in the future.

However, Hijink warned, “ASML’s biggest concern is that the current restrictions on Chinese technology may even spur it and cause bigger problems in the long run.” He noted that the less China gets these Western lithography tools, the more Chinese companies will be inclined to build their own. “Even though they may not be able to compete with ASML and Nikon now, in the long run you will see more room for growth and R&D funding.”

However, Hutcheson believes that without large government subsidies, no company can challenge the monopoly of ASML lithography machines. "Five years ago, I proposed this maxim: The number of unsubsidized competitors that a market can sustain is equal to 1/5 of the total market size divided by the cost of developing a new generation of technology, and then subtract 1.5 times that cost," Hutcheson said. “In order to attract another competitor in the market, the risk threshold must be low enough so that there is enough available R&D to support the development of at least 1.5 more tools than what is currently on the market at the time.”

In these 5 years, this maximum value has remained around 100, or there are more and more sub-markets of semiconductor equipment, which explains the market consolidation, from more than 20 semiconductor equipment manufacturers in the 1980s to now There are only a few. Hutcheson explained.

As a veteran of the semiconductor industry, Hutcheson has witnessed how ASML has survived the vicissitudes of the past four decades through better management and technological development. Hijink added that ASML has survived because despite being on the verge of bankruptcy multiple times, it outspent on R&D during low cycles and cleverly invited customers TSMC, Intel and Samsung to invest and co-develop EUV machine.

Through this investment, ASML successfully obtained funds to acquire Cymer, a light source company headquartered in San Diego, and established a strategic alliance with German lens manufacturer Zeiss. These key components and electron beam technology create a recipe for ASML's success that cannot be replicated by competitors.

However, repeating such success is becoming increasingly difficult. Hutcheson quoted Nvidia's John Chen, TSMC's first R&D president: "No one company can do this alone because R&D is becoming more and more expensive." Industrial collaboration. "

Hutcheson estimates that developing a new EUV tool from scratch would cost trillions of dollars, equivalent to the gross domestic product of some countries, and that's assuming all intellectual property barriers are bypassed. "In a way, it levels the geopolitical playing field. So maybe EUV should get a Nobel Peace Prize because no country can attack any other country unless they have EUV," Hutcheson quipped.

ASML's high numerical aperture EUV is making headlines in semiconductor industry newspapers as Intel, Samsung and TSMC aim to mass produce 2nm chips in the next few years and move towards 1.4-1.6nm.

As always, ASML shipped the $300 million worth of high-NA machines to customers according to the schedule agreed upon with them, and worked with customers to make painstaking adjustments and corrections. The media has been describing high numerical aperture EUV as the key for technology chasers to leapfrog incumbents. However, Hutcheson delves into the complexities involved in the complex processes of semiconductor manufacturing beyond lithography.

"There are a lot of technical issues in the 2nm and 1.4nm era. We have to see if they can succeed and turn it into an advantage." Hutcheson explained that lithography is not the only key factor, there are other factors. There are problems with the gate-all-around (GAA) structure due to leakage issues. "There are also reliability issues, and the epitaxial layer is very poor. So it's actually more of a materials and systems engineering issue than a lithography issue."

Although Intel seems to have bought up all the high-NA machines available this year, TSMC is in no rush to apply them to its 1.4-1.6nm process. However, ASML said that all customers purchasing EUV will upgrade to high numerical aperture EUV.

"I think Intel has to do more and try to overtake TSMC, which is a bold move, but it's only part of Intel's challenge," Hijink said. "Their biggest challenge is not the technology, it's that they want to start a foundry business. , because they have to create enough wafer footprint or enough scale to be able to economically compete with huge competitors like Samsung and TSMC. So this adds to the problem of adopting new technologies that foundries have little experience with. "

Hutcheson warned it was too early to judge the outcome of the technology race. But looking at the history of the semiconductor industry, decades of research, blood sweat and tears, and fixing a lot of bugs, coupled with making the right decisions at the right time and building deep trust with customers, has made ASML what it is today. and TSMC.

Through acquisitions, ASML brings together European, Asian, and American cultures, and it can be a great teacher for TSMC, which is just beginning to expand into the United States and Europe. With such a strong connection between the two, and the fact that artificial intelligence is a major driver of future growth in semiconductors, you can be sure that there will be no shortage of exciting stories to come.

Public account ID: icbank