Welcome everyone. Good morning, and good afternoon for those watching in Asia. Welcome to this press conference about the fourth quarter and full year 2021 at ASML. We are here today in the Veldhoven auditorium, which is a new auditorium, and we would have loved to have you all here today to interact with us and to see you again, because also last year we weren't able to have the media here. You're all home, you're all listening or in the office. I would like to give you a glimpse of this auditorium, though, before we start, and so you can choose the spot that you want to have next year, because I'm assuming that next year you will be here.
If I can ask the camera people to show the room a little bit. We have a few colleagues in the room to, you know, to have a little bit of an audience, but hopefully we'll see you next year. My name is Monique Mols. I'm responsible for media relations at ASML. With me today are Peter Wennink, CEO of ASML, and Roger Dassen, CFO. Together, we will talk you through the results, our business updates, and an outlook for this year. If you are a journalist, you can submit your questions. If you go to the website, and you are on the website because you're listening in, you will see this little window where you can submit your question. Just write down your name, your publication name, and submit your question.
My colleague Sander Hofman will take the question and make sure that I get it on this iPad. Later on, we'll go through them, and we'll answer them as best as we can. Let's start. Peter, can I ask you to come here on stage?
Yes. Thank you Monique. Good morning, good afternoon, good evening, wherever you are. Welcome to the press conference, annual results 2021, fourth quarter 2021. As Monique said, I'm going to give a bit of a high-level overview of the market or basically the innovation drive that the company continuously has. Roger will also talk about the financials and the business update. Forward-looking statements. I'm not going to ask you to read all of it, but please do, but I'm not going to read it out, but it's for you to read through. I would like to talk about the key messages. The key messages, I'd like to refer back to our Capital Markets Day.
This is effectively what we said at the Capital Markets Day. At the Capital Markets Day, we talked about the global trends. We'll talk about them. I will repeat them again. Growth in the industry, our product portfolio, what it means in terms of our potential growth numbers, and you know, where we are today and the growth percentages that we currently see ahead of us. Also our ESG KPIs, which is becoming more and more important. Having said that, I think what we've said in November, I can only repeat, and I said it at the time, that I don't think those are aggressive assumptions. The reason why I'm going to go into next couple of slides, 'cause it is about innovation.
I think the semiconductor industry has fueled innovation over the last decades, 40 years, and we'll keep doing that in the decades to come. Why is that? It is because we will talk about this later, the value that we create is significant. Because we see that value curve coming, the innovation push that we have is relentless. It will continue. We see it around us. You know, it will create a virtual reality, artificial intelligence, high-power compute solutions that we're all going to use, and it's going to create value. The chip demand is, as we say, insatiable, but it's also because we've underestimated, structurally underestimated the value that has been created by our customers and the customers, and the application of that technology in day-to-day life. Yeah.
I think it's in every corner of day-to-day life we find semiconductor technology, and we certainly start to realize this. I said this before. I said it in an interview this morning. It's the 1973 revisited, the oil crisis. I said it before. Oil was a given in 1973 until it wasn't there. The same with semiconductors in 2021. Semiconductors were a given until they weren't there. Factories close. It becomes a geopolitical issue because the value of semiconductors is so high that new applications and new services are developed every day by tens of thousands of companies. It also means that the wave of demand is growing very fast, and especially for the logic chips. They do now, let's say, easier calculations to very complex high-power compute calculations.
I mean, wave of demand will grow, which also means that the number of machines that we need to ship will grow. Now, I actually always like this slide, and I, you know, this is, perhaps it's repetitive, and people who have seen it, I apologize but I like to talk about this for the simple reason that this slide basically paints the picture. You know, you have, the centralized computing in the cloud, yeah, which basically has the big cloud, which is served through the data centers and the cloud at the edge. And it's all connected through this big pipe called 5G. Yeah, which is the result of public networks, but also private networks. Why is that important? Because it actually takes away a lot of latency between transferring data from one point to the other.
Extremely important that you have the highway, the big data highway, where at maximum speed, data and information can be transferred to the cloud and from the cloud, but basically to what we call the edge. That's all these applications that are working, connected in the internet, yeah, or through the internet, using the big 5G pipe when it's mobile, yeah. Whether it's industrial, medical, traffic, surveillance, all through the smartphone, robotics, you name it. It's everywhere. I think this is an important slide. Why? Because when you look at the right-hand side, it's the edge. It's where we say what's called privacy, performance, and personalization. All these services and all these solutions are actually driven by sensing technologies. Sensors. Now, sensors are very important because they capture data. It's...
That data, that analog data needs to be transferred into digital. You have an analog-to-digital chip converter node. You actually need that. You need an RF device, radio frequency device. You need a power IC, because it's very often mobile, so you want power to be managed well. You have sensor technology, image sensors, movement sensors, chip technology, and microcontrollers. Yeah? That's not the most advanced logic chip actually deal with that data, yeah? It's logic chips that can be 10 years old or 15 years old or 20 years old. In other words, a sensor in your washing machine doesn't need a five-nanometer device. Yeah? It needs a device that from a logic point of view is perhaps 20, 30 years old. Yeah?
That is why we currently see this insatiable demand for semiconductors across the entire product line in every technology. This is what's happening today, and it's noticeable. Like I said, you know, 1973, when we suddenly found out that oil wasn't a given, it became a geopolitical issue. 2021, when chip and semiconductor technology goes to the heart of all these value solutions, and suddenly it's not a given, it becomes a geopolitical issue. People start to realize across the globe and governments that it's a vital part of the economy. This is why you get geopolitical focus on making sure that this vital technology is available, you know, in the big continents around the globe, in Asia, in the United States, in America, in Europe.
This is why governments suddenly start to realize that it's important to have a certain amount of semiconductor manufacturing on local soil. This is why we get these big investments and these CHIPS Acts around the globe. You know, if you read the papers, I mean, there's a European Chips Act in the making. There's a U.S. CHIPS Act in the making. There are CHIPS Acts around the globe. If you add it all up, it's hundreds of billions EUR that will become available for the industry. Now, you can argue that's a lot of money. You'll create over capacity. Well, to be honest, the industry will grow, and we will support it. By the way, growing the industry, the expectation is that this industry will double at the end of this decade.
At the end, the industry of our customers, the chip industry, to about $1 trillion. Yes, we need capacity. Government support is helpful and will also help to incentivize our customers to build fabs for which we will provide the solutions. What are those solutions? It's not only lithography. When you look at what we provide, we provide the lithography machine that actually creates the pattern on the chip, but we also provide them with metrology and inspection systems to make sure that we can control the process better, because it's all about affordability. You know, Moore's Law, the law to continuously shrink or let's say at the same cost, provide more value on the chip is a law of economics. Yeah? It's about affordable scaling, making things smaller, and yield. Yeah?
Why do I want to talk about DUV? DUV is the instrument, it's the technology, it's the machine that provides this mature technology, and we talked about this. Mature technology, especially at the edge, supporting the sensing technologies and the applications resulting from sensing technologies, that's done with deep UV, deep ultraviolet machines. In 2016, we actually thought that by 2025, the number of systems would go down. How wrong could we be? I mean, what we currently see is that the Internet of Things actually creates a huge demand for DUV systems. Yeah? The markets are very strong. Yeah? It also means that DUV will stay there. When we look at 2021, 75% of our DUV business was systems, and 25% was installed base.
We have a lot of machines in the field, and we do upgrades and productivity enhancement packages. By 2025, we expect it's still going to be 70% while the market will grow. Actually, we will ship more DUV systems, not less. Five years ago, we thought it was gonna be less. Now we think it's gonna be more. This is why we, our customers and the customers of our customers structurally underestimated the value and the growth of this industry. Deep UV is going to stay extremely important. A good example is that what we innovated in 2021 is our NXT 2050. That's a new volume manufacturing machine, and if you look at this slide, this information, it shows two graphs at the bottom, and one basically shows the productivity of that machine.
It actually is a good example of the maturity of that machine, because that machine was newly introduced, and we got over 5,000 wafers per day in less than 18 days after shipment. It's the fastest ramp we've ever seen. Now, you have two color bars there, the dark blue and the lighter blue. The lighter blue is the previous generation, which was already a very advanced machine. The dark blue bars are the new system. You can see how much faster we ramp to yield. It's productivity. It's more wafers out. With the chip shortage being what it is, you can imagine how important that is for our customers, but also the reliability. In 13 weeks, we're close to 200 hours of continuous working without the system having to stop once.
Deep UV, extremely important for our customers and for our business, and we are working off a high maturity base, which helps our customers to get more wafers out with an ever-increasing complexity in the production process, where Deep UV will also help. You know, it's what we call the overlays, the positioning accuracy from a Deep UV layer to an EUV layer is getting tighter and tighter and tighter, and also there we make big progress to help our customers manage the complexity. Now, I talked about wafer capacity, talked about output. Now, what do you do in a market where we are significantly under shipping the demand of the market? We have to basically give the customer more wafers. How do we do that? We do that by shipping more systems and by making the systems more productive.
When you look at this slide, you look at Deep UV, it's on the left-hand side, and EUV on the right-hand side. We're planning to increase the number of units from 2020 to 2025 with 50%, but also make them more productive so that the number of wafers out is double. With EUV, we're going to get twice the number of systems out, and increasing the wafer capacity threefold because the productivity goes up. Now, these are numbers that we are internally challenging. We might have to do more, so we're looking at increasing capacity, quite significant, capacity increase, not only at ASML, but particularly also in the supply chain.
We have to build faster, we have to hire more people, we have to increase production space, so we need to build square meters, and we have to increase the productivity of the machine, get more wafers out per machine. 'Cause we all know the chip shortage has to do with the number of wafers that our customers can produce. Also in EUV. In EUV, we introduced EUV in 2017. It wasn't high volume. High volume really started 2019, and beginning of 2019. And you see the dark blue bars here that the average availability of the installed base is growing.
This is the average of the installed base, which is a mix of the most advanced EUV systems, which have an availability of over 90%, and the more mature systems that we shipped two years ago that have still a lower level of productivity. Those systems, the lower-level productivity systems, we are upgrading to make sure that they get to the high level of productivity of the most advanced and the newer systems. You will see this bar chart going to the right, going up and up and up over the next couple of years, where our target is to be over 95% availability to get more wafers out, productivity. That's for now. That's for tomorrow. What about the day after tomorrow? It's about next generation EUV.
When we introduced EUV and EUV became high volume, we had to think about the next generation that brings us way into the next decade, which is called High NA, the next EUV version, which our customers will use to drive output and the newer generation of chips in the second half of this decade into the next decade, beyond 2030. We're building the first machine in our factory in Veldhoven as we speak. We've started the integration. The first major big modules have arrived, and we're starting to integrate the system. We receive orders. We have 5 orders for, let's say, the pre-high volume systems that customers will use to do R&D. Last week, we also received the first order for the next generation High NA tool, which will be the high volume system.
That will be the system to crank out more wafers with even more advanced chips in the second half of this decade, going into the 2030s. The future is bright. The market will grow very fast. I think the market will grow faster than we currently think. That's what my expectation is. We need to follow by adding capacity and by relentlessly investing in innovation and R&D to make sure that we provide our customers with the solutions to drive value for these end products way into the next decade. I'd like to turn over to Roger for the financials. Thank you.
Thank you Peter. Good morning, good afternoon everyone. Peter, as he said he took us on this last slide, he took us to the day after tomorrow. Unfortunately, it's the fate of the CFO to take you back to yesterday. That's where we're gonna start, but I won't only keep it to yesterday. I promise you that I will actually you know also look at today and look at tomorrow a little bit as well. Let's start with yesterday, and let's start with the Q4 results. I promise you, I won't take you through every single number that we have here on the slide or the slides to come because that would be a bit of a boring session. Just to give you a flavor of what Q4 looks like.
I think it's fair to say Q4 was a good quarter. It really landed as far as revenue, as sales was concerned, it landed within the guidance. It's also fair to say that if you look at the composition of the sales, it changed a little bit to what we expected three months ago. Three months ago, we had a higher systems sales that we expected, and we had lower Installed Base Management. It all has to do with what we talked about three months ago. Three months ago, we talked about chip shortages also affecting ourselves. It's a bit cynical, but it's the way it is, right? Even ASML is confronted with the chip shortages in its supply chain.
We also talked about a new logistics center that we commissioned and some of the issues, the startup problems that we had as a result of that. I think it's fair to say that those issues combined had a little bit more of a bearing on Q4 than we anticipated three months ago, and that led to a slightly lower shipment of systems than we anticipated at that point in time. In order to make sure that customers, you know, still get productivity upgrades, that customers still have the potential to have more capacity given the constraints that they operate in, given the very, you know, significant demand that they are confronted with with their customers, we provided them with upgrades.
These are the upgrades that Peter talked about upgrades to really, you know, make sure that the installed base, the systems that are already out there at the customer, that they operate with a higher productivity in order to really give customers a higher capacity for output. That is what we did, and that really is reflected in these numbers. Slightly lower system sales, but you know, significantly higher installed base management in comparison to what we thought three months ago. That also has an impact on the gross margin, because a lot of these upgrades really are software based upgrades, because that's a very fast way to give customers this productivity enhancement. Typically, the gross margin that we enjoy on software upgrades is fairly high.
That's why the gross margin is quite a bit higher than what we anticipated three months ago. That really permeates throughout the entire sequence of numbers that you have here, leading to higher operating margins and actually, you know, pretty high net income and earnings per share. Also wanna highlight the net bookings of EUR 7.1 billion. Pretty high number, but it's also interesting to look at the composition because you've seen high bookings numbers in the past. The total bookings number for the full year was a whopping EUR 26 billion. Typically, you would see that a very high portion of the bookings number actually is related to EUV.
What you see here is, you know, very solid intake of orders for EUV, but also, you know, a pretty high contingent of the rest of the business, primarily DUV. That goes back to what Peter talked about, because a lot of the demand that is out there in the market is related to, for instance, Internet of Things, related to you know, to more mature technology that also demands, you know, continued increase in capacity, and that really also drives our DUV business. You know, in that way, it's.
It really is clear from these numbers. It's not just about EUV. It's also DUV that has gone through very high levels of appreciation in the industry and continues to be, as far as we can see, a very strong driver of the demand for our systems. Included in there, as Peter also referenced, the one High-NA EUV tool. That's it for Q4. Let's look into 2021 and the highlights of 2021. Again, not gonna talk you through all of the numbers. A really good year.
I mean, a year ago when we stood here, and when the question was asked, "What do you think next year is gonna be?" We talked about low double-digit growth, 'cause that's what, you know, that's what the market signaled at that point in time, low double-digit growth. Well, 33 is not our equivalent of low double-digit growth. I think we far exceeded our own expectations, with the performance of this year. 33%, and really, if you look at it, if you look at the composition of the sales, it is very clear that ASML is really firing on all cylinders. You know, if you look at every technology that we have, all the end markets, it really is, they all contribute in their own way to a very strong growth that we experience.
If you just quickly look at these numbers, EUR 18.6 billion of total sales, 52.7% gross margin, which is very high indeed, really also echoing the value that we bring to customers. Because the only way that we are able to get to these high numbers of gross margin, and we will see that also when we make a four or five year comparison, you will see that that number goes up, and that is because of the value that we provide to our customers. Leading to a net income, which absolutely is an all-time high for ASML. It's a record number at EUR 5.9 billion, with an earnings per share of over EUR 14.
If you look at it all, everything contributes to it. We talk about Installed Base. Installed Base is the service that we provide to customers, but also the upgrade business, the productivity upgrades that we talked about, including the software, but also hardware upgrades. Very strong business for us, growing very, very nicely with 35% this year. We talked about EUV growing 41% this year on the systems side. Peter referenced the good progress that we're making on the High-NA program. We talked about Deep UV, where a couple of years ago, people started talking about that the market really declining.
Well, that's not really what you see, and it's also not really what we expect for the time to come. Very strong growth, and that is also what we continue to see into this year. Then the application business, which you know also had a very healthy growth of 47%. Also noteworthy there, the fact that we completed the first eScan 1100 multibeam system that is very close to shipment to its initial customer.
Also noteworthy I think is that we returned approximately, well, close to EUR 10 billion, EUR 9.9 billion, as you see here, back to shareholders in terms of share buyback and dividend. I'll give you a bit of a breakdown on that later on. Very strong year. Not gonna take you through this one. Look at it at your leisure. But in essence, it tells you that both on end markets, on technology, and on region, you see that, you know, every one of those components really contributed to the very strong growth that ASML enjoyed in 2021. Same thing here.
This really also illustrates to you graphically that, you know, really 2021 was a year where we have enjoyed quite some growth in the past couple of years. It's very clear that 2021, building on the drivers that Peter was talking about, you know, is a growth acceleration for the company. Wow, what about this slide? Well, I'll take you through it a little bit without referencing every single number once again. There are a few interesting trends here. If you take the 2021 to 2017 comparison, so a four-year growth comparison if you like, it's clear that, you know, on the net sales line, we're doubling. We're actually more than doubling.
You see the gross profit increasing from, you know, a little shy of 45% to 52.7%. Again, this is primarily because of the value that we bring to the customers. The only way that we can do this is because we achieve and we succeed in bringing more value to the customers with the tools that we provide today. R&D cost, you also see actually more than double in this period, which is, I think, you know, a very clear indication that we continue to see a lot of potential out there in the market, but also a lot of potential within our team to continue to bring technology to the customers, which is appreciated and is conducive to, you know, helping cater to the demand that Peter was talking about.
SG&A, below doubling, which every CFO loves because that means that we're efficient in our overhead operations, so, you know, good performance on that front. You see all of that really developing to a net income that nearly triples in a three-year period, and an earnings per share, which literally triples in three years, also as a result of the fact that through the share buyback, we have been able to reduce the number of outstanding shares, which boosts the earnings per share a little bit more in comparison to the net income growth. Once again, look at the booking number, the net booking number of EUR 26 billion. We're currently looking at an order book of EUR 24 billion.
EUR 24 billion, if you recognize that, the expectation that we have for this year in terms of system sales is a little short of EUR 17 billion. That tells you that there is you know, more than EUR 7 billion in the order book that really relates already to 2023. That, I think, is a clear illustration of the very, very strong demand that ASML is currently enjoying based on the waves that Peter was talking about. You know, the benefit of that is to the customers. The benefit of that is to all of the stakeholders of ASML, but also to the shareholders. As I mentioned, EUR 9.9 billion was returned to shareholders.
EUR 1.4 billion in 2021 in terms of cash provided through dividend, and EUR 8.6 billion in shares that were bought back during 2021. The recommendation we have to the annual general meeting of shareholders is to double the dividend, the total dividend. Dividend last year over 2020, the total dividend was EUR 2.75, and we recommend to the AGM to increase that to EUR 5.5, so as I mentioned, doubling it in that regard. Outlook. CFO closes the looking at yesterday chapter, and now we go to today and actually also tomorrow. Let's first look a little bit at today, at Q1. Net sales expected between EUR 3.3 billion and EUR 3.5 billion.
I know you're gonna look at that and say, "Wait a minute, didn't I just see EUR 5 billion for Q4?" You're right. The 3.3 indeed is a number that might strike you as low. You need to watch out here because of a change that we've really introduced this year. We did a little bit in previous years, but this year we have made a major change in the way we bring systems to our customers. What we do, we ship systems to customers, and we really do you know, the significant part of the testing at the customer site.
We're not doing it in the factory, as we used to do, but most of the testing and then the final acceptance happens at the customer site. The reason we do that is because by doing that, you actually skip a couple of weeks in the cycle time in the cabinet. It was one of the levers that Peter was talking about in terms of, you know, how do you reduce the cycle time, and how can you actually increase productivity for ASML. This was one of the levers. That's why we did this.
We ship the systems to the customers without them having been extensively tested at the factory and, you know, the remainder of the testing program and the final acceptance happens at the customer site.
That means that we ship systems, but we cannot enjoy the revenue of that. So accounting-wise, we can only take the revenue once the final acceptance, the formal acceptance, if you want to call it that way, happens at the customer side. That means that we have about EUR 2 billion of output of shipments that we have this year that we cannot recognize as revenue. So if you add that to the EUR 3.3 billion-EUR 3.5 billion, you're looking at a healthy EUR 5.3 billion-EUR 5.5 billion worth of output that we would be shipping this quarter. That I think is once again in line with how we would be looking at the full year. I'll come to that in a moment.
Included in that installed base of EUR 1.2 billion, gross margin also is affected as a result of the fact that, EUR 2 billion is taken out. So if you adjust for what we call the fast shipment effect and the gross margin, you would see gross margins back again at the levels that I showed you that we had for last year. R&D once again growing because of the very aggressive agenda that we're pushing there, and SG&A at around EUR 210 million. Finally, how does that translate into the full year? We expect the full year and at once again 20% or once again very strong growth at 20% for the full year.
That would get you to approximately, you know, EUR 22.3 billion. That's the number that you would end up with if you add 20% to the EUR 18.6 billion that we had for this year. We expect 55 EUV tools to be shipped. Six of them will be fast ships. For those we will not recognize the revenue, so we will recognize revenue for 49 systems. Also, the DUV in the applications business will continue to grow in a very healthy fashion. We expect it to be north of 20% growth in comparison to last year. Again, the company runs on all cylinders.
Also, in terms of end markets, the growth is both in in logic and memory, and we also continue to see a strong growth in the installed base business that we referenced. Peter said the future is bright. Absolutely. Also 2022 we expect to be very bright. With that, Monique, I would like to hand it over to you for Q&A.
Thank you Peter. Let's see if there's any questions coming in, because I thought that was pretty clear.
No questions.
It was pretty clear. Let's start with the first one. I think you both addressed this question, but maybe there is something you want to add to that. Debby Wu from Bloomberg had a question, "In the morning interview on Bloomberg TV, you mentioned 40%-50% higher demand versus your max capacity. Can you provide more details? Is it both for DUV and EUV?" I think the last question you already answered.
Yeah. I said that this morning because we say we're capacity limited, that we're working at our maximum capacity. We, as ASML and the supply chain, don't forget, you know, most of our modules come out of the supply chain. We don't have a factory that builds those modules. Yeah, I think, you know, when you look at the demand, which is of course a function of, you know, the disturbances due to COVID, it's the underlying growth trend. It's, you know, also the focus of governments, pushing our customers to build more creates this massive demand. Yeah, I think we're going to be at 40%-50% above what we can make. That's what they That
Now, you could argue, because people know that we are somewhat limited that they want to put in some extra demand. That's all probably logical, but it's across everything. As what I said earlier, it's across our mature products, our advanced products, and our very advanced products. It's across the entire product line for all industry segments. For memory and for logic, it's across the board.
Yeah. Thank you. Eva Schouten, she read our press release about the information about the Berlin fire. She says, "You said that the Berlin fire has no significant impact, but what impact does it have?
Sorry.
The fire in Berlin did not have a significant impact.
Yeah.
What impact did it really have?
Oh you know first of all, it scared me to death. That was an impact. Yeah. So,
Yeah
You know, you don't want to start the year January 3, and you get a phone call saying, "Yeah, the factory's on fire." I mean, that's the last thing you want to hear. It did have an effect, but it also means that you don't know at that time, you don't know what the consequences are. Yeah. I must say, you know, when I got the first pictures of the factory, I thought it was going to be very significant, but it was just a gut feel. It turned out that, you know, huge work from our people in Berlin, I mean, a tremendous creativity, the way that they look at it.
Within a week they could say, "We can secure the Deep UV modules that go into the DUV machines." We need more time for the EUV. It was basically it's a clamp, it's a wafer clamp that holds the wafer in the EUV machine. It's a very complex and very critical part of the machine that the fire broke out in that department. There were machines that were destroyed. There were machines that seemed untouched, but you know, smoke and water. That to do that analysis really took some time. I think it wasn't until the end of last week, I think, or over the weekend that we really understood the consequences that we could manage this.
Mm-hmm.
We could actually manage this and that, yes, of course, there is no production of those wafer clamps now, but we have inventory. We have work in process. We have repair actions. We have stuff that we can do which actually buys us time to repair the production flow, and then so we think we will not have a significant impact on the EUV shipment this year.
Okay.
Yeah.
Yeah, it definitely had an impact then, but first of all on myself when I heard.
Yeah.
Monique, what we had to do is to, you know, so late last week we had an impression of, you know, what is the damage? Then we had to take a fresh look again at our entire output and production plan. How can we include it in that? That was really, you know, working against the wire to get that done. I think also the fact that we're now doing fast shipments gave us a little bit of flexibility. The fact that, you know, rather than having the 55 systems completely shipped, have them gain some time on those 6 shipments as a result of that, I mean, that's the kind of buffer that we also needed.
Yeah
... to get the impact of Berlin addressed.
Yeah, yeah. A lot of creativity was
Absolutely.
A lot of creativity
It's a good news then.
A lot of hard work.
Yeah. Absolutely.
You know, around the clock, 24/7.
You bet.
I think, great people, yeah?
Yeah.
A question from Max Smolka at Frankfurter Allgemeine Zeitung for Peter. Could you give an update about the export of EUV to China?
Yeah. Well, the update is business as usual. I mean, there's nothing more to say, you know? I think it is with the governments, you know, EUV is export-controlled technology.
Mm-hmm.
Ultimately, you know, it's up to the government to say, "I'm going to issue an export license or not." And that's still-
It's still a no.
Where it is today.
It's still a, yeah.
We'll just wait. Now, we do get some question, what does it mean for a chip production? Well, you know, chips are being sold into China, but the manufacturing of those very advanced chip doesn't happen in China. The manufacturing happens somewhere else. Happens in Taiwan, in Korea, in the United States. That's where we ship our EUV machines, and I said this before, I don't believe that is going to be a major impact on our shipment pattern. We just ship them somewhere else.
Yeah.
Yeah.
Okay. Question for Roger from Sandra Olsthoorn, Het Financieele Dagblad. I'd like to hear more about installed base growth. How predictable is the growth level here? Will it grow on the current level? Can you give an indication on the gross margin compared to the margin on machines?
Yeah. Installed base is very much a mixed bag, and that's in terms of all the things that Sander was referencing. It's a mixed bag in terms of gross margin, right? Because if you just, you know, peel the onion of the installed base, what is in there? First off, you have service, regular service to a tool, and then you have upgrade business. Upgrade business, some of the upgrade business is hardware, some of the upgrade business is software, so that's the mixed bag in terms of activities. Then obviously you have a mixed bag in terms of, you know, is it DUV? Is it EUV? Is it in the application business?
That makes it a very different composition in terms of many of the things. In terms of predictability, I would say that obviously the service business is far better predictable than the upgrade business, because the upgrade is discretionary. Service is, you know, kind of mandatory. If you don't service the tool, then it will break down at a certain point in time. That's fairly predictable. The upgrade business is discretionary. And that's up to the customer to decide, you know, when am I going to give an upgrade? And that's not a very evident case for customers.
Mm
Because they always have to say, "If I do an upgrade, that means that I need to power down the machine for a while." They have to make the trade-off, how much time, how much machine time do I have to give to ASML? What is then the benefit that I get in terms of capacity and output expansion as a result of that? That's a trade-off. A lot of emphasis that we've put this year is to try and come up with upgrades and do them in a fashion that really reduces the machine time that the customer needs to give to ASML, and there the software upgrades, of course, come in handy.
In terms of profitability, upgrade business, particularly when it has a software component to it, is very high margin business. I think the service business historically has been below the corporate gross margin level of ASML. We are, you know, providing those services now in a way that it really provides a lot of, you know, a lot of value to the customer, and that we are able to drive up the gross margin there as well. Our ultimate goal, I would say in the 2025 timeframe, that's also what we said at the Investor Day, is to try and get the installed base margin, you know, close to the corporate gross margin level.
Thank you. You can still submit your questions, right? If you want to ask a question, please use the website and the little window we have there. A question from Nikkei, Cheng Ting-Fang. How many people are you looking to recruit this year?
Mm.
Do you see talent shortage?
It's a good question. I mean, if you asked if we had that question last year, I would have said 1,500, and we did 6,000, so it was a factor of four more. That was 2021. Now, I think we are, we've actually hired for our capacity plans in the production area. I think we have hired the number of people that we want. I don't think we'll hire 6,000 people this year.
Mm-hmm.
I think if we look at the plan and look at the expansion of R&D that Roger talked you know, talked about, this is people, yeah? I think we will increase our R&D base and our customer service base, because we will have to ship more, so we need more people in the field. I think it's still going to be a few thousand people, yeah? Anywhere around 3,000 people, 3,000-4,000 people. Because also have some attrition. Yeah, why would you leave a great company like ASML? It does happen. Yeah. I think it's probably going to be more like 4,000 people that we're going to have to hire. You're absolutely right. I mean, it's a war on talent.
You know, it is getting more complex around the globe.
Mm-hmm
to find the right level of people for the simple reason that we're competing with our customers and with other players in the industry. I mean, if everything, if everybody's growing so fast, I mean, in Taiwan, we're competing with TSMC. You know in Korea, we're competing with Samsung and SK hynix. In the U.S., we're competing with Intel, with Micron, and we're competing with Apple and with Qualcomm. So, just an appeal also to governments, yeah, to absolutely focus the education system on educating young people and basically driving the interest for science, technology, engineering, and math-related studies. Yeah? I think that is critical. Yeah?
Mm-hmm.
I think we'll do fine. I mean, you know, last year we received, I think the latest number I got, 295,644 applications. Yeah. That we hired around 6,000 people, so you could say you guys still have a choice.
Mm.
You know, yeah, it's true, but we want really good people, yeah?
Yeah.
you know. Then these really good people help us with things like the fire in Berlin.
Yeah, absolutely.
You know, we're very lucky.
There's a lot to do. Just a tip, on our website, we have a full list of our vacancies. There's a question from several reporters, so I don't have a name here, but several people asked, so it's apparently an urgent question. Can you share more color on the High-NA efforts? How is assembly going-
Yeah
When are you shipping?
I think we're shipping end of 2023. That was going to be one of those five systems, those five R&D systems, which we'll finalize throughout 2024, and then we will basically start the first shipments of the High NA high volume tool, the one that we received an order from Intel, will be at the end of 2024, and then we're moving into 2025, 2026 also with other customers. That means we do the assembly now.
Mm.
You know, when you walk into the clean room here in the Netherlands, in Veldhoven, and you walk into the High-NA area, I'm just flabbergasted by the size of the, you know, machine. It's big machine, yeah? It's getting together. I think the most important thing in High-NA was the optics. You know, we basically have a bigger lens, you could say. The optical technical requirements of the lens were so high, were so different than what we do with Low-NA, that was really a major technical project with a lot of risk. Now, kudos to the best optical company on the planet, which is Carl Zeiss, yeah?
Yeah.
You know, they pulled it off together with us, you know, because we created very big measurement machine, which is basically a factory, you know, and where we were able to actually measure those optical systems, and they gave us the results that we wanted, which was as again, it's an extreme project.
Mm.
You know, extremely good people get extremely good results, yeah?
Yeah.
That's what happened. Thanks to our colleagues and partners at, you know, ZEISS, and with the collaboration with our own people, we're there, you know? We're going to make this machine happen.
Yeah. The picture we just saw, that was just one part of it, yeah? It wasn't the entire.
Just one part. When you look at that picture, just, you know, go back and look at the picture, just look at the people in front of one of the modules. There's a frame.
Metro frame.
Yeah. Which actually, the actual machine is even bigger than that.
Mm-hmm.
Yeah, it just gives you as an indication of the size of.
Yeah
the machine.
Oh it's impressive. Peter van Ammelrooy of de Volkskrant, a question for Peter. Your machines also have chips on board. Is ASML impacted by the chip shortage as well?
Yeah. It's a daily fight, you know? We are directly impacted, but not that we buy chips, but our suppliers do. I mean, we have thousands of electronic components in our machines, so it's not strange that when you look at our supply chain, our first tier, they buy from a second tier, they buy from a third tier, that, you know, they buy components, as we call them, electronic components, which have chips in them. Yes, there are shortages, yeah? This is a day-to-day fight. What we actually do, we gather it all together, and we combine it at ASML in a scarcity center, and then we figure out, you know, which semiconductor manufacturer makes those chips, and then we pick up the phone and we call them up.
You know, we basically say, "Can you help us?" You know, "You need to help us because we want to ship those machines, and you need those machines. So how can we, you know, find it either in the distribution system, at the semiconductor maker, with the brokers?" It's a constant fight.
Mm-hmm.
Until now, we're able to manage.
Okay.
Yeah. But that's not a given. It's a challenge, but you know ASML, we like challenges.
Absolutely. Do you want to add anything, Roger?
That's it.
You're fine? Okay. Question from Joe Miller at Financial Times for Peter. Do you think reshoring parts of the chip value chain will prevent future bottlenecks, or is it more important that regions double down on their expertise?
Both. Yeah, I think we need to do both. You know, if this industry with the current growth rates that we're looking at, I mean the industry of our customers, the chip industry.
Mm-hmm
will double to $1 trillion by the end of the decade, yeah? That means we need to add, you know. It took us 40 years to build the capacity to do $500 billion plus of sales, so we need to go $1 trillion, we need to perhaps not double the capacity, but go a long way towards that. We need more capacity. Now, to say why don't we just expand that capacity on the few places on the planet where we do this today in large volume, which is predominantly Asia, might, from a geopolitical risk point of view, might be a bit too much, huh? Why don't we then say, the government says, "Yeah, we need to bring some of that back to our continents." Yeah, because the dependency becomes so very high.
Just give you an example. 20 years ago, around 25% of the chip production came out of Europe. It's now 8%.
Mm.
What if the market doubles again? I will tell you, it's less than 4% because these were euros, and the most advanced chips are being made somewhere else. You have less than 4% of the world market in chip manufacturing, while chips are so important for your economy.
Mm.
It's only logical that governments around the globe say, "Wait a minute. 1973, what happened again? The oil crisis, you know? 1921, chip crisis, you know?" It's geopolitical important, yeah? This is why we need to both. It also means that you need to invest heavily in knowledge. I go back to the education system, you know. We really need to focus on enticing, stimulating our younger people to say, "I want a career in this exciting industry." Yeah? We have to invest in R&D and in innovation. It is not only about the mature chips where we have big shortages. We will have big shortages in advanced chips also. We need to do both. Yeah?
If there are companies around the globe that say, "Get your point on the geopolitical reliance, I wanna come to your region, I wanna build a advanced factory," we call it a fab, yeah, which costs $15 billion-$20 billion, why don't we help them?
Mm.
'Cause it's a risky project. You know, it's. They could easily build that fab in the area where they are today, just add one fab, 'cause they already have the ecosystem, the infrastructure. Now they come here, it's risky, yeah, because they have to start from scratch, create that new ecosystem, which I think is going to be important for Europe, for instance, if you think about Europe, because it's going to be a magnet. It's going to be a talent magnet.
Mm
Which is important for R&D. Yes, of course we need to support. Also we need to support the part of the industry that says, "We need Deep UV capacity," yeah, because, you know, it's sensing technology, Internet of Things that we talked about. It's going to also need more chips. Yeah? It's both.
Both. That's always a good answer. There's a question for Roger from Paul van Gerven, Bits&Chips, "Is EUV profitability now on par with DUV profitability? Will ASML take an initial hit on gross margin for High-NA like it did for 0.33 NA EUV?" That's always a very difficult one to pronounce.
We, with the NXE:3600D tool, which is the tool that we introduced last year, now have the EUV gross margin at the level of what we call the corporate gross margin.
Mm.
I think we've achieved that milestone. We've said that you know, with the successor to the NXE:3600D, that's the point in time where we're really gonna see it cross with DUV profitability. That's still a small step to be made, but getting there, closer and closer. Again, because of the value that the tool provides in terms of productivity overlay quality to the customer. In terms of High-NA, you know, High-NA, as Peter pointed out, is a massive exercise. But it's also evolutionary. It's not.
Mm
... a kind of revolution that you had going from immersion to EUV. This is an evolutionary step, a big evolutionary step, mind you.
Mm-hmm.
It's still an evolutionary step, and that means that a lot of the components that we have on High-NA, by the time High-NA really gets introduced, have already been, you know, tested in many ways on Low-NA, as well. Actually, one of the key things that we're driving is commonality between a number of parts on the Low-NA machine and the High-NA machine. I think in that way, it's, you know, less, I would say, of a, you know, of a venture in terms of what it does to, you know, get into high volume manufacturing than it was with Low-NA, which really was a paradigm shift. Long way of saying that this will be, this will not be a gross margin level from day one.
Mm.
For sure, you know, there will be a learning curve that we have to go through together with our customers. I think this will be accretive to gross margin, I would say, in the 2026, 2027 timeframe.
Yeah.
That's the current expectation that we have. We'll be dilutive in the initial years, but by 2026, 2027, we should be able to get it to the levels that we want it to be.
Yeah. Just in case anyone who's watching will want a High-NA system in their home, how much is that?
Sorry, how much?
How much is a High-NA system these days if you want to buy one?
You mean in terms of euros?
If you want to buy one.
Okay. Well, first off, you need, if you talk about home, you need a bloody big living room, right to Peter's.
A very big living room.
... to Peter's point. A significant bank account to go with that.
Okay. Okay.
The R&D tool that Peter talked about is EUR 270 million.
Okay.
And for the high... and the high v-
Yes
... and the high-volume manufacturing tool, the EXE:5200, where we just got a first order for, we talk about significantly higher than 300.
Okay. You know, there's a question from Marc Hijink at NRC Handelsblad, "How do you assess the risk for overcapacity and the effect on ASML revenue-
Yeah
when all geographies fire up their plans to build big fabs at the same time?
Yep. No, I think it's a good question. I can only say, well, you know, we've structurally underestimated the growth of this industry. I don't see any reason why, you know, we would be so much better at guessing where this industry will go. I mean, we currently believe this industry will go to $1 trillion, yeah, by the end of the decade.
Mm.
It's almost doubling. Yeah? We need that capacity also.
Mm-hmm.
If you then people say, "Well, but there's so much government funding," well, if you add it all up, it's $500 billion, $500 billion euros or dollars, let's say dollars. Yeah? Probably even a little bit more. Well, this industry is $1 trillion annual sales, so of our customers' industry, $1 trillion by the end of the decade. This is only 50% of annual sales, yeah, by that time.
Mm-hmm.
Now, of course, we will pre-invest. Yeah? Of course, this industry is also known for these kind of shorter term cyclical movements, because there are big capacity step-ups, you know, as a big semiconductor factory will create a lot of capacity in one go. Well, you will grow into it, but it will be short-lived.
Mm-hmm.
Yeah? I think ultimately it is. It's about do we together believe society needs more technology solutions that are driven by semiconductors? We all believe that. I actually, we see the proof of it. Yeah? Then we need capacity.
Mm-hmm.
Will it, you know, when it all happens at the same time, will it give some short-term disturbances? Yeah. You know, we're used to that. Yeah. Longer term, I don't see an issue.
Okay. There's a question from Toby Sterling, Reuters. It's a question we get regularly. How long will it take for China to develop EUV lithography that will compete or substitute for ASML's technology, Toby?
Yeah. It's a good question. You know, EUV technology, over the last 15, 20 years, many people said, "What you guys are doing is completely irresponsible." It's not what we're doing, it's what we're doing with our partners and doing with our knowledge institutions. It cannot be done. Actually, we achieved it because it's this massive ecosystem that is helping us to do that. Actually, one example is ZEISS, but I can name another dozens, you know.
Mm-hmm
of companies that are absolutely world-class in what they do. We acquired Berliner Glas, which is now ASML Berlin. Why? They make this wafer clamp, which is a module that takes 52 weeks to make. But the technology that's in there is decades of knowhow. It's not patents, it's knowhow.
Mm-hmm.
That's spread in this ecosystem of hundreds of companies.
Mm-hmm.
You need to replicate that.
Mm-hmm.
I'm not saying it cannot be done, because laws of physics are the same in China as they are here.
Yeah.
I know it's the combined effort, the synergy between all these companies.
Mm.
Yeah, bring it all together in this system approach, which is basically our core competence. Our core competency is system integration and system architecture. Yeah. That, I think, will be very difficult to replicate, because you need to go to all these places where they do these very complex things that nobody else on the planet can do.
Mm-hmm.
Yeah. That's going to be a massive exit. Never say never.
Yeah.
They'll certainly try, trust me, they'll certainly try. It's going to be difficult. What do you do? What do you do normally? You just go faster.
Mm-hmm.
Yeah. How do you deal with this level of competition and threat? You go faster. You innovate. You go to High NA. You go beyond High NA, which by the way, we're also doing in our research. Yeah. I think this is what you do.
Mm-hmm.
Relentless investment in innovation.
Yeah. We still have a little bit less than two minutes. Koos Schwartz from Trouw. ASML receives innovation subsidies from the Dutch government. Do you need those subsidies, Roger?
Well, I think they are definitely helpful in doing the things that we do, right? I mean, they do allow us, you know, to invest in the future, and that's, I think, you know. The fact where we are today.
Mm-hmm
The position that ASML enjoys today, I think we would not have been able to enjoy without the support that is out there. I think it has been very helpful to get us to where we are today, and I think it also helps us to focus all of the efforts that we're doing in Veldhoven. I mean, if you look at the way we concentrate all of our R&D efforts here in the Netherlands is pretty unique. I don't think there are many companies where you have such a unique focus and a concentration of R&D work as we have it here in
No
The Netherlands, and that's not primarily as a result of subsidies, but you know, definitely, it is a reflection, I think, of the very helpful ecosystem that we find ourselves in.
Mm-hmm.
I think that ecosystem you know, with the government supporting that, with many of the people being attracted to such an environment, I think that really helps us to do the work that we do here.
I'd like to add to that because I think it is particularly the innovation box, the real subsidies that's.
Yeah, I think that's what he
It's really the innovation box.
Okay.
It's a kind of a reduction on your tax. We look at these activities also from a cost point of view. You know, what's the most cost-effective way to do research and development? To Roger's point, yeah, around 40% of our R&D bill, yeah, is in the supply chain.
Yep.
It's a multiplier for this ecosystem that we have created in the you know Netherlands and partly Germany. Yeah. It's that incentive to keep the cost of R&D lower.
Mm-hmm.
It isn't only to our benefit, it's to the benefit of hundreds of companies in our ecosystem, which is Netherlands and a bit of Germany.
Yep.
Yeah.
I see that our time is up. No further questions. Reach out to us if you have any additional questions. You know where to find us. We're on the website. Sandra and I will be happy to talk to you, and if you are an investor, you can talk to my colleagues in Investor Relations, obviously. Thank you very much for listening. Thank you very much for participating and for sending in your questions. Again, a promise, we'll see you next year here. Thank you very much.
Thank you very much.
Thank you.
Stay safe.
Yeah.