ASML Holding N.V.

Good day and thank you for standing by. Welcome to the ASML 2024 fourth quarter and four year financial results conference call on January 29th, 2025. At this time, all participants are in a listen only mode. After the speaker's introduction, there'll be a question and answer session. To ask a question during the session, you will need to press star one and one on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press star one and one again. Please be advised that today's conference is being recorded. I would now like to hand the conference call over to Mr. Jim Kavanagh. Please go ahead.

Thank you, operator. Welcome, everyone. This is Jim Kavanagh, Vice President of Investor Relations at ASML. Joining me today on the call are ASML CEO Christophe Fouquet and our CFO, Roger Dassault. The subject of today's call is ASML's fourth quarter and full year results 2024. The length of this call will be 60 minutes. and questions will be taken in the order that they are received. This call is also being broadcast live over the internet on www.asml.com. A transcript of management's opening remarks and a replay of the call will be available on our website shortly following the conclusion of this call. Before we begin, I would like to caution listeners that comments made by management during this conference call will include forward-looking statements within the meaning of the federal security These forward-looking statements involve material risks and uncertainties. For discussion of risk factors, I encourage you to review the safe harbor statements contained in today's press release and presentation found on our website at www.asml.com and in ASML's annual report on Form 20F and other documents as filed with the Securities and Exchange Commission. With that, I would like to turn the call over to Christophe Fouquet for a brief introduction.

Thank you, Jim. Welcome, everyone, and thank you for joining us for fourth quarter and full year 2024 resource conference call. Before we begin the Q&A session, Roger and I would like to provide an overview and some commentary on the fourth quarter and full year 2024 resource, as well as provide some additional comments on the current business environment and on our future business outlook. Roger? Thank you, Christophe, and welcome, everyone.

I will first review the fourth quarter and full year 24 financial accomplishments and then provide guidance on the first quarter of 2025. Let me start with our fourth quarter accomplishments. In the fourth quarter of 2024, total net sales were 9.3 billion euros, which is above the high end of our guidance, primarily due to install-based revenue. As expected, it also includes revenue recognition on two high-end A systems. Net system sales were at 7.1 billion euros, which includes 2.9 billion euros from EUV sales and 4.2 billion euros from non-EUV sales. Net system sales was driven by logic at 61%, with the remaining 39% coming from memory. Install-based management sales for the quarter were above guidance at 2.1 billion euros, primarily driven by additional upgrade business. Gross margin for the quarter was above guidance at 51.7% due to a combination of additional upgrade business and lower than planned costs associated with the new product introduction of high NA systems in the field. On operating expenses, R&D expenses came at 1.116 billion euros in line with guidance, while SG&A expenses came in above guidance at 318 million euros, due to year-end payroll adjustments and a pull-in of IT costs. The effective tax rate Q4 was 21.5% as a result of a one-off tax expense related to a historic tax position, bringing the full-year ETR to 18.6%. 2025, we expect an annualized effective tax rate of around 17%. Net income in Q4 was €2.7 billion, representing 29.1% of total net sales and resulting in an EPS of €6.85. Turning to the balance sheet, we ended the fourth quarter with cash, cash equivalents and short-term investments at a level of €12.7 billion. We ended Q4 with a free cash flow of 8.839 billion euros, which is significantly higher relative to last quarter, with the majority of cash coming in at the very end of the quarter. Moving to the order book, Q4 net system bookings came in at 7.1 billion euros, which is made up of 3 billion euros of EUV and 4.1 billion euros of non-EUV. Net system bookings in the quarter were driven by logic with 61% of the bookings with memory accounting for the remaining 39%. At this point, I would like to briefly comment on the reporting of our bookings going forward. As we have said in the past, our order flow on a quarterly basis can be lumpy and does not necessarily reflect our business momentum accurately. Our sales guidance is primarily based on all our customers as part of our planning cycle. With this in mind, we will continue to report bookings on a quarterly basis through 2025, but will no longer report on bookings thereafter. As of 2026, we will report the total systems backlog on an annual basis. Looking at the full year, net sales came in at 28.3 billion euros, with a gross margin of 51.3%. UV system sales realized from 44 systems, including I&A, were 8.3 billion euros, 9% lower compared to 2023. DPV system sales grew 4% to 12.8 billion euros. Or metrology and inspection system sales increased 20% to 646 million euros. Looking at the market segments for 2024, logic system revenue was 13.2 billion euros, 17% lower than 2023. Memory system revenue was 8.6 billion euros, 44% higher than 2023. Mental base management sales were 6.5 billion euros, 16% higher than 2023. We concluded 2024 with a net systems backlog of around 36 billion euros. In 2024, we continued to invest in innovation across our full product portfolio, increasing R&D spending to €4.3 billion in 2024, or about 15% of sales. SG&A increased to €1.2 billion in 2024, which was about 4% of sales. Net income for the full year was €7.6 billion, 26.8% of net sales, resulting in an earnings per share of €19.25. In 2024, we generated free cash flow of 9.1 billion euros. We returned 3 billion euros to shareholders through a combination of dividends and share buybacks in 2024. With that, I would like to turn to our expectations for the first quarter of 2025. We expect Q1 total net sales to be between 7.5 billion euros and 8 billion euros. We expect our Q1 installed base management sales to be around 2.1 billion euros. Gross margin for Q1 is expected to be between 52% and 53%. This is primarily driven by a positive effect from no high in a revenue recognition in the quarter, partly upset by lower immersion volume. The expected R&D expenses for Q1 are around 1.14 billion euros, and FG&A is expected to be around 290 million euros. In Q4, ASML paid the second quarterly interim dividend over 2024 of 1.52 euros per ordinary share. ASML intends to declare a total dividend for the year 2024 of 6.4 euros per ordinary share. The third interim dividend of 1.52 euros per ordinary share will be made payable on February 19, 2025. Recognizing this third interim dividend and the two interim dividends of 1.52 euros per ordinary share paid in 2024, this leads to a final dividend proposal to the general meeting of 1.84 euros per ordinary share. In Q4 2024, no shares were purchased. With that, I would like to turn the call back over to Christophe.

Thank you, Roger. As Roger has highlighted, we finished 2024 with a strong quarter. We are extremely thankful to the whole ASML team that worked very hard to realize these results. In 2024, we have also successfully achieved a number of technology milestones, including the release of a number of new products critical for our customer technology on-apps. On our low NA EUV technology, the NXE3800E, we demonstrated the full system specification in our factory with 220 wafers per hour throughput at a new record overlay. We are on track to deliver a new system at full specification and start upgrades for the system already at our customer during the first half of 2025. We continue to work with our customer to drive the maturation of the system to support their ramp to high volume manufacturing. On INA EUV, we completed installation and customer acceptance on two systems in Q4. Customers have now run over 10,000 wafers on INA system, and their feedback has been very positive. They are seeing major performance benefits in imaging, overlay, and contrast, which also provides significant cost reduction opportunities for both logic and DRAM processes. We continue to work with our customer to define the exact insertion point for INA in their processes. We also shipped a third system in Q4 that is now undergoing install and qualification. On DPUV, we shipped the first NXT870B, the latest generation KRF system, capable of throughput of over 400 wafer per hour, and the NXT870B, 2150i the latest generation immersion dpv system capable of achieving throughput over 310 wafers per hours and overlay performance of equal or less than one nanometer finally in application after close collaboration with multiple customers we have successfully and recognize first revenue from a number of E-scan 1100 multi-beam inspection system. All in all, our product pipeline is pretty strong, supporting the roadmap requirements of our customer and driving our overall competitiveness. We will share more performance data at the SPIE lithography conference in February. Looking to 2025, we see full year revenue between 30 billion euros and 35 billion euros and gross margin between 51% and 53%. Consistent with our view from last quarter, artificial intelligence has become the key driver for growth in our industry at this moment. As we have witnessed in 2024, AI has created a shift in the market dynamics that is not benefiting all customers equally in the short term. If AI demand continues to be strong and customers are successful in bringing on additional capacity online to support that demand, there is a potential opportunity towards the upper end of our range. On the other hand, there are also risks related to customers and geopolitics that could drive reserves towards the lower end of the range. Looking at market segments, we currently expect logic to be up versus 2024, with the ramp of leading-edge nodes, while we expect memory to remain strong, similar to 2024. With respect to our install-based business, we expect revenue to grow versus 2024, driven by both service and upgrades as part of a growing install base, to which EUV contribution is continuing to grow. Our China business in 2023 and 2024 was relatively high because of our ability to execute on the backlog that was created after low order fill rates in previous years. For 2025 and beyond, we expect our China business to go back to a more normalized percentage of ourselves. Looking longer term, overall, the semiconductor market remains strong, with artificial intelligence creating growth, but also a shift in market dynamics, as I highlighted earlier. These dynamics will lead to a shift in the mix of end-market products towards more HPC and HPM, which requires more advanced logic and DRAM. For ISML, we anticipate that an increased number of critical lithographic exposure for this advanced logic and memory processes will drive increasing demand for ISML products and services. As a result, we see a 2030 revenue opportunity between 44 billion euros and 60 billion euros, with gross margin expected between 56% and 66%, as we presented in the investor day 2024. With that, we would be happy to take your question.

Thank you, Renee and Christophe. The editor will now instruct you all momentarily on the protocol for the Q&A session. Before, I would like to ask you to kindly limit yourself to one question with one short follow-up, if necessary. This will allow us to get through as many calls as possible. Now, operator, could we have your final instructions, and then the first question, please?

Thank you. As a reminder, to ask a question, you will need to press star one and one on your telephone and wait for your name to be announced. To withdraw your question, please press star one and one again. We'll now go to the first question. One moment, please. And the first question comes from the line of Joe Katracki from Wells Fargo. Please go ahead.

Yeah, thanks for taking the question. I had a question on your 2025 outlook. You mentioned that if customers are able to build a capacity for AI, then you think you can hit the high end of your revenue guide. How do we think about that relative to your lead times? I know you've pre-built some EUV tools, but shouldn't we have a pretty decent idea in the first half of this year if that high end of the target guide is possible?

Well, I think, you know, as we go into the year, I think that we'll have indeed more visibility on what the final expectation is going to be. I think right now we are in our supply plan taking into account some of the upside we see, so we are preparing for that. And I think you're getting it right. As time goes and these opportunities become more concrete or less concrete, we will decide to build or not build those tools.

Okay, that's helpful. And then in terms of the gross margin dynamics, you know, the cost of high-end A installs this past quarter was a bit lower than expected. You know, as we look in the second half and you talked about, you know, gross margin maybe a little bit higher in the first half, how do we think about the expectations of just the cost of installing those tools in the first half relative to the kind of outperformance that you saw this past quarter? Is there any kind of update there?

Yeah, Joe, so indeed, the costs were a little bit lower in the last quarter. But, you know, in coming up with our projections for 2025, I would say that those learnings are also embedded in there. So we are recognizing in 2025. that we have lower costs. So I would still see that as a one-off benefit to the quarter. I wouldn't necessarily read into that that we're going to have significant savings that were not recognized when we provided the 51% to 53% guidance. But you are right. In the second half, we did say that we expect the gross margin to be a little bit lower because the revenue recognition on high day is to be skewed towards the second half. Thank you.

Thank you. Your next question comes from the line of Tammy Q from Barenberg. Please go ahead.

Hi, thank you for taking my question. So, firstly, on high NA, since CMD, you have been mentioning that the progress at customers has been ahead of where you were expecting. So, currently, of course, you are negotiating with them the time frame, et cetera, but Can you comment if there is any potential for them to use it quicker because that helps them saving the cost of manufacturing?

Well, Tammy, this is Christophe. So I think that, you know, there's a few conditions for customers to use a new tool like INA. The first one is the one you repeated, which is, you know, the performance has to be attractive, has to be good. I think that's most probably one requirement we check at this point of time. The second one is going to be the maturity of the platform. And there, as you know, we typically try to ship a few early tools. These are our 5,000 so that we can start to work aggressively on maturity. We are going to start shipping the 5,200, which is a tool more fit to high-volume manufacturing. And then we'll have to spend a bit of time to – demonstrate that the majority of the research that customers are comfortable to use it in production. So that process typically takes 12, 18 months. It depends a bit on where we start with the tool. And the discussion is a discussion we have, I would say, on a regular basis because, again, depending on the progress, of course, the appetite to move faster or slower is going to be adjusted.

Okay, okay, thank you. I guess I will circle back in 12 months' time. And also, just in terms of China, so China, I noticed that they have been kind of pulling in some of the orders, maybe because of the pending U.S. restriction, which is supposed to be implemented quite soon. I mean, the new wave. Do you actually see your customer pulling in their order? As a result, do you see China order decline quicker in the coming months because of everything has been pulled in? And also, I know that you mentioned in the pre-recorded video that China's expectation on U.S. has been falling into expectation. So do you think there is potential for China to have some downside or upside in this year as you have everything clearer to you versus three months ago?

I think to be honest, our view on China hasn't really changed in comparison to when we last spoke, right? So when we last spoke, we said the key dynamic on China is that, and why China was so very strong in 2023 and 2024 now, because for a number of years, we have been building up a very significant backlog because all the fill rates for China was so low. That is the reason why the China sales in 2023 and 2024 was so high. And that is also the reason now that around this point in time where we have eaten significantly into that backlog, we say that backlog is being normalized. And that also means that we believe that the sales to China will be a more normalized percentage. And as we said, we think that is a low 20% number that China will represent in 2025 in our total sales numbers. But that is what is going on. You know, the order intake that we received from China in the fourth quarter, I would call it healthy. But it's not, you know, an order intake that is, you know, in any way dramatic or skewed or whatever. That's not the case. It is, again, a relatively normal order intake for China. And therefore, I think the expectation that we articulated last time, China to be a more normalized percentage of our sales, I think that's what you will see in 2025 and frankly beyond.

Thank you.

You're welcome.

Thank you. Your next question comes from the line of Sarah Russo from Bernstein. Please go ahead.

Thank you very much for taking my question. So we've had quite a few conversations around your decision to phase out bookings information. And as we understand, it only provides a backlog of you annually. Can you talk us through the decision to do that? And I guess what are you expecting to be the reaction from investors? And, you know, is this a response to the increased focus you've been seeing on bookings that seems to be taking place over the last 12 to 18 months?

I think we explained it in the video. And just to recap where we are, the way we come up with our annual guidance is really based on the review process that we have with our customers. We have ongoing dialogues with our customers. And based on that, that gives us the insight into their business and that gives us the basis to then articulate at the start of the year what we think the total sales level for the year is going to be. As you've seen in previous quarters, and as we have highlighted multiple times, order intake can be very lumpy. And that is because, you know, the total size of the orders per customers are pretty significant. You know, given the tool prices for high-end tools, for low-end tools, which, as you know, are pretty substantive. You know, if customers take on a number of tools, which they typically do, you get to very, very high numbers. And that means that if you have a few large customers that place an order in one quarter, they typically wouldn't do it in one quarter and then get another massive order intake in the quarter thereafter. So as a result of that, you see sometimes very high order intake in one quarter and then lower order intake in other quarters. That does not necessarily reflect the business momentum that we enjoy. And that's why we think that the market is better off With us having a robust discussion and a robustly underpinned guidance that we provide at the beginning of the year, we think that is more meaningful than these swings in the order pattern that you've seen before. If you look at past quarters, you do see that the market looks at these orders, and sometimes the reactions, both in terms of positive and in terms of negative order intake, the reactions can be quite significant. And that has been the reason that we say we have been telling the market that order intake is lumpy. We have been saying that it doesn't necessarily reflect the business momentum accurately, but we do see very significant market response to it. And as a result of that, we have reached a conclusion that maybe the market is better off with what I just gave you, including an annual update of the backlog. From that regard, we came to the conclusion that less is actually more. You know, this is something that we discussed with quite a few of you. When we had the investor day, we discussed this quite extensively with many investors. And, you know, the conversations that we had there gave the confidence to Christophe, myself, and the full team that the decision that we just took is the right decision for the company and for its stakeholders.

Great. Really appreciate the additional context there. And then maybe just a quick follow-up. One of the pieces of guidance that you were talking about in the last set of results was, you know, you were probably at the bottom end for EUV bookings to see the low end of guidance. And since you've had a very strong quarter of bookings now, $3 billion of EUV, are you sort of more confident in the midpoint of guidance, or are you starting to see 2026 EUV orders come through as well, given the typical 18-month lead times?

Yeah, what we said last time is indeed that we needed about 2 billion of EUV orders to get covered for EUV on the midpoint of 2025. So with the 3 billion that we have in there, we've reached that point. So it's clear that we're either starting to build on 26 or we're starting to build into – into the higher end of the guidance, but that really is the timing question that is dependent on it. But you are right, we said we need to get 2 billion order intake for EUV, and the fact that we have been recording three gives us, I would say, good confidence on the EUV part on the midpoint. On DPV, we're not yet fully covered, but I think the coverage there is quite high for the midpoint of the guidance as well. But as also Christophe mentioned, of course, there are always timing issues, and customers always have the ability during a year to either pull in or push out from the last quarter.

Great. Thank you very much.

You're welcome.

Thank you. Your next question comes from the line of Francois Bovigny from UBS. Please go ahead.

Thank you very much. My first question is, you see a strong demand for EUV logic driven by AI. Christophe, you mentioned in your video, presumably mainly from one large customer. How do you think about the capacity of 2nano, given it's going to be potentially a very significant node, Could we see a front-loaded, you know, capacity build in a way that 25 will be a big build to prepare for 26? So, in other words, I mean, how do you think about the phasing of this strong node as first year could be bigger and then you add on, you know, a bit less in 26, 27, or is it going to be really a granular buildup?

Well, I think, first of all, as you know, we don't have the details of our capacity customer ramp fully. I think we don't see any strange pattern in the ramps. I think we shared in the past the ramp is definitely starting in 25, will extend into 26, most probably into 2027, in fact, with potentially a mix of two nanometer and some nodes. That's what we see, so a normal pattern. The other thing we mentioned in our commentaries is we see that maybe there is an opportunity to ramp a bit faster if the demand remains strong. So I think that's another discussion we're having with the advanced logic customer. But I think we don't see any kind of pattern that you described.

And Francois, this was actually also confirmed by CCUA on the call that they had. So he made specific comments on the ramp profile of N2, so you might want to refer to that as well. That's consistent with what Christophe just said.

Thank you, Claire. I was wondering on the lead, too, maybe, given the lead times, but, I mean, as I said, that's very fair. Thank you for that. And the second question is, I mean, AI is very strong, although, I mean, the market might have some doubts in the recent, you know, days. I'm not going to ask you, you know, the impact on deep seek and the demand for AI. But you described some upside on the AI potential if we get the capacity in the second half or if the demand still remains strong. But similarly, if you look at the iPhone or Apple or iPhone, which supposedly is a very significant part of your demand on the leading edge nodes, do you think it's You know, AI is big enough to offset some downside in the, you know, if we see downside on the iPhone, at least in the short term, because in the long term, you describe at your CMD that, you know, it's centralizing. But I was more wondering in the short term, so 25, 26, you know, the dynamic of high AI and slower smartphones.

Well, I think in the short term, I think that we try to explain also with the upside. I think that the demand is strong to really drive, I would say, the utilization of whatever capacity would be built initially on advanced logic. I think, by the way, this will not be fully or so mostly dedicated to AI. So I think, yes, the rest of the market is not doing spectacular, but it's still there. I think that I will also refer to the TSMC call where they were reporting also for the first time to some recovery on mobile. So I think nothing is spectacular there. But, you know, as you go for the ramp and you know that the ramp will build up capacity over time, so you don't get the peak capacity before quite a bit, I think we – we think this part is solid at this point of time. So we see more upside than risk, at least with all the elements we have at hand. Great. Thank you very much. You're welcome.

Thank you. Our next question comes from the line of Chris Sanka from TD Cowan. Please go ahead.

Thanks for taking my question. Christopher, the first question on the decline in China sales from 47% in September and 27% in December. Is that all a function of digestion by Chinese customers? Since I'm assuming that the incremental export control rules did not impact your last quarter, so I'm curious, when do you expect this digestion to probably meet a steady state or bottom out?

Yeah, I think, Rish, we were already signaling in Q3, I think, that this trend might happen. And I repeat what I just said on China. I think this is the main reason why we're looking at a normalized percentage for China is the reason that I gave before. So, yes, export controls are a part of that. But the lion's share of the reason for the normalization of the demand is what I said earlier, which is that the eating into the backlog that was built up all the way through 22 is now coming to an end. And now we're really seeing a normalized pattern.

Got it. Got it. And then a question on memory. You kind of said that this year logic would be up, but memory kind of flattish. On the memory front, you're more exposed to DRAM than NAND, and you keep hearing about HBM capacity increase, but it seems like you're not seeing better DRAM dynamics this year. Can you just tell us a little bit about what you're seeing in the memory ecosystem? Sure.

Well, I think you almost summarized it in your question. I think that high bandwidth memory is driving today, I would say, also an aggressive capacity addition, at least for some of the customers. I think on the normal DRAM, I would say my comment is similar to the one on mobile for 4logic before. I think there is also nothing spectacular, but there is some recovery, which also calls for more capacity. That's why we still see DRAM pretty strong in 2025.

Thank you.

Thank you. Your next question comes from the line of Chris Casso from Wolf Research. Please go ahead.

Yes, thank you. Good morning. I guess and, you know, as I listened to you this morning, it sounds, like a situation perhaps similar to how you start out last year and correct me if I'm wrong with some pretty good backlog coverage to start the year. And last year there were some changes that backlog middle year. So I understand your prudence. I guess one is that accurate depiction of what we're seeing here. And then secondly, What are the sort of plausible things that we should be watching for and concerned about that would justify the lower end of guidance? Is it, you know, export control restrictions? Is it China? Is it perhaps some of the concerns about AI? You know, what are some of the scenarios which would sort of point you to the lower end?

Let me take the first part and then Christophe can look at the second part. I'm not entirely sure I follow your logic on that we are today where we were a year ago. And also I would point out that a year ago we gave a guidance that I think we pretty much nailed during the year, right? So I think we delivered last year exactly on the guidance that we provided in the year. The reason why I think the situation is also not comparable is because Last year, we started the year saying we think this year will be comparable to 2023. I think this year we started the year with a clear growth perspective at the midpoint, somewhere around 15%. So I don't think the situation is similar. Yes, we have a strong backlog. Yes, we had a strong order intake in the last quarter. But I think Christophe discussed it on the call, and he will provide the color and context. Now we see a clear path within that, you know, within that, within that guidance, maybe you can.

Yeah. And I think on the law, and I think even Roger mentioned it already. I mean, we, we have seen last year, I think this was one of the low light of Q3. We've seen major push out of some of the capacity for some of the customer. I think Q3, You know, through our guidance this year, we see that this risk is a lot lower in terms of, you know, in terms of volume. But I think, Roger said it, that's always possible. So I think we are very transparent and we share with you the fact that, you know, if we look at the total dynamic, that some of those push-outs may not be completely excluded. That's what defines our range. So I think that's not more than that, I would say. Geopolitics, we also mentioned that. It's even more difficult, of course, to quantify because we just had a new export control being released, so there's no expectation of anything on our side, but this is also out of experience, this call a bit for credence on our side. So that's a bit what we try to reflect in the lower hand of the guidance, not more than that.

Sure. Very helpful context. Thank you. The inevitable kind of follow-on to that is how that affects the view for 26. And given the lead times and the customer forecast, imagine you have some degree of visibility now, although it's very early. At this point, would you expect 26 to be a growth year? And what are the variables that you're thinking about with regard to 26?

Yeah, Chris, that's as much as we've said. You know, we are looking at 26 as a potential growth year for ASML. That's how we look at it. But it is exactly as you said, it's way too early to provide any direction or magnitude on that.

That's great. Thank you. You're welcome.

Thank you. Your next question comes from the line of Didier Simama from Bank of America. Please go ahead.

Yes, good afternoon, gentlemen. Thank you for taking my question. I wanted to go back to the order intake in EUV and DUV. So maybe first on DUV, can you maybe just help us understand how much of the DUV orders was related to DUV capacity build for the non-critical layers, e.g. non-China or any color you can give us? And I'll have a follow-up on the EUV order breakout. Thank you.

Yeah, we typically don't disclose that, DJ, as you know. We don't break down between customers or between regions. But a significant part of the DPV order intake for the quarter was indeed related to larger customers and for larger customers for the non-critical layers, as you mentioned it. But we're not putting a percentage on that.

Perfect. That's good to hear that it's confirmed. On the EUV order intake, obviously largely driven by foundry logic. Well, I think you said 50-50 perhaps earlier on. On the DRAM side and on the foundry side, maybe any color you can give. So obviously on the foundry side, you've got three major customers. On the DRAM side, you've got three major customers. Can you just obviously discuss the various customers' sort of health? but is there any change with regards to the trajectory of the various buckets, e.g. is there any bucket looking a bit better? It looks like DRAM is looking a bit better, given the diversity of the customers in there. It would be soundly not changing very much, but it would be much more interesting hearing your views.

Well, I think we're looking at each other, basically. I don't know which one of us have to tell you first that, well, these are not the type of detail we share because we would be even adding speculation on some of the market situation. So the only thing I would say is, again, no major change compared to what we have discussed in the previous quarter. That would be my input to that question. Amen.

Okay. Well, I'll allow myself to follow up then. Fair enough. I don't blame you for trying, right? You know, French people. No, on high-end, just maybe give us a little bit of insight into that third shipment. Is that for a North American customer or is that also for maybe a Taiwanese or Korean customer that would be helpful?

I think we said it's not for the North American customer, but we have not said and will not say where it went.

Okay.

I'll go back to the first two.

I've tried.

You did.

Thank you. Your next question comes from the line of CJ Muse from Cantor Fitzgerald. Please go ahead.

Yeah, good afternoon. Good morning. Thank you for taking the question. I guess first question, in the video when you discussed your 2025 outlook, you talked about obviously AI supporting the high-end, other markets low-end. I would have thought you would have completely de-risked the other markets. So can you touch on the level of de-risking embedded in your assumptions for those customers where you've seen push-outs to 26? And for the high-end, is there sufficient clean room space that those foundry and HBM customers need? to actually take tools by the Q4 timeframe?

Yeah, I think on the first part, I think Roger indicated when it comes to booking, I think that mostly we have the risk, the large part of EUE. So the risk, I think I'm going to repeat myself on this one, but the risk is just on some potential push out. I think that's what it is. So we're just a bit cautious based again on some of the experience we all went through last year. That's the first part to your question. The second part, I think that's the right question. I think, you know, most probably today, you know, we're looking at opportunity to get more clean room space. I think that's what we mean by building up additional capacity. And if this is possible and if this happened, then we can most probably put some of our tools there. But that's exactly what we're looking at with our customers.

Very helpful. And I guess as a follow-up, you talked about the ongoing shift led by AI across HPC and HBM. And curious, you know, in your discussions with your customers across both those fronts, Can you comment on kind of what you're seeing in terms of EUV layer counts, whether it's A16 or whether it's going HBM3, E4, et cetera? I would love to hear kind of your thoughts on the trajectory there.

Well, I think we tried to explain that in November. So the need for a high-power computer, the need for high-bandwidth memory, so basically the AI-driven chips, are more demanding when it comes to advanced process. I think what we said is we see those new product mostly calling for a bit of an acceleration of Moore's law and therefore our customer being more aggressive when it comes to technology transition. I think in some extent you see a bit of that already with two nanometer and mostly what will come after that. And I think we also said in November that historically This has always driven more demand for advanced lethal. That's, of course, part of the discussion we are having as we speak with our customers. So, you know, part of the shift we see, I think that's one that we describe as a positive one for us back in November. And I think when it comes to the trend, the trend is there. And, well, you're asking very detailed questions on, you know, future nodes. We don't have the detailed answer yet, but the trend is definitely there. Thanks so much. You're welcome.

Thank you. Your next question comes from the line of Andrew Gardner from Citi. Please go ahead.

Good afternoon. Thanks for taking the question. I had another one on the customer visibility side of things. It's encouraging, obviously, to see that you finished 2024 strongly in terms of orders. But if I look at 2024 overall, you've had a fairly low level of orders from the logic community, lowest since 2020. Now, I understand the multiple reasons why it's been weak in recent quarters. So really, I'm sort of interested in a more forward-looking view. You've already explained that 2025 is essentially fully booked. But as I look to 2026, assuming growth, as you've described loosely, Roger, It's something like 30 to 40% booked for next year, depending on one's assumptions. And so then if I think of your lead times and the points you were making about the stated customer plans for their capex, the two nanometer migration, and in particular, the high volume ramp in the first part of next year, don't you need to see all the commitments come in for EUV over the first half of this year in order to be able to meet those customer needs on time?

Yeah, Andrew, I think that's fair. So we should be seeing, you know, order intake in the first two quarters, particularly when it comes to EUV. I think that would be logical. As Christophe described, we do have flexibility, obviously, in the way we approach 2026. So we do have flexibility in our supply chain, also in our manufacturing capability. So it's not like at a certain point in time that's where the music stops. But definitely we should continue to see good order intake in the first two quarters. I think that's a fair assumption.

Thanks. I mean, related to your answer there, Roger, the strategy of pre-building, is that continuing? Have you changed that at all given the shift in customer dynamics we've seen over the last six to nine months?

So really we're doing this also, Andrew, to make sure that we have a level loading of our factory, right? And that's one. And second, in order to create sufficient flexibility for us. So we're reviewing that. We're obviously reviewing that based on the latest insights that we get from customers. I reiterate the ongoing discussions that we have with customers rather than just the POs. And based on what we get there and based on the insights that we get there for this year and for next year, that also determines what our pre-built strategy is. But, you know, that is an option that we continue to have and that we continue to use in order to, you know, to better optimize the loading of a factory.

Thank you very much.

Thank you. Your next question comes from the line of Mehdi Husseini from SIG. Please go ahead.

Yes, thanks for taking my question. Two follow-ups. Roger, can you give me an update where we are with the throughput for NXE 3800E? Where is it today and how are you going to close the year? And then for the team, I want to better understand how you're planning internal capacity, especially looking beyond the 2025. When we had a strong cycle a couple of years ago, your backlog was in the $40 billion range. Do you see more flexibility with your capacity to bringing the lead times down, especially with higher EUV mix? and concentration of customers, or would you be more prudent and try to build a longer backlog? I'm just trying to understand and balance your internal planning for the next couple of years versus your ability to get customers committed, especially given there is more concentration of customers for EUV.

Yeah, Maddy, so on the throughput, the systems that we shipped in last year were all on a better throughput than the throughput of the 3600, but not yet at the maximum throughput. That maximum throughput has been demonstrated, and we're now in the process of rolling that out. So the newer tools that we're now shipping to customers will gradually meet that specification for the tools that have been installed with the customers we're going to make sure that in the course of this year that those tools also meet that higher expectation. So therefore, you will continue to see also us reporting on the improvement and also the revenue that we recognize from the fact that we get those higher throughputs achieved with customers and also get recognized for that. So that's the status of the 2021. has been demonstrated now in the process of porting that out for the new tools and also, you know, backporting that to the 3800s that are currently in the field. In terms of capacity beyond 2025, the key thing that we're doing is to make sure that what we call is the long lead time items that they are in place, i.e. that we have the flexibility to respond to, you know, to higher demands. and that we have the long lead time items that we have those in place you know if you look at what we're projecting in terms of growth for the second half of this of this decade including what we're projecting for 2030 and it's clear that capacity needs to be built so the infrastructure um you know we we have already built and continue to do that so that the long that the infrastructure is in place and that then together with light chain and with our hat count we can respond once we get get the very clear signals from our customers that demand is accelerating. That's the way we approach it.

So does that mean that the backlog would be kind of in the 18-month range, or are we going to go back to like 22, 23 when backlog was extended due to supply constraints? Yes.

No, that would not be. I think a normal backlog is ultimately what we want to be able to offer to our customers. We want to be able to offer to our customers a normal backlog because that will also help them respond to their business. I mean, that's ultimately what you want to get, that customers have a backlog and an order lead time that is manageable for them and that allows them to respond in a good and flexible way to their demand fluctuations. And that's our ultimate goal. And that's why we want to be as flexible as we can. And that's why We put in the infrastructure such that we are able to give our customers a normal orderly time. That's the objective that we are driving.

Got it. Thank you for the details. You're welcome.

Thank you. Your next question comes from the line of Stefan Uri from OdoBHF. Please go ahead.

Yes, good afternoon. I would like to come back on the guidance for 2025 last quarter. I think you gave us some elements to rationalize how you could get to the middle of the range, and you gave notably two elements. One was on the non-EUV, non-China that was supposed to grow largely by double digits, let's say 50%. So can you maybe come back on that and confirm if that's still what you expect? And the second element was about the install-based management revenues that you set It could be at about $7.5 billion, but if you take the first quarter where you get it for $2.1 billion, it gives a run rate, if you multiply it by four, much bigger than $7.5, $8.4, $8.5, so $1 billion above. So is there any seasonality we should have in mind about installed base management? Thank you.

So let's take the last one first. So the 7.5, that's the number that we still include in our guidance. And, of course, over the quarters, you've also seen that last year there can be some fluctuation. To a very large extent, that's dependent on upgrades. So that fluctuates from one quarter to the other. I would also say, and you know that, our visibility into the upgrades is, of course, less than the service businesses. But I think for now, I would encourage to still use the 7.5 billion as the right number. When it comes to the non-China EUE business, I think you mentioned 50%. I think that's on the higher end. I think that's a little higher than what I had in my model when we were having the conversation. So 50% is a little too high, I think. I think we said it's sort of on par with what you would see in the EUV business, and that would get you to approximately 40% increase. And that's sort of where I think you're going to land. And this is a change, right? What was that? No, it's not a change. It's not changed. Okay. No, that has not changed. That is still consistent with the analysis that we provided after Q3. Thank you very much. You're welcome.

Thank you. Your next question. comes from the line of Michael Rourke from Digoff Petercam. Please go ahead.

Yes, good afternoon. Can I ask a question about one of your slides from the Investor Day 2024? We'll do all day. Okay. Well, now you're still laughing. Whoa! It's a presentation by Mr. Harchandani. Slide number 13. And that is the slide with the 7M markets and the total semiconductor sales. I'm sure you'll be able to visualize that one, correct?

I have it in mind. Yeah, yeah, yeah.

Okay, good. Well, first of all, it's a pity that you no longer show the historical data, which you did in the past. So now my whole model is missing a few years. But now my question is the following. Do you believe that your customers will be able to generate $1 trillion in sales in 2030?

It's not a customer's right. When you talk about the semiconductor industry, this is for customers.

And that is precisely the answer I want to hear. Because the Gartner market definition doesn't represent your customers. It represents customers. your customers, the IDMs, and the fabulous companies that are customers of your foundry customers. However, if I look at these seven end markets, then my perception is that the difference between vendor sales and manufacturer sales, which is a huge gap for the total market, that that difference is biggest for smartphone and data centers. because that's where the IDMs have the smallest positions. So you've changed all your scenarios with data center scenarios going up, most other markets scenarios going down, which optically looks great. But from a manufacturer perspective, I don't know if anything has changed actually.

To be honest, I think you're losing a bit of the audience here, and you're looking at a slide that no one else is looking at. So I would suggest that we have a follow-on conversation on this topic.

Perfect, perfect. However, you know what I'm talking about. I'm surprised that everybody keeps showing the Gartner data, which are not a reflection of the equipment customers. Yeah.

I appreciate it, but I think we have the right follow-up action.

Let's take it offline. One tiny follow-up, a whole different topic. You will no longer show the quarterly bookings. Is perhaps a 12-month rolling booking figure something that would be smoothing out all those lumpy intakes?

But to be honest, I think a backlog per annum gets you awfully close to that, right? Because if you have the backlog and you have the sales numbers for the quarters, I think it doesn't take you too much to kind of recalculate that. True, true.

We only see it once a year, don't we?

Yes, you do. Yeah, yeah. And I think, as I explained before, we believe that's an important data point for you, and it's good enough.

Okay. All right. We'll take it offline, and then more questions about these slides will be in the next quarter and the one thereafter, because I've got many more questions.

Thank you.

Appreciate that. Thank you.

Okay, we've got time for one last question. If you've been unable to get through on this call and still have questions, please feel free to contact ASML Investor Relations with your questions. Now, operator, can we have the last caller, please?

Thank you. Your last question today comes from the line of Sandeep Deshpande from J.P. Morgan. Please go ahead.

Yeah, hi, good afternoon, and thanks for letting me on. I have a quick question as a follow-up on high NA. Christophe, you talked to a response to an earlier question that you will know in 12 to 18 months the timing of the insertion of high NA, but in terms of your backlog, I believe you already do have high NA tools in your backlog for shipment in 2026. So how do you put the two and two together that you will know whether the customer is going to use these tools in high-value manufacturing only later in the year or next year, but you're already possibly building these tools for the customer. So is there a risk that the customer says to you at the end of the year that, well, actually we're not going to implement the tools into production next year, and that these orders don't happen in 2026, which are there in the backlog today?

Well, I think, so you mentioned the backlog and I think you know that indeed in our backlog, we have quite a few tools that are not EXE 5000, which was a tool basically planned for R&D. So indeed our customer have upfront already ordered some 5200, which are tools intended to be used in high volume manufacturing. So this means that The logic of insertion has been there for quite a while. You start with R&D, then you plan for, I would say, a reasonable insertion. So those tools in the backlog are indeed going to support that. And I think we look at a few tools per customer, which are sufficient, basically, to start going. And that's what we call, you know, usually the first node insertion. And after that, of course, when we look at new booking on INA, we'll be most probably looking at the bigger insertion in the node after that, which is a bit later. But the logic you describe is the logic customers have put in place together with us already, in fact, a couple of years ago. And I think they are sticking to that logic today.

Okay, since you've run out of time, I will leave it at that. Thank you so much.

Thank you. All right, and so now on behalf of ASML, I'd like to thank you all for joining us today. Operator, if you could formally close the call, that would be much appreciated.

Thank you. This concludes the ASML 2024 Fourth Quarter and Four-Year Financial Results Conference call. Thank you for participating. You may now disconnect.