ASM International Aktie

Good afternoon. This is the Chorus Call conference operator. Welcome, and thank you for joining the ASM Second Quarter 2026 Earnings Call. [Operator Instructions]

At this time, I would like to turn the conference over to Mr. Victor Bareño, Head of Investor Relations. Please go ahead, sir.

Thank you, operator. Good afternoon, and thank you for joining our Q2 earnings call. With me today are our CEO, Hichem M'Saad; and our CFO, Paul Verhagen. ASM issued its second quarter 2026 results yesterday at 6:00 p.m. Central European Time. For those of you who have not yet seen the press release, it's available on our website together with our latest investor presentation.

As always, we remind you that today's conference call may contain forward-looking statements in addition to historical information. For more details on the risk factors relating to such forward-looking statements, please refer to our press releases and financial reports, all of which are available on our website. Please also note that during the call, we will refer to profitability metrics, primarily on an adjusted basis. Reconciliations to the reported numbers can be found in the press release and in the investor presentation.

And with that, I will now turn the call over to our CEO, Hichem M'Saad.

Thank you, Victor, and thanks to everyone for attending our second quarter 2026 earnings call. We will follow the usual agenda for today's call. Paul will begin with a review of our second quarter financial results. I will then discuss market trends and our outlook followed by the Q&A session.

I will now turn it over to you, Paul.

Thank you, Hichem, and thanks also, everyone, for joining our call today. So let me start with the Q2 financial results. Revenue in the second quarter of 2026 amounted to EUR 1 billion, above our guidance of EUR 980 million. And on a constant currency basis, revenue increased by 24% year-on-year and by 15% compared to Q1 '26.

Equipment sales increased by 22% year-on-year at constant currency and were driven by record high ALD sales. Spares & Services continued to deliver a very strong performance with a 34% year-over-year growth in global currency. This reflected the continued success of our outcome-based services and also strong demand for spares in the current environment of elevated customer fab utilization rates.

In terms of customer segments, revenue was again led by logic foundry, which represents the largest customer segment. Sales in the leading-edge logic/foundry segment increased strongly compared to Q1. [indiscernible] related sales accounted for the largest part, while sales in the 3-nanometer to 7-nanometer nodes also showed a nice uptick.

After the acceleration in Q1, mature logic/foundry sales remained at a solid level in Q2, with China continuing to account for the majority of these sales. Memory sales increased sequentially compared to Q1 and were mainly driven by HVM-related DRAM applications. With these Q2 results, we have now disclosed for the first time the equipment sales breakdown by customer segment for the first half year. In the first 6 months of the year, logic/foundry was by far the largest segment accounting for 77% of total equipment sales, both the leading edge and the mature segments had a solid contribution.

Memory contributed 15% of the total in the first half. This is slightly below the 16% contribution report for the full year '25, primarily explained by the phasing of shipments. We expect memory sales in the second half to be substantially higher than in the first half, driven by strong demand for advanced DRAM solutions. The remainder of sales consisting primarily of power analog and wafer, represents a relatively low 8% of total equipment sales in the first half. All the power/wafer and analog revenue increased compared to prior year, it was from a low base, reflecting the continued impact of softer market conditions.

For the second half, we expect the contribution from power/analog and wafer to increase. Turning now to profitability. Gross margin in the second quarter amounted to a strong 51.9%. Gross margin benefited from a favorable product and customer mix, including a continued strong contribution of the China market and also the results from improved efficiency and productivity initiatives. For the full year, we expect gross margin to be around 51%.

SG&A as a percentage of revenue improved meaningfully to 7.9% in Q2. This reflected solid operating leverage from higher revenue levels and our continued focus on cost discipline. For the full year, we expect SG&A as a percentage of sales to be below 8.5% compared to 9.2% in prior year.

Net R&D increased 22% year-on-year in constant currency in Q2. We continue to invest heavily in innovation to support customer road maps at future technology nodes and to advance our expanding portfolio of growth opportunities. Despite the increase in spending, net R&D as a percentage of revenue declined slightly to 11.1%. For the full year, we intend to keep net R&D within our target range of a low double-digit percent of revenue.

Adjusted operating profit increased by 27% year-on-year at constant currency, and the adjusted operating margin remained at a very strong 33%, in line with the record level achieved in Q1. If you look at the main movements below the operating line, financial results included a currency translation gain of EUR 22 million in the second quarter compared to a translation loss of EUR 60 million in the second quarter of last year. As a reminder, we hold a large part of our cash and receivables and payable positions in U.S. dollars and related translation differences are included in our financial results.

Our share of income from investments, reflecting our approximate 24.6% stake in A&P, amounts to $9 million in the second quarter, up from EUR 4 million in the year ago period.

Let's now move to the balance sheet and cash flow. ASM's financial position remains on a strong footing, and we ended the quarter with a cash position of EUR 1.2 billion. Free cash flow increased to a record of EUR 355 million in the second quarter, driven by strong profitability and an improvement in working capital days. In Q1, we still saw working capital cash outflow, reflecting the strong ramp-up in activity levels and the back-end loaded nature of that quarter sales. Days of working capital improved to 50 at the end of June compared to 69 at the end of March. We believe working capital remains well under control, although it will continue fluctuating from quarter to quarter.

CapEx amounted to EUR 63 million in the quarter. And for the full year, we continue to expect CapEx to be above the higher end of the guidance range of EUR 150 million to EUR 250 million, with the largest part related to the construction of our new site in Scottsdale. In short, the quarter once again demonstrates our ability to combine strong growth with continued investment in innovation while maintaining excellent profitability.

And with that, I'll turn the call back over to Hichem.

Thank you, Paul. As Paul discussed, we delivered strong results with quarterly revenue exceeding the $1 billion milestone for the first time, despite increasing strength across the semiconductor supply chain. Supported by robust end market demand and ongoing industry capacity expansions, customers continue to place a high priority on securing the equipment required for their growth plans. I'd like to thank our teams for their execution and tireless efforts to deliver on our commitments in this demanding environment.

Over the past several years, we have invested ahead of the curve to expand our manufacturing capacity in our key manufacturing sites of Singapore and Korea. Today, we are well positioned to increase output to support customer demand.

As supply chain conditions become increasingly stretched, we remain focused on working closely with both suppliers and customers to meet shipment schedules and help enable our customers' success.

The demand environment remained very favorable in the second quarter. Hyperscalers continue to invest aggressively in AI infrastructure to support rapidly growing AI workloads. Advanced semiconductors are a critical building block enabling this expansion. And the rapid increase in compute demand is driving the need for both additional semiconductor manufacturing capacity and continued technology innovation. As a result, investment activity across the semiconductor value chain remains strong, supported by both capacity expansion and ongoing leading-edge technology transitions.

Let's first review the trends in logic/foundry, our largest market. In advanced logic/foundry, we continue to see strong momentum across multiple technology nodes. Capacity expansion has been 2-nanometer node remains the largest driver of investment activity, supported by the ongoing capacity ramps and increasing adoption of gate all around technology for advanced logic devices.

At the same time, we are seeing an uptick in investment activity in the previous generation, leading-edge nodes of 3 to 7 nanometer, consistent with the trend that we first highlighted during our first quarter earnings call. Growing demand for advanced CPUs and emerging agentic AI workloads in tightening available capacity and driving increased demand for the 3- to 7-nanometer nodes following a period of relatively limited spending levels for these 2 nodes in the last -- in the past couple of years.

While the 3- to 7-nanometer nodes are no longer the industry most advanced technology generation, they remain ALD and FEMSA's nodes where ASM continues to hold a strong share of wallet. Looking ahead, leading customers are preparing for the industry's next major technology transition at 1.4 nanometers. Customer engagement remains high, and we continue to project the first contribution in the second half of 2026 as customers start investing in 1.4-nanometer pilot lines. This node is expected to deliver another mini fold step forward in device performance and power efficiency, enabling the next generation of AI and high-performance compute devices. Some customers have commented that they view the 1.4 nanometer as a potentially larger opportunities than 2-nanometer, which itself is expected to exceed the scale of the 3-nanometer node.

As we have discussed previously, we expect our served available market at 1.4 millimeter to increase further as customers increasingly deploy additional process steps and performance-enhancing there to unlock the full potential of the next generation of gate-all-around architecture. Next to a solid increase in our SAM, we remain confident that our market share in 1.4 nanometer will further strengthen compared to name both in ALD and in API. We are also very pleased by the recent wins for our modeling ALD offering at the 1.4 nanometer node. These trends position leading edge logic foundry to remain a key growth driver for ASM over the coming years.

Let's now discuss the mature logic/foundry market. In mature logic assembly applications, particularly in China, demand remained strong in the second quarter, following the acceleration already seen in the first quarter. Customer appetite in the China market for capacity addition continues to supported by many of the same secular trends underpinning investments elsewhere, including the growing demand for AI-enabled devices and infrastructure. Our strong sales development in China also reflects our company's continuous competitiveness with customers valuing the combination of leading performance and attractive cost of ownership of our equipment.

Looking at our China sales in total, mature logic/foundry continues to be a sizable part but we are also seeing increasing demand from a small base in the memory segment and a gradual recovery in power/wafer/analog.

[Audio Gap] DRAM market. And during the quarter, we were selected by another DRAM customer for our epitaxi solution. Looking further out, we remain very positive about the strategic opportunity in DRAM with the transition to 4 square cell architectures and FinFET-based peripheral sites, which are expected to move to production in the 2028, 2030 time frame. Since F-based peripheral [indiscernible] is expected to deliver further improvement in performance and speed, while the transition to [indiscernible] square cell architecture and vertical channel structure is targeted to enable a higher bid density and continuous scaling.

This technology transition increase process complexity and are expected to drive additional ALD and Epi intensity, creating an attractive long-term growth opportunity for ASM. We reiterate our forecast that these transitions will increase our DRAM served available market by USD 400 million to USD 450 billion over the next 2 notes, supported by expanding customer R&D engagement in 4F Square and Finery, we are targeting an increase in our DRAM market share.

Innovation remains a key focus for ASM. As AI-driven demand continues to increase the need for more capable and energy-efficient semiconductors. We continue to invest heavily in R&D to help enable key technology transition, including next-generation gate all-round architecture and for 4S Square DRAM. We also see advanced packaging emerging as an attractive medium-term growth opportunity as chiplet-based architecture and heterogeneous integration increase the importance of materials innovation, bonding and interface engineering.

Beyond ALD and Epitaxi, we continue to invest selectively in areas where we can bring differentiated technology to customers. One example is our plasma-enhanced CVD patterning solution that's gaining encouraging customer engagement due to its excellent gas sales capability relevant to many applications. Although still in the early stage of adoption, it illustrates how as a material discovery company, we can translate innovation in materials and process technology into future growth opportunities and gradually broaden our served market.

Let's now discuss the outlook. As communicated in our press release, we expect Q3 revenue to increase to EUR 1.1 billion. For the second half, we project revenue to be up by over 20% compared to the first half at constant currency. The key driver will be the advanced logic/foundry business, including solid sales in the 2-nanometer node, the 3- to 7-nanometer nodes as well as the first meaningful contribution from the 1.4-nanometer node.

We also expect our memory sales to show a substantial sequential increase in the second half, supported by a record high quarterly orders in the segment in the second quarter with the phasing of shipments this year more second half weighted. We expect this increase to be driven primarily by advanced HB and DRAM and to a lesser extent, by an improvement in memory demand in China.

In the power/wafer/analog segment, we expect sales to increase in the second half from a lower base in the first half. Growth in this segment remains selective and is primarily linked to AI-related applications, particularly technologies supporting the increasing power requirement of data centers. The only segment expected to be down is mature logic/foundry, reflecting the first half weighted nature of sales in this segment this year, mainly from our customers in China, as discussed last quarter. We nevertheless expect our overall China sales to remain at a solid level in the second half with growth in power/wafer/analog and memory marginally offsetting the decline in mature logic/foundry.

Looking beyond our outlook for the second half of 2026, our confidence in the longer-term growth trajectory of the business has continued to strengthen. Since our Investor Day in September 2025, market expectation for WFE spending has increased significantly. Supported by strong order momentum and customer visibility, we now expect our 2027 revenue to exceed the top end of the EUR 3.7 billion to EUR 4.6 billion range we shared last year.

Thank you, Hichem. Let's now move to Q&A. To accommodate many callers as possible, please limit your questions to no more than 2 at a time.

Operator, can we have the first question, please?

[Operator Instructions] First question is from Sandeep Deshpande, JPMorgan.

My question is, you've indicated 2027 is now going to be above top end of your guidance, how should we be thinking of the granularity there? I mean when we look at your guidance in the fourth quarter, you're close or rather implied guidance in the fourth quarter you're looking at, say, EUR 1.2 billion or so of revenues. Would this be going up sequentially into the first few quarters of the year? And how do you see that trending? And I have one quick follow-up.

I'll take the question. So we are very positive on detail about our 2027 revenue projection. The reason why we are very positive is because of our interaction with our customer. And they have given us actually very early on their equipment plan for 2027 and some of them actually even in 2028. So if you look into our revenue in 2026, we have said that the second half of the year is going to exceed 20% in the first half, making your [indiscernible] our revenue was going to exceed EUR 4 billion. So the projection that we have made before in the investor meeting in September, where we said our revenue is going to be between EUR 3.7 billion to EUR 4.6 billion. The lower end doesn't make sense from that point of view because we're going to grow in 2027 and 2026 would be at EUR 4 billion. So that's really one of the reasons we mentioned we really wanted to make sure that you guys understand that we have to talk about this.

And then, okay, we're talking about the higher range, the EUR 4.6 billion. I think based on the momentum that we see right now, okay, we see 2027 to be a very strong year for ASM and also for the industry. Well, it's very early to give really specific guidance, okay? But let's discuss the trends that we are seeing right now for 2027. So if you look into advanced logic and foundry, it's going to be really the key driver for us in 2027. This is supported by 2-nanometer node. And as we mentioned just earlier, we also see 3-nanometer and 7-nanometer node to be actually strong in 2027. In addition, we actually expect the 1.4-nanometer node to contribute meaningfully for ourselves, really strong -- 1.4 nanometer is going to be really strong in 2027 because leading customer is preparing for high volume manufacturing in 2028. As we have all around said that 2028 will be the 1.4 nanometer HVM. So we see orders happening in -- for 1.4 nanometer into -- right now. We're shipping in the second half of 2026, and we're going to actually ship even more in 2027 because customers are very serious about the 1.4-nanometer node because as we have mentioned, the benefits for this node in efficiency -- in power efficiency and performance are second to none.

So we see customers really being very excited about that. But also, we see a strong growth in our memory business because customers are increasing their investment in new DRAM. They're putting more DRAM capacity online in 2027 because of the very tight supply/demand condition right now. We also expect benefits from our expanded position in DRAM. I mean DRAM is good. And as you guys know, we are coming from a small base, but we have seen some good wins lately. And we have some wins. And because of that, we are very positive about DRAM progress for us in 2027.

Also, in the power/wafer/analog, we actually expect a recovery. We see some recovery happening right now. It's really starting, and we see it to continue in 2027. And this is really driven by power solutions for that. Even silicon carbide, I mean, we also have seen the past couple of months some good activity in silicon carbide Epi from that point of view. So all in all, we expect 2027 to be a very strong year. And I think when we arrived in 2027, we probably will be able to give you more visibility on when -- what's the number greater than EUR 4.6 billion.

Regarding -- you talked about on the release about this moly win that you had. Is this an expansion of your current position where you have already had some logic wins in the molybdenum market? Or is this part of those wins you already had?

No, this is new wins. I mean we talked before that we had some wins in moly. And actually, in the past quarter, we had actually new wins in molybdenum. We feel very excited about this market. This is the first time that ASM is looking for metal deposition market. And customer likes the solution, and we're really excited about the latest wins that we have.

Next question is from Nigel van Putten, Morgan Stanley.

First question would be on the mature logic/foundry segment. Unlike all the other segments, I think the entire industry is not growing into the second half. And I guess, interesting the same trend we've seen last year, so just trying to understand order behavior from those customers. Would it be fair to say there's a seasonal pattern there? And maybe on that, would it be fair to assume that revenue could -- or maybe should recover in the first half of '27? Or do you see a reason to caution against that? That's my first question.

Yes, maybe on mature logic/foundry, and Nigel, this is Paul speaking. What we see unique in the first half, actually, in Q1, we saw an acceleration -- strong acceleration and although we never know for sure, we mentioned that this could be related to the potential new export controls that, of course, are being debated, but are so far still not clear if they will come and in what form or shape they will come. But for sure, we believe that plays a role why we see customers accelerating orders.

In the second quarter of this year, we saw actually again a very strong quarter. So basically, as I think, already said in the Q1 earnings release that we would expect a stronger mature logic/foundry in the first half compared to the second. That's exactly what we see now, both in orders but also, of course, then revenue that will follow. But the good news is that Hichem already indicated that we see that actually compensated through growth, although both from a low base in memory in China and in [indiscernible].

So the share seems to be trading at a discount, both to the historical valuation also peers. So to me the obvious decision would be to acquire shares in a meaningful way, but clearly, you seem to have a different opinion. So I guess my question would be, is you're looking at more sizable opportunities from an M&A?

And then we'll see going into next year how we deal with excess cash. But you're right. We are looking at M&A. And if we can find opportunities, we will act on it. But there's nothing now at this moment that I can talk about.

Next question is from Didier Scemama, Bank of America.

So my first question is really for Hichem. Can you help us understand how you think about ASM revenue growth over the course of '27 and '28 relative to WFE. I think consensus expectations are for around 30% WFE revenue growth over the next couple of years. I would have thought that given your increases around 1.4-nanometer with high ALD and Epi layers, your new wins in moly metalization and maybe in 28, the beginning of a benefit in 4S Square, you'd be comfortable to be at least in line, just wanted to hear your thoughts around that. And I've got a follow-up.

Yes. Okay. Thank you very much for your question. I think that based on what I mentioned really earlier, we are very, very positive about 2027. We really are. So whatever the market is going to grow, we actually is going to grow at that market even higher than that. I think that we are very positive about the -- our position in leading edge logic and foundry. Our expanding market share in 1.4-nanometer was actually happening. And in 2027, we are very excited about our growth in DRAM with new applications and wins in both ALD and Epi taxi. What can I tell you? If the market is going to grow 30%, then we are at least going to go at that level. There's no question about it.

Makes sense. The other question was about 4S Square. So it may be a bit early to talk about that, but I think at least some of your customers are really investing in 4S Square transition towards the end of '28 for maybe 2029, 2030 type of accelerated ramp. Some people talk about even more optimistic assumptions. But what I wanted to hear from you, Hichem, is how should we think about your market share in ALD and Epi in the transition to 4S Square? Historically, as you mentioned, you've got a, let's say, weaker competitive position in DRAM versus leading logic/foundry. But obviously, you have a very strong position in single-wafer ALD and taking share in Epi. So would it be -- it's fair to have something in between these 2 market share? Or do you think you can even hope to get, let's say, single-wafer ALD or Epi market share consistent with leading edge logic?

I think time will tell. But what I can tell you here from this point of view is that we have a very strong interaction with all the memory customer for 4S Square for both our ALD and Epi technology but even more than that in some of the leading technology that we have. Customer really working with us on these applications. We understand that for 4S Square will be started in 2028, continue through 2030. So we're -- yes, we're coming from very small market penetration, but we're really excited. I mean if I look into ALD, I mean, definitely, there's more ALD layer happening in the FinFET. I think 4S Square -- with 4S Square, also there's going to be more FinFET. We're very excited also about the architecture for the architecture, which needs some ALD, both terminal ALD and PLD happy definitely. We are -- we have gained share, and we think that our solution is being accepted by our customer.

We're already in HVM in Epi with one customer, and we're getting there with the other customers. So overall, things are very positive from that point of view.

Right. Can I just squeeze in a quick one. I wondered you mentioned the strength of [Audio Gap] started since 1998. We have a very good understanding of precursor and chemistry and because ALD depends on that benefit. But also some of these materials that we do are actually providing some benefits, for example, energy efficiency. We're taking also our material know-how and expanded not only to ALD, but also to other parts of our business. like epitaxy and PECVD and we see significant benefit for us. So what can I tell you? I'm very excited. I think we're using our core competitors which we have for ALD in precursor knowledge and so on, and we extend it to other parts of our business, and we see significant acceptance of that from our customers.

Next question is from Francois Bouvignies, UBS.

I just wanted to come back on the memory comment. Hichem you said that H2 Memory will accelerate in the second half of the year, but it seems that it's mostly volume-driven capacity increase driven. Now in the last 2 quarters, you interestingly announced 2 new Epi customers on the DRAM side. So I was wondering when do you think these layers will come through will be visible, Okay, and which application, just...

Yes, I think that -- I think we're not going to talk really about really the implications because it's really customer specific. But let me tell you, it's a large application.

Okay. And my follow-up is a bit a follow-up to Didier's question and digging a bit more on '27. So if I look at 26 and your guidance, even if I take a conservative numbers, you're going to grow 35% at constant currency, most likely, which is above most likely this year, which is quite remarkable given the memory lower exposure you have. Now if I put all of that together for next year and you described many times, these AP layers now you're just that you will have to ramp up next year with 1.4 nanometers. Is it fair to say that -- I mean, it's not at least we are talking about it, the gap, the outperformance of WFE should be much wider in the next 2 years given the mix is going more into your favor? And on top of that, you have the memory layer count boosting on top? So is that a fair representation? Or am I missing something?

I think you're not missing anything, and you just heard me say that I'm -- WFM are very optimistic and positive about 27% and beyond. We're really excited about...

Next question is from Adithya Metuku, HSBC.

My first question is just on the performance you talked about in the last couple of answers. WFE numbers, if you look at it, depending on whose numbers you look at. You're basically looking at 30% to 40% growth in '27, potentially another 20% after that in '28. So if I follow on from the answer you gave to the previous question, you're essentially talking about potentially maybe 40%, something like that in terms of revenue growth in 2027. I just want to understand, I can do mathematics correctly, how you're thinking along similar lines. And I've got a follow-up.

Yes. What we've said, Adithya, is that as a minimum, we expect to grow in line with WFE and most likely more Isis explained the trends that we see in I'm not going to do the math for you, yes. There's also other, let's say, elements that play a role. I talked about China. Although today, we are still positive about China. Based on everything we know today, it looks actually quite good. As I just mentioned to, I think it was Nigel's question. But at the same time, there is low visibility. So things can still change. But overall, I can only repeat what we've said already. We are very positive. We had some nice wins. The trends are looking good, are looking in our favor and yes, indeed, if we grow more than 30%, then you can do the math, you take 26x 1.3 something. So you're correct, yes.

Okay. And then just as a follow-up. Some of your peers have been talking about potential delays to the 4F Square transition, partly because your memory customers want to focus on adding capacity at 6F Square to meet the very strong demand that they're seeing. And the worry being that if you transition to 4F Square, you might have yield issues initially at least. Is that something you've also heard from your customers? Just any color on what you're seeing there on that transition. And if that's going to the plans that people had in place 6 to 12 months ago? Any color there would be helpful.

Yes, Adi, I will take this question from you. Yes, whenever you transition to any new technology node, architecture, you might see some hiccups and so on and so forth. Yes, we are very close to our customers, and we see some of them having some issue for the transition to 4F Square. But for us, to be honest with you, this is not consequential or from that point of view because even in the 6F Square right now, technology customers want to have performance and we see penetration both in both ALD and Epi taxi right now in the 6F Square technology node.

So performance is needed. To be honest with you, maybe sometimes you need it more before we transition to second generation. I mean if you're not getting the benefit from architecture, you need to get benefit from materials. So it's good for us either way.

Next question is from Stephane Houri, ODDO BHF.

Yes. Actually, my first question is about the 1.4 nanometer and the moly recent win in ALD that you have discussed. And I'd like to understand if we're talking about something that could be sizable already in 2027. And if you can maybe come back on your global market share at 1.4 nanometer, if it's just a small improvement or something more significant? And I have a follow-up.

So what I -- thank you for your question, Stephane. I mean the way to answer your question is that first, we -- yes, we have incrementally won a couple of more applications the past quarter in molybdenum, which we are very excited about. And this is going to happen in the 1.4 nanometer technology nodes. As I mentioned that in previous calls, I mentioned that molybdenum is going [indiscernible] is going to happen gradually from gen 1 generation to the other. And because metalization you have tens of layer. So I mean -- so when you win these ones in antis there, yes, it's very good. It's beneficial. But it's not a very huge market. For us, it's really significant because this is a market that we've never been there. And every layer that we qualify is very exciting for us, especially at the 1.4-nanometer node because this node is going to be very significant starting in 2028. So overall, this is an exciting time, but also at the same time, molybdenum just starting in the industry. And with more and more generation, we're going to see more and more implementation and proliferation of molybdenum into the load.

So incremental, this is incremental benefit for us. and it's going to add to our revenue. And I mean, this is really -- for me, this is very exciting. I think that's our strategy to move into metalization. Metal deposition is working. And we also feel very positive in the future. We were developing also new precursor, new technology for moly, which is going to be even more and more differentiated in the future. And we're working with our customers for this. So things look good from that point of view.

Okay. Okay. And the second question is about the gross margin trajectory because you have always been a bit conservative with your pretty wide guidance from 46% to 51%, and you've been more or less constantly above. So you explained this with the size of China, which is quite easy to understand. But when we heard the conference call of ASML, they didn't talk about the price increase in the market because the market was so hot that they wanted to benefit a little bit more from the added value extracted from this market. So are you thinking about expanding your gross margin above the high end of the current guidance, which is 51%? And are you thinking about price increase? I mean reasonable price increase, but still price increase that would help the gross margin.

Thanks for the question, Stephane. On the margin, you've seen it in this quarter, we guided around 51%, which indeed is at the high end of the range. It could be slightly higher, could be slightly lower, but around 51%. And -- you have a few questions. Are we, let's say, implementing price increases? The answer is yes, where possible. We have some targeted price increases amongst orders also to deal with some of the cost inflation that we see happening from our supply base. So as a minimum, we want to pass that on into supply chain.

Two, yes, we still do value-based pricing. We still believe for the medium to long term, it's the best way to do it. But what you see is one of the reasons why the margin is so good. It's not only China. China is definitely still a part of it. It's accretive, but also because we have a relatively high share of advanced products, which typically, not always, but typically have higher margin, which is value-based because the more complex certain deposition layers become, the more complex that tools becomes the higher the value we can offer. We, of course, trials reflected in our pricing. That's another reason why we have been actually at the higher end or maybe even above the higher end of the range. So that you should also take into account.

And last but not least, we talked about a number of initiatives in prior calls on the standardization of platforms on merchant transit. So we also put a lot of focus and effort on becoming more efficient and working on our costs. So that's another element. So if you add it all together, you get what you see now. And of course, we will try to continue to do that and see if we can get it structurally at a higher level. But for now, we're not changing the guidance other than that we said for this year, we will be around 51%.

Next question is from Jakob Bluestone, BNP Paribas.

Earlier this week, we heard about China making progress in DV. So I was wondering if you could maybe give us a little bit of an update on what is the state of Chinese local competition that you currently face?

Okay. So to answer your question, I think that, yes, we heard the news about DTV in China. We also -- we have -- we also know that with the competition in China from different players. We -- in China, actually, we're working on -- like we mentioned, we're working on mature logic node and also we work on memory and power/wafer/analog. And not, of course, we don't ship tools for the leading-edge devices. We see our position to be good in those markets. I mean the competition is there. But I think we've been -- we see some wins and continue to really to do well from this point of view. I think by keeping really continued on innovation, which we are doing uncertainty for -- to really compete in the China market, which is very cost competitive from that point of view.

So we have to improve our cost of ownership, and we're using innovation, technology and process innovation to really achieve a lower cost of ownership and we've been able to win in that area. And I mean, that's really one of the reasons that in the mature node, which is you can think of it as very, very cost competitive from that point of view. We're still holding our own, and we see -- we still -- we are still very competitive and we like what we see right now.

So right now, everything is -- from that point of view, we are competitive. We see our competitiveness be there. We understand that China market is -- China players, there are many China players coming in from that point of view. But if we continue to innovate and which we have done both in technical benefit and also in cost of ownership reduction, I think we should be able to continue to do well.

Understood. And just a quick follow-up. Paul, I think you mentioned export controls potentially earlier. Just interested, are you seeing any sort of -- or are you currently seeing in the ordering extra inventory build because of that? I'm not sure if that's a thing you can comment on?

Yes. We -- as I said, we see actually very strong demand in China, and we believe that one of the reasons is in the export controls. And yes, we also see some accelerated ordering. It's not excessive, but there is some of that. but not excessive in any way or form. But yes, that there is speculation on new controls that typically supports acceleration of orders and delivery towards Chinese customers. So there's some of that but not excessive.

Next question is from Timm Schulze-Melander, Rothschild & Co Redburn.

First one, I just wanted to talk about was just on the technology road map and sort of capital allocation. I think, Paul, you talked about maybe looking for some further acquisitions. You've done silicon carbide P, CMP as a tuck-in. I just wanted to ask, is there strategically an asset or a capability that you don't have right now that you think would fit very well? And number two, just an update on how the integration of those is going? And then I had a follow-up.

Is there a strategic capability? Yes and no. What you've seen in the past mainly is that we accelerate access to certain technology. Quite a few things we could have done ourselves. But for a number of reasons because there was an opportunity, we decided to do it in organic. You might see that in the future as well. On the last acquisition, CMP, one of the reasons is there's a lot of complementarity with some of the deposition that we do, but also it helps us in our strategic objectives to grow in advanced packaging. So there was another reason to do it. But again, there was also a clear leverage with some deposition that we do. So it's always important. So we have typically a choice to do things organic, but of course, doing things organic takes a longer time. But yes, if we don't see, let's say, the right inorganic opportunities, we will -- we might start actually some inorganic -- sorry, some organic development for certain, let's say, capabilities or certain technologies that we want or think would be supportive to our gross strategy. That's I think how you have to see it.

Very clear. And then just looking at the revenue mix, I mean, I can't imagine it's ever been any better than it is right now in terms of just the strength in these key segments that are all showing incredibly strong growth into next year. So you talked about flowing through the revenue guide, surely a very strong tailwind on gross margin for '27, '28, just given the size of those numbers. What are the offtakes, because I mean, you talked a lot there about productivity? Is there anything in the supply chain, lead times of suppliers? Anything that we should think about this is going to sort of curtail the enthusiasm for the gross margin trajectory on a kind of 1- or 2-year view?

Yes. So you're right. I mean the trends that we see in the market and our position in that market looks really good as Hichem talked about it. You name one by one, so I don't have to repeat that. At the same time, I talked about cost inflation. So there is definitely cost inflation that we need to offset through pricing. It sounds very simple. It's not always even not in this environment because when you gain share, we still have to compete. And you can imagine that some of our competitors are -- they don't put this on -- they don't, let's say, give it away. They will fight for it. So at the same time, increasing prices is not always easy, but sometimes we can, again, depending on the value that we deliver. So there is some of that.

Yes, the whole product mix, of course, is important. But given how the market is developing, you will see a lot of advanced products, ALD, as an example, which is typically good for the margin. There is some operating leverage, not a lot, but still every year that we grow, there is some of that. And if you add it up over a number of years. It also starts to count. I talked already about China and export controls. There could be some of that. There is some level of uncertainty. But yes, today, it looks good. But yes, you'll see what will happen there. We don't know. So overall, things look pretty good. I mean, I cannot say it otherwise.

Next question is from Tammy Qiu, Berenberg.

Sorry, technical issue. So firstly, on your 1.4 nanometer outlook, did you see more customers getting more aggressive on 1.4 nanometer from a time line and volume perspective comparing to last quarter?

And also the second question is, can you talk about your China business mix potentially into 2027, i.e., mature edge foundry/logic has been very strong, driving China in this year. Do you see memory in China picking up and sustain the strong momentum China has?

I think that it's very clear that for the 1.4 nanometer node, it's already public, but there's more than one supplier for that node. And as such, the -- and that's number one. Number two, we mentioned again that the 1.4-nanometer node is a node that's significantly better in both performance and energy efficiency, which is the name of the game right now for all these AI application and data center, if you can reduce energy usage, your hero. So based on that, based on the fact that you get more performance. And will be one of the first tools that you need to order in a fab from the point of view since it's on the transistor lever that you see that. So we have good visibility. And with our very strong position there, we are very confident about what's going on the 1.4 nanometer node.

And maybe on China, Tammy, what we see today, I mean, first, maybe the disclaimer because China, there's always low visibility. So the more details we go, the more swings you will see. But based on we see today, what we expect next year is, again, the build will be mature. Henry, but yes, there will be further growth in memory somewhat, but from a low base. We're not very strong in memory in China so far. But yes, we expected -- we would expect it to grow. Yes, baring unforeseen circumstances, again, a potent you never know. And we would expect the same in Poway analog, where we see now, let's say, also the start of a recovery, still from a low base, but we would expect that to continue into next year. -- in the previous quarter, we already knew that we already knew which customers we would, let's say, ship 1.4 nanometer.

So nothing changed and maybe other than that maybe it has increased a little bit overall. But no, no, we're working with all customers as is already said, and that already was known also last quarter.

Final question is from Robert Sanders, Deutsche Bank.

I was just wondering about your supply chain and whether your backlog, in particular, as to how much of your 18-month deliveries over the next 18 months are kind of already sold out? And is your supply chain now a potentially bigger limiting factor than clea room availability? And have a follow-up.

On supply chain, it's a good question, Rob. It's definitely we see the stress levels increasing there. We already talked about it last quarter because we have a shared supply base and the whole industry is ramping. We are doing everything we can to work with our suppliers to make sure that we get the allocation that we believe we should get. We are, of course, needed also developing dual and triple sourcing. We have people at suppliers where there is stress. So far, we can manage it. But yes, there is definitely a level of stress there, but so far manageable.

Can I squeeze one last question, just on the very aggressive ramps in China and DRAM next year. Are you going to be able to participate in a meaningful way in those ramps? Or is it still a bit early days?

Yes, I would not talk about it gradually today. We don't see that yet. We do expect further growth again from a low position in memory in China, but it goes too far to say that, that would be an aggressive growth. If that's coming, we haven't seen it yet in our projections.

There are no more questions registered at this time.

Okay. Thank you, everyone, also on behalf of Hichem and Paul for attending our call. Goodbye.

Ladies and gentlemen, thank you for joining. The conference is now over. You may disconnect your telephones.

Good afternoon. This is the Chorus Call conference operator. Welcome, and thank you for joining the ASM First Quarter 2026 Earnings Call. [Operator Instructions]. At this time, I would like to turn the conference over to Mr. Victor Bareño. Head of Investor Relations. Please go ahead, sir.

Thank you, operator. Good afternoon, and thank you for joining our Q1 earnings call. With me today are our CEO, Hichem M'Saad; and our CFO, Paul Verhagen. ASM issued its first quarter 2026 results yesterday at 6:00 p.m. Central European Time. For those of you who have not yet seen the press release, it is available on our website together with our latest investor presentation.

As always, we remind you that today's conference call may contain forward-looking statements in addition to historical information. For more details on the risk factors relating to such forward-looking statements, please refer to our press releases and financial reports, all of which are available on our website. Please also note that during this call, we will refer to profitability metrics, primarily on an adjusted basis. Reconciliations to reported numbers can be found in the press release and in the investor presentation. And with that, I will now turn the call over to our CEO, Hichem M'Saad.

Thank you, Victor, and thanks to everyone for attending our first quarter 2026 results conference call. We will follow the usual agenda for today's call. Paul will begin with a review of our first quarter financial results. I will then discuss market trends and our outlook followed by the Q&A session. I will now turn it over to you, Paul.

Thank you, Hichem, and thanks, everyone, for joining our call today. Let me first walk you through the Q1 financial results.

Our revenue in the first quarter of 2026 amounted to EUR 863 million, which was at the high end of our guided range of EUR 830 million plus or minus 4%. On a constant currency basis, revenue increased by 16% year-on-year and by 26% compared to Q4 '25. Equipment sales increased by 14% at constant currency and were led by ALD. Spares & services continued to deliver a very strong performance with a 23% year-on-year growth at constant currency. This was the result of continued expansion of our outcome-based services and strong spares demand in an environment of elevated set utilization rates.

In terms of customer segments, revenue was led by logic/foundry, which accounted for the clear majority. For the full year, advanced logic/foundry sales are expected to show significant growth this year. However, due to quarterly phasing, they were down from the very strong first quarter last year. Mature logic/foundry for the large part from customers in China increased compared to Q1 last year and rebounded strongly compared to the relatively low level in Q4.

Memory sales showed sequential growth compared to Q4 last year and also expected to grow significantly for the full year, mainly in DRAM. Sales in the memory segment were predominantly driven by applications for high-performance DRAM in HBM-related applications. Sales in the power analog wafer segment increased compared to the first quarter of last year, mostly in silicon-based solutions but from a low base.

Gross margin in the first quarter amounted to a strong 53.3%. This was virtually unchanged compared to 53.4% in Q1 of last year, up from 49.8% in Q4. Gross margin was supported by a favorable product and customer mix including an increased sales contribution from China, which rebound strongly compared to the lower level in Q4. The gross margin also benefited from a gradual impact from cost reduction programs that we have been implementing over the past few years. We expect the gross margin to be at the higher end of the target range of 47% to 51% for the full year.

SG&A expenses increased by 8% year-on-year at constant currency, mostly due to higher variable expenses, but dropped slightly as a percentage of sales, demonstrating our ongoing focus on cost control. For the full year, we continue to expect SG&A as a percentage of sales to drop below 9%. Net R&D increased 11% year-on-year at constant currency in Q1. We continue to step up R&D investments to support customer transitions to next-generation nodes and to advance our expanding pipeline of opportunities.

For the full year, we intend to keep the net R&D within our target range of a low double-digit percentage of revenue. Operating profit increased by a solid 21% year-on-year at constant currency, and the operating margin reached a new record of 33.1%. If you look at the main movements below the operating line, financial results included a currency translation gain of EUR 10 million in Q1 '26 compared to a translation loss of EUR 40 million in the first quarter of last year.

As a reminder, we hold a large part of our cash in U.S. dollars and the related translation differences are included in our financial results. Our share of income from investments, reflecting our stake of approximately 25% in ASMPT amounted to EUR 7 million in the first quarter, up EUR 2 million in the year ago periods.

Next, the balance sheet and cash flow. ASM's financial position remains [ solid ], and we ended the quarter with a cash position of close to a EUR 1 billion. Free cash flow was EUR 48 million negative mainly reflecting the working capital outflow in the quarter marked by a sharp ramp in activity levels. Days of working capital increased to 69 at the end of March, up from 45 at the end of December. The main driver for the increase was higher accounts receivable due to strong sales increase compared to the relatively low level in Q4 as well as back-end loaded distribution of sales during the quarter.

CapEx amounts to EUR 38 million in the first quarter, up from EUR 30 million in the same quarter of last year. And for the full year, we expect CapEx to be around or to be somewhat above the higher end of the guided range of EUR 150 million to EUR 250 million, with the largest part related to the construction of a new site in Scottsdale, which remains on track for completion in Q1 2027. And with that, I'll turn the call back over to Hichem.

Thank you, Paul. Let's now continue with a review of the market trends. The first quarter, again, confirmed that AI is the main driver of semiconductor demand. Customers continue to add capacity to support the ongoing expansion of AI data centers and the broader infrastructure build-out.

This is keeping demand strong in the areas where we are most exposed, especially logic/foundry and we saw this demand strengthening further during the quarter. We have also noted a continuing proliferation and diversification of the AI workloads into the CPU and the power markets. For this reason, we see AI driving strength in all segments of our business: advanced logic/foundry, mature logic/foundry, memory and especially DRAM and, to a lesser extent, power, wafer, analog market.

Looking ahead, our strategic view remains unchanged. As AI adoption broadens and demand continues to scale, compute capacity is increasingly the limiting factor. In semiconductors, this is translating into tighter capacity needs for advanced logic/foundry and memory devices, driving higher investment intensity and increasing the urgency of tool deliveries.

Against this backdrop, our focus is on execution as we continue to support our customers' expansion plans. The pace of demand is putting additional pressure from the supply chain. But so far, we have been able to manage these rising challenges in close cooperation with both suppliers and customers, reflected in the sharp step-up in our quarterly sales from EUR 700 million in Q4 of last year, to a level approaching EUR 1 billion projected for Q2.

Turning now to customer segments. Logic/foundry again led our performance in Q1 supported by continued strength at the advanced nodes and a sequential rebound in mature logic/foundry demand. Our view is unchanged that logic/foundry will be a strong driver of our sales in 2026 and also going into 2027. The structural outlook for this segment remains strong.

AI-driven compute requirement and the ongoing shift to more complex 3D device architecture and new materials continued to increase ALD and epitaxy and density. As we progress through the year, we expect momentum to build further with ongoing capacity additions as the 2-nanometer technology node accounting for the largest part of advanced logic/foundry sales in 2026.

This first generation of gate-all-around device technology is shaping up to be a large node, enabling new applications in high-performance compute, including AI as well as advanced mobile and other leading applications. We continue to benefit from the step-up in our served available market at 2-nanometer supported by a broader position in Epi and sustained strong market share in ALD.

In addition, we have seen a healthy uptick in demand related to the nodes from 3-nanometer to 7-nanometer, driven by agentic AI. The demand is outstripping supply which has led to renewed capacity investment. Looking ahead to the industry's next node transition to 1.4 nanometer, we expect pilot-line investment to begin later this year. We are deeply engaged with key customers as they prepare for that transition, and we expect the first meaningful contribution to our sales in the second half of 2026.

As we have highlighted before, we expect the SAM uplift at the 1.4 nanometer to be even larger than what we saw at 2-nanometer node. At 2-nanometer, the industry's main priority was to get the first generation gate-all-around architecture and to high-volume manufacturing with gate-all-around now in production and ramping, customer have more room to include additional performance boosters.

And for ASM, that translates into more functional there in the transition stack to further optimize power and performance, including additional dipole layers to enable Multi-Vt options. Alongside the higher SAM opportunity, we have already secured several key product penetration which supports our expectation for a higher ALD market share in the 1.4 nanometer node, public disclosure from some leading customers suggest that the 1.4 nanometer node is designed to deliver clear improvement in performance, power efficiency and density versus today's 2-nanometer node.

This is well aligned with ever increasing AI token demand and the associated compute and power constraints in data centers. As our customers move toward higher volume manufacturing in 2027 and 2028, we expect 1.4 nanometer to become a meaningful driver for ASM.

Next, looking at memory. Demand in Q1 was solid, with robust momentum in the most advanced DRAM technologies used in HBM-related applications. Continued investment in AI infrastructure is keeping demand for high-performance memory strong and supporting ongoing expansion of advanced DRAM capacity. For the full year, we continue to expect healthy growth in our memory business.

Looking further out, DRAM remains a meaningful and strategic opportunity for ASM. From a technology perspective, our customer R&D engagement in memory continue to expand, including development work around new ALD and epi applications that support the transition to 4F² and very fantastic DRAM. As we highlighted at Investor Day, the transition to 4F² is expected to drive a step-up in ALD and Epi intensity and expand our served available market by approximately USD 400 million to USD 450 million based on 100,000 wafer start per month capacity.

Turning over to power, analog, wafer market segment. The contribution in Q1 remained relatively low, reflecting the soft market condition in broader parts of automotive and industrial. That said, we have seen some pockets of strength in selected areas, particularly in power applications for AI data centers. For 2026, our view is unchanged that this segment should recover gradually from a low base. We remain well positioned to benefit once demand conditions improve more broadly.

Moving on to China. The increase in Q1 was largely driven by the mature logic/foundry segment, where we saw higher activity across a broader set of customers, reflecting improving market conditions and to a lesser extent, the power, analog segment. In addition, I'd like to highlight ASM's ongoing success in winning new positions which also contributed to our strong performance in China. This demonstrated the continued competitiveness of our solution and the strength of our local team. Based on current visibility, we expect sales in China to increase for the full year with a stronger contribution in the first half.

Now let's talk about advanced packaging. As we have discussed during the Investor Day, we are looking into advanced packaging as another midterm growth area for ASM. We believe that this market is ripe for disruptive solution in new materials and interface engineering playing into ASM's strength. We are engaged with multiple customers on advanced packaging, and we are seeing some encouraging traction for our innovative solutions.

That brings me to the outlook. At current currency, we project revenue to increase in Q2 2026 to EUR 980 million plus or minus 5%, and we continue to expect revenue in the second half of 2026 to be higher than in the first half. As mentioned, China sales are expected to be first half weighted. This means that our other business segments are expected to strengthen from the first to the second half, including continued solid momentum in advanced logic/foundry higher sales in memory and a gradual recovery in power analog.

While it's too early to provide specific guidance for the full year, based on our guidance in Q2 and a further increase in the second half, it should be clear that 2026 is going to be a strong year for ASM. And with that, we have finished our introduction.

Thank you, Hichem. Let's now move on to the Q&A to ensure that everyone has an opportunity to participate please limit your questions to no more than two at a time. Operator, we are ready for the first question, please.

[Operator Instructions] First question is from Andrew Gardiner, Citi.

Hichem, just sort of pick up on the point you were making at the end of your prepared comments there. You're saying you will have growth in the second half of the year versus the first half, but obviously, the visibility is perfect to quantify it for us yet. Previously, you've been willing to talk about your performance relative to the wafer fab equipment market broadly and that ASM would outperform that. Clearly, WFE expectations are moving quite rapidly as well at the moment. Could you give us an update on how you see the broader market in terms of WFE? And can you confirm that you will still outgrow that in 2026?

Thank you very much for the questions. Yes, we talked about that in our previous conference call that we're going to at least perform as good as the wafer fab equipment market or better. Yes, we have seen improvement in the WFE market. I mean we follow very closely what Gartner and VLSI are talking about. And we can reconfirm again that our growth in our market, in our revenue in 2026 will at least outgrow the WFE market again. So as I mentioned, we see strength in the market and our revenue is strengthening, and we are very confident that we'll be able to at least grow at least at the WFE market or beat that in 2026.

Okay. And just a clarifying question, the point you were making on China. In the second half, so is that China down second half on first half? Or down year-on-year or perhaps it's both?

No. I think China is really up year-on-year. So the -- what we talked about that we see right now that China is lower in the second half of 2026 versus the first half of 2026, saying this, and I want to repeat it again. China visibility is not that great as we talked about it, okay?

So from that point of view, if there is anything, that second half China business that we see right now might also increase eventually. But right now, what we see very strongly that the second half would be a little bit slower than the first half. But again, that might strengthen in the second half. We don't know.

Next question is from Nigel Van Putten, Morgan Stanley.

I want to follow up on the previous question on China. Perhaps for the full year, there are some limited visibility. But can you provide us a revenue or China revenue as a percentage of overall revenue for the first half at least? And how that maybe compares to the full year '25 when you said it's going to be -- or it came in a little bit over 30%. That is my first question?

Nigel, thank you for your question. Nigel, maybe there's a misunderstanding that about the visibility -- low visibility of China. Right now, okay, our visibility for 2026 is very good, okay? China or no China, okay? Because I mean, it's really clear everywhere in our market, okay? So that's why we are really confident about the market. If there's anything in China, the revenue is going to increase further in the second half, okay, from where we see it right now. So, but China business has been good, and we feel very confident about it.

So maybe to add to what Hichem is saying, what we see in is now, at this moment, at least, is an accelerated demand. And in China, we have a higher H1 expectation. And H2, which might still change, we don't know, as Hichem has indicated. Possibly, that is because of concerns on export controls, we don't know.

One thing is for sure that the overall sentiment is very good, like in the rest of world, also AI related. That's itself positive. Two, we also won some more layers in itself is a positive. But yes, there is clearly an acceleration going on, which, of course, customers are on tariffs, but which could be triggered by concerns around export controls and how that will develop further, I think, at this stage, and nobody knows. Then on the full year, based on everything we know today, I think the equipment revenue as a percentage of total sales will be similar to last year. But again, it's really too early to tell, so this might change because for all the reasons that we already mentioned.

Got it. That's really helpful. Then now maybe switching to the advanced logic customers which I understand are providing increased visibility maybe 8 quarters on a rolling basis. Question would be, do you see any sort of broadening on the horizon, sort of it's clear that the main customer remains very strong, but how are the other two doing maybe today? And how do you see that developing into the second half of the year?

No, I think, Nigel, I think we see -- we're working right now with all customers in advanced node for both of the 2-nanometer node and 1.4 nanometer node. And then we see that gate-all-around is a technology that's going to be adopted by more customers. And we feel very confident that that's going to be the case. Of course, okay, some customers have better yield or performance than the other ones. But we think that gate-all-around is going to be really a broad technology node and for a variety of customers.

Next question is from Didier Scemama, Bank of America.

Just a follow-up actually to the previous question on the boarding of the customer base in advanced logic. Obviously, your largest customer is doing terrific. On the two smaller ones, is that supposed like expected to strengthen in Q2? Or is that more of an H2 driver? And I've got a follow-up.

Yes, let me take that question. I think what I can say on that. I don't want to be specific on Q1 and Q2 or Q3 when it comes out both down to customers. But what I can say is at least that based on current visibility that all those customers are expected to grow year-on-year. And of course, there is a significant difference with regard to the size of the various customer and the absolute amount of growth as a result of that. But we expect all three for the logic part, all three of them to grow year-on-year.

Okay. And for my follow-up for Q2, would you expect China to be up sequentially or flat? Or how should we think about that relative to your overall sequential growth guidance?

What we've said indeed is that for next quarter, we expect EUR 980 million plus or minus 5%. We also said that for H1, we see an accelerated demand for China coming in for various reasons I just discussed in the call before. So I think it's reasonable to assume that also Q2, China will be pretty good.

Should we expect, therefore, the gross margins in Q2 to remain at sort of above the long-term guidance given the mix?

You know that I'm not going to specifically guide on a quarterly basis for the margin, but the margin will be good that I can say China is accretive, as you know. But also, I think what is also not unimportant. I also want to highlight that is that the other product mix that we've seen actually in the last few quarters has been very strong.

So that also helps. And last, but not least, the structural cost improvements that we're working on, which will every year add a little bit also play a role. But having said that, yes, higher share of China typically is accretive, yes.

Next question is from Francois Bouvignies, UBS.

I have a question for 2027, actually. So if we look at your '26 growth drivers. I mean if I look at the different drivers, I don't see much layers increase in '26 as a growth driver, because I think it's mostly capacity, [indiscernible] was already adding a lot of capacity last year. So from a year-over-year point of view, you don't have a lot of incremental layers.

Now if you look at '27, it looks like you will have a lot of layers opportunity that you laid out at your Capital Markets Day. So I was wondering, if we think about this dynamic of layers increasing, is it fair to say that '27, if we assume the same capacity increase that '26, that should be a higher growth than '26? You have more drivers on top of the capacity in '27 than you had in '26. Is that the right way to look at it, if you understand my question?

Yes. I think we understand your question. I think it really depends both on the end demand from that point of view. But we -- as the technology node transition from 2-nanometer to 1.4 nanometer, we see the adoption of 1.4 nanometer starting in the second half of 2026. And we see the 1.4 nanometer bias to increase in 2027 for final production in the first half of 2028. And with the 1.4 nanometer node, there's more ALD and more Epi. And as we mentioned, these ALD layers are mainly in the front end of line for performance level. And that's where we have many more -- a lot of strength, and that's where we're going to have many more ALD layers.

So we are really very happy with -- we'll be very happy with the 1.4 nanometer transition, because of the higher ALD intensity. Also, we have more ALD layers in molybdenum. I think that as we mentioned in our last press release that we are very happy to be in production, high-volume production at the 2-nanometer node with our moly ALD. And with the transition to the 1.4 nanometer, we also have won some process of record layer in molybdenum. So overall, the transition to 1.4 will be very accretive to us, and we'll be very excited with that transition in the future.

And the memory side -- Yes, go ahead.

I think you said it, but I want to make it a little bit more explicit that just for you guys to be clear that already in this year with the pilot for 1.4, which is also, of course, increased layers, as you know, we already see a very, very meaningful contribution of 1.4. So that's not only in '27, but it's already starting in '26.

Good to know. And maybe you didn't address maybe the memory layers and maybe for '27. And then you mentioned market share higher in A14. So can you maybe explain a bit the higher share here? I mean, is it because just your time is getting higher than the others? Or you just have more layers than you expected? -- more than before?

Yes. So the 1.4 nanometer, what's the difference between the 1.4 nanometer node and the 2-nanometer node. So they are both gate-all-around. But for the 2-nanometer node, that's an architecture change. So customers didn't want to be very aggressive in putting many functional layers because of the change in architecture. But once we move to 1.4, they have added many goodies, which we call performance layers. And those layers are really mainly ALD layers. And they are all in the front end of line, where we play significant. That's where our strength is. Yes. So definitely, we see more ALD layers in 1.4 nanometer.

Second thing, as you shrink, those -- and with the gate-all-around structure, as you shrink those layers become much more difficult because of the 3D nature and the shrinking. And with that, the since every layer becomes much more difficult, that also slows down the process. And with that, you need more equipment from that point of view. The other thing we see is that also there is a higher epitaxy intensity going forward. So overall, that's very positive.

Next question is from Stephane Houri, ODDO BHF.

Yes. To come back on the Q2 guidance, which is about EUR 100 million of what the consensus was expecting. So I'm just trying to understand what led to this acceleration, if it's more advanced logic or memory. And if you could comment also on the lead time at the moment if they are increasing? And is there a difference between the two different segments? And I have a follow-up.

I think that the acceleration is happening in mainly in the 3-nanometer to 7-nanometer node. in addition to the gate-all-around node. So what we have seen lately is that Agentic AI is becoming more important. And with that, that tends to favor using the CPU instead of GPU. So the 3 to 7-nanometer node is really mainly driven with CPU. And we see much more demand from our customer in that node, and that's really happening super fast at this point in time. We also see strength in memory continuing. So overall, the market is really strong in the leading edge, both logic and mainly logic and foundry, that's really the highest part of the market. Second is really also DRAM is also increasing.

And about the lead time, sorry.

So regarding the lead time, I mean, lead time has increased because of the supply chain constraints right now. I mean there's a huge demand everywhere. So yes, the supply chain has increased, and that's really the customer specific. We've been able to expect that to happen. That's why we can -- we have increased our capacity to from like EUR 700 million per quarter in Q4 of last year to about EUR 1 billion per quarter this year. And it's going to continue to increase in the second half, as we have mentioned.

Okay. And that's exactly my follow-up. I mean you're going to be at least EUR 1 billion per quarter in the second half run rate and there's probably some additional growth coming in 2027, given what you said and what we see in the market. So at what point will you fill all your plants and notably the Singapore plant and that you will have to again increase the capacity?

I think our manufacturing capacity is -- can take care of our business. I think we have expensive manufacturing capacity in Singapore and Korea. So we're ready for much higher volume. I think what's limiting -- if there's any limit is really the supply chain that's limiting the capacity than anything else. But I think that we'll be able to manage that in the second half. So that's why we're confident of increased volume in the second half of 2026.

Next question is from Sandeep Deshpande, JPMorgan.

Maybe you can give a comment on what has changed in your customer behavior versus what you were -- you had seen from your customers the last time you reported in -- reported your results. Has something substantially changed given your very strong guidance into the second quarter? And then I have a small follow-up.

I think that the market is really strong all over. Has there been any significant change? I think the change that we have seen is really on the PC part where for -- on the CPU part where it used to be that AI is mostly driven by GPU, but we see that CPU part becoming more important than before. And we see that's the strength we see in the 3-nanometer to 7-nanometer node, which was not there before. So that's really the strength we see. It's mainly the CPU-driven part for artificial intelligence.

And then when you look at the WFE, I mean you had said 15% to 20% at last results. I mean, given your guidance for the second quarter and your indication on the second half of the year, it looks like you're going to grow well over 20%. So what is your perception on WFE at this point for this year? And I mean, despite your lower exposure in the memory market, you are growing incredibly well. And so is this mainly associated with the second half ramp also with 1.4 nanometer where your content is growing, your number of layers you have is growing very substantially. So this is essentially share gain in the WFE market?

Yes. Let me take that, Sandeep. So yes, to give you a very short answer, that's part of it, absolutely. But also basically, I think as Hichem already said, but maybe in different words, we're firing in all cylinders. Every segment of the market is growing significantly. I mean, advanced logic/foundry, mature logic/foundry, memory of which, in particular, DRAM, we see a high growth and even power with analog for power-related AI data center applications from a low base, but as a percentage, still high growth. And of course, also pilots 1.4, that I started with, adds a decent amount for this year already, yes.

Next question is from Adithya Metuku, HSBC.

Firstly, I wanted to talk about 2027. I know you gave these targets of EUR 3.9 billion to EUR 4.6 billion, top line at a EUR 125 billion WFE number. So call it EUR 4.2 billion midpoint. If you look at WFE numbers now, people are depending on whose numbers you take 40% to 50% higher than that EUR 125 billion in 2027. So my first question is, should we assume that, that EUR 4.2 billion could be maybe 40% to 50% higher from 2027? What are the nuances we need to keep in mind when we think about where WFE is going and how your revenues might go in 2027, you've clearly talked about outperforming WFE, I presume that will continue.

So just any pointers you can give around how we should think about these targets you gave at the CMD 40% higher, 50% higher? And I've got a follow-up.

Yes. Let me take that. So indeed, I think we said EUR 3.8 billion to EUR 4.7 billion at CMD, where we assumed EUR 120 billion WFE, which today's view is indeed significantly higher, but there's one big difference. The assumption that we took at that time, which was somewhere September last year on the composition of the mix is very different from what we see today. So we had by far the largest part of the total WFE basically logic/foundry, while now the relative share of memory is significantly larger than what we assumed.

And although we grow a lot in memory, but still our relative share of memory in our business is still relatively small. So that's why you will not see the full benefit of that increased WFE dripping down into our numbers. Having said that, based on everything we see today, we believe that '27 will be a strong year. But adding 40% to 50%, I would not recommend you to do that. That would give some distorted figure. At the same time, it's a very wide range, EUR 3.8 billion to EUR 4.7 billion is almost EUR 1 billion range. So also even within that range, there's still a lot of room to maneuver. And more than that, at this stage, I don't like to say.

Got it. Okay. We'll leave 40% or 50% of side go with 30% then. And just quick follow-up. On the MATCH Act, can you give us some color on how you're thinking about any potential impact for you guys as you think about your China revenues? Yes. Any color you can give around how you might be affected? I know it's hard to quantify numbers, but any qualitative color would be great.

Yes. So the MATCH Act indeed is being discussed as we speak. If it will happen or not is uncertain. It might or it might not. In what shape it will happen is also uncertain because at the end of the day, it is important, literally the point and the commerce are very important there, especially in relation to how to interpret what is exactly restricted. We're in, of course, discussion with relevant authorities, as you can imagine. So it's very hard, and I would love I could give you some more color to give decent color at this stage. Obviously, if something like that were to happen, it's not a positive, that might be clear. But how much, I'm really not in a position yet. It's too -- it's literally too unclear and too uncertain still on what might happen. So I don't like to speculate on that.

Next question is from Tammy Qiu, Berenberg.

So the first one is regarding your very strong short-term momentum. You mentioned that just now it's all driven by the CPU-related incremental demand. I just want to confirm that, have you seen any customer from both logic and memory perspective, pulling forward? Are you asking you to accelerate the shipment of equipment because end market demand is coming so dramatic in the short term. So therefore, it's like a pull forward from 2027 at all?

I think every customer wants the tools now instead of tomorrow. I think the demand is really high. And for us, it's which customer we ship to first than the other one. So I think like we mentioned, we are fully booked for this year. From that point of view, we have a strong demand in all parts of our business, really every part of our business very high demand. And yes, we see customers the demand is even increasing. So I mean, we -- our book is full. So we have to do our best to be able to satisfy the demand that we're getting right now.

Okay. And the second one is, last quarter, we discussed that the 1.4 nanometer is mainly driven by one customer versus others have been having discussion with you, but still a bit distant away from pilot production, et cetera. I'm just wondering where is the status of those remaining customers? Are they getting closer to make the decision on pilot production? Or are they still further down the line?

So as we mentioned that we see -- we are working with all customers to the 1.4 second generation with a 1.4 nanometer technology node. And we see that business strengthening in all the customers from that point of view. Some of them is at a marginal increase and the other have a higher increase, but I'm not here to speculate on which customer, which, but we see at least a marginal strength in some and a significant strength in other customer. But saying this, I think more likely, like I mentioned that 1.4 nanometer would be more than one customer.

Just to confirm, have you seen any progress during the quarter, i.e., all of them have moved forward or just one of them moved forward comparing to last quarter.

Can you repeat your question, please?

So basically, the time line of the 1.4 nanometer, last quarter, you mentioned that one is active preparing for pilot production, remaining two is still in discussion firmly at this stage. I'm just wondering, this is three months after, have you actually seen other customers together with a leading customer or moved forward in the time line for 1.4 nanometer? Or just one customer has moved forward instead of all three of them?

I'm going to repeat my answer, where we see 1.4 nanometer strengthening broadly with some strengthening marginally in some customers and significant increase in other customers.

Next question is from Jakob Bluestone, BNP Paribas.

I want to come back to Adi's question around your ability to sort of take part in growth in memory. And my question is, when do you anticipate the transition to 4F² and FinFET for the cell periphery in DRAM to impact your revenues? So is this something that would impact in '27? Is it '28? Or do you think it's further out?

I can take that question. Yes, I think because I think last time already, we mentioned that the pace of adoption customer by customer is different. There might be even a customer that might completely skip it. We don't know yet, but that's to be seen. And I think for us, based on what we see and think we know today, I think you should take into account '28 as the first year where we start to see a positive contribution related to 4F². [ Might ] -- maybe a little bit earlier, I don't know yet, but I would -- I mean time line is still a little bit uncertain and very different from customer to customer. So I think the best color I can give right now is in '28.

So add to what Paul has mentioned here, we see a strength in memory in 2026 and also increasing for us in 2027 and beyond. The biggest increase for us will happen really in the move into 4F², where we have more ALD layers and more also Epi intensity. But also we've seen some customers put in FinFET in their node in their road map. And with that, we're working with them and we might -- and since we have been very prominent in our FinFET technology in logic. So that we see some customers really pulling in that technology node. And with that, we probably will get some more layers as customer put in their FinFET technology node. So the biggest increase would be '28 and beyond, but also we see some increase in 2027.

Understood. If I can just ask a quick follow-up as well. You mentioned a few times the sort of pickup in 3 to 7-nanometer transition, and I don't know if you can give any color on whether that's your largest customer or kind of more broad-based?

Which transition, are you talking about, sorry?

3 to 7?

Yes, I think it's really broad based. That's really broad-based. It's not only one customer, it's very broad-based.

Next question is from Ruben Devos, Kepler Cheuvreux.

I just had one on Epi in HBM. I believe you talked about significant Epi engagements with another HBM customer and expect good news this year. So of course, curious whether two months on has anything firmed up on that additional qualification? And would that be, let's say, fully incremental to your memory plan in '26?

Okay. So to answer your question, yes, we talked about that, and we are engaging with our customer on epitaxy. There's really nothing else to say right now, but we'll let you know if there's any news from that point of view. It's really working with customers on a couple of customers on epitaxy. And hopefully, we can share some good news with you in the next investor call meeting.

Okay. And then second one, really to just get a feel of maybe the aftermarket sales, right? I mean you've had a stretch of very good performance in the last few quarters, again, 23% up the past quarter. Outcome-based is about 25% of the mix. So it looks like, I mean, the target you said at the Investor Day of 12% CAGR is becoming more of a floor.

I was curious whether you could talk a bit about, yes, the extended visibility you might have now in aftermarket sales. And I can imagine a margin uplift to realize if you manage to make a transition more towards outcome-based. But also besides that, are you able to sort of have the customer pay more per tool for the servicing packaging in general?

I think that for service market, okay? What I mentioned, the service market is really good as you transition in a newer technology node because of process complexity. It's very important to -- a customer need more support from us and for the more advanced node. And with the advanced node also, we see a transition to much tighter specification on wafer-to-wafer and also repeatability on chamber-to-chamber matching and also on system-to-system matching. And with that, we have to provide a new solution to customers to improve the uptime and the availability. So we are very really -- we think that the surface business is going to increase in the future as you transition to tighter and tighter technology node.

And we see that happening in the area of automation, in the area of robotics, in the area of optics. And those are really the solution that we're providing our customers. So the growth is going to be good in that part of the market. You mentioned that 12% growth. To be honest with you, right now, every part of the market is growing a lot. This year, the market is growing over 20%. I mean, latest, you see Gartner talking about 25%. So everything is great. It's really just spending my time, okay, to make sure that we can execute on getting customers the tools in time and make sure that the availability and the execution is top notch.

Next question is from Timm Schulze-Melander, Rothschild & Co Redburn.

First one for Hichem, please. Just looking at the technology execution and just trying to scale maybe how much upside there is to that? If I look at your long-term revenue guide, the high low range is kind of 20%, 25% between the low and the high. Obviously, part of that is the strength of the cycle. Maybe part of that is also conversion of existing evaluations and layer wins. Maybe could you just share how much of that is upside potential from layer wins? So if you could just think about that in the context of your go-forward revenues? And then I had a follow-up.

You said the percentage I didn't hear you well on the percentage, which percentage are you talking about, please?

Yes. So if we look at your EUR 3.8 billion to EUR 4.7 billion revenue guide, so part of that is going to be cyclical. Part of that's going to be your execution in terms of technology wins. I'm just trying to think is that half off, but some kind of scale of that?

If you look into the business and where we are, one thing I can tell you, I'm really very excited about our technology road map. I think that things are going in our direction. If you look into logic, you see more and more layers coming in with 2-nanometer and also with 1.4 nanometer. ALD intensity is increasing and [indiscernible] intensity is increasing, and we're winning share in that part of the market. If you look into memory, memory is moving more and more into FinFET more and more in 4F², which needs more Epi, needs more ALD. So we're going to have more layers, and we feel very confident about it.

And if you look into logic -- if you look into power, wafer, analog, we are really -- if you look at the power, wafer, analog, the power part of the market is the only part of the power, wafer, analog that's strong right now. And that's being driven by data centers, power devices for data center.

If the wafer part and the analog part goes up, it's going to be even accretive to us. So the service business is also good. In the service business, we're going more and more by automation. And we really -- we're getting some -- getting into even robotics and that customer and leading customer, we're even selling them, okay, some robots to improve the system availability and so on.

If you look into advanced packaging, that's an area that we mentioned that we have entered last year. It's a new area for us. I can tell you that we have so many -- believe me, so many interactions with customers. And we have to prioritize which one to do and some customers tell them, guys, maybe we don't have -- we cannot really help you there. And -- but with the customers that we're really engaging right now, I can see that they really like to work with us as a company because we're looking into the advanced packaging through a different option.

We're looking into that as, okay, what can we do to disrupt the technology? What can we do to provide a solution that's better than what it is right now. How can we -- a solution to make sure that, okay, we reduce the -- to reduce the thermal mass on advanced packages coming with a new material that improve thermal conductivity.

We're working with customers to make sure that we can seal the devices much better. So there is no moisture going in the packages. We're working with customers to actually improve the speed of connection between one chip to the other one, working with them on some innovative photonic layers.

So I'm sure that with all of this really, we feel very confident where things are are going to go from that point of view. And depending where the market is going to go, I'm very confident that we're going to at least match the WFE market growth or actually have a higher growth than WFE. It's a great time for ASM right now. And we -- I see customers really want to work with us. I think our execution has improved. I think our competitiveness is getting even better than before. And what I can tell you, it's the best time to be in semiconductor.

A very impressive runway. Maybe just a quick follow-up for Paul, just some housekeeping, actually. You talked about rising utilization rates, but actually Q1 aftermarket sales were down sequentially. And on your guide, I think last quarter, your guidance range was plus minus 4%. This time, you've widened that range to plus minus 5%, which doesn't maybe sit that well with a sort of improving visibility. Just wondered if there's any color you could share in terms of what you're seeing.

Yes. So actually, the range is, I think, already referred to is related to supply chain challenges. So far, we've been able to manage it. But at the same time, we have to be on top of it to make sure that we get what we need to deliver what we need as per our customer preferred COD customer request date. So that's a little bit where the range comes from, Timm. It's not so much demand. It's more what can we deliver on time given the supply chain constraints that so far manageable again. But yes, we have to be on top of it and nothing can go wrong here.

So that's what was in the aftermarket in Q1 and maybe there's some catch-up in Q2?

I don't know if there's catch-up in Q2. I mean I think had a very good Q1. I think we delivered more or less what we wanted to deliver, and we will target to do the same in Q2.

Thank you, Timm. We still have a number of participants in the queue, but we are running out of time. So let's take one final question. Operator, can we have the last caller?

Final question is from Javier Correonero, Morningstar Equity Research.

In the interest of time, I will just ask one. So your Axus acquisition 3 months ago, it is small, but I think there is a lot to unpack there when you think longer term. So Axus is specialized in silicon carbide processing. So I was wondering if you could explain a little bit more what's the rationale of the acquisition here? Is it like more silicon carbide content as we move into the 800-volt data center? Or is it TSMC potentially adopting silicon carbide interposers in the next few years or both? And of course, it is very early and small acquisitions, but do you have an estimate of what service of addressable market this acquisition could open once it is properly integrated with ASM [indiscernible]?

Okay. So thank you very much for the question. So yes, we have acquired this company called Axus Technology, which is -- we're very excited about the acquisition in CMP. They have a very great CMP technology and very innovative, to be honest with you. And the -- like we mentioned, we have acquired this for the advanced packaging market because advanced packaging is -- needs more and more CMP layer, many, many, many more CMP layer. So there is room for another player. Also, it's a technology that's all about interfaces. And I think that we have some -- we do have some knowledge in interface engineering so that we will be able to really put our print there. It also CMP helps us with our new materials that we're developing for advanced packaging that I just talked about a few minutes ago because I mean, you deposit the film, but also you need to CMP. So we want to understand what's the interaction about the material that we're depositing the new material that we're depositing and the CMP, because CMP also has a slurry.

There's a new -- with a slurry that means we're talking about new chemicals and so on and so forth. So it would help us also develop better materials in ALD, but at the same time, also good polymerization, which is extremely important for advanced packaging. So that's really why we made that acquisition. And then we're working right now on developing the product for advanced packaging, and it's going to increase our SAM absolutely. It is going to increase our SAM and we're in the process of doing R&D and so on in this part of the market.

Okay. That concludes the Q&A. Thank you all for attending our call today, also on behalf of Hichem and Paul. Thank you. Goodbye.

Ladies and gentlemen, thank you for joining. The conference is now over. You may disconnect your telephones.

Good afternoon. This is the Chorus Call conference operator. Welcome, and thank you for joining the ASM International Fourth Quarter 2025 Earnings Call. [Operator Instructions]

At this time, I would like to turn the conference over to Mr. Victor Bareño, Head of Investor Relations. Please go ahead, sir.

Thank you, operator. Good afternoon, and thank you for joining our Q4 earnings call. With me today are our CEO, Hichem M'Saad; and our CFO, Paul Verhagen. ASM issued its fourth quarter 2025 results yesterday at 6:00 p.m. Central European Time. For those of you who have not yet seen the press release, it is available on our website together with our latest investor presentation.

As always, we remind you that today's conference call may contain forward-looking statements in addition to historical information. For more details on the risk factors related to such forward-looking statements, please refer to our press release and our financial reports, all of which are available on our website.

Please also note that during this call, we will refer to profitability metrics, primarily on an adjusted basis. Reconciliations to the reported numbers can be found in the press release and in the investor presentation.

And with that, I'll now turn the call over to our CEO, Hichem M'Saad.

Thank you, Victor, and thanks to everyone for attending our earnings call. I'll start with a few of the highlights. Even with the revenue at a lower level, results in Q4 remained solid, and the quarter marked a reacceleration in demand. For the full year, our sales increased 12% at constant currency, our ninth consecutive year of double-digit growth and operating profit increased by 17%. Strategically, we strengthened our position with key customers with the current generation of gate-all-around going into high-volume manufacturing and with strong traction in R&D engagement for the next nodes.

To power ASM's next phase of growth, we continue to invest in our people, our global footprint and in innovation. I want to thank all our people for their relentless dedication and collaboration, which contributed to another successful year for ASM. As for the agenda today, it's the standard format. Paul will start with a review of our financial results. I will then discuss market trends and provide our outlook followed by the Q&A session.

With that, I'll hand it over to Paul.

Thank you, Hichem, and thanks, everyone, for joining our call. Let's start with the review of the fourth quarter results. Revenue in the fourth quarter of '25 was EUR 698 million as preannounced on January 19. This represents a 7% year-on-year decline at constant currency, but came in above the guidance range of EUR 630 million to EUR 660 million, which we provided with the Q3 results.

Logic/foundry was our largest customer segment in the fourth quarter. Within this segment, advanced logic/foundry accounted for the majority with sales approximately flat compared to the third quarter and somewhat down from a very strong level in Q4 of '24. Mature logic/foundry sales, mostly from the Chinese market dropped sharply as anticipated, both compared to the prior quarter and to Q4 of '24.

Memory sales were relatively steady, both year-on-year and compared to Q3 with solid advanced DRAM sales, offset by lower NAND. The contribution from the power analog wafer segment remained at a fairly low level overall. Our spares and service sales were up 22% year-on-year at constant currency. This represents a very strong performance, especially considering the tough comparison with Q4 '24 when sales grew roughly 50%, driven by accelerated demand in China.

Gross margin of 49.8% in the fourth quarter was down from 51.9% in Q3, but still at a solid level, supported by a favorable mix. SG&A expenses were down 1% year-on-year, while net R&D expenses increased by 6%, largely due to phasing of R&D investments. Operating margin dropped to 25% in Q4, explained by lower revenue and related gross margin, partially offset by lower OpEx. The results from associates increased to EUR 26 million in Q4, which was for a large part, explained by a one-off benefit in ASMPT's results.

Our Q4 net earnings dropped compared to Q3, mainly as that quarter included a noncash reversal gain of EUR 181 million related to the recovery in the market value of ASMPT. Our new orders in the fourth quarter amounted to EUR 803 million, up 19% year-on-year and also better than indicated Q3 results. This was driven by very strong advanced logic/foundry orders. Mature logic/foundry orders from Chinese customers were relatively soft, but showed an acceleration in demand towards the end of the quarter.

Memory orders were steady compared to Q3. Power, analog and wafer orders showed some recovery and reached the highest level in '25, but we're still at a relatively soft level. Please note that starting in 2026, we will discontinue reporting of quarterly bookings. This change reflects the high volatility of quarterly orders, which has been driven more by timing effects than by underlying demand trends.

We will continue to disclose the year-end backlog as part of our Q4 results. In addition, beginning in '26, we will report sales by key customer segments, logic/foundry, memory and other on a half yearly and annual basis. Let's now have a look at the full year results. At EUR 3.2 billion, our sales increased 12% in constant currency to a new record high. In terms of customer segments, logic/foundry accounted for the largest part of equipment sales.

Within this segment, advanced logic/foundry represented the clear majority. Gate-all-around related sales increased very strongly as customers stepped up in investment in 2-nanometer high-volume manufacturing. Mature logic/foundry sales, mostly from the Chinese market also increased but at a more modest pace compared to the leading-edge segments.

In memory, sales dropped to 16% of total equipment sales, down from 25% in '24. Advanced DRAM for HBM-related applications continue to be solid and accounted for the large majority of memory sales. However, this was offset by a normalization of memory-related sales in China. As discussed in previous quarters, memory in China is typically a small market for ASM. But in 2024, it showed an unusual high demand.

The overall drop in memory sales was also explained by lower 3D NAND sales, which were still at a relatively higher level in '24. Power, analog and wafer sales dropped for the second consecutive year. As part of this, silicon carbide sales, which were still resilient in '24, dropped by more than 50% in '25, reflecting the sharp deteriorization in this market. At constant currencies, equipment sales increased 10% in '25, primarily driven by strong double-digit growth in ALD. Spares and service sales grew 18% at constant currency, an excellent performance, which was driven by strong growth in our outcome-based services.

Gross margin for the year increased further, rising from 50.5% in '24 to 50.1% in '25. This improvement was primarily driven by a stronger product and customer mix, including a resilient contribution from the Chinese market. Additionally, the margin benefited from the gradual impact of cost-saving initiatives such as more move to common platforms and ongoing cost optimizations across our manufacturing and supply chain operations.

Gross R&D increased 9% in '25, reflecting the continuous growth in our pipeline of new opportunities. As a percentage of revenue, net R&D expenses increased slightly to 12.5%. Our target remains to keep net R&D in a low double-digit percentage of revenue. SG&A expenses decreased 7% in '25 on the back of disciplined cost control as well as the benefits of earlier investments made to scale the organization for growth. As a percentage of sales, SG&A decreased from 10.6% to 9.2% in '25. For 2026, we project SG&A to show a further decrease as a percentage of sales.

Operating profit increased 17% in '25, thanks to improvements in revenue and gross margin, SG&A discipline and with continued growth in R&D investments. The operating margin increased from 28% to a record 30.2% in '25.

Now turning to the balance sheet. ASM's financial position continued to be in good shape. We ended the year with a cash slightly north of EUR 1 billion and no debt. Excluding M&A-related cash payments totaling EUR 181 million, free cash flow increased 12% to a record of [ EUR 615 million ] in '25. This growth was driven by improved profitability and lower working capital, partially offset by higher CapEx.

Working capital decreased to EUR 347 million at the end of '25. This was mainly due to the phasing of revenue during the year with Q4 2025 sales at a relatively lower level, together with very strong cash collection. CapEx increased from EUR 168 million to EUR 280 million in 2025, fully in line with our guidance range of EUR 200 million to EUR 250 million. This increase is for a large part driven by spending related to the completion of our new Korean facility and ongoing construction of our new facility in Scottsdale.

2026 will be a year of continued investments for a large part related to our Scottsdale site, which remains on track for completion in the first quarter of 2027. Regarding cash spent on acquisitions, in December '25, we acquired Axus Technology, a provider of CMP solutions for EUR 81 million, net of cash acquired, along with a potential earnout up to EUR 30 million tied to performance targets over '26 and '27.

In addition, we paid EUR 100 million in earn-outs as part of the earlier LPE acquisition and as already communicated with the Q3 reporting. In terms of shareholder remuneration, we spent close to EUR 300 million in cash on dividends and share buyback in 2025. And with our Q4 press release, we announced a new share buyback program for an amount of EUR 150 million as well as a proposed dividend of EUR 3.25 per share, up from EUR 3 in the prior year.

And with that, I'll hand over to Hichem.

Thank you, Paul. Let's now review the trends in our markets. In 2025, the semiconductor market continued to be driven by AI, reflected in a wave of new AI data center and infrastructure expansion plans from hyperscalers and other leading industry players. This drove solid capacity investment in leading-edge logic/foundry and in advanced memory, areas where our ALD and Epi technologies play an increasingly central role. At the same time, several other end markets, including smartphones, PCs, automotive and industrial remained relatively soft due to persistent macroeconomic and geopolitical uncertainties.

Looking ahead, the fundamental technology drivers remain firmly intact. Demand continues to rise for faster, more power-efficient semiconductor devices capable of supporting the massive growth in data and compute intensity. This will further accelerate the industry's transition towards more complex 3D device architectures and the introduction of new materials. These trends increase the number of ALD and Epi layers required at future nodes, supporting healthy long-term growth in our key markets.

The main engine behind our growth in 2025 was the continued strong momentum in leading-edge logic/foundry. Our gate-all-around related sales rose substantially as customers ramped 2-nanometer capacity and started to move into volume manufacturing. At our Investor Day, we reconfirmed the significant expansion of our served available market by about $400 million in the transition to first-generation gate-all-around. We also highlighted the increase in our Epi layer share from 22% to 33% and the reinforcement of our leadership position in ALD.

Our product penetrations included new applications such as moly ALD and area selective deposition entering high-volume manufacturing at the 2-nanometer node. In 2026, we expect customers will continue investing in 2-nanometer expansion, supported by rising end market demand across AI, high-performance computing and advanced mobile applications.

Based on public commentary from several of our customers, the 2-nanometer technology node is expected to be large and long-lasting. While 2-nanometer will continue to represent the majority of leading-edge logic/foundry investment in 2026, we have also seen an uptick in 3-nanometer related demand. The pace of innovation is not slowing down, and customers are already advancing toward the 1.4 nanometer node with pilot line investments expected to start in the second half of 2026 and volume production in 2027 and 2028.

This transition is projected to expand our served available market by a further USD 450 million to USD 500 million. A significant driver of this increase is the rising importance of functional layer in the transition area, which is a core strength for ASM. As also highlighted during Investor Day, we expect these transition-related layers to increase to roughly 60% of all ALD layers at the 1.4 nanometer node, up from about 50% at the 2-nanometer node.

Based on the breadth of R&D engagement and the production tool of record selection secured so far, we expect to gain further market share as the industry moves to the 1.4 nanometer node. Let's now talk about memory. Our sales in the Memory segment decreased in 2025. And as discussed, this reflected a normalization in China after an unusually strong 2024.

At the same time, momentum in the advanced segment of HBM-related DRAM remained robust. AI-driven data center investments continue to require high-performance DRAM. And in this segment, ALD high-k metal gate has become essential to achieving the performance and power efficiency levels customers demand. During the year, we strengthened our position with new ALD wins for layers that are expected to ramp in 2026 and 2027.

And we also recorded our first Epi win in the DRAM segment. We expect healthy growth in our DRAM sales in 2026, even though memory is likely to remain a smaller share of our business than logic/foundry in the coming years. Looking forward further out, DRAM scaling presents a significant long-term opportunity. The transition to the 4F2 architecture will require more complex 3D channel structure and additional ALD and Epi steps, expanding our served available market by USD 400 million to USD 450 million.

Let's now look at the power/analog/wafer. In 2025, the power/analog/wafer market remained in a cyclical downturn. In 2026 and based on the early signs of stabilization, we expect for this segment a modest sales improvement. This recovery will be limited to silicon-based power and analog applications. The silicon carbide market will take longer to recover, but we remain well positioned with a strong portfolio, including our PE208 platform for 200-millimeter applications.

China remained an important market in 2025. After 2 years of exceptional growth, we had anticipated a period of normalization. Revenue from China did decline in 2025, but the decrease was milder than expected and mostly supported by continued robust activity in the mature logic/foundry segment. Sales softened in the second half of the year, particularly in Q4, but we saw demand accelerating towards year-end. Based on this momentum, we now expect higher sales in China in 2026, an improvement from our earlier forecast of a double-digit decline.

As Paul already discussed, we continue to invest in R&D and CapEx to capture the opportunities ahead of us in logic/foundry, in DRAM and also in new areas such as advanced packaging. In 2025, we completed our new expanded innovation and manufacturing center in Downtown, Korea. Combined with our key manufacturing sites in Singapore and ongoing efficiency improvement in our supply chain model, we believe we have sufficient capacity in place to support our growth well into the next decade.

I'm also excited to see the progress at our new Scottsdale facility, which will enable substantial expansion of our ALD and Epi product development activity in the coming years. And last but not least, in December, we announced our intention to invest in a new site in the Netherlands, which will house our new global headquarters and a state-of-the-art clean room. In December, we also acquired Axus Technology, a provider of differentiated equipment for chemical mechanical polishing focused on markets such as compound semiconductors and more than more manufacturing.

CMP fits well with our capabilities in chemistry and interface engineering and plays an increasingly critical role in emerging technologies such as 3D integration. In 2025, we made further progress in accelerating sustainability, which remains one of ASM's strategic priorities. We maintained 100% renewable electricity for the second consecutive year. We also deepened collaboration across our value chain, including a new initiative to support suppliers with energy efficiency improvement and renewable energy adoption.

In addition, we continue to advance product sustainability through initiatives that improve the energy efficiency and precursor consumption of our tools, contributing not only to reduction in Scope 3 emission, but also helping our customers lower operating costs.

Let's now look into the outlook for 2026. Let me recap the key points of our guidance as included in yesterday's press release. We expect advanced logic/foundry to be our strongest business in 2026. In memory, we anticipate healthy sales growth. In power/analog, we expect a modest recovery from a low base. And for China, we expect sales to increase in 2026. For the first quarter, we expect revenue to increase to a range of EUR 830 million, plus or minus 4%, with a further increase projected in Q2 compared to Q1. And we anticipate our revenue in the second half to be up from the first half.

With that, we have finished our prepared remarks. Let's now move on to the Q&A.

Thank you, Hichem. We'd like to ask you to please limit your questions to no more than 2 at a time, so that everyone has the opportunity to participate. Operator, we are ready for the first question.

So the first question comes from Tammy Qiu of Berenberg.

So first one is on 1.4 nanometer. So I remember that last time when you said you are expecting to ramp up pilot production in the second half of the year, you were talking about only one customer, but hoping for another one. I'm just wondering what is the progress? Did you get more interest of more customer ramping up 1.4 nanometer than just one?

Okay. Thank you for the question. I think that 1.4 nanometer is a technology node that all key customers are looking for. And right now, I cannot really be more specific on whether 1 or 2 or 3 customers are ramping up the 1.4 nanometer node, but we see interest from all leading suppliers for the high-end logic and foundry to work on 1.4 nanometer. So we are very excited about the opportunity, and we see our customers really continuing to increase their investment in the next-generation gate-all-around. The other thing that we see is that the customers are very serious about going into HBM in the 1.4 nanometer node in 2027 and 2028.

Okay. And the second question is on China. So your peers also talked of China comparing to their expectation from Q3 last year. But your comment from down year-on-year to growth year-on-year. And also, we all know that China has been a lower visibility market comparing to the rest of the market. Is that something you've seen -- significantly changed over the past few months made you to have this call at such an early stage of the year? Or is that just basically customer conversation has been giving you the confidence that it will happen?

I think that -- I think what we have mentioned before, it was very tough for us to really give a very clear projection for China. We mentioned very, very often that it was very tough because of many factors. One of them is the changing trade restrictions and also the funding releases, which really are very unpredictable from our customer.

But we did see that at the end of -- actually at the end of the year that the customers are becoming -- Chinese customers are becoming much more bullish about their business in 2026, which really was very exciting to us, and that's why we feel that 2026 is going to increase with us. For us, we see visibility with them, and we see a strong momentum from that point of view.

The next question is from Andrew Gardiner of Citi.

Sort of related one on China, but really as it pertains to your business mix and how you think that will shape up over the course of 2026. If I go back to how you were framing things in October and indeed at the Capital Markets Day in September, you had cautioned us that business mix as it pertains to gross margin would deteriorate in 2026 and the decline in Chinese revenue was going to be a large part of that.

Today, as you just mentioned, China is going to grow, maybe not quite as quickly as the advanced logic demand, but still it's going to grow. And so your mix isn't going to deteriorate as much as you had previously suggested it might. How should we, therefore, take that into consideration when looking at gross margin for 2026?

So I think for our gross margin, it's -- like you mentioned, it really depends on product and also customer mix. And China would actually help our gross margin. So with the change in the mix, then the gross margin will also depend on that. And we see that in 2026, I think we have made -- I think we have made in 2025 and 2024 significant improvement, not only in -- actually, we made significant improvement in reducing our cost structure in our tools with the commonality.

So we have initiatives to commonize more and more of our products, giving us economies of scale and more leverage with our supplier base to reduce cost. And I think with the way 2026 will materialize, I mean, if the customer mix goes further and further positive on the China-wise, then yes, we expect our gross margin also to be positive from that point of view. But I think overall, I think that we have made improvement in our basic gross margin in the past few years with better cost control. And I think China also would help in gross margin in 2026.

Just so that I'm clear, Hichem, so -- I mean, you're suggesting with less of a change in customer mix year-on-year and with those structural improvements you've made, maybe we shouldn't see much change at all in gross margin relative to last year.

Yes. Let me take that, Andrew. Basically, confirm what Hichem said. So the mix, given that China is now better than what we anticipated before is, of course, a positive plus all the structural measures that we've taken. I think it goes too far to say that it will be the same as this year or maybe even higher at this stage. But I think what is -- what we're confident to say at this point already is that it will be at the higher end of the range that we guided for. And you know that we guided for is 46% to 51%. So at the higher end of that range, I think, is a reasonable assumption to take given what we know today.

The next question is from Didier Scemama of Bank of America.

I've got 2 questions. First, I think if we take the sort of baseline WFE growing 15% to 20% constant currency, are you comfortable to tell us that you will at least be in line with that at constant currency? Or do you see any reason why it should be better?

So I think that -- okay, thank you, Didier, for your question. I think that for our market -- WFE market in 2026, we see that our growth will be at least at the level of the WFE growth. So if the market is going to grow by 20%, WFE, then we grow at least at that level at the 20% level. We are very, really upbeat about our position and our growth this year and actually in the future.

Didier, as you already mentioned in your question, but I want to repeat that because there is quite a difference, of course, between the current rate that we, of course, project going forward and, of course, the average rate of last year. So the question was right at constant currencies.

And sorry, I don't want to use a follow-up for that. But Paul, since you opened the door, can you tell us what your average was in '25 so that we know what the starting point is?

Yes, I think it was around 112.

112. Okay. My follow-up is, I think -- there is a bit of confusion in the market as to what you're actually trying to say when it comes to the outlook for '26 because it feels like you're saying, yes, foundry/logic or advanced foundry/logic is going to be the key growth driver for the business. And then memory is going to have healthy growth. So implicitly that memory will grow less than advanced logic/foundry. So I guess, is that the right way to interpret that? And if that's the case, why wouldn't memory outperform given how bad it was in '25 and given that DRAM WFE looks pretty healthy, at least from a top-down perspective?

Yes. Let me take that question, Didier. So basically, what we said is that we expect growth almost on all fronts, to be honest. But the base from which the growth starts is very, very different. And I think that's the key. So of course, by far, our largest business is logic/foundry. And within logic/foundry, the largest part is advanced logic/foundry and then in China, in particular, mature logic/foundry. So especially advanced logic/foundry, we continue to see strong investments supported by investments, continued investment in 2-nanometer, but also 1.4 pilot, as we mentioned mature, you just heard our comments on China, we expect to grow. So that's a positive.

Memory from a much smaller base, in particular DRAM, we also expect good growth, but it will still be the much smaller part of our overall business, as you've seen also in '25. And in addition to that, although it's still also from a lower base and modest, we also expect this year a modest improvement in power/wafer/analog, excluding silicon carbide. Silicon carbide is still -- that will still take longer, as Hichem already explained, but also power/wafer/analog, which is partially in China, by the way, and partially outside of China.

Okay. So your commentary was based on euro incremental growth as opposed to percentage of growth. Is that correct?

That's correct. I think that's really what we're trying to say here, Didier. Logic is a higher base. But in percentage-wise, the growth in DRAM is higher than -- percentage-wise than logic/foundry. Percentage-wise, DRAM is higher, but we're starting from a very much lower base.

The next question is from Francois Bouvignier of UBS.

My first question is, Hichem on your comment on the AP, you mentioned an AP win on the DRAM side. Can you give more color on what this win is, when it's going to start kicking in? And was it competitive? Just more color on this comment, that would be great.

Okay. Thank you very much for the question. I think that we're really excited of getting win in DRAM, and this is really in HBM, which we have realized in 2025. And it's going to be incremental to our revenue starting this year in 2026.

And is it going to be for multiple customers or just one?

So it's -- as I mentioned, okay, so we had the win in 2025 for this particular customer, but -- and HBM is happening this year. But at the same time, we have significant engagement with other customers in high-bandwidth memory with epitaxy, and we expect some good news also happening hopefully in this year, too.

Great. And maybe on the advanced logic side, I mean, as the pilot line is getting in order this year, I just wanted to check your market share. You mentioned some moly ALD and area of selective deposition design wins as well potentially. Can you maybe give some color on your market share here? I mean, on the pilot line, how do you think it's trading? Is it higher? Is it similar? And selective deposition, I mean, I thought it was something a bit later on. So I was a bit surprised to see selective deposition in your comments. So just is anything happened? Did anything happen on the selective ALD front to accelerate the road map?

Okay. So let me answer your question first about ASD and then talk about molybdenum. So if you look into ASD or area selective deposition, I mean, to be honest with you, we are -- myself, I'm very pleased that we got some win in area selective deposition in the gate-all-around area. And the reason we saw this win is because actually it's higher yield that's experienced by customers. So I think the simplification of the process flow and the reduction in the number of process steps have given rise to improvement in yield.

So we see that happening. And actually, we see that also going to happen more and more into the future, and I see even acceleration in that. A very exciting area, I mean, and there's lots of possibility going on in this realm. And when we talk about the moly, we're also excited about winning HCM capability at the 2-nanometer node. I think we have mentioned very often that the moly adoption in the industry is going to happen, but it's going to be happening in a very slow pace.

We mentioned that it started with 3D NAND. The second logic adoption is happening. And third is going to happen in DRAM. The change in the -- for us, this is penetrate -- this win is the first time that ASM has moved into the metal deposition area. So we're very excited about it. It's an area that we have -- we didn't have any experience about many years. We didn't have any experience because we didn't know what -- how to integrate metal layers. We don't know how to characterize that.

But right now, I feel this win shows that, okay, yes, not only we can develop this new area, but actually, we can achieve HVM capability at the customers. So we're building the expertise within our development team and our teams, and this is something that really is exciting for us. But as I mentioned, the moly adoption in logic is going to take a while. And I mentioned before, it's going to start a little bit at the 2-nanometer. You're going to have a little bit more at 1.5 4-nanometer node, a little bit more at 1.0 nanometer node, but it's not going to be at the level of what tungsten and copper is. But this is really exciting for us, to be honest with you.

The next question is from Adithya Metuku of HSBC.

So I had a couple. Firstly, just on the CMP acquisition, I just wondered, Hichem, if you could give us some color on which end markets you want to focus on, what sort of applications and end markets and whether you intend to have any partnerships with your CMP tool, especially when it comes to advanced packaging. There's a lot of debate there. So any color there will be helpful. And then secondly, I wondered, Paul, if you could give us some color on how you're thinking about your OpEx growth in 2026 with a focus on R&D and SG&A.

Yes. So thank you for the question. We have -- when it comes to CMP, we when we looked into this market, okay, we talked in our Investor Day that we have focused on also some M&A and especially in an area where it can do 2 things for us. It's an area that's strategic to ASM and it's an area where we can add value to it. And also, we mentioned, okay, that when we do some M&A, we really want to make sure that there's some technology component to it. And this is really what we saw in this CMP acquisition. The company has a very good technology, very exciting technology to be honest with you. That's very unique by itself.

They have a great footprint and also very good cost of ownership, very competitive cost of ownership. And we think that this technology, CMP is actually -- is complementary to our strength in interface engineering. As you know, that when you do the bonding at the end of the day, it's really to put -- it's like to bond interfaces together. So after you do polishing, you have a new surface. And if you can make the surface better and more -- you can actually control the engineering of the surface, that can help you also do the bonding. And also, it's also complementary to our deposition technology. I mean we have CVD deposition in advanced packaging and CMP would be also complementary to that.

So from all this point of view, we think that we can add value to this technology. We are going to see how it works for us in the future. This is a very small acquisition, about EUR 80 million acquisition. And we're going to test it. We mentioned that we're going to test it in the advanced packaging area, and we'll see how it materialized. But definitely, the technology is differentiated. We like the tool architecture. And we think that also cost-wise, it will be very competitive.

Yes. And then on the OpEx, what I can say is that let's start with R&D. We will continue to invest in R&D. So that will grow further. And our net R&D will grow at a higher pace than our gross R&D simply because of the increased amortization expense because more and more films that we have been working on in the last few years have entered into HCM. So we start to amortize those. But I think net R&D at constant currency again, because also I'm not going to give you a percentage now, but a decent chunk of R&D cost is also dollar-based and another decent chunk is also euro-based.

So net, around 10%, I think, is a good guidance to take into account and gross will be slightly below that because net will grow faster than gross. On the SG&A, we will continue to be very strict on SG&A. There will be some increases, of course, annual inflation, merit, but also we invested in -- as you know, we are the global big bang of our new ERP system that costs under IFRS need to be capitalized. So we will start amortization of these costs as well. We might have some higher variable expenses in '26. But overall, I think if you take into account a few percent increase, that's what we try to manage. We will be very strict on SG&A. And as a percentage of revenue, for sure, that will go below 9%.

Got it. Understood. Maybe, Hichem, just to clarify on the Axus acquisition. Is this -- do you have a specific focus on hybrid bonding? Or is it more generally bonding applications?

Yes. I think we -- really we're looking into the advanced packaging as not only hybrid bonding. So it's really packaging as overall market from that point of view.

The next question is from Robert Sanders of Deutsche Bank.

Maybe just if you could just discuss a bit about what you see in terms of clean room constraints at your customers. Obviously, we've seen some companies talk about that. Obviously, your tools don't take up as much footprint. And -- but in terms of looking into '27, do you see significant capacity opening up? And how does that set you up for next year?

I think that we see -- if you look into 2026, -- we do see there's a constraint in fab space in multiple areas, which limits the really expansion for our customer. I mean it's -- I think it's very clear. It's very public information. And also, we see a growing sense of urgency from our customers really to get some of the tools. So based on that, we see that there is a good momentum as more and more capacity come in, in 2027. So we see momentum really continuing in 2027. So not only we are excited about 2026 and very positive about 2026, but we see the fact that more and more capacity -- the fact that there's a constraint right now in fab space, but that fab space is going to open up in 2027 means that also 2027 is going to be a very good year for our company.

Great. And just a quick follow-up. Just a clarification on foundry/logic in '26, are you saying that's going to grow? Or are you just saying it's going to remain the largest part? I didn't quite understand. Maybe I missed that if that was asked already.

Both, yes and yes, it will remain the largest part, and it will also significantly grow.

The next question is from Stephane Houri of ODDO BHF...

I just wanted to come back on the decision to stop giving the orders on a quarterly basis because, yes, indeed, some of your -- other players in the industry have done the same. But it seems to me that the volatility was not such a high volatility as for others and that given your lead time, it was a good indicator. So what are you going to give apart from more granularity on the current level of sales? And I have a follow-up.

Yes. Let me take that question, Stephane. Actually, we -- multiple, let's say, stakeholders that we have discussions with gave suggestions that we should maybe like our peers, stop with quarterly bookings because the risk of an overreaction is there. As we said already, it's not always demand driven, but it's timing, just phasing, nothing else. So we see overreactions up if it's really good or overreaction down, if it's really not good, which is not helping, of course, the stock increases volatility. So that's a key reason why we're stopping.

On the other hand, I think with the hopefully improved transparency on segment information, that might help. We will continue with revenue guidance, of course, like we do today and maybe some qualitative guidance if we believe that is necessary. So with that, I hope that you guys have enough to model and to come to a view on how we will develop in the coming period. But these have been the considerations. And based on that, yes, we have made this call.

Okay. And what about the evolution of spares and services it has been outgrowing the equipment parts in 2025. So what is your view on 2026?

I think on 2026, I think with the fact that the fabs are at maximum capacity, we expect continued growth in 2026. The other thing, as the market moves more and more into more advanced nodes, we see the service part of our market also growing even higher than the rest because of the complexity of the equipment at the very high-end node, which favor outcome-based services or solutions that are much more valuable and add more value to the customer. So I'm very optimistic also on the service market this year and in the future.

The next question is from Jakob Bluestone of BNP Paribas.

Just on Axus, could you maybe just help us understand how you plan to cross-sell CMP tools? And if you have any idea what share of your customers are currently already using CMP in their processes?

I think that for -- I just want to make very clear for Axus as a company, I mean, it has a very low revenue. I mean we're talking about revenue between USD 20 million to USD 30 million per year. So this is the latest revenue they have in 2025. So this is a very small acquisition from this point of view. And we're going to leverage the -- our expertise to help this technology, bring this technology to a larger customer from that point of view. But as I mentioned, okay, this is all about a technology buy where we think that, okay, we can add some of our strengths into the CMP market.

I mean if you look into CMP, it's systems for chemical mechanical polishing and that chemical part that really means chemistry. And that's where we play with. I think we have some idea on how to make the chemistry better, especially when you go into 3D integration whereby the chemical part of the CMP becomes most predominant than the mechanical part, especially as you go to 3D and the structure becomes more and more fragile from that point of view.

So we think we can play a role there. We're going to see how it goes. But again, this is something that we think that we can improve. It's a very small acquisition. I think it plays on our street and going into a market that's growing a lot, which is the advanced packaging. So we are very excited to look into how can we make it even better and improve our penetration into the advanced packaging in the future.

Great. Maybe just a quick follow-up as well just on CapEx. Can you provide any commentary on CapEx for '26, I guess, particularly in light of the expansion in Almere?

Yes. No. The expenses as per our guidance from the Investor Day, I think it was EUR 200 million to EUR 250 million in the years where infrastructure expansion. So in '26 this year, it will mainly be CapEx related to Scottsdale still. And then very likely, as we see it today, then in '27, you will start to see the first more material CapEx for Almere.

The next question is from Sandeep Deshpande of JPMorgan.

My question is back to the M&A you've done. I mean has the policy of ASM changed at all with regards to M&A? Those of us who have covered the company for a long time, I mean this -- the company did a lot of M&A, then made a lot of exits. Now you've started doing M&A in a small way again. So has the overall policy towards M&A changed at ASM? And are there more areas apart from now the CMP acquisition that you plan to do? And does the company plan to become stand-alone players in this? Or is this just addition to existing tools, which is probably a less risky proposition. And so I just want to try to understand your thought process behind the M&A.

Yes. No. So did it change? At least in the last 5 years, it did not, although indeed, we made 3 acquisitions in the last 5 years. And in the 10 years before that, we made none. So from that point of view, you could maybe think there is a change. But I think there's not really a change because also before that, what I understand from my predecessor, they've looked at certain opportunities, but for whatever reason, they never materialized.

So we look into M&A if we see an opportunity where we see clear value-creating opportunities and that helps us to grow and build our position further in certain areas that we have labeled as important/strategic to us, then we want to act. And we did that now 3 times. There's always a very clear link to strength that we have. It leverages, let's say, our strength of the capabilities of the company that we buy. It can build and leverage on our global network that we have. So yes, the logic at least for the last 3 have been actually exactly the same for Reno, for LP and now for CMP, Sandeep.

And we will continue to scan the market. We have continuously said that. Our first priority in terms of capital allocation is growth. Number one is organic growth. That's why we continue to invest in R&D, very important and in infrastructure expansion, as we explained. But the second is also inorganic. If we see true value-creating opportunities, we try to do it in a very disciplined manner. We're not throwing money away because we have it, no. We only do it if we truly believe that there is a medium- to longer-term strategic play that can create a lot of value to us based on the capabilities that we have in combination with the targets. So that did not really change as far as I'm concerned, Sandeep.

And a quick follow-up. I mean, I think a quick follow-up. I mean, in terms of the earlier question on your improvement being seen in the logic/foundry market. Earlier last year, you had talked about a slow start to '26. So did something change in the last few months in terms of the slow start that some key customers changed how the trajectory of how they're taking delivery of the tools? Or was this slow start is what you have already guided? This is the guidance and it was underplaying what the market expected. The market was underplaying what you expected, sorry?

No, I think, no, absolutely it changed in the last, whatever, 2 to 3 months. You've seen announcements from some of our customers that have significantly increased their outlook, especially a large foundry customer, which I think that's where it started with. We just explained the improved sentiment in China in combination with a pause of some of the export control measures that were initially put in place, but then paused. Some customers will take advantage of that. But at the same time, also clearly improved sentiment there.

You read about the hyperscalers and their investments in data centers and infrastructure, hundreds and hundreds of billions. It's definitely a different situation in the last 2, whatever, maybe 3 months than what we thought before. We always thought '26 would be still a good year, but starting in the slow, as you said, and then accelerating more towards whatever the second half of the year. But that acceleration that we actually had expected maybe somewhere in the course of the year, literally starts now. So there's clearly a change, yes.

The next question is from Timm Schulze-Melander of Rothschild & Company Redburn.

I had 2, please, one for Hichem and one for Paul. The first one is just on the CMP business model just with respect to consumables, slurry and pads. I know it's a small business, but is that going to be something that you provide? Or is that going to be provided by an external or a third party? And then I had a follow-up.

Okay. So to answer your question, okay, regarding the CMP part of the business and the acquisition. So the -- once -- as the technology in packaging moves more and more into high end, from -- it's going to move from TCB to hybrid bonding in the future. Then what happened is that we're going to go to lower temperature processing and the surface of the interface becomes a very significant in the hybrid bonding part of the advanced packaging.

So for such, interface control is very important. We have solutions, organic solutions from our ALD know-how to engineer interfaces and engineered surfaces. But also CMP is part of that whole the whole process flow. And by definition, CMP also affects the surface of the deposition layers that deposited film. So it's important for us also to understand how that interface from CMP works with our deposition films that we developed in CVD and ALD to engineer a very clear interface.

So I hope that's very clear from where we stand. This is a new market for us. This is a new market, and we try to understand this market very well. We have -- as we mentioned, we have organic offering there. This organic offering are in ALD in the area of ALD. This organic offering are also in the area of CVD like PECVD, but also this offering, the organic offering is also in the area of epitaxy and silicon photonics, where we also have some traction in those things.

So CMP plays a significant role in engineering the interface. It's complementary to our deposition technology. And it's very important for ASM to really know how CMP also engineers the surface and interface in addition to the offering that we have in deposition, both CVD and ALD. The next thing regarding the question that you have asked about slurry and so on and so forth. As I mentioned, the CMP part is moving more and more into the chemical part. So you have CMP, you're trying to polish. So polishing both with force, okay? That's the mechanical part, but also the chemistry, which is the slurry and so on and so forth.

And that slurry thing or the chemical part is becoming much more important than the mechanical part because of the 3D drive that's happening in our device. And when you talk about advanced packaging, you're going to put things on top of each other. And you also have wafers that are very thin, they are very brittle. So you cannot put too much force. So the chemical part becomes much more important. We are a company that knows a lot about chemistry, and we have know-how and knowledge in that, which we have applied for ALD and other parts. And we think we can do the same for the CMP part of the business.

Great. That's very clear. Just moving on to Paul. sincerely appreciate the improved disclosures. For one, I'd probably request for a quarterly rather than a semiannual, but the disclosure improvement is much appreciated. I just wanted to ask about the cost saves and the run rate and just kind of get a sense as to kind of what the exit rates were or are for '25 coming into '26. And maybe just trying to think about the cost savings contribution and how that might scale or how that sizes relative to the increase in R&D, which I think you're guiding is going to, on a gross level, rise by about EUR 40 million, EUR 45 million. I just wanted to get a sense of maybe the extent to which cost saves might offset how much of that they might be offsetting.

Are you specifically referring to SG&A and R&D or also to cost of goods?

I'm referring to the broad A to Z cost savings and efficiency programs that you guys have across the company and just trying to scale those relative to the specific cost increase that you're guiding for in the gross R&D spend.

Okay. So on the real cost savings, it's more, let's say, margin related where we have explained before on the -- for instance, the standardized platforms. So we have more and more products now that are qualified by customers based on standardized platforms, which have a better cost structure, lower cost structure, more common parts, et cetera, which leads to cost reduction, but also to a reduction of complexity in terms of logistics will lead to somewhat improved inventory simply because of more commonality.

I'm not going to give you a number there, but it's -- yes, it's a meaningful improvement, let me say it like that. The other part, but also that will go slow. So every year, you will see some benefit there is the MIT that we talked about before, the merchant transit, where we don't have everything come to Singapore first, assemble it, test it, but have the platform go straight to the customer, the process chamber that comes from Singapore then straight to the customer and assemble it there and test it there, which skips one big step, which is also a big improvement.

Of course, we have value engineering initiatives on our products continuously. We have material cost savings, commercial savings. So there's across the board savings going on. On the R&D part, here, I mean, the name of the game is selecting the right priority from the many priorities and many opportunities that we see, which, of course, to a large extent, are based on the, the overall market opportunity that we see and whether or not we can have a differentiated proposition or not, but it's not so much about saving costs, although we try to be very efficient, of course, in what we do there.

For SG&A, it's literally doing more with less. So we grow and grow, but we want to automate more. We want to make our processes better. We want to leverage AI better. So there instead of just adding people more and more, it's all about doing more with less and do maybe more with the same, to be honest, to support this growth without adding too much cost. So every line has a different dynamic, if you wish.

The final question is from Marc Hesselink of ING.

First its a follow-up on the market share in memory. I think now you very clearly stated that whatever the industry of WFE is doing, you expect to be growing faster. So does that also imply that in the more advanced parts of the memory market, your market share is getting closer to the market share that you have in the advanced logic/foundry.

Let me take this, Marc. No, absolutely not. I wish because then we would be looking at very different numbers. No, no. So what I think we try to say is that on the small revenue base that we have today in memory, especially in DRAM, we expect significant growth -- significant growth even more as a percentage than in advanced logic/foundry.

Having said that, for '26, we also expect significant growth, in particular, in advanced logic/foundry, maybe a little bit less maybe as a percentage than DRAM, but still very high. If that growth would, let's say, change significantly during the year, the percentage so that advanced logic foundries, we don't expect that this scenario would grow much lower and DRAM would suddenly go even more and stronger than we are today, then the statement that we make is maybe -- would become maybe invalid because if all the growth would be in DRAM, of course, we would not be saying what we were saying. But also, we say what we say because we believe that also the growth in advanced logic/foundry will be very significant. That is the reason why we say what we say.

Okay. That's clear. And then my second question is on -- it's more of an organizational question. So you're adding R&D capacity. You're adding -- you have introduced a new IT system. You have now a new Board member. With the growth of the company, you're adding this kind of, let's call it, another layer of professionalization within the company. Is that the way to look at it? Is there more to come? And is there more that you have to scale in the coming years given the high growth that you have going forward and what you had in the past?

Maybe one small adjustment, Marc, we did not add a new Board member. It's a new ExCo member. So senior leadership, but it's not a Management Board member. The carrier that we talked about is an ExCo member. I think what we're doing is we have -- actually, this is something that's happening for many, many years in a row. Of course, we're trying to professionalize the company and trying to get ready to scale the company in an efficient and effective manner.

So for instance, the new ERP system, which we have globally implemented through a big bang is, let's say, the foundation for further, let's say, growth in a more automated fashion, in a more productive way than we would have been able to do without this because with the previous system, we had to do much more manual work than what we can do today as an example. Also AI applications, we can leverage better with the foundation that we put in place now than what we would have been able to do without this foundation. So yes, it's all about scaling. Yes, it's about professionalization of the organization. So I'm not sure if that answers your question, but this is what I can say.

The question was also a bit, is there more to do on this side? I mean, I think we had quite some announcements over the past few years. Is that -- are the large part behind it now? Or are you still taking another steps here?

Yes. I'm not sure you're referring to because so many announcements I don't think we've had. I mean, we have an Expo. We have a number of KPIs. We professionalize, I mean, our way of working, which I guess every company does. So we will continue to do that. We will not stop. It's not like, okay, from now on, you will not see any announcement anymore. But yes, I think what we do is for the reasons that I just tried to explain is to scale the company in a controlled, professional and highly productive manner. That's what we are trying to do.

That was the final question. I will turn it back over to the CEO for any closing remarks.

On behalf of Paul and Victor, I would like to thank everyone for attending today's call. We hope to meet many of you guys very soon in the upcoming investor events. Thanks again, and goodbye.

Ladies and gentlemen, thank you for joining. The conference is now over, and you may disconnect your telephones.

Good afternoon. This is the Chorus Call conference operator. Welcome, and thank you for joining the ASM International Third Quarter 2025 Earnings Call. [Operator Instructions] At this time, I would like to turn the conference over to Mr. Victor Bareño, Head of Investor Relations. Please go ahead, sir.

Thank you, operator. Good afternoon, and welcome, everyone, to our 2025 Q3 earnings call. I'm joined here today by our CEO, Hichem M'Saad; and our CFO, Paul Verhagen. ASM issued its third quarter 2025 results yesterday at 6:00 p.m. Central European Time. The press release is available on our website.

With our latest investor presentation, we remind you, as always, that this earnings call may contain information related to ASM's future business and results in addition to historical information. For more information on the risk factors related to such forward-looking statements, please refer to our company's press releases and financial statements, which are available on our website.

Please note that the profitability measures mentioned in this call today will be primarily based on adjusted non-IFRS figures. For the reported results as well as the reconciliation between reported and adjusted results, please refer to the quarterly results press release.

And with that, I'll now hand the call over to Hichem M'Saad, CEO of ASM.

Thank you, Victor, and thanks to everyone for attending our third quarter 2025 conference call. First off, I'd like to thank all our investors and stakeholders who joined us for Investor Day last month. It was great to see so many of you. For today's call, we'll be following the usual agenda. Paul will begin with an overview of our second quarter financial results. Next, I'll discuss the market trends and outlook, followed by the Q&A session.

I will now turn it over to you, Paul.

Thanks, Hichem, and thanks, everybody, for joining our call today. Let me start with revenue. The revenue in the third quarter of '25 amounted to EUR 800 million, up 8% year-on-year at constant currency. And compared to the second quarter, sales were flat at constant currency. This was at the high end of a guided range of flat to down 5%. Equipment sales increased 10% year-on-year at constant currency and were led by ALD followed by Epi.

Spares and service sales were up 2% at constant currency. Year-on-year growth in spares and services was lower than in the past few quarters, and this is explained by the accelerated above trend demand in China in the second half of last year. Growth in our outcome-based services continues to be healthy.

In terms of customer segments, revenue was led by logic/foundry, followed by memory and then power/analog/wafer. Logic/foundry continued to account for the majority of sales. Advanced logic/foundry sales for the largest part 2-nanometer related were up substantially compared to the third quarter of last year and approximately similar to Q2. Mature logic/foundry sales mostly in China, were up year-on-year and down from the second quarter.

Memory sales decreased compared to Q3 of last year and were roughly similar to Q2. of this year. ALD sales for advanced HBM-related DRAM applications represent a larger part of our memory sales. The year-on-year decrease was mainly explained by some lumpy and relatively high sales and orders from memory customers in China in Q2 and Q3 of last year, as we also discussed in previous quarters.

Sales in the power/analog/wafer segments were up slightly but still at relatively low levels, reflecting the continued downturn in these markets. Gross margin in the third quarter remained strong at 51.9%, roughly similar to Q2 and up from 49.4% in the third quarter of last year and again supported by a positive mix. As mentioned in the press release, we expect a less stable mix in the fourth quarter, which should bring the gross margin to around 51% for the full year.

One of the mix factors was revenue from China, which decreased both year-on-year and compared to Q2 but still represented a solid level in Q3. We still expect China sales in the second half to be lower than in the first half with a more substantial drop in Q4. As discussed in the last quarter, our forecast for the gross margin excludes the impact from potential new U.S. tariffs. Our industry is currently still exempted, but it remains unclear what any new tariffs will be. We have several contingency scenarios in place to help mitigate potential direct impacts including the option of expanding localized manufacturing in the U.S.

SG&A expenses were 10% lower year-on-year. This reflects lower variable spend and our continued cost focus. For the full year, SG&A is still expected to be somewhat below prior year. Gross R&D was up 10% in Q3, reflecting ongoing increases in our R&D programs. This increase plus the inclusion of a EUR 4 million impairment was partially offset by a relatively higher increase in capitalized development expenses, leading to an increase of 3% in net R&D expense. At 30.9%, the operating margin continues to be strong, supported by the solid gross margin and the year-on-year decrease in SG&A.

Below the operating line, financial results included a currency translation gain of EUR 11 million, and this compares to a currency loss of EUR 60 million in the second quarter and a loss of EUR 48 million in the third quarter of last year. As a reminder, we hold the largest part of our cash in U.S. dollars.

Let's quickly switch to ASMPT. Our share in income from investments, reflecting our stake of approximately 25% in ASMPT amounted to a loss of EUR 7 million in the third quarter, which is explained by one-off restructuring costs taken by ASMPT in the quarter. Our net results in Q3 also includes an impairment reversal of EUR 181 million, driven by a recovery in the market valuation, our stake in ASMPT in the quarter. With that, the impairment charge of EUR 250 million that was recognized in Q1 of this year has now been fully reserved.

Let's go back to ASM now. Order intake. Our new orders amounted to EUR 637 million, a decrease of 7% compared to the second quarter and a decrease of 7% compared to the third quarter of last year at constant currency. With the Q2 results, we already indicated that book-to-bill would be below one in Q3. As an expected rebound in advanced logic/foundry orders will be offset by a sharp drop in China orders following a very strong first half of the year.

In our update last month, we indicated advanced logic/foundry orders would still be up, but not as strongly as previously expected due to very mixed trends per customer and also that power/analog/wafer orders came in somewhat lower than expected.

Looking at the breakdown by customer segments, logic/foundry was the larger segments, followed by memory and then power/analog/wafer. Logic/foundry orders decreased slightly year-on-year and compared to Q2. As just mentioned, a solid increase in advanced logic/foundry compared to Q2 was offset by a sharp drop in mature logic/foundry orders mostly in China.

Memory orders dropped compared to last year and were relatively steady compared to Q2. And the largest part of memory orders was for advanced DRAM applications.

Let's turn now to the balance sheet. ASM's financial position remains in good shape. We ended the quarter with EUR 1.1 billion in cash, up from EUR 1 billion at the end of June. Days of working capital dropped to 37 days at the end of September, down from 43 days end of June. This level is relatively low and also below the longer-term target range. As previously explained, this is due to a relatively high level of contract liabilities for a large part in China, which is expected to gradually normalize over time.

CapEx amounted to EUR 38 million in the third quarter due to phasing of investments for our new Arizona facility, CapEx in the fourth quarter will be higher. And for the full year, we still expect CapEx of EUR 200 million to EUR 250 million. During the quarter, we also paid EUR 100 million in earn-outs related to the acquisition of LP in 2022. In total, free cash flow amounted to EUR 139 million in Q3. Excluding the earn-outs, the free cash flow amounts to a stronger level of EUR 239 million in Q3 and EUR 628 million in the first 9 months.

During the third quarter, we spent EUR 109 million on share buybacks as part of our EUR 150 million program that was completed on July 25. And lastly, let me recap the 2030 financial targets we shared during the last month's Investor Day. Growth prospects for ASM remains strong on the back of rising ALD and Epi intensity in the logic/foundry and DRAM markets and new opportunities in, for instance, PECVD and advanced packaging and continued double-digit growth in spares and services.

We expect the revenue to grow to more than EUR 5.7 billion by 2030, and this represents a CAGR of at least 12% from 2024 through 2030, twice the rate expected for the wafer fab equipment market. We raised our gross margin target to a range of 47% to 51%. And as part of this, we discussed a number of initiatives that will drive efficiency and productivity improvements.

One of the key initiatives I'd like to highlight again is the successful launch of a new ERP and PLM systems. We went live 3 months ago and the transition was executed smoothly without any disruption to operations. This milestone lays a solid foundation for future efficiency improvements, including the rollout of real-time analytics and other digital transformation efforts. We will remain disciplined regarding operating expenses.

Combined with the operating leverage effects, we expect SG&A to drop to less than 7% of revenue by 2030. We intend to increase R&D investments. This is our lifeline as the opportunities continue expanding in the next nodes. The target is to keep net R&D in a low double-digit percentage range of revenue. This will all lead to solid operating margin of 28% to 32% in the coming years and from 2030 onwards of more than 30%.

In terms of CapEx, we expect an annual level of EUR 150 million to EUR 200 million during years of infrastructure expansion and EUR 100 million to EUR 200 million in years without such an investment. Combined with improving profitability, we project free cash flow to increase to more than EUR 1 billion by 2030.

During Investor Day, we also reiterated our capital allocation policy. #1 priority remains investing in the growth of our company. That includes R&D and infrastructure investments and also M&A in case of attractive opportunities. In addition, a strong financial position remains important with at least EUR 800 million in cash as we remain committed to our dividend policy and to return excess cash in the form of share buybacks. And with that, I'll turn the call back over to Hichem.

Thank you, Paul. Let's now continue with a review of the market and business trends. Starting with the end market conditions. The overall picture continued to be mixed, similar to the previous quarters. In various parts of the semiconductor market, the recovery continues to be held back by uncertainties around the economic outlook and geopolitics. It's clear that AI remains the bright spot across multiple sectors and markets, adoption of AI is being accelerated as a key driver of innovation and productivity gains. The surging demand has been highlighted by recent strategic partnerships and announcements from industry leaders aimed at expanding AI data center capacity.

In terms of wafer fab equipment, the growth in AI is expected to drive significant and structural growth in the advanced logic/foundry and DRAM markets. These strengths play to ASM's strength. If we first look at our advanced logic/foundry business, overall demand continues to be healthy even though trends by customer has been very mixed. As already mentioned, these mix trends had some impact on Q3 bookings and will also impact sales in the second half. For the full year, we still expect a very strong increase in our gate-all-around related sales.

The 2-nanometer transition continues to be a strong driver for our company. In the Investor Day last month, we reconfirmed the $400 million SAM increase in the move from FinFET to first-generation gate-all-around. We also reconfirmed that in this transition, we at least maintain our ALD market share and expanded our share of Epi layer accounts from 22% to 33%.

Our customers continue to report strong demand for 2-nanometer for both AI and smartphones. We expect this to support continued investment in 2-nanometer capacity expansions in 2026, including new sub nodes, such as the backside power distribution. At the same time, customers continue to progress steadily in the development of the upcoming 1.4-nanometer node. In the second half of 2026, we expect the first 1.4-nanometer pilot line investments, followed by the start of volume production in 2027 and 2028.

As also shared in our Investor Day, we expect a SAM increase of $450 million to $500 million in the 1.4-nanometer transition. Based on the intensity and breadth of our R&D engagement, we expect again to at least maintain our market share in the transitions to the next 1.4-nanometer node. We expect new ALD layers in [ backside ] power in NIMCAP and metal ALD layers in the middle-end-of-line.

However, the biggest area of increasing ALD intensity continues to be in the transistor area, the front end of line. This is the heart of the chip where the overall device performance is defined by functional materials such as the high-k and electric dipole layers for multi-DC and work function layers. As a percentage of total ALD layers, we expect the number of layers in front end of line to increase from 50% in the 2-nanometer node to 60% in the 1.4 nanometer node. This is an area where our company holds strong market share position.

Let's now review the memory business. High-bandwidth memory, HDM-related DRAM continues to be the main driver. Fueled by strong demand for AI data centers, customers are expanding manufacturing capacity for the most advanced DRAM devices for HBM applications. These devices require ALD High-K Metal Gate technology in which ASM has a leading position.

Looking at the year-on-year performance, despite the good momentum in high-end DRAM, it's important to note that our memory sales last year included elevated sales from Chinese customers, which are not repeated this year. As a result, we still expect our memory sales to be lower than last year at less than 20% of overall equipment. The outlook for CD NAND, which is the smaller part of our memory business, has been improving somewhat, and we continue to be well placed with our ALD death field solutions with key customers.

We expect DRAM investment to further increase in 2026. In the next couple of years, we expect a further gradual increase in the number of ALD layers in advanced DRAM. In the press release, we highlighted new wins in Epi and ALD dipole and work function-related layers in DRAM HBM for most expected to ramp in 2026 and 2027.

Starting in 2028, we anticipate a significant increase in our SAM in DRAM driven by 2 major technology transitions. The move to 4Fsquare architecture and the adoption of FinFET in the periphery. In 4Fsquare, the channel structure becomes vertical, and producing a more complex 3D architecture. This shift will require additional ALD layers for gap-fill, oxides and metals and we'll also increase the role of Epi as an enabling technology. Shortly thereafter, the transition from planar to FinFET in the periphery will further increase demand for logic like ALD and Epi layers.

At our Investor Day, we quantified the SAM expansion in DRAM at $400 million to $450 million as a result of this multi-node transition. In addition, this presents a compelling opportunity for ASM to grow our ALD shares in the DRAM market and to accelerate the expansion of our memory business.

Next, the power/analog/wafer segment continued to experience weak market conditions. While there were early signs of end market recovery at the end of Q2, it became evident over the past 3 months that investment level in this segment will not rebound in the second half of the year. Assuming no adverse economic development, we expect spending in these markets, starting from a low base to gradually improve over the course of 2026.

Thanks to innovative products that we launched in recent years, we are well positioned to benefit from this recovery. One example is our Epi Intrepid ESA tool which has helped us secure several new customers and position in 300-millimeter power and wafer applications. It's important to note that our outlook for gradual recovery excludes the silicon carbide market, where market conditions remain more challenging.

Looking at China. Revenue was still at a solid level in Q3, but bookings dropped significantly. And as Paul already mentioned, this was the main reason for the sequential drop in our overall bookings. China bookings were still strong and very much concentrated in the first half of the year. On top of this, we incurred some additional impact from the export restrictions that were announced earlier this month. The impact on a total annualized sales is expected to be around 1% to 2% negative. With a stronger drop in Q4 sales, we project sales from China to be lower in the second half compared to the first half.

Equipment sales from China will also be lower in the full year of 2025 and expected to account for approximately 30% as a percentage of total ASM revenue. Looking forward, we expect a gradual normalization in China demand in 2026 and subsequent years, in line with our previous view. This follows on a number of years of very strong spending, particularly in the mature logic/foundry segment. For 2026, the contribution from China is projected to remain meaningful, although sales are to decline by double digits.

Before moving to the guidance, I'd like to repeat a few more of the takeaways and strategic priorities we shared in our Investor Day. ALD and Epi remains key growth markets for our company. We project a CAGR of 9% to 13% for both markets in the period of 2024 through 2030, which is clearly ahead of the 6% growth expected for the WFE market. This growth is driven by increasing complexity and increasing ALD and Epi layers to address challenges related to more 3D structures and new materials in these nodes, both in logic/foundry and DRAM.

Advanced packaging is emerging as a key midterm growth driver for ASM, with the market projected to grow at an attractive CAGR of 15% through 2030. Although it currently represents a smaller portion of our business, upcoming generations of advanced packaging featuring finer pitches will demand more sophisticated solutions were aligned with our strength in chemistry, innovation and surface preparations.

We've recently secured new ALD wins for TSV liner applications, and we are currently pursuing initiatives aimed at doubling our served available market by 2030. At the Investor Day, we also introduced growth targets for our spares and services business. For the period of 2024 to 2030, we expect a continued strong CAGR of more than 12%. The main engine of this growth is our outcome-based services, which we target to account for more than 50% of our sales and service sales by 2030.

These innovative services deliver guaranteed outcomes to our customers, such as improved tool performance and availability. One example is our new dry cleaning solutions for refurbishing critical tool parts. Compared to conventional cleaning technologies, this approach improves defectivity performance and extended parts lifetimes, thereby reducing costs for our customers. It also contributes significantly to sustainability with a 67% reduction in CO2 emissions versus traditional wet cleaning methods.

Another example is the use of automation in services by developing robots to place replacement parts in reactors, we achieved far greater precision than manual placements. This enables customers to operate our tools with action level precision, essentials as geometries shrink and nodes become more complex. Last but not least, we remain focused on driving operational excellence, maintaining a flexible footprint and as Paul also emphasized delivering improved financial performance.

Let's now have a look at the guidance as outlined in our press release. For Q4 2025, we expect revenue to be in the range of EUR 630 million to EUR 660 million. For the full year 2025, we continue to expect revenue growth at close to 10% at constant currencies. Despite the projected slow start in 2026, we expect ASM revenues to grow in 2026.

In terms of orders, we expect the trend to bottom out in Q4 at a slightly higher level than Q3. And looking at next year, we project quarterly orders to pick up again as 2026 progresses.

With that, we have finished our introduction. Let's now move on to the Q&A.

We'd like to ask you to please limit your questions to not more than 2 at a time so that as many participants as possible have a chance to ask a question. Operator, we are ready for the first question.

Thank you. This is the Chorus Call conference operator. [Operator Instructions] The first question is from Didier Scemama, Bank of America.

Maybe a first question for Hichem. Can you maybe give us a little bit more color as to what's already in your backlog? Because we've seen, of course, over the course of the last few weeks, a significant improvement in the picture for AI CapEx or logic chips, but also HBM, but also commodity DRAM. We've seen, of course, one of your customers this morning talking about a substantial increase in CapEx next year. So is that already in your backlog? Or is that yet to come and sort of give us confidence that your bookings have to materially improve from here? I've got a follow-up.

Yes, Didier, I'm going to have Paul answer your -- this question, okay?

Yes, Didier, thanks for the question. You've seen our backlog, which came down further on the back of book-to-bill below of one, to be precise 0.8. So what's in the backlog, as you know, we have a relatively short-term backlog, if you compare it to let's say, one of our companies here in the southern part of the Netherlands, 3 to 6 months, typically. We also said that we expect a relative soft start of 2026. So not everything that you are referring to is already in our backlog. Of course, some elements are -- I'm not going in detail what is precisely in or out. But given the relative low bookings that we have, which we also guided for after the -- or at the Investor Day, you can imagine that not all of that is in the backlog at this stage.

Okay. Got it. And I think in the last 12 months or so, commodity DRAM and commodity NAND, which historically are reasonably small part of your revenues, have been very, very low in terms of capital investments. If we were to see greenfield capacity addition for 3D NAND, but also investments in DDR5. Do you think that could become a meaningful driver in '26? Or is your participation in those markets fairly de minimis?

Yes, I can take this question. I think that as DRAM devices will -- will improve, increase, there is more and more ALD layers. And with that, we expect an increase in our business from that point of view.

I think you said 25% is memory? Is it like a really small sliver of your revenues today that's basically ex-HBM?

So most of the revenue right now is actually in HBM, Didier, okay? That's what...

The next question is from Tammy Qiu, Berenberg.

So the first question is on GAA. I remember earlier this year, you were saying that GAA order would be up quarter-on-quarter in 2025. So now we have a little bit of timing-related issues. So going to 2026, would you say that GAA would be still grow year-on-year versus 2025 level? And how should we be thinking about this pattern on the -- from an order perspective?

So we see that 2026 is going to grow in GAA versus 2025. I think in 2026, there's going to be 2 things happening. First 2-nanometer production will continue in gate-all-around. But also, you see the 1.4-nanometer node will also start in pilot production in the second half of the year. The other thing that we see happening is that also there is going to be more customers in a 2-nanometer technology node in 2026 versus 2025, which is also driving some new business, especially in the U.S.A.

Okay. And was 1.5 -- sorry, 1.4-nanometer, you mentioned, the first batch of order for pilot line should be starting to be seen in next year. So is that coming from all the customers? Or this is only from one customer? And also, when would you expect the volume kind of ramp-up phase for 1.4 nanometer start to be seen?

So to answer your question. So right now, for the 1.4 nanometer in our numbers, what we looked at is one customer in 1.4 nanometer in the second half of 2026. I think the -- we hope that there's going to be another customer in 2026, then that would be also a better business for us in 2026.

And -- sorry, the volume ramp-up time frame?

So the volume ramp-up is going to be very small. It's the pilot production in the second half of 2026. That's what we put in. And we -- the 2-nanometer node that we mentioned is a very long node. So 2-nanometer is going to continue in 2026 and also 2027, it's a very long growth. As you know also the 2-nanometer node has sub-nodes and with different structure -- like different structure like a midcap that backside power distribution and so on and so forth. So for 1.4 nanometer, I think the big production start will be 2027 and 2028.

The next question is from Robert Sanders, Deutsche Bank.

Maybe a question on the gross margin. You're looking for a double-digit decline year-on-year in 2026. When I plug in a kind of estimate for China gross margin, that means that the consensus gross margin looks too high by quite a big margin. So is there anything that could mean that the gross margin is not well down next year as I think about next year?

I'm not sure I understand the question. What you say, a double-digit decline in gross margin?

In China sales.

You put in your release a double-digit decline in China sales. I look at the consensus. And the consensus is pretty optimistic on gross margin for next year despite the China mix being a headwind. So I was just wondering if there's any reason why the gross margin ex China would improve?

Improve compared to what?

Calendar '25.

Yes. Then I think I have to disappoint you because I don't see it improve compared to '25, to be very honest. We have -- I mean, in '25, we have a few things that are very important. One, we have a pretty still strong level of China sales over the full year, with H2 below H1, Q4 below Q3, et cetera. But still, if you look at full year, a pretty strong level of China sales, and we expect to meet a double-digit decline next year, which everything else equal will have impact.

Two, we have a very strong product mix, especially with, of course, 2-nanometer ramping to have a lot of leading-edge products and relatively spoken, lower, let's call mature products, power/wafer/analog, et cetera. So next year, when we also expect, of course, still continued growth in leading edge as Hichem just explained, but we also expect growth in, let's say, the power/wafer/analogs are relatively spoken, the rate of that segment will increase a little bit. That doesn't mean that every product in that segment has a lower margin and leading-edge product. But on average, I think it's fair to say that leading edge, we can demand higher margins than in that segment. So if you add it all together, lower China sales and a slightly different mix relatively spoken, you very likely get to a lower margin than in 2025. How much lower? I'm not going to tell you.

But that's also at the end of the day, really, it all depends on customer mixture, and product mixture. And I think that as Paul has mentioned, okay, we have made some efficiency in our business and also our cost structure in such a way that we fundamentally improved our gross margin from previous years. So it's still well -- we're going to see what's going to happen in 2026.

But we are very happy with really what we achieved this year. For us, 51% for 2025, this is a record for ASM, really a record. And that's an indication of the -- first, our position in the market, our competitiveness, but also in our cost control and better operational efficiency. And we will continue to really drive that in the future. We're not stopping right now.

And maybe to add, Rob, that was also the reason why we have increased our margin guidance during the Investor Day from 46% to 50% to 47% to 51%. So overall, indeed, we see improvements based on everything that Hichem also just once more emphasized.

The next question is from Francois Bouvignies, UBS.

So my first question, Hichem, on your new wins in Epi and ALD dipole and what function that you talked about in the DRAM HBM. So can you just provide more details on that design wins? Did it happen in the quarter? Was it competitive? And I mean you said in your remarks that it was for '26, '27 time frame, which I thought it would be more like '27. So is it like earlier than you expected? Maybe these wins and how many layers are we talking about? So I know it's many questions within that, but basically, giving more color on this Epi and ALD dipole design wins you had this quarter?

Yes. And we're very excited about the wins that we have made in both Epi and ALD in the last quarter. So this is really work we've been doing with our customer for the past couple of years, and we've been able to become POR for this business. And with starting HVM high-volume manufacturing in 2026 and beyond. As I mentioned, we're really working with many customers and on more and more layers and products. I think what we have seen right now in the industry is really pull-in in a way in some of the high-performance. I think the AI market is becoming super-hot on that point of view. And we see that the customer really want to have a higher performance. So with a higher performance, we see some of these things have been a little bit pulled in.

But I think the majority of the business, the bigger business for us really would happen when the move of DLM to FinFET, I think that would be great. But also we see many of the memory customers also very much using advanced packaging. And we're working with many of those right now on some of advanced packaging applications. But I think we see that 2027 and 2028 where things will become much more positive from that point of view. So 2036 will be the start and '27, '28...

How many layers did you win specifically for that this quarter?

[indiscernible] how many layers, but to be honest with you, we have many customers right now. We're actually working with all the DRAM customers. And you will hear in the next few quarters, more and more wins as those materialize.

That's great. And maybe China. I mean China, you forecast double-digit percentage growth. It seems that everybody is seeing the same thing, the double-digit decline, sorry, percentage next year for China business, LAM is seeing the same. [ ASML ] is seeing the same. So how do you build your forecast, I mean, out of interest?

Because I mean, my understanding is China is quite low in terms of visibility right now. You had some restrictions that only impact 1% to 2% of your sales, obviously, it's more bigger of your China business. But how do you build this forecast out of interest? Because it's very difficult to know where China is going to be next year. And when I look at the retail imports are increasing significantly in the second half of the year versus H1 this year. I would think that you should see decent year from a deposition and etching point of view. So just wanted to understand how you forecast China?

Yes. Let me take that question, Francois. Actually, the very short answer is customer intel. So we have people on the ground. We get, of course, we try to get, of course, as much as we can insights into the plans various customers have into new fabs that are being built or not being built. So every year, we have an idea, but you're correct, there is limited visibility. That's also true for next year. So there's no change from that point of view, but still based on the number of new fabs that you think might start based on inputs in that respect that we get from customers.

We had one year, it's maybe higher than the other year. That's an important input for us. Also last year, we had that input for this year. We're actually at the beginning of the year, we were -- I mean we were maybe a little bit prudent. Looking back now, China did a little bit better than what we anticipated but not to the extent that it was better than the year before. So also for next year, I mean, it could be slightly higher, slightly lower than what we currently think. But in any case, what we see based on all the intel that we have is what we guided and what we said in the press release. So that is, let's say, the best guidance we can give you based indeed on the limited visibility we have, but it's still supported by as much as possible customer intel that we can get.

You next question is from Jakob Bluestone of BNP Paribas Exane.

I had a slightly similar question actually. I mean, you say that you expect the order trend to bottom out in Q4 at a slightly high level than in Q3 and then sort of gradual recovery through '26. And I guess, just interested in the sort of broader business, where do you get the confidence from that? Is that what you're starting to hear from your customers? Or was that just sort of more from the various announcements that have been made? So just to get a sense of how concrete is your confidence on that trajectory that you've laid out for the improvement in orders?

So I'd like to -- what Paul has mentioned earlier, we are very close to our customers, especially we are extremely close to our logic customers. Since we're working with all the top larger company, leading edge -- in leading edge logic and foundry. And then so that's the information really we're getting it from them. I think they have their investment plan for 2026. And based on that, okay, we are talking to them, and we know what kind of business we're going to achieve from them. So we feel confident from that point of view, okay?

Understood. And if I can just ask a quick follow-up as well. Just on lead times. Can you comment on whether the lead times, particularly for advanced logic are changing or have they stayed the same?

Lead time for us as a company, we always said that our lead times like used to be 6 months. But as I mentioned during the Investor Day, actually we have made a significant improvement, whereby we can reduce our lead time to about 3 months right now. So we took -- like we mentioned, we have made significant efficiency in our business processes, in our manufacturing and operations in such a way that we can be very fast in really being able to meet customer expectations.

The next question is from Nigel van Putten, Morgan Stanley.

I guess another question on your sense of growth into '26. So what are the areas you are seeing the biggest certainty and uncertainty in terms of materiality both on positive and negative. Now my guess is that the advanced foundry is pretty predictable and a strong positive into next year. And it seems to be that maybe more advanced logic and also power/wafer/analog are maybe a little bit more uncertain. Am I in the right ballpark here? And also, do you think that you could still grow if these 2 areas do not show growth next year? That's my first question.

Yes. So I think you're directionally correct, Nigel. So I think where we -- as Hichem also said, where we have, let's say, the most reliable -- maybe too strong, but we work the forecast with customers and especially the large customers and especially with the leading edge logic/foundry, the quality of the forecast that we get is better, I would say, than with quite a few other customers. So that's one. So there, we have, I think most confidence. That does not necessarily mean that things cannot change. Things will always change. For sure, pull-ins, pull-outs, et cetera, will always happen. But will also happen next year. But -- and that's the area where we have most confidence.

Two, I think in DRAM, given everything that is happening there, if you read all the market intel, I mean there is capacity shortage, I would say, demand is higher than supply. I think we're relatively confident on what we can achieve there. So there, we have quite some visibility. We just talked about China. We do believe China will come down, as we said, there is limited visibility, but it's not that we steer completely in the dark. We have still reasonable customer intel, but not as good as what we have from the logic/foundry customers.

And then for the power/wafer/analog, I would say it's maybe most uncertain. That market is now 7 quarters and by the end of the year, 8 quarters in a cyclical downturn. So it is more based on the fact that at a certain moment in time, that market should also start to see a turning point. But there, we do not yet see that in orders coming in, but we do expect that to come in and partially also based on, I guess, on customer intel, but that, I would say, is maybe the most uncertain one, but we would expect that to happen somewhere in the course of next year.

Got it. I'm going to use my follow-up then to still maybe press a little bit on sort of the dynamics you've also pointed out this quarter that there is still quite a bit of difference between one and the other. So maybe just for your forecast, are you sort of assuming multiple customers to grow next year in a material way? Or is that not your base assumption at the moment?

Yes. Basically if you talk leading edge for customers at this moment. I think 2 of them, and I'll leave it up to you to guess which one. We are reasonably confident that they will grow. One of that is still uncertain, and it's not an important one. But yes, we'll see what will happen there. We have, of course, a certain view baked into these projections that we have given, but at least 2 and maybe 3.

Yes. But I think if I add something to what Paul has mentioned, I mean we see customer concentration has increased in the recent quarter for advanced logic and foundry and we think it will continue in 2026.

The next question is from Stephane Houri, ODDO BHF.

Yes. Actually, I have a question about 2027. I know you gave -- just gave sense of guidance for 2026, with a very low point apparently and so it means that there is a need for an acceleration in the second half. Overall, we should expect probably a growth kind of low growth next year or mid-single-digit growth, I don't know. But it means that you reach your 2027 target, you need to have a double-digit growth in '27 at least. So what are the pieces of the puzzle we should look at to understand if you're in the good trajectory or not?

Fair to say, Stephane, I think firing on all cylinders is the right description here. So one, as I think Hichem already said at the beginning of the call, end of this year or middle of the second half of this year, we expect pilot -- investments in pilot for 1.4 and then going into HVM in '27. That's of course, a big driver. You've seen the SAM increase that goes with it.

Two, of course, we would expect memory to continue to be good on the back of, in particular, AI, which now, by the way, is driving both HBM also, let's say, conventional more high performance in DRAM, in particular, which should be good by '27. I mean, if power/wafer/analog by then is still not recovering then, okay, I don't know what is happening there, but we would expect logically that's also there. By that moment in time, you would see a turning point. So -- and also spares and service business will continue to grow as we -- based on outcome-based services, as we said, during the Investor Day. So firing on all cylinders would be, I think, the right description here.

Yes. One thing I would add to what Paul has mentioned is also what we see in 2027 is in logic, you're going to have 2-nanometer continued investment in capacity, but also the capacity increase in 2027 for the 1.4 nanometer node, okay? So you're going to have really leading edge 2-nanometer and 1.4-nanometer significant investment in those 2 technology at the same time. Because 2-nanometer is actually a very strong -- it's a very strong node, it's a very long node. That's what our customers are telling us. And as you know, there's many sub-nodes in 2-nanometer. So investment in 2-nanometer is still going to continue at a very healthy level in 2027. At the same time, you have the 1.4 nanometer expansion in that same year.

Okay. Okay. And I just wanted to come back on China where you see double-digit decline next year. Are you sure that this is only the market and that you are not facing an increased level of local competition. There is more and more noise about the efforts they are making and the quality they are obtaining. So can you maybe describe the situation there and the sustainability of your market share?

That's a good point. I think it's a combination of both things. One, China, we've seen everybody actually also appears to have seen a very high level of investment in the last 2 years or so, 2.5 years. So we've already communicated, I think, end of last year that we would expect this to gradually normalize, whatever that means.

Two, with recent, let's say, announced export controls, that also has impacted us to a certain extent, but also our peers, although it's maybe not materialized 1% to 2% of our annual revenue globally, but it's still a few percent, of course, of our China revenue, which again leaves a vacuum for local competition to step in. So yes, local competition will for sure, benefit from this, will get on a quicker learning curve because, yes, the unfortunate reality is that because of all these restrictions, yes, there is a playing field -- an unleveled playing field where local competitors can step in with what we would say inferior products compared to our products, but at the same time, learn at an accelerated pace compared to the situation if they would not have been able to get their products into customer fabs. So it's a combination of these 2 things, I would say.

And one thing I would add to what Paul has mentioned, okay? We think that we are very competitive in China vis-a-vis the Chinese competitor. We don't see our position to be really worse than what it was before. I think our products are very good, and we continue innovation unabated which really gives us an edge from that point of view. In China, what really -- what 2026 shows for us right now, we have really very low visibility on how some of the part of the market is going to materialize.

I talked today in my prepared remarks that we have won some significant business in Epi Intrepid ESA in the power/wafer/analog and some of those actually businesses is happening in China. So depending on what that -- so that shows really our competitiveness in that market. And I think if the market -- if the power/wafer market analog recovers, that's going to be also very positive for us in the future. And as you know, in the power/wafer/analog, visibility is very -- it's not very long. That market can go down immediately and go up at the last minute. So that's really a question we're going to find out in 2026, but we are very competitive. We think we can compete very well in that market. And it's an important market which we're going to continue to address.

The next question is from Adithya Metuku, HSBC.

I had 2. Firstly, look, when I look at your growth, you've always tended to outperform WFE given the company-specific growth drivers. As I look out to 2026, is there any reason why you will not outperform WFE next year? I just wanted any -- is there any headwind that we should keep in mind? Any color there would be helpful. And then I've got a follow-up.

I think on the outperforming the wafer fab equipment, we have mentioned that in our Investor Day that we will outperform the wafer fab equipment when going from '24 to 2030 -- from 2024 to 2030. It doesn't mean, okay, we're going to outperform every year. So still I'm sure about that. I mean we already announced saying that the 2026 for us is actually a growth year, which we are very confident about it. But to answer the -- whether we're going to outperform it in 2026, that's too early to say. But what we can tell you that from 2024 to 2030, we will outperform the WFE market.

Got it. Maybe just on that, I mean, if you were to underperform given where you sit today and your visibility into the different end markets that you talked about on this call before, if you were to underperform what would be the reason? I struggle to see what -- I don't see any reasons, but I'm just trying to see if I'm missing something here.

It depends really on the WFE mix. What's the mix of products, memory versus logic versus power/wafer/analog. It's really mix dependent.

Okay. Got it. And just as a follow-up. Paul, I just wondered if you could give us some color on OpEx in 4Q. I know you commented on SG&A, but I don't think you commented on R&D. And also if you could give any color on how we should think about OpEx into 2026? That would be super helpful.

Yes. So R&D for Q4, I think, there will be similar to Q3 gross R&D and most likely also net R&D. SG&A, Q3, we mentioned that there was relatively low variable expenses, where we made an adjustment based on certain accruals that were made. So I think the Q2 is more of an indication than Q3 going forward and into '26, and we've given guidance in the Investor Day. We will continue to invest in R&D, which is a lifeline. So there, you will see gradual increases compared to this year. And SG&A, we will keep very tight. And as a percentage of revenue, with revenue growth, we would expect that to come down a little bit further.

Can we have the next caller, please? Operator?

The next question is from Timm Schulze-Melander, Rothschild & Co Redburn.

My first one just very big picture, maybe a question for Paul. 2026, do you -- should we expect the aftermarket side of your revenues to outgrow system sales? And then I had a follow-up.

I have a view, I'm not going to tell you because it's too early to tell. But what I can say is that we do expect a healthy growth in our spares and service business in '26 compared to '25. It's too early to already say it will be higher or lower. But we do see what we believe will happen is that we will continue to see a very healthy growth in spares and service business.

Okay. That's helpful. Maybe just one other one. If we just -- you talked a lot about how important mix is to the outlook in 2026. Could we just when we think about your ALD business, I know you've talked about it being more than half of your system sales, but could you give us a kind of 5% to 10% range kind of what that ballpark looks like for 2025?

No, we were not going to give very specific guidance other than what we've given. We've obviously had ALD is more than half of our equipment business, which indeed it still is and maybe this year more so given the ramp-up 2-nanometer. It's also one of the reasons why the margin is where it is. Also given that, again, the more power/wafer/analog market is down. So relatively spoken, a lower percentage of the overall mix. And of course, still China is strong, although below last year, was still strong. So adding that all up, that's what you get, what we have, but we're not going to give specific guidance within a few percentages, what -- where we stand with ALD.

Okay. But given your prior comment about power/analog and some of the other mix, then as a percentage of sales, ALD might be flat or down as a percentage of your equipment sales in '26?

Depending on what we assume and depending on how much the relative markets grow, there might be indeed relatively spoken, a -- but now we're really talking scenarios, there might be relatively spoken, slightly lower share of ALD compared to power/wafer/analog if -- and it's a big if, if power/wafer/analog will grow more fast or faster than ALD, but that's still to be seen as well. So it's really too early tell. I really don't know. We have, of course, certain scenarios and assumptions, but it's too early to give external guidance on that.

And the last question is from Marc Hesselink, ING.

Yes. I have 2. Actually, on 2 bit smaller categories. Firstly, advanced packaging that you also now point out again in the press release, also at the Capital Markets Day spent some time of it. The fact that you're really focusing on it, does it show that this is something that can be material as well over the coming years? And how do you expect that then to ramp into your revenue numbers?

So I think we mentioned -- thank you for the question. I think we mentioned that we have made some wins in advanced packaging the past quarter, which we are really very excited about. We really think that advanced packaging is very enabling. We think that as a company, we can provide meaningful innovation and disruption to the market in the area of new materials and in the area of surface preparations and based on our experience in both ALD and also chemistry. We have a significant engagement in the past actually a few quarters with some key customers in both memory and logic, high-end logic to develop some of those films. We're very excited about this part of the business. And hopefully, we see more and more wins in the next few quarters, and we'll keep you updated of those when time comes.

But maybe to add, is it then material? Or is just the first start of something that can be material in the future?

It's really the first start. It's really the start right now. And that's why we're excited about the future of the company, and I think that would be something that, let me say, very excited about it, to be honest with you. I think there's -- we see customers putting in very hard. I think the -- as you know, in both memory and logic, heat generation is a big problem. So we're developing some films that really would reduce the hotspots with higher conductivity capability. We see films in actually the microphotonics area team developing films that can reduce light dispersion. So it's really one area that becomes very important for both the logic and the memory customers. And that's an area that, hopefully, is going to be very accretive to us in the future.

Great. And the second question is on -- actually on LPE. So you paid the earn-out in the quarter. So you did hit the milestone for that one. I mean I think in the press release, you can still read that it is very, very weak at the moment and also no recovery into next year. So can you then still talk about the building blocks, why did you still reach the milestones? And when do you -- are you still confident on this one to pick up maybe in the longer term?

Simple answer. The milestones are based on 2024. And 2024 was a star year for us with almost, I would say, explosive growth in silicon carbide. So if it would have been -- if the milestones would have been based in 2025, they would not have been met. But yes, the reality is that they were based on '24 and '24 is a very strong year for silicon carbide.

Okay. And no visibility on that improving in the beyond '26 period?

We, of course, expect that markets to come back at a certain moment in time, not yet for next year, at least we see no evidence for that to start to happen next year. But we still believe that there is a market for us, a good market for silicon carbide that we will start seeing come back, hopefully, also in 2027, but that's as again, it's too early to tell.

Gentlemen, there are no more questions registered at this time. I turn the conference back to the management for any closing remarks.

Thank you all for attending our call today and also on behalf of Paul and Victor. We hope to meet many of you again in the upcoming investor conference and other events. Thanks again, and goodbye.

Ladies and gentlemen, thank you for joining. The conference is now over, and you may disconnect your telephones.

Good afternoon, everyone. Welcome to ASM Investor Day. Thanks for joining us today. It's great to see so many familiar faces today in the room. My name is Victor Bareño. I'm the Head of Investor Relations. Also a warm welcome to everyone on the webcast. A few housekeeping items. In case of emergency, we will have emergency exits here on the left and in the back of the room. And in an emergency, please follow the instructions of the hotel staff. Kindly mute your phones, put them to silent. And let's now have a look at the program. So we will start with our CEO, Hichem M'Saad, who will talk about strategy and our priorities throughout 2030. Hichem will be followed by Vamsi Paruchuri, our Corporate VP of Technology Innovation and Market Research.

He will talk about the market outlook and technology inflections. Vamsi will be followed by Jason Foster, our Corporate VP of the spares and services business unit. He will talk about growth and innovation in our Services. After the break, we will continue with Eric Shero. He will talk about ALD technologies and a number of exciting new applications. Next, Paul Verhagen, our CFO. He will provide a financial update and present the targets for 2030. Hichem M'Saad will wrap up the presentations, and then we will open the floor for Q&A for questions from the audience here in the room with all 5 presenters back on the stage.

I have to show you the forward-looking statements, very important because today's presentations will contain forward-looking information about future business and results. The event today will be webcasted. A recording of the webcast will be available shortly after the end of the program on the Investor Day website as well as the presentations. And with that, let's get started, and please join me in welcoming Hichem M'Saad to the stage.

Thank you very much, Victor, and I would like to welcome each one of you for -- in our third Investor Day meeting. We had our first one in -- during COVID in Amsterdam in 2021. And the second one, we had in this very beautiful setting. Today, we have a very full agenda. We're going to talk technology. We're going to talk marketing. We're going to talk sustainability, and we're going to talk about our projection and business projection and also all the cool stuff we've been developing in ALD, epitaxy and other products.

And we'll show you how our growth through innovation fits in all of this. At the end, after my presentation, also you're going to see a presentation of my colleagues. And at the end of my colleagues' presentation, we're going to have our CFO, Paul Verhagen, to talk about our projection for 2030. But before we get started, I would like to show you a video highlighting the lifetime changing impact our technology has on the world. So where is the video?

[Presentation]

Okay. So I hope that you guys love the video. I'm going to start my presentation, and the title of my presentation is growth strategy to 2030. The key takeaways of my presentation today is that we have delivered on our strategic objectives that we have set in 2021 and 2023. We have outgrew the WFE market. We have maintained and also, we have expanded our ALD and Epi share in the transition from FinFET to gate-all-around. We also grew our spares and services business. We have developed many ALD applications, including clustered multiprocess application like area selective deposition and some of it are actually in production at the 2-nanometer technology node.

We are very well positioned in ALD and Epi for the upcoming transition in both logic and in memory. In logic, we're going to have second-generation gate-all-around, third-generation gate-all-around and CFET. And in memory, we're going to transition to 4F2 or the vertical channel transistor and 3D-DRAM. We also have developed AI and ML technology on our common platform. And this AI/ML is becoming very important, and we use it for 2 applications: one, to speed up the innovation for our customer; and two, to help in addressing HVM concern. Another thing that we're really very excited about is advanced packaging. We think an advanced packaging is another midterm growth area for ASM.

Application in advanced packaging will benefit from the chemistry know-how, the interface know-how and the surface know-how that we have developed for the past few years. To scale our company, as you guys know, we've been like growing -- doubling growth every 4 to 5 years. To scale our company, we actually focus on talent development. We try to develop talent for the next -- for our new products. We also focus on product commonality to make sure that all our products can have the same platform and also on flexible manufacturing footprint, especially in this day and age where tariffs is a concern. And also, we have upgrading our ERP and PLM systems for improved operational efficiency so that we can become much more profitable. Sustainability is also a very key part of our DNA.

And we use -- sustainability is actually integrated in our way of doing things, that's in a way of working. And we actually -- sustainability is very important for 2 reasons. It's good for the environment, but most importantly, sustainability is good business sense because with sustainability, you actually reduce the total cost of ownership for our customer. Last but not least, because we're really excited about the opportunity that we have in the future. We have targeting 2030 revenue of greater than EUR 5.7 billion with operating margin of 30% and free cash flow of greater than EUR 1 billion. So let's go and see our journey from 2021 to now.

In the previous strategic objective that we have talked about in both our 2021 and 2023 Investor Day meeting, we have these 6 strategic objectives: maintain leading share in ALD and expand in memory, increase share in Epi, grow selectively in vertical furnace and PECVD, grow spares and surface revenue, accelerate progress in sustainability and then also drive continuous operation performance. So how did we do in each one of these? Let's first start about ALD, where we need to maintain our leading ALD share in logic/foundry and expand in memory. We are very pleased to share with you that we actually have increased our share in ALD to greater than 55%.

The transition from FinFET to gate-all-around really helped us increase our market share. And here, we can show you that our market share is actually greater than 55%. You can see on the right axis, the number of ALD tools at customer sites. And you see from 2016 to 2020, we actually have doubled the number of our installed base. And from 2020 to 2024, we doubled again. So every 4 years, we actually were doubling our ALD installed base at our customer sites. Second, Epi market share. In Epi, we have increased our market share to 25% in 2024 for the leading-edge epitaxy. In 2020, our market share was 12%. And in 2024, our market share in leading-edge Epi is 25%. On the right axis, we actually took our share of layers ASM layers in leading-edge logic and foundry.

So in leading-edge logic and foundry, this is taking all the 4 customers that are working on 2-nanometer technology node, and we count all their layers, and this is really the layers that we have at the 2-nanometer node. As you can see the 2-nanometer node, we actually have increased the layers that we have business in to 33% from FinFET, where we are at 22%. Third, we want to grow selectively in furnace and PECVD niche application. And here, our performance was average. If you look into the vertical furnace, we're showing you the revenue from 2020 to 2024. You see the revenue has increased up to 2023 and came down in 2024. Our vertical furnace business depends a lot on wafer/analog business.

So it's really -- that's the more than more business. And more than more has grown to 2023. And in 2024, we came into a downturn. That's why our revenue has came down. But as you can see, when we grow to -- when the market was growing to 2023, our revenue has grown because of the new products that we have in furnace like the SONORA and the DUO. In PECVD, our performance was worse than vertical furnace. Our revenue did not increase the past few years. Why? Because we didn't have any new products in PECVD. But in PECVD, we actually -- we are more optimistic about the future. And we've been using our chemistry knowledge where we have developed, I'm showing you here, a new technology called flowable carbon.

This flowable carbon film is a film that actually used in patterning application in high aspect ratio, both vertical, as you can see, and actually lateral structure. And we are very excited about this technology that's actually being implemented for the gate-all-around technology node. But we think that this technology has even more application in advanced packaging in the wafer-to-wafer and die-to-wafer bonding. We have said we want to increase our spares and service business. And here shows that we actually have doubled our spares revenue from 2020 to 2024 from EUR 277 billion (sic) [ million ] in 2020 to EUR 547 million in 2020. So EUR 277 million to EUR 547 million, which really corresponds to a 19% CAGR growth in our business. And the reason we've actually been growing this high in spares is because of the outcome-based product that we have.

So what's the outcome-based product? The outcome-based product is a product that actually we guarantee performance to the customer. We tell the customer, we're going to give you this kind of products, spares products, but we guarantee uptime, we guarantee availability for you. And that's becoming a very important part of our business as equipment becomes much more complicated because the device structure is getting also much more complicated. So I'm going to show you 2 examples on spares, whereby actually we have reduced cost, and we actually improved performance. Our vision for the spares business is actually to increase the availability and uptime out with our tool using renewable and actually sustainability.

So -- and that's really what we're looking at. So on the left axis, you see that how we've been able to using dry processing as an alternative to wet processing. So in spares business, if you have -- once you want to refurbish a part, actually, you refurbish it with water and use lots of nasty chemicals like acid, sulfuric acid, phosphoric acid, very corrosive. We said, okay, let's get away from these nasty wet processes and do dry processing. So actually, we're refurbishing right now our tools with dry processing. And as you can see here, using dry processing, we actually can reduce 95% of CO2 emission. With the wet processing, you can achieve 85% reduction. So it's more reduction. But most importantly, because dry processing gives you better selectivity, the parts that you refurbish actually last longer, and that's how you can reduce the cost to your customer.

On the right side, we show that how we can use dry texturing of the parts. So what does dry texturing means? It means you want to put roughness on the parts because when you put the roughness, then you have better addition of your deposition and doesn't [ fall ]. And once you have better addition, then you don't have particles going down on your wafer. Before, the way you get texturing is through bead blasting of the part. But once you do bead blasting, you have many issues with defectivity because then you need to clean them. So you use lots of wet processing. With dry texturing, you don't do that. And with it, as you can see, you can actually extend the lifetime of your part. The other thing we want to do, accelerate progress in sustainability, our #5 objective.

And here, we show you recognition by many parts of the business to our sustainability effort. We're really very excited that we are one of the leaders in sustainability in the semiconductor industry. We got AA from CDP. We also got the lowest risk rating in Sustainalytics, and we're actually #1 in the semiconductor industry from that -- in Sustainalytics. And then if you look into the RE100 or the Renewable Energy 100, we joined this consortium in 2023. And in 2024, we've actually been able to generate 100% of our electricity with renewable. So in 2024, all our sites work on renewable energy. And that's why we have achieved the best newcomer award. And this year, TIME magazine has recognized ASM as being one of the world's most sustainable companies. Number six, we need to drive better financial performance.

And here, on the left axis, you see that we actually have outgrown the industry from 2020 to 2024. So this is really looking into the equipment part, not service, but only equipment. We grew about 20% in our equipment business, while the industry grown by 13%. And because of this growth, we've been able to generate much higher profit and our operating margin hovers between 26% to 28%. This is not world class, but I want you to know that during this year, we have spent lots of money into building our infrastructure, especially in our manufacturing infrastructure. So I get many questions from the analysts like yourself, how come you guys have been able to compete to be successful in this business when you have very many strong competitor.

And here, I would like to address some of the concern that you guys have. Number one, we have great technology, and everybody knows technology, we're a technology company. But technology is not enough. It's really -- technology is necessary to be successful, but it's not sufficient. And there are many other reason why we are successful, which I'm going to try to address in the next few slides. So first one, we need to talk about technology. And when you talk about technology, you need to start about patent. And this one shows our patent portfolio by LexisNexis, where LexisNexis compares our patent portfolio, our strength in our patent versus our competitor.

So competitor A to competitor D. So on the X-axis, you see the portfolio size. And on the Y-axis, you have the competitive impact. And the size of each dot, that's really an indication of your patent strength. And it's actually the product of the competitive impact by the portfolio size. Second, we have actually a very global footprint. We are everywhere. We are in the Americas, we are in Europe, and we are in Asia. Wherever our customers are, we, as a company, we are actually very close to our customers. Second, we actually have been investing a lot in the past few years, and we've been actually continuing investment. And this investment here is actually investing in Dongtan in Korea. As you can see, there are 2 buildings here.

The one, which is the new building, we call it, Dongtan 2, but next to it, that's our first Korea building. And this is a new manufacturing expansion that we have. And in Q4 of this year, in December of this year, we're going to have our grand opening of this newest manufacturing facility in Korea. And with this opening of this manufacturing facility, actually, we will be able to cover our production to up to 2030 or beyond. Second, we're actually investing in R&D. And here, the investment in R&D is a new site in Arizona. ASM is actually one of the first company -- semiconductor company in Arizona. We've been there over 50 years.

And we are very happy to welcome some of you guys next few weeks in the SEMICON West, which is for the first time, is going to happen in our neck of the woods. We are very excited to expand our R&D capability in our premier site in Arizona and especially in Scottsdale, and we're going to have our opening in Q1 of 2027. So we're very excited to do that. I'm personally excited because it's not very far from where I live. The third thing that really we have done in the past few years is really to make sure that we take our product strategy in place and to especially commonize our product strategy because we had our products okay on different platform. But here, since 2010, we have shipped our flagship XP8 platform.

This XP8 platform has been running ALD wafers at customer fabs since 2010, and we have thousands of these platform, XP8 at customer sites. But we need to get XP8 platform ready for the future. And what's the future? The future is really making sure that your platform is ready for ML and AI. It needs to be compatible to that. So to do that, we had to do a couple of things. We need to make it more modular. We have to put a new software in the platform, but also, we need to make sure that the data transfer is actually very fast. So we have to change also our data transfer and so on, and we make it compatible with that. But also we need to make it compatible for the advanced ALD and CVD processes.

These advanced ALD and CVD processes need integration with surface clean, with surface modification, with selective etch, with treatment. So that's why we took the most -- our best platform, and we expanded to the XP8E, which has 6 facets. So XP8E, E stands for extendability. And I'm going to show you right now that we can use -- how we're using this platform in, for example, area selective deposition, which an advanced ALD process. So let's take an example, whereby we want to deposit dielectric film on dielectric. So we call it DoD. And here, our XP8 platform has 3 different reactors. Al2O3, that's the dielectric inhibitor reactor and a clean/trim reactor. So here, we want to deposit ALD. But to deposit selective ALD, you have to have different reactor.

Here, you need to have an inhibition reactor, you have to have a clean and trim in the same reactor from that point of view. So this is just an explanation about what we do. So when the wafer comes in, the wafer is very dirty. And as you can see the dots there, that's very dirty. So what we do, we take the wafer in the clean reactor to remove the dirt. And then we do inhibition because we want to deposit on dielectric and we do the inhibition, which really -- this is a carbon film, which we want to make sure that gets deposited on the metal. So what's orange is metal. But also, a little bit of it gets deposited on the dielectric. So what you do, you actually trim it. You do trimming and you remove the inhibitor from dielectric.

And right now, you can deposit your ALD, aluminum oxide on dielectric. You have the inhibitor, what you do, you take it to your trim reactor, you put hydrogen plasma through it. You have carbon-hydrogen plasma, [ you form CfCH4 ] and you take it out [ and voila ], you have dielectric film on dielectric film, DoD. So this way, and you can see the yellow on the TEM micrograph, SEM micrograph below, the yellow, that's the aluminum oxide on top of silicon oxide. So this is really how these kind of what you call them advanced processes are happening. The other thing that we talk about is really AI and ML. And we have a very strong program in using artificial intelligence and machine learning in our products.

And we use them for 2 reasons. How can we speed up innovation at a customer, so for R&D, but also, we use them for HVM, how can we keep our tool up and available to the customer as much as possible. And here, I'm going to give you an example in HVM, whereby we look for anomaly detection. So when you're in a fab, what the customer cares about? You want to make sure that your process is very repeatable, okay? And every deviation is a problem. You need to -- we don't want to have any deviation. For these AI guys, they don't talk about deviation. For them, deviation is anomaly, okay? So that's why the terminology of anomaly is here. And if you have an anomaly of 0, that means everything is great, and anomaly of greater than 0, then there is a problem.

On the left axis, as you can see, we actually have been running wafers, like 50 wafers. And on the -- on the Y-axis, you look into the anomaly score. So as you can see, the dots in the bottom, you see anomaly score of 0. That means everything is good. And then at Wafer 30, up, there's an anomaly happening at -- which is like 0.4. And then at Wafer 39 or something, there's another anomaly. And Wafer 46, 47, you see anomalies happening there. So there's some deviation. And it's really the customer will set up the threshold in anomaly. For example, customer said, okay, everything above 0.4 is a problem. It's a big problem. I need to stop the tool. But if the anomaly is less than 0.4, this is just an example, then the tool is okay. The yield is still okay. So the way our software works is that you can really choose any of these anomalies that -- in the data.

And you can -- you want to find out what gives rise to these anomalies, okay? And here, we're showing you the 5 top contributor to the anomaly. That means you really immediately in the software, you just press the button and it tells you, well, Sensor #1 is the problem. It has the highest problem, okay? And the sensors here are actually the process parameter. In a reactor like ourselves running ALD, the process parameter can be temperature, can be pressure, can be RF power, can be flows and so on. And we can have actually as many as hundreds of these sensors. And if you take Sensor 1 and you look what's going on in Sensor 1, you can see actually the anomaly there. So we can detect very small anomalies. You can -- we showed it here for Sensor 1 and also Sensor 5.

So with this, you actually are able to help the customer find out the problem immediately and fix it as soon as possible. So this is like once you run the tool, you can actually immediately find the problem instead of running your wafers, very expensive wafer and find out later on what the issue is. So we're really excited about this capability. And as I'm telling you here, we're using it for HVM, but also we can use it for -- we actually use it for R&D to speed up R&D. Our success does not depend on only us. It really depends on also our partnership with university, with also research institution. And we are very happy and very pleased to have very strong relationship with imec in Leuven, Belgium, with TU Eindhoven in the Netherlands and also with University of Helsinki in Finland, lots of collaboration between them on ALD development.

But to win, you really need to win customer trust. Everything else is secondary. To win is really winning customer trust. And here, we're showing you many of the awards that we have received from our logic, foundry and memory customer. So let's stop talking about the past and let's talk right now about the future. What's going to happen in the next 5 years? And here, I'm going to show you what are the things that really the trends that's of concern, not concern to us, the trends that we are very much interested in. Number one, the market. The market is actually moving to high-performance computing in logic and memory, which is driven by AI. That's really the #1 thing.

And with that, we see also some changes in the technology. We see things moving more and more into 3D. We started from FinFET going to gate-all-around, but now gate-all-around going to second-generation gate-all-around, third-generation gate-all-around and CFET, which is going to happen after 2031. And also in memory, as I talked about, we see 4F2 is going to be in production in 2028. And then the move to 3D-DRAM is going to happen, but it's going to happen later on in 2032 and beyond. We also see device scaling and DTCO happening. So the device scaling with channel Epi, with the dipoles for multi-Vt, which is actually significant business. New metal interconnect, for example, using molybdenum, contacts are becoming a big deal. Before contacts is not a big deal.

But right now, since you have 3D structure to make a contact, it becomes much more difficult. So there's lots of contact business happening in ALD. And DTCO or design technology co-optimization, and that's backside power that you guys heard about and MIMCAP are becoming really key drivers in our devices. The other thing that we have seen is advanced packaging. And advanced packaging, we see the conversions between back-end processing with front-end processing. And actually, back-end processing is moving to front end. So advanced packaging is coming to us. And we think there is an opportunity for us to grow into this business, which I'm going to talk about in a few minutes. So let's first talk about the technology landscape.

This pie chart is a pie chart that was presented by TSMC in April of this year during the North America Technology Symposium. It shows that by 2030, 45% of the semiconductor business is in high-performance computing and AI, and 25% is in smartphone. Adding these 2, that means you're going to have 70% of the $1 trillion semiconductor business market in 2030, driven by leading-edge logic and memory technology. That's really what's important. So let's look here into the logic technology. I'm going to let my colleague, Vamsi talk more about both logic, but also talk about memory. So in the logic technology, here, we're showing you the structure of the different gate-all-around technology node. We start with the gate-all-around first generation.

Production, that's 2 nanometer production this year, 2025. As you can see, there is the GAA area in the bottom, okay? And the interconnect and you see the MIMCAP. So the GAA is very small. And we love the small because when something is very small, you need to use something thin, that's where you need ALD. So don't really get confused with those interconnect that you're going to have lots of ALD. It's mainly CVD and -- CVD and PVD, but some of it is actually moving into molybdenum ALD deposition, but not the top ones, actually the bottom ones because as you go down, this device structure is smaller and smaller. So in metal zero, that's where you see some of the metallization happening.

So we love the gate-all-around, okay? That's where everything is happening. And then you see on the top, the MIMCAP for signal. And the MIMCAP, you have lots of ALD to hafnium oxide and the metal is dinitrate . And then you move to the second generation gate-all-around with the 1.4-nanometer node. And this generation will be in production in 2027. And with it, you see backside power distribution happening. With the backside power distribution, there's also some ALD layers. And then in 2029, we're actually going to move to the 1.0-nanometer technology node. And with it, in addition to the backside power distribution, we're going to have MIMCAP at the bottom. So you're going to have MIMCAP at the top and the bottom, much better performance.

And CFET was going to happen at the 0.7-nanometer node in the year 2031. And in the CFET, you're going to have more shrinking because the PMOS and NMOS, which in the gate-all-around are separate, going to be together. So you're going to have the PMOS and NMOS on top of each other that's going to have -- even make it much more complicated, which we're going to make it -- you need much more ALD in the future. So I want to just show you this. And this wheel graph shows you that ALD intensity is going to increase from one gate-all-around generation to the other. Let's give you a little bit more detail on this. Here, we will look into an internal emulation of the gate-all-around structure with backside power.

And as you can see, we have the back end of line, we have the contacts, we have the transistor, and we have the backside power. One thing that really comes to mind very easily, look under the transistor. That's the number of ALD films that you have. You have the most. You have high-k dielectric, hafnium oxide, you have dipoles, N-dipoles, P-dipoles, that's lots of ALD. You have the work function metal. You have the patterning films in those, which you don't see, okay? But you need patterning because why is lots of patterning there? Because the device structure is very tough. To deposit your film, you have to do lots of patterning to get them there because of all the gate-all-around happening there.

And also, you have dielectric gapfill like silicon nitride, ALD and so on. In back end of line, you see PVD, CVD moving to ALD gapfill. And in contacts, you have ALD silicides and also, you're going to have area selective deposition. And I talked to you about -- guys about area selective deposition earlier. And backside power, you have ALD liners, but also you're going to have ALD metal gapfill, that's ALD moly gap fill. So the message here, there's lots of ALD in the transistor area. And we see in the transition from actually the 2-nanometer node to the 1.4-nanometer gate-all-around node that the number of ALD layer are increasing. From 3 nanometer to 2 nanometer, the increase is about 20% and the same thing is happening from the 2 nanometer to the 1.4 nanometer.

The number of ALD layer is increasing over 20%. But most importantly is really the wheel graph you see on the right-hand side. Here, we're looking into the mix in ALD layer. So we took the front end of line by itself, which is the transistor and the rest, that's middle end of line, back end of line and BS means backside power distribution. At the 2-nanometer node, 50% of the ALD mix is in front end of line. And when you go to the 1.4-nanometer node, 60% of the ALD is happening in front end of line. We love it because that's where we are the strongest. So when we go from 2-nanometer to 1.4-nanometer technology node, our ALD market share does increase because the increase is happening in our neck of the woods.

Let's now talk about advanced packaging. In advanced packaging, as I mentioned earlier, there is a convergence between back end of line, back-end processing and front-end processing, not back end of line, back-end processing and front-end processing. And here, we look into the wafer fab equipment for advanced packaging from 2025 to 2030. As you can see, the market doubles from $5.6 billion to about $11.5 billion in 2030. So significant growth of about 15% CAGR. So let's look into the 2030 market segmentation. And this is -- actually data for the market segmentation is the data from both TechInsights and also our analysts. So when we talk about advanced packaging, everybody talks about bonding, but maybe we're missing it.

The biggest business in advanced packaging is PVD and the second biggest is ECP. ECP and PVD, what do they do? It deposit metal. And that's really why you need to have packaging because you need to get the signal going from the CPU to the GPU to the DRAM. And the signal needs to go to the metal and goes with electron and the electron becomes -- when you have electron movement, you have collision and becomes very hard. That's why in advanced packaging, we talk about heat dissipation, okay? Because electron are moving into the metal. And so 40% of material in -- 40% of the market in actually advanced packaging is metal deposition. And then you see next, etch, you see CVD, you see litho and also wafer-to-wafer bonding and die-to-wafer bonding. We, as a company, we play in the CVD because we have PECVD.

We play in ALD, 3%. And also, we play in the surface preparation, which we have knowledge about it, and we have product in that area because in surface preparation, you need to make the surface clean. You need to modify the hydrophobicity, the hydrophilicity of the surface so that you can have lots -- much better bonding pretty much, so much better adhesion. So 15%, if you add 3%, 3% and 9%, that's 15%. So right now, actually, we cover 15% of the SAM for advanced packaging. And we we're looking to extend our SAM to greater than 30% of the TAM by 2030. So looking ahead to what's next for us to 2030. I showed you, okay, that the technology is moving. I showed you what we've been doing and what's our new strategic objective for 2030.

Number one, in our strategic objective is to grow our ALD business. We need to maintain our leading share in ALD for logic and grow share in DRAM and HBM. ALD intensity is expanding. We are an ALD company, and our #1 focus is really ALD. How we're going to be able to achieve that? We still need to continue our innovation in new materials, in new chemistry, but also having AI and ML, okay, to help our customer speed up innovation and also make sure that the tools in manufacturing are working very well.

Second objective is we want to continue to grow in epitaxy. And the way we need to grow there is to capture new Epi layers in both logic and also in DRAM, especially high-performance DRAM. Third, we want to grow our application into advanced packaging, like I talk. Right now, we have PECVD, we're working on PECVD films, [ TiOs ], SiCN, edge, high films, et cetera, et cetera, ALD and surface prep.

Number four priority for us is to grow high-value outcome-based services. You're going to have a very cool presentation later on by my colleague, Jason Foster, about our service business. Service business is becoming very important for ASM. As you have seen, we have actually doubled our revenue in the last 2 years, but we have many more products in this part of the business. Why? The reason because as the technology becomes more complicated by going to 3D, as our ALD layers come thinner and thinner, every monolayer matters. So we're actually going into the Angstrom level accuracy in deposition. With the Angstrom level accuracy, you need micron accuracy into your tool. Before in the tool, you can actually put something -- you can put a part and with your eye, you can position it in your part. That's what we call with the eye, you get millimeter level accuracy in part placement.

When you go to Angstrom -- when you need to have Angstrom level precision into your ALD film, that doesn't work anymore. You need to do micron-level accuracy. And you can only achieve micron-level accuracy with automation. So there's lots of automation in the surface business among others.

Our fifth priority is to accelerate prompt in sustainability. As you have seen, we have a very good program in sustainability that we have achieved the last 5 years, but it's not enough. The most important part of sustainability is actually in the product sustainability. And as we go into 3D devices, we're using much more deposition. When you use much more deposition, means you're using much more chemicals. So we need to reduce these chemical usage. We need to make our processes much more efficient and we need to recycle those chemicals. It's good for the environment, but most importantly, it's actually good business sense because it reduces the cost to your customer and makes you much more competitive.

Last but not least, we need to continue in our operational efficiency, our operational excellence. We need to have be flexible in our footprint, both on manufacturing and also on R&D. And it's very important right now because of all the geopolitical issue that's happening. We need to be flexible. We need to be ready to move anywhere. And that's really what we have done in the past few years in our infrastructure. But also with this, we need to achieve a very strong financial performance to our customers.

Yes, we have 28% operating profit right now, but that's not enough. That's why our target in 2030 is actually to achieve 30% operating profit. And because we are so excited about all the opportunity there is in both ALD and epitaxy, we are willing to say that we can achieve over EUR 5.7 billion in revenue by 2030. Thank you very much for your time.

So let's go right now and have Dr. Vamsi Paruchuri talk to you about all the opportunity and growth area that we will have. Thank you, guys.

Thanks, Hichem. So today -- my name is Vamsi Paruchuri. I'm the Corporate Vice President for Technology Innovation and Market Research at ASM. So Hichem slightly touched on the shifting landscape in the semiconductors. So today in my presentation over the next 20 minutes or so, I will show you how technology inflections are transforming the semiconductor landscape, creating new opportunities for innovation and growth, and most importantly, of course, how ASM is positioned to lead these changes going forward.

So first, the key takeaways. Essentially, secular growth trends. The semiconductor market is facing a very strong growth trends with indications to reach USD 1 trillion by 2030. I think this is not new news, everybody in this room should be aware of that more or less with the new cycle that's going on every day surrounding AI and other technologies. But the most important thing here is that this market growth is essentially driven by leading-edge logic and DRAM technology. So that's the important point where it comes to for ASM.

Secondly, this logic and DRAM technology scaling is increasingly dependent on materials and adoption of more complex 3D structures. So dimensional scaling has slowed down. So essentially, any performance you want to get, you have to go either 3D or you need new materials to drive the performance of these technologies, both in logic as well as increasingly in DRAM as well. And then the third is -- so based on these trends, we expect the ALD to outgrow the WFE market, wafer fab equipment market. It has been outgrowing the WFE market, and it will continue to outgrow the WFE market within the time frame that we are talking about.

The market will reach -- ALD single-wafer market will reach $5.1 billion to $6.1 billion by 2030, reflecting a CAGR of 9% to 13%. Similarly, Epi is going to play a very strong role as well in this leading-edge technologies, used to be predominantly in logic/foundry. But going forward, we see Epi playing a very critical role in DRAM scaling as well as we see some of the new inflections coming into play in the very near future. And it will also drive a very strong growth in the market of single-wafer Epi to $2.5 billion to $3.2 billion by 2030.

So coincidently, a CAGR of 9% to 13%. But most importantly, it will have a higher CAGR in the leading-edge Epi segments essentially. So putting all that together, from being an ALD company focused on logic/foundry, we are very excited as we see the opportunities going forward to see that we are now having ALD and Epi growth, driving both in logic/foundry as well as in DRAM going forward. And that presents us a very significant SAM increase combining the ALD and Epi opportunities. We see -- from 2-nanometer gate-all-around to the second generation or the 1.4-nanometer gate-all-around, we see an opportunity of $450 million to $500 million SAM. That's the nodal revenue increase, so that's for every 100,000 wafer starts per month that we'll see.

And also, the second and most -- also very important is the growth in DRAM. DRAM as it transitions from 6F2 to 4F2 cell architecture, and also importantly, the peri-CMOS, which used to be planar high-k/metal gate will transition to FinFET high-k/metal gate and that will drive a huge opportunity in SAM increase. So combining both those transitions or the inflections in DRAM technologies, 6F2 to 4F2 as well as peri transition from CMOS planar to CMOS FinFET, we see a total SAM increase of $400 million to $450 million opportunity for us as nodal revenue growth. This will be over multiple nodes from D1d to D1b as different companies transition over the time frame, but that's the opportunity that we see in front of us.

Also, Hichem talked about advanced packaging. Advanced packaging is another growth area for us that we are targeting, especially because we see with the advanced packaging road map, moving to finer pictures in the TSV and finer structures. We see a lot of opportunity where chemistry innovation and surface preparation, where ASM has excelled at in the past in the leading-edge technologies of logic/foundry. They come into play in the advanced packaging as well going forward. And with that, where we excel, we intend to fully utilize those opportunities and go after those opportunities in advanced packaging as well.

So now let's take a look at industry. So let's dive deep into how the semiconductor industry is being driven in the next few nodes. So we all know about the AI transformation that is happening, the growth of AI, how the market is developing for that. But if you look at the underlying technologies of the data centers that AI is building and all, it's essentially the leading-edge logic/foundry technologies and the memory technologies that comprise into the -- that go into these data centers. And if you look -- further zoom into the -- what is driving those things, it's the logic technologies and the memory, especially DRAM technologies that drive these markets going forward.

So if we zoom further, these technologies have now -- so what is the building block, the smallish building block that is driving the biggest AI revolution is essentially the monolayers that compromise of this gate-all-around devices, the DRAM transistors, that is going to be the driver behind all these markets. And that's where leading-edge logic/foundry and DRAM technologies are becoming very important to fuel this growth going into 2030.

So with that in perspective, Hichem touched on this again, the market is essentially focused on -- 70% of the market is focused on leading-edge logic/foundry and memory technologies. And by focusing on these areas, which are in line with where ASM has been predominant is essentially placing us for growth in the highest growth areas of the market as well. So we are well aligned from that perspective to march down the growth path. So now let's look at briefly the logic technology road map. So I'll briefly spend some time on the logic technology and the DRAM technology before going into the opportunities.

So in logic, FinFET was introduced in 2012 into high-volume manufacturing. And then it has given a few generations of it for the last 12, 13 years, which means that there was some dimensional scaling going on which helped with the technology forward. And now we are in the gate-all-around device being introduced this year, 2025 in high-volume manufacturing. And gate-all-around provides better scalability as well as better performance, hence, the transition to gate-all-around.

We see few more generations of that happening from 2 nanometer to 1.4 nanometer to 1 nanometer entering into sub-nanometer or the Angstrom era of technologies going forward. And then the last part is the CFET where the horizontal scaling completely is hitting a plateau and then we go vertical with respect to the CMOS structures being placed on top of each other. And that drives -- that is called the complementary FET or the CFET, which will happen in early 2030s. So if you look at these transitions from FinFET to gate-all-around to CFET, these are not just small transitions in geometry. They present a leap of complexity in process integration, leap in materials that needs to drive this as well as geometries. So all these transition present a big inflection point for opportunities, especially when you are focused on atomic scale process where every monolayer matters.

So similarly, in DRAM technology, 6F2 architecture is the current architecture. It's more planar architecture with capacitor built on top of it, bit-line and world-line are more in the planar dimensions with capacitor on top of it. And also most importantly, shown is the top part is the DRAM transistor with the capacitor. The bottom part is the periphery. They are not on top of each other. That's why it's intentionally shown them not being connected. They are next to each other on the same plane. But going forward, we'll transition to 4F2 architecture, but also at the same time, these 2 DRAM transistor and the CMOS peri being next to each other will come on top of each other.

And that drives a lot of new opportunities for us, and that also enables the transition from -- in the periphery of transition from planar to FinFET devices going forward. And that will continue for a few generations before 3D-DRAM, which is equally exciting opportunity from a deposition point of view when the 3D-DRAM comes into picture, and that's more like a 3D-NAND where deposition will play a huge role going forward when that happens in the 2030s.

So with that said, now let's move into the markets. So this is a graph with -- sourced from TechInsights, and this is the basis of our modeling that we have going forward for the numbers that I will show in the rest of my presentation. So the most important thing to take away is USD 155 billion market in 2030 for wafer fab equipment.

And the wafer fab equipment revenues -- wafer fab equipment spend by segment is shown here. Again, logic/foundry showing the largest growth and with sustained growth in DRAM as well. So before I go into the future, let's see where we are baselining ourselves right now with respect to the ALD market as well as leading-edge Epi market, and Hichem touched on this. We are at greater than 55% of ALD, single-wafer ALD market share now. The total single-wafer ALD market share, we have more than 55% of it in 2024, and we have 25% of leading-edge Epi market. So that's leading-edge logic/foundry as well as DRAM. And then the most important part here is essentially we have maintained, and this is very important, we have maintained our leading ALD market share in transition from FinFET to gate-all-around.

And that's a step function increase in ALD, and we have maintained our lead in the ALD transition from FinFET to gate-all-around. And we also grew Epi market share. And the key point there is that we are now the DTOR or PTOR for all GAA channel layers, whereas only one company was using it previously in FinFET. Now every company uses channel layers, and we are the PTOR for every -- one of those layers.

So now leaping forward, let's dive a bit more into the technology inflections that drive the single-wafer ALD and Epi markets. So the market we see as growing from $3 billion to $5.1 billion to $6.1 billion in 2030 with a CAGR of between 9% to 13% against a WFE CAGR of 6%, as I have shown in the previous slide. The growth drivers are: one, increased number of layers in the technology. So as you may go from 2 nanometer to 1.4 nanometers, the number of layers increases. And similarly from 1.4 to 1 nanometer, the number of layers increases. Same for DRAM inflections that I talked about, and we'll touch on subsequently as well in the next few slides. But also most importantly, as the geometries become more complex, the deposition complexity also goes up. So now it's a compounding effect between the number of layers going up and the complexity of deposition going up. So it means there's more opportunity basically for every single layer.

And by 2030, our outlook is to -- our target is to maintain -- still maintain a market share of greater than 55% in the single-wafer ALD market, while maintaining leading share in logic/foundry, but also more importantly, going forward, gaining share in memory, specifically DRAM. Now looking at the Epi market. Essentially, the market is going to be $2.5 billion to $3.2 billion per our estimate going from -- up from $1.5 billion in 2024. Similar CAGR, 9% to 13%, but the growth now is also coming from not just GAA applications, but also in DRAM going forward. So DRAM from 6F2 to 4F2, there are opportunities -- from planar to FinFET, there are opportunities for Epi. So those are the things that will drive the market expansion for Epi.

So if we look deeper into the logic/foundry segment, we have -- first of all, I want to baseline, going from FinFET to gate-all-around 2-nanometer device, we have forecasted back in 2023 that we'll have a SAM increase of $400 million, and we can now confirm that with 2 nanometer GAA ramping across all customers, we can now confirm that our SAM -- nodal SAM increase is within that range. So positively delivered on our forecast from back in 2023. And I would like to reemphasize, we at least maintain share in ALD and gained share in Epi through this transition.

So we have -- with the opportunities increasing, we are also increasing our market share, maintaining the share in ALD and gaining share in Epi through this transition. And if we look at the forward-looking from GAA to 2 nanometers to 1.4 nanometers, we are forecasting from our analysis, an additional USD 450 billion to $500 million nodal SAM increase. So it's incremental to the $400 million SAM increase that we saw from FinFET to GAA, and we see the continuous transition -- continuous growth of opportunity in the transition from 2 nanometer to 1.4 nanometers going forward.

And these are the number of all the layers that are included in this analysis. It's either higher number of layers, it's more complexity included or it's the same layers that are included here and also processing on the wafer frontside as well as wafer backside. If you look at what this means and where the growth is coming from, let's just look at the 2-nanometer is on the left side, the 1.4-nanometer fabrication steps are shown on the right side. This is the GAA process that we are seeing right now. And then that is the most complex, as Hichem said, the smaller the feature sizes, the greater the opportunity for us. And once that is in front end of line, there is a lot more complexity there, more ALD layers being added.

And once the -- once you proceed beyond the contacts and the first level interconnects, it's spacious, wide features, the same things get ported over from the previous technologies and goes on. So the action really happens at the transistor level where there is no space and you need more complexity introduced to make the things work. And you just saw that the wafer got flipped after the frontside processing is done. And then the backside power is put in place, and now the wafer is flipped again for the backside contact. So that's the final structure, and this shows how the 1.4-nanometers is, and I have all the layers listed for what's changing from 2-nanometers to 1.4-nanometers. Again, not everything is new. It's increased, more complex or new layers, everything is included in this analysis.

So if we look at the layer -- ALD layer count by node, just to give you guys what is in this analysis. So from 2-nanometer to 1.4-nanometer, we still see equivalent growth of ALD layers, just layers count this is. And the mix of layers is essentially 60% still happening in front end of line. Front end of line is basically from silicon wafer all the way to the first contacts and then everything else is either middle of the line, back end of line or backside, and that's about 40% of the total layers. So then again, if we -- when we go into the third generation and CFET, we see continued growth in ALD and Epi intensity. I'm not going into specific details right now, sticking to the time horizon that we have until 2030.

Similarly, we see DRAM, a lot of opportunities, accelerated opportunities in leading-edge DRAM and that's essentially driven by what's happening with the inflection. So this is the 6F2, as I mentioned before, the top part is the cell -- DRAM cell and the capacitor, the bottom part is the CMOS peri. And if we just focus on what will happen, the first inflection that is going to happen is essentially the transistor part is going to change from the 6F2 to 4F2. That is essentially going vertical in DRAM cell, essentially, the DRAM transistor. So which means you will have to create a channel that is conducting, which can be grown Epi-channel. And then you will also have ALD layers for the bit-line and word-line deposition and essentially additional process for channel Epi, contact Epi, ALD oxides for low temperature because now we are doing capacitor first.

ALD dielectric gapfill to make the structure stand in the vertical fashion that they need to and also new metals that will come into play in back gate and front gate devices. So the DRAM transistor is fundamentally changing as you go from 6F2 to 4F2. And the next set of transition happens from planar to CMOS. This is something that we have lived before in the logic technology. And that was when actually ASM started to gain a lot of market. When that happened in the logic technology and that's -- those are the same opportunities that are now coming back in the DRAM with the transition from planar to FinFET CMOS in the periphery. So if we put all that together, the -- for the just the FinFET -- planar to FinFET, these are the opportunities, same material, same set of materials where we have a large market share with high-k, dipoles, work function metals, patterning materials and also Epi comes into picture.

So if we put all that together, what is happening, so if we take the 6F2 with planar last node and if we put the opportunities together for both the 4F2 transition as well as the planar to FinFET transition in the peri, we see that we go from -- where we are in D1d to about adding another $400 million to $450 million in nodal SAM. So that's where the number comes from that I shared earlier. And then in the longer term, as we go to 3D-DRAM, this takes another step function increase in the ALD and the Epi intensity to make these structures going forward. So we see, again, a large increase of opportunities for ALD and Epi intensity moving from 6F2 to 4F2 and even larger in the future past 2030 when we transition to 3D-DRAM.

So last but not least, advanced packaging. So this is something that we have set our strategic objective for mid-term growth. Hichem touched on this. We have our current SAM of -- if you mix all the CVD, surface prep and ALD layers, it's about 15%. And what is that -- why are we so excited about it? If you see the -- this is a typical application of advanced packaging for a system on integrated circuit. So this is a typical. There are many different flavors of this application. I chose one typical thing to show you where the opportunities are. So it's essentially the bonding layers, gapfill, die edge fill, which is where some of the flowable carbon applications that Hichem talked about could have an impact, TSV liners as the TSV scale, ALD is what is needed to make those structures happen.

So based on the road map of advanced packaging, which used to be very large features, not so easy to do structures but to more narrow structures and difficult aspect ratios, et cetera, and also needing some functional materials to make this happen going forward. We feel very excited about the opportunities here. The surface preparation, the materials innovation, they all come into picture in the advanced packaging as well. And putting that together, the growth in our current SAM as well as the new opportunities that we are looking at, we intend to double our SAM to greater than 30% by 2030. So that's our objective.

And then I just leave you with the key takeaways again. Very excited about what is in front of us, both on a secular big picture end market, semiconductors wise, the growth is going to be there. And also for ASM where we are leading in the logic -- leading-edge logic foundry and the new opportunities that are coming from DRAM side, which plays into our strengths more or less. It drives us towards a good growth area and a lot of continued opportunities as we look forward to 2030 and even beyond, it's even better. So that's all I will leave with here, and I will hand over to my colleague, Jason Foster, to talk about our innovations in services.

Thank you, Vamsi. Good afternoon, everyone. It's a pleasure to meet you. My name is Jason Foster. I'm the Corporate Vice President and Head of our Spares and Service Organization. I'm also responsible for the Global Quality Organization at ASM as well as the Technical Training Center at ASM. I'm very pleased to speak today to you about how we're enabling our customers with Angstrom precision through service innovation and automation.

Let me start off with our key takeaways. So there's one thing -- if there's one thing that I want you to take away from our key takeaways is innovation. Innovation is absolutely required for supporting our products in high-volume production in the latest generation technology nodes that our customers are putting into production today. So we're bringing innovation to the spares and service business such as outcome-based services, and these solutions are creating measurable value to our customers and bringing -- which drives growth to ASM. How are we doing this? We're leveraging our core competencies in chemistry and surface engineering. So ASM is widely known throughout this industry as being a leader in ALD.

We didn't get to that position in leadership in ALD without having core competencies in chemistry and surface technologies and surface engineering. Those are key factors in developing those products. We have a long history in that area. We're now leveraging that capabilities and that know-how in developing our solutions that are used for outcomes-based services in the spares and service business. So these outcome-based services, as Hichem mentioned, really, our focus is guaranteeing tool performance to our customers. We have the know-how on how to keep those products into production. We now have the know-how of how to deposit those films on wafer in high-quality, and we're using that to -- that know-how to guarantee our performance going forward, key performance metrics such as tool availability and uptime, we're improving on wafer performance, and all that through driving innovative, environmentally friendly solutions.

There's two key products that I want to introduce to you today. One is a new dry cleaning solution. This new technology enables 10x selectivity compared to typical sand blasting and wet cleaning solutions that are used in the industry today, all while driving a sustainable manufacturing solution and automation. Automation is becoming more and more critical as we drive towards serving the technology nodes that Vamsi and Hichem just introduced, right? As the complexity of these applications become more and more critical to achieve Angstrom-level control in ALD and epitaxy, it requires micron level precision in part placement within our reactor, and automation is the key to that. So I want to touch on our vision that drives the way of working within my organization.

So we're empowering our customers with innovative, environmentally friendly solutions that deliver high tool availability and better performance to our customers. This is really what guides the guiding principles that guides the way we work on a daily basis. Why is innovation required? So as Vamsi and Hichem had indicated, our customers are putting more and more complex devices into production on a regular basis. Very complex architectures such as gate-all-around, the 2-nanometer technology node place very stringent hardware requirements on our reactors. So parameters such as surface roughness, parallelism, planarity, hole dimensions, all of those are critical parameters on our parts that deliver the performance that's required on the wafer to meet the requirements in gate-all-around.

Simply put, our customers cannot run high-volume production without having a sustainable solution for maintaining those parts throughout the life of the part. We have two key areas of innovation that I want to introduce that we'll touch on in this presentation. Dry cleaning solutions, we're implementing a new technology for cleaning. Hichem touched on it, right? The new technology is a dry cleaning solution that is eliminating the need for blasting -- sand blasting and wet cleaning solutions. So we're going to deliver superior defectivity performance, improved selectivity and driving better sustainability by enabling -- extending the life of the -- of our critical process kit parts to much longer.

And then automation. So we're levering advancements in automation to enable the precision that's necessary to operate in these advanced technology nodes. So I want to introduce next the two areas of our business in the spares and service organization. For the history of our company, we've been focused on transactional-based products, and this is having the right parts at the right place, at the right time, to support our customers in high-volume production. Also having the right service engineers in the field to support and sustain our products. So that's our transactional-based product lines, and that's been the foundation of our business for many years.

However, as mentioned, we are transitioning more and more of our business towards outcome-based business. This is where we're introducing guaranteed performance. So again, we're guaranteeing the output of our products by reducing -- we're committing to reduce variation and predictable output to our customers. And it's where we're introducing the sustainability concepts of reduce, reuse and recycle in our business. So let me talk next about innovations in outcome-based services. So I have an example here how we're creating value through surface technologies. Typically, in an ALD reactor, we're depositing film -- an ALD film on the wafer. That is the value add that we bring with our reactor.

But in doing so, we're also depositing a film on -- we're also depositing film on the walls of the reactor. So I show here a cross-section of an ALD reactor from the life cycle of this reactor when we have new parts, parts are pristine. There's no deposition on the walls. But as we process thousands of wafers, we build up accumulation on the walls of the reactor, and that degrades performance over time. So I show here an example of defect performance. Defect is a critical process parameter that impacts device yield for our customers. So as the life cycle of the reactor goes, defect performance starts out very stable, but then we run -- we typically have an in-situ based clean that restores defect level back to baseline performance. But gradually, over time, this performance degrades to a point where we can no longer meet specs. And at that point, that's where the outcomes-based services comes to play.

This is where we take -- we -- our customers go into a maintenance event. We take the chamber down out of production, and we take the dirty parts, the dirty process kit parts out of the reactor. We implement our cleaning solutions, our kitting and testing of those parts to bring them back to as new conditions. So we have -- the benefits of outcome-based services is we have improved on wafer performance. We're improving the defect control, thickness uniformity control. Not only that, we provide better tool availability. So our customers benefit from higher operational efficiency within their fab by implementing outcomes-based services. And it improves sustainability. We're able to use the parts longer and reuse them over and over again.

Overall, this lowers the cost of ownership for our customers by enabling the reuse of these components over and over again. Let me talk about new innovations in our surface technology area. So I'll talk about dry cleaning more in the coming slides, but I also want to indicate that ALD coatings. ALD coatings is a -- we are the leader in ALD deposition on wafer. We're taking that know-how in developing ALD coatings that can be placed on our process kit parts to extend the life and to enhance the performance of those parts. So we're taking that leadership in ALD wafer deposition and enhancing our capabilities at depositing ALD films on parts.

Surface modifications. So it's very critical that our parts maintain pristine surfaces to improve the performance on wafer. Adhesion control is very critical. Hichem touched on it in his presentation briefly that how well the surface roughness of our parts are maintained, controls the defect performance on wafer. So we've developed a new dry surface modification technique, that allows for very uniform surface roughness modifications of parts that is much better than bead blasting and blast technology today.

Thermal uniformity control or thermal control is also critical. In an ALD reactor, thermal uniformity on the wafer drives the performance of the film that's deposited. Our new surface modification techniques can greatly improve our ability to maintain the emissivity, uniformity within our parts. So these are 2 key areas, adhesion control and thermal control through surface roughness control and emissivity control of our parts that's being implemented. Not only that, we have to guarantee the performance of these parts after we've cleaned and modified the surfaces. We've developed novel metrology techniques such as measuring the flow uniformity or thermal uniformity on these parts. It allows us to guarantee the performance going forward.

So again, our outcome-based services is minimizing part-to-part variability, preserving process integrity and extending the life of these components. Let me give an example here of our new dry cleaning solution that is bringing significant business benefits as well as sustainability benefits to our customers and to ASM. So I have an example here of the environmental impact. So if we take -- simply take a dirty part out of a reactor and replace it with a new part, we're not recycling it. It's equivalent to 1,700 tons of CO2 emissions on an annualized basis in a 500 kit example. If we look at implementing wet clean. We dropped that down to 85%, simply by recycling the part. But then our new novel dry cleaning solution is, again, it's eliminating blast media and cleaning solutions that are necessary. We're able to drop that down 67% further from our wet clean by utilizing our dry cleaning solution. So it's a huge improvement from a sustainability perspective.

But not only sustainability, really, the big improvement here with our dry cleaning is on selectivity. Selectivity is important. If we take this example, it's a titanium process kit part. In many of our ALD reactors, we're depositing titanium nitride or titanium oxide type film. Selectivity refers to how we remove that film while maintaining the base material without impacting the base material. So we want to just remove that film without removing any of the base material. If we remove the base material, we degrade those critical parameters of the part. So we degrade the performance of the part if we impact the base material.

Selectivity allows us just to remove the material we want to remove and not touch the base material. So we're greatly enhancing the selectivity by utilizing our novel dry cleaning solution, which allows us to extend the lifetime of the parts by 5x. We have better dimensional control of the parts. And there is no hazardous chemicals that are used by implementing this dry cleaning solution. Again, the sustainability impact is greater than 95% compared to [ none. ] And we have a great improvement in business impact by -- we're lowering the cost of ownership to our customers by more than double.

If we look at another example here of where the dry cleaning improves. On the top, I'm showing a video of sand blasting. This is where we use sand blasting to remove the byproducts from parts. Following up on sand blasting, we need to wet clean using acids and DI water to clean the part. The graph on the right here shows how using blasting media and wet clean, we're only -- we still have micro-level contamination, sodium, as shown here in the graph, which impacts the part performance. So we have to spend a lot of time clean and recleaning these parts after blasting them clean. With our new dry cleaning solution, which is shown on the lower -- the video on the lower right is we're able to achieve very clean surfaces after removing the material and not requiring a wet clean afterwards. So it's a very effective cleaning solution that delivers a very pristine part post cleaning.

So our new dry cleaning, the main thing is we're enabling precision selectivity, extending the useful part lifetime by maintaining the critical dimensions clean after clean. I want to touch on the reason why selectivity is even becoming more and more critical, as we transition to more complicated ALD films. Today, we're using single element ALD films. We're transitioning to 4, 5 or even 6 element films in the future. And simply put, what chemistries cannot deliver the selectivity that's necessary to remove those type of films. So that is where this dry cleaning technique is becoming absolutely critical for our future ALD films and of course, the sustainability impact by extending the life of these parts. So highlighting a little bit more on how we're accelerating the sustainability impact by implementing reduce, reuse and recycle concepts.

So as I mentioned many times in my presentation, we are reusing parts by implementing our novel cleaning solutions, but also the byproducts that's removed from those parts can be reused. An example is silicon carbide. We use -- we have silicon carbide reactor. We have to remove silicon carbide from the parts on the reactor. That silicon carbide material can be reused in the industry. Silicon carbide is an abrasive that's used in industry -- widely used in industry. So that's an example where we can reuse those byproducts that come off of our parts.

Recycling. Of course, recycling parts, recycling the scrap that happens during machining of parts, et cetera. Reuse -- or reduce chemicals, DI water, blast media and bulk material. All of this is possible -- made possible through our dry cleaning solution, our high-quality coatings and our high selectivity cleans and advanced metrology.

So next, I want to talk about automation. How and why is automation important? I touched on it earlier. Automation and maintenance is needed to deliver micron-level part placement control. And why is that necessary? It's necessary because of the advanced generation devices that our customers are using. We can no longer support that through our highly experienced field service engineers that place these parts by hand. They cannot get the accuracy that's necessary to drive the Angstrom level precision that's required by our customers. I show here in the video, an example of an epitaxy reactor, where we're using a PM service bot. This is a robot using closed-loop vision system to place parts precisely in the reactor. Again, we're looking at driving efficiency in the fab, but not only in the efficiency, we're looking at how do we improve the performance of the reactor by better precision in part placement.

So again, achieving micron-level placement precision is required for the Angstrom-level control that's required for the processes that we serve. Next, I want to touch on -- I'll touch on our revenue growth in spares and service sales, and it's coming from outcomes-based services. So we're targeting continued growth in spares and service business, greater than 12% compound annual growth rate from '24 to '30. And that -- of course, we have a growth of our installed base that's going to help us get there, but that growth of our installed base is coming from our newer products. As we're selling more and more products to serve the advanced technology nodes, those require more and more of our outcomes-based services. So we have higher market share of outcomes-based services for our newer products that's going to be sold in the future.

And by 2030, we're expecting more than 50% of our business in the spares and service organization to come from outcomes-based services. So that's a summary of my presentation today. If you come away from anything, it's -- I really want to emphasize the importance of innovation in the spares and service organization to bring the solutions that are necessary to allow our customers to operate in high-volume production. We're leveraging our core competencies in chemistry and surface technology to create outcomes-based services. We have 2 key areas of focus for new products that we've developed, our new dry cleaning solution and advancements in automation. Thank you very much.

Thank you, Jason, for everyone on the webcast. At the start of the program, we experienced a technical issue. We are very sorry for this and thank you for sticking with us. So now we will have a break of approximately 20 minutes. We will continue with the second part of the program at 3:20. And for our guests here in the room, please join us for a coffee or a tea in the foyer. Thank you.

[Break]

Welcome back, everyone. We will now continue with the second part of our program. The next speaker will be Eric Shero. Eric?

Good afternoon. Welcome to the art of atomic layering. My name is Eric Shero. I'm Vice President and ALD Key Product Unit Head for [ MX films. ] With the short time we have together today, these are the key takeaways that I want you to leave with. One, the ALD is an essential technology for advanced 3D structures. You really can't grow these -- you really can't make these devices without ALD. In growth, as we've already mentioned, ALD is growing at a tremendous rate, and we expect it to grow over the next 5 years at double the WFE broader market. ASM has always been a leader in logic/foundry inflections, and we plan to expand our presence in the memory. Legacy of ALD at ASM is strong.

Basically, we can trace back our roots almost 50 years to the invention of ALD. And we continue to innovate and keep our ALD technology ahead and staying ahead of what's next and our competitors. We have a new common platform, which has been mentioned, which drives enhanced clustering and productivity. It allows us to couple our surface clean technology and deposition solutions. And last, ALD+. ALD+ means advanced materials, new chemistries that are differentiated and unique and technology solutions to tackle our customers' most high-value problems.

Okay. So as we look at logic and memory road maps, the one thing you'll see is that the complexity of the device is increasing tremendously. Everything is exploding up into the third dimension with increasing aspect ratios, increasing surface area enhancement, all while making the CDs smaller, which makes it much more difficult to get chemistry in and out of these structures to satisfy that enhanced surface area. So ALD is becoming even more important. Also, there's a narrowing process windows. Our customers are requiring tighter thickness uniformity, tighter composition control. And of course, node-over-node, the electrical specifications get much more challenging. So all of these things are trending toward ALD. ALD is what's going to be needed to move technology forward. And of course, we think our brand of ALD is superior.

So as been mentioned, ALD adoption is increasing. So for real-world illustration of this growth, I think we can look at a familiar leading-edge consumer product in Apple's iPhone. So we all kind of generally know from iPhone generation from one to the next, what the capabilities are, how it's improving node over node or generation over generation. The one thing you can see is that the ALD content packed into that phone is increasing as well at actually a very tremendous rate. If we look at what happened between the iPhone 13 and iPhone 15, we packed about 50% increase in ALD films inside of those 2 devices. And as we move toward the newly released iPhone 17, we expect it will also have double-digit growth in terms of ALD content.

So I want to make sure everybody is on the same page when we start here. So what are the key steps in the prototypical ALD cycle? Well, first of all, just remember, ALD is a surface controlled layer-by-layer process that deposits thin films, basically one atomic layer at a time. So interfaces are very important. Surfaces are very important. So what we need to do is we need to start with a controlled surface. So in step 0, really before we even begin, we want to make sure that the film -- the substrate is properly functionalized. So it has the right groups on the surface for ALD to basically nucleate and grow properly. Now we start kind of the normal steps. So step 1, we exposed the substrate to a first precursor or to a first precursor pulse. That precursor reacts with the surface, chemisorbs. So it binds through chemical bonds and then it releases byproducts.

Step 2 is repeated twice. It's very important. It's the purge step. Now why is this important? Everything is occurring in the surface. So why is the gas phase important? Well, the gas phase is important because the way we index the precursors and the reactor, we need to keep them separated spatially and temporarily to make sure they do not react with each other in the gas phase and cause parasitic CVD. So that parasitic CVD is bad. It takes you away from ideal ALD character. So we want to make sure that when we design our reaction chambers, they're designed to have a very efficient purge and turnover of these gases.

If you go to step 3, we expose the substrate to a second precursor. This is also called the co-reactant and we convert the surface through this step. I want to note here that this is where thermal ALD and plasma-enhanced ALD can differ from one another. If we use plasma energy, whether it be radical or direct plasma form. And the reason you do that is to lower the overall deposition temperature or improve the quality of the film by using plasma. So step 4 is again another reactor purge step. We want to make sure that we get rid of the second precursor and any of its byproducts, and that completes the first ALD cycle. So what we do is we just cycle this over and over again to grow a thick film, essentially one monolayer at a time. Okay.

So ALD is geared for a 3D world. Now why do I say that? To understand that better, we need to look at the kind of the competitive thin-film deposition techniques. The first is PVD. PVD is highly directional, it's like when you play billiards, you hit a ball, it goes in a straight line until it interacts with something else. That's what PVD is like. So PVD, the material from the target travels to the wafer, and it tends to be deposited thicker on horizontal surfaces to the direction of motion. So the top of the structure, that uses shaped structure, they can get thick and then the very bottom. And so it's not what we call a conformal film. It's not covering the surface everywhere 100% uniformly.

Next up is chemical vapor deposition. So this is less directional than PVD, which is good. It's fast. But the problem is there's kind of a chemical cloud over the wafer, and your film is essentially dropping out of that chemical cloud and depositing on your substrate. So in this case, we do get some film everywhere. We get it on the sides, we get it on the bottom. But again, the surfaces that are closer to this chemical cloud tend to get thicker, and we still have what we call non-conformal growth. So it's like frost on a leaf. It tends to -- it kind of condenses everywhere, the frost, but it's much more heavy on the top side of the leaf.

Then last, that brings us to ALD. So it's sequential. It's self-limited. It's nondirectional and it's very conformal. So really, we're basically creating a wall, one brick layer at a time. We're stacking it to create our overall film. So you can see why ALD is really geared towards these advanced 3D structures. Okay. So ALD is an indispensable 3D technique. And to be honest, it's becoming so ubiquitous that we kind of lose track of how amazing it is as a process. So it's not atypical for us to deposit a 1-nanometer ALD film completely uniform across the 300-millimeter wafer to within 0.1 angstroms thickness uniformity. And we do that not only on the wafer surface itself, but on billions and trillions of high aspect ratio structures that you can't even see with the human eye.

So for some people, it's like, no, I don't understand what 1 nanometer is or 0.1 angstroms. I don't really know what these things look like. It's not that impressive. For context, put it in everyday dimension that's equivalent to covering a modern city like Dubai, with a 5-millimeter coating over everything, every structure you can see, including the Burj Khalifa, the tallest structure in the world manmade, or a city even full of them. We can do that 5 millimeter coating to then the accuracy of a human hair. So basically, what that means is perfect conformality citywide over all the structures, I think that's amazing.

And in 3 years' time, if that's not impressive enough, the device aspect ratio will exceed stacking the equivalent of 50 Burj Khalifa towers on top of one another going from ground level all the way into the stratosphere and ALD will allow us to deposit that 5-millimeter film all the way conformally. So as Hichem said earlier, a lot of people ask ALD or ASM, why are you the market leader in ALD? What makes you stay ahead? How are you going to stay ahead of your competition?

It really comes down to five things, five take aways. First of all, we are -- we have a 50-year legacy in ALD starting from 1974 with the invention of ALD by Dr. Suntola through our key acquisitions of Microchemistry and the Pulsar reactor through other technology acquisitions in the plasma enhanced ALD space, all the way through to today, where we're introducing DCM and QCM chambers like the Prominis and integrating those onto our advanced XP8E platform. So 5 decades of innovation in the ALD space.

Second, although theoretically, ALD is very straightforward, I'll just explained it to you in maybe a minute. It has challenges like any technology. And those challenges can be divided in kind of three sections. So first our reactor challenges. Basically what -- how we design our products, how we design our reactor. There are a lot of things that can go right and there's a lot of things that can go wrong, and you have to be very cognizant of what's going on and make sure that you design things properly.

Precursors. The precursors have to work in collaboration with the reactor to deposit these ultrapure high-precision films. And then the services we deposit on, these are somewhat outside of our control, so we have to take what the customer gives us and we have to turn it into something that we can control and do very good ALD film deposition. So how are we different?

Well, first of all, our reactors are conceived for ALD. We're not re-purposing a CVD reactor and trying to make it an ALD tool. Second, our reactors have very small volume, and we have different types of reactors, cross-flow reactors like the Pulsar and shower-head type reactors. Third, we integrate our pulse valves directly proximal into the reactor space. Now why is that a big deal? Because that's just extra volume, you want to minimize that. You want the valves as close to the deposition space as possible. And then last but not least, we have Tykon EVC plasma control, which we got from our acquisition with Reno Systems. And that has ultrafast impedance matching which gives you less variation and less damage to the services that you're trying to deposit on and around.

If you look at next, our precursor innovation sphere has been very well established. That means we have the best network of in-house and external partnerships in terms of chemical scientists, and they develop on real-world equipment. So when they're developing a process with new chemistries, they don't want to dump it on a lab scale reactor or a bench-top reactor, they develop it on our actual HBM hardware, which is co-located with these scientists.

On the tools in terms of precursor delivery, we offer the best temperature uniformity. So you have to make sure that you don't get the lines to -- the gas lines too hot because you'll start to decompose your precursor, which leads to Parasitic CVD or if it's too cold, it will condense and that also causes issues. So you have to be in that Goldilocks zone, where you have a very tight temperature uniformity.

And then last, on our equipment. We have the highest number of chemical sources. So that enables our customers to develop unique multicomponent processes. So some of our competitors have 1 or 2. We have 3 or 4 chemistries, which is really necessary for some of these advanced DRAM capacitor dielectrics and oxide semiconductor materials. Last, in terms of integrated solutions, we need to condition the wafer surface, and we need to control the environment around it. And the way we do that is through a suite of clean treatment and deposition products, all integrated on the XP8E platform, which is our newest and most advanced.

ASM has the largest and broadest ALD product portfolio in the industry. We cover plasma and thermal deposition in single, dual and quad form factors, addressing the range of applications across various fab volumes. And our new XP8E platforms are built on a foundation of AI and machine learning, and that allows our customers to accelerate their R&D development times and then also deliver HBM operation, which is key. It's not just getting customers, but it's maintaining those relationships through the quality HBM systems they can rely on.

Next, Okay. There's been tremendous growth in the materials and elements that can be accessed by ALD. There's essentially no limitations now versus in the past where PVD could deposit metals that ALD couldn't, or CVD could access certain materials that ALD couldn't. There's no limitations anymore. Also, ASM has developed reference processes for 70% of the elements cited in the literature to date in the ALD literature. So we have these [ BCAMs ] that we can offer to our customer at any time. And the way we've achieved this is through a strong R&D collaboration with the University of Helsinki and our chemical innovation network worldwide, with, again, co-located with our HBM equipment.

As Hichem mentioned earlier, we have a very prolific impactful IP portfolio. That's not per us. That's per [ LexisNexis ]. And one of the other things we have to mention is that when IP ages out, we have 1.5 to 2 patents ready to take their place. So we're continuing to innovate in our patent portfolio is growing in size and impact.

Okay. So we've heard a lot today about how sophisticated the gate-all-around transistor is, and these are very challenging. And with each generation, it gets more challenging. Not only geometrically in terms of the space, the spacing that we have accessible for our chemistries to enter the structure, but also in terms of more stringent electrical requirements. So this trend is going to lead to a few things.

First, it increases the complexity of our existing ALD films in the gate-all-around on structure. So as Vamsi said earlier, so we have high-K films. We've had that in FinFET. Those are positioning in the gate-all-around structure, but they're getting more challenging because of these physical constraints. But more importantly, it's spurring the growth of new ALD functional materials and also more passives of sacrificial ALD materials that are needed to pattern those and place those directly where you want. So it's just an explosion of both in terms of ALD opportunities. And of course, we're innovating our products to go after the technical needs of our customers in certain key areas, that being MIMCAP, as Hichem mentioned, where we have metal electrodes and high-k dielectric deposition, in the gate all-around transistor area with higher-k gate oxides. Vt tuning materials like dipoles and work function metals, we're also going into molybdenum metallization, silicides, Hardmasks, Etch stop layers, and more importantly, in these structures, gapfill materials are becoming even more critical.

And last but not least, area-selective deposition processes, which can simplify the integration of our customers and also deliver some technical differentiation that they could not otherwise get with traditional patterning schemes. So more ALD materials, more new materials, more passes needed with gate-all-around scaling.

In Memory, the transitions are also driving ALD adoption. So as Vamsi mentioned, DRAM is scaling from 6F square to 4F square to much more compact design. In the 4F squared cell, the capacitor, high-K and the liner are very critical layers that must absolutely be deposited by ALD because they're high aspect ratio. They need to be very pure and conformal and they need to deliver very low leakage and high capacity and density. Also, the capacitor of metal electrodes, those will transition over. But interesting, back gate and front gate will require metal more than likely molybdenum. And then dielectric gapfill again is important. There's an opportunity in 4F squared or DCT for new channel materials deposited by ALD. Those are oxide semiconductors. And of course, those oxide semiconductors need new gate oxide materials as well. So a lot of opportunities in the cell.

In the CMOS area, for 4F square, they're transitioning to FinFET, which is obviously one of ASM's great strengths. We're the leader in FinFET technology, even logic foundry, and we expect to enable our DRAM customers similarly with High-K, dipoles, work function metals and then again, patterning materials and spacers.

In 3D NAND, stacking is continuing vertically. So of course, that means there's even more of a need now for gapfill materials, and I'll speak to that in a minute, even more critical. And then the peri transistor and 3D NAND will be transitioning from 2D SiO2 to high-K hafnium oxide. And of course, ASM will be there to support the customers' transitions in NAND. So a lot of opportunity in the peripheral circuitry as we go to 4F square and also 3D NAND.

So ALD is essential for 3D scaling, both functional and gapfill materials, and we feel like we excel in solving a lot of these high-value problems along with our customers. Unfortunately, today, I don't have enough time to go into every film and every application where ALD would be required, but I do want to kind of cherry pick a few to kind of do a deeper dive and showcase them.

Okay. So as we mentioned, gate-all-around complexity is increasing and it's only going to be met with cutting-edge ALD. So our customers want to access more Vts. So you can see from 2016 to 2027, number of Vts is increasing. Now why do they want to access more VTs. VTs stands for threshold voltage of the transistor. And basically, it means at what voltage will the transistor channel open and device turn on. And the reason why customers want as many VT levels as they can possibly pack is they want to be able to optimize power consumption versus performance. So low Vt is for increased performance and high Vt is for reduced power consumption. So they want to be able to target basically every transistor on the device and make sure it's optimized with best blend of performance versus power consumption. So how do they do that?

The best way to do that, the most efficient way to do that due to the density in the limited space in the gate-all-around structure is through dipole technology. And dipole is what we call a 0-volume solution. So first of all, there are only 1 or 2 or 3 angstrom stick as deposited. But after all the integrations done and the dipoles are driven in, they basically disappear. They don't occupy any space. So it's a really nice solution as we shrink the nano-sheet spacing, and we pack more density and dipoles are going to -- are really invaluable, and that's why they're picking up in terms of the number of passes, both P&N.

But the challenge that comes is, we only have a certain band gap in silicon. And if customers want more Vts, that means the Vt levels get very close together. So we need to develop new dipole materials that can span that entire Vt range from band edge to band edge. But then also, we need to be able to control the dipole films very precisely because ultimately, the Vt levels are only separated by 1 or 2 ALD layers. So we need these new dipole materials and we need atomic skilled precision by ALD in order to shift and have the requisite amount of distance between Vt levels because, obviously, if two devices have the same Vt, that's a fail.

Now along with these dipole layers, there's multiple ALD pattern assist layers that are needed to properly place these dipoles. So for every dipole pass, there could be 2 or 3 more ALD passes, which is, again, great for ASM since we're a market leader. So we're advancing our multi VT solutions, with products specifically designed to control these dipoles layers very thin and very uniform and conformal over these complex devices.

So if we want to continue device-scaling the via and trench lines, the resistivity of via and trench lines has to be reduced, and customers have already made the decision to transition to moly in most cases. So our solution is very flexible. We have two different types of deposition, PEALD molybdenum and thermal ALD molybdenum. And basically, we couple these numbers on powerful XP8E platform, it enables low temperature, bottom-up, seam-free and highly productive solutions, non-damaging to the areas around. So that flexibility and deposition technology, combined with our pre-clean chamber Formion -- or Formis depending on whether it's ion or thermal-based basically gives the customers everything they need in one platform to address not only logic applications, but memory applications with very high productivity and very low resistivity as a result.

Okay. Kind of the last key challenge I want to talk about is applicable to both logic and memory customers. It's filling structures with materials such as SiO2, without seams or voids in very challenging geometries. So you could see here, in this case, we're looking at filling a lateral cavity, and of course, the customer wants to have essentially no seams or voids. Now why do they want to get rid of seams and voids. These are weak points in the film. There's a lot of downstream process and that's not shown here, obviously.

So when a seam is exposed to a dry etch or a wet etch, it can open up and when it un-zips, it basically exposes the rest of the device to that chemistry and they can destroy the overall device. So it's very important to limit the number of seams and voids. And the way we've done that up until now is using what we call a kind of more conventional gapfill process, goes through an ALD deposition, then a gas exchange and there's a plasma inhibition step, and then another gas exchange. And basically, what that does is it kind of inhibits the growth around the openings of the structure, so that the chemistries can get further in and kind of start filling from the inside out or the bottom up. And that's ideally what you want to have.

Well, our solution to that conventional gapfill problem is basically to leverage our scientific community, like our worldwide innovation network of chemists. And they've come up with a very innovative novel molecule that has self-inhibition properties. And what that allows us to do is basically to curtail or to shorten the overall ALD deposition process. We don't need that inhibition step anymore because it's self inhibited. And this molecule is amazing. As you can see, lateral gapfill. We eliminate all seams and voids. This is SiO2. You can see in the lateral -- lateral fill is great, its seamless, and then even in vertical gapfill applications like, for example, reentrant structures with narrow CDs or very tall 3D NAND stacks, we can get that same level of gapfill with no voids or seams. This is tremendously enabling for our customers. And the great thing is it's done on a very productive HBM-proven tool set, our Arius deposition system. Okay.

So that brings me to key takeaways. I don't want to spend too much time on this, but just real quickly, ALD is essential. You can't anywhere without it. You cannot innovate. Every customer needs ALD, that's why it's growing so fast. ASMs legacy is really unparalleled. We've been -- we're not a Johnny come lately. We've been engaged and bought into ALD from the very beginning from its inception. You could trace our roots all the way back to Dr. Suntola's invention of ALD. We continue to innovate. That's how we stay ahead of what's next, clustering. We have a new common platform that allows us to put all of our products together and give one solution overall to the customer.

And then ALD+, it means advanced materials, unique chemistries and technology solutions and tackling the highest value problems of our customers. Thank you.

Thank you, Eric. It truly is the art of atomic layering. It's not always easy to follow, but it's incredibly interesting, at least from my point of view. I think this one is maybe more easy for you to follow. So what I'm trying to do here is to tell you what you've heard this afternoon, what does it mean in terms of financial terms. So it's all about delivering long-term value and investing for growth.

A few key takeaways. First of all, we believe our growth through innovation strategy is delivering value for all stakeholders. We have updated our 2027 guidance, adjusted revenue for currencies and improved margins. You've seen that. We also introduced new guidance for 2030. So revenue more than EUR 5.7 billion, outgrowing WFE with a CAGR of 12%, at least 12% improved gross margin range by approximately 100 basis points compared to what it was. And operating margin points increased or improved from 26% to 31% to 28% to 32% and from 2030 onwards, more than 30%.

Operating expenses. So as we've always said, you will see quite some operating leverage and also efficiency and digitization initiatives in the SG&A line. We expect that to be below 7% by the end of 2030, and you've seen a lot of technology this afternoon. So that's why the R&D part, the net R&D part will be low double digit. It used to be high single to low double. We change that to low double. It's incredibly important. We continue to invest in R&D, which we consider to be a lifeline. And I hope with all the presentation you've seen this afternoon that has been -- that we have been able to also give you some comfort that it really is.

Capital allocation policy unchanged. And as we've seen, I think, in particular in Spares and Service presentation of Jason, sustainability is not only a license to operate, but there's also a clear business sense. In '21, we had our first -- ASM had its first Investor Day ever, and that was on the back of 2020. So I want to give you a few data points, what we've been doing in the last 4 years. So we returned cash to you, to shareholders of EUR 1 billion. The accumulated free cash flow was EUR 1.6 billion. Average return on capital was 36.5%. The CAGR was 22%, basically strongly outgrowing WFE. We had an average gross margin of close to 49%, operating margin 27%. Since almost 2 years now 100% renewable electricity, which is a big thing. And that, of course, has contributed a lot to the reduction in Scope 1 and 2, which has been reduced by 85%.

'25, unfortunately, not every growth journey is one, let's say, linear line up. There will be some bumps along the way. And of course, we also have experienced a small bump, so we had to communicate this. So I think the key points of what we communicated this morning are as follows: Q3 still very much in line as we updated you with Q2 earnings update. Q4, unfortunately, is lower than what we expected when we did the Q2 earnings update for a few reasons. One is that the overall leading-edge revenue will come in somewhat lower than we thought, in particular, because we see a very mixed picture per customer. Some do really well. Others struggle a little bit more. Of course, we have taken that into account. But what we see is just even more of that, which has resulted in a lower-than-expected revenue for Q4.

In addition, Power, wafer, analog, we saw some encouraging signs in the end markets by the end of Q2. We did not plan for a recovery, but we assumed a slightly better picture in H2. But based on what we see today, based on everything, orders, et cetera, also that is not yet happening. It will come at a certain moment in time. This market is already 7 quarters in a cyclical downturn. When it will come back, we don't know yet, but at least not yet in this year. And for that reason, second half revenue will not be flat based on constant currencies like we communicated, it will be down 5% to 10% compared to the first half. And that has also translated into lower bookings. So also for the second half in total, so Q3 and Q4, we expect a book-to-bill below 1. That's what's basically communicated by the bottom statement on this slide.

That means for the full year, we will still grow, but not at the midpoint of the 10% to 20%, but at the low end of that range, that is our current expectation. And I want to repeat that also in 2025, demand for leading-edge logic/foundry is very strong. We've seen significant growth in leading-edge logic/foundry. We also communicated orders actually would increase in Q3, Q4 compared to Q2. That's also happening. So that's all still very much in place. But adding everything together, what I just said, you get to this picture.

As you know, capital allocation, excess cash returned to shareholders. On the left side, you see the dividend. You all are aware of that, basically an annual increase with one pause in 2022 when we did 2 acquisitions, again, prioritizing growth over returning capital. But since then, again, further increase. And if you look back 10 years, most of the cash returned to shareholders was through share buybacks, more than EUR 1 billion, close to EUR 1 billion dividends and EUR 200 million through a reduction in paid-in capital.

This is how the share price developed in this year, not so good, unfortunately. But still, if you look over a -- yes, since January 2020, over whatever, 4.5-year period, we're still seeing a return -- total shareholder return in excess of the semiconductor index, but also in excess of the main cap index in Amsterdam, where we are listed.

So maybe the '27 guidance update, again, it's only an update for currency, nothing else. That brings the -- based on the U.S. dollar euro, you see it in the bottom of the slide, we've been very explicit this time of 1.17. That brings the top line from EUR 4 billion to EUR 5 billion to EUR 3.7 billion to EUR 4.6 billion. We've increased the margin a little bit, 100 basis points to 47% to 51%. SG&A, high single digits, no change. R&D, you've seen where we want to invest and it's paying off in a pretty decent return on capital. You've seen that over the last 4 years. And effective tax rate basically will stabilize around the low 20s.

I think more important now because this gives you a longer, let's say, view on what we think where we are heading is the targets which are new for 2030. First, revenue, a CAGR of 12%, which compares to a -- in our view, at least a CAGR of WFE of 6%. We believe we should at least grow 12%. For this time, we did not do a range because what happens if you give a range, you guys go to the midpoint. So we thought let's not do that. Let's make your life easier. So as a minimum, 12%. And what is driving that growth? That's actually very important. And then you can also form your own assumptions, of course. First and of all, as we always say, growth of end markets. If end markets don't grow, it's hard for our customers to grow and also for us. But if end markets do grow, and I believe the trends are in place to grow, we will grow with it.

The second one is the growth and the composition of the WFE market. So where is the growth? So EUR 155 billion WFE market, if that's 100%, let's say, trailing edge, that's not so good for us. If it will be 100% logic/foundry leading edge, it would be super for us and then everything in between. So composition of the market is very important. You've seen in all the presentations increased ALD and Epi intensity going forward based on all the trends that we see, which is positive because it plays to our strengths. We've added this time growth in advanced packaging. We do believe that there's great opportunities based on, let's say, the strengths that we have and also based on some other things that we're looking at to further expand our SAM. And Jason, I think, has given you a very good view on all the innovation that is going on in our spares and service business and automation and as a result, we also believe we will continue to see as a minimum 12% growth in spares and services.

This is, let's say, the revenue in the last 4 years, equipment revenue. So 60% in the last 4 years was logic/foundry, so the majority. Memory, 19%, relatively small, but you've seen in Vamsi's presentation that that's one of our growth areas. We want to improve our share in memory, in particular, in DRAM. It's very important given the growth that we see happening there. And others is basically PECVD, vertical furnace, silicon carbide, et cetera, where in vertical furnace, you've seen a nice growth with the market coming down, of course, with the cyclical downturn. PECVD, we have not done well. We didn't grow, but with global carbon, we believe that we have a great opportunity to hopefully also start growing that business. Then in terms of technology, so 68% of our revenue over the last 4 years was in single-wafer ALD, I guess, known to you, 13% in epi and 19% in others.

Then gross margin, we've increased the range to 47% to 51%. When you see this trend, it's not so hard to imagine why we did that. We have not been below 47% in the last 4 years. And also this year, we will not be below 40%. So I can safely say that for 5 years in a row, we've been north of 47%. So we have increased it from 47% to 51%. You see the drivers here of margin. Of course, sales price is a critical one that goes one-one to your margin improvement or reduction, it's one-one margin. Application and customer mix that also includes China. And as you know, China is accretive to our margin. But we believe in the years to come, China will gradually come down. So the relative share of China in the business is coming down because you've seen 2 super strong years, let's say, for China. And we don't think that, that is the new normal. It will normalize more.

Cost efficiencies, I will come back to that. We have quite a few initiatives to improve our cost base. Operating leverage because growth in itself will help to improve margin, but gross margin, not a lot because the vast majority of our cost of goods is variable, it's purchase material, systems, subsystems, et cetera. U.S. dollar-euro, it doesn't help our margin. So a 5% decline of the dollar relative to the euro measured in euros will, let's say, reduce our margin by whatever, 0.1%, 0.2% by a little bit. So it's not helping. But the moment when the dollar comes back, of course, that should be a benefit. Maybe to remind you, when we did our first Investor Day in 2021, guess what the U.S. dollar-euro rate was. You're all right, 1.17, same as today. So the dollar goes up and down as we know. So it's nothing new.

What is not included, I also want to mention that is all the uncertainty around, let's say, tariffs. At this moment in time, we cannot tell what will happen. Today, we're still exempted. Maybe it remains exempted, maybe not. Maybe the impact is big, small. Obviously, any impact we will try to pass on in the value chain. But of course, that also depends on what competition is doing, but that's for now not impacted -- not included because, yes, it's -- we just don't know what will happen.

SG&A. Here, we see a steep decline. Already in 2021, we communicated our target for '25 is high single digits, which will happen. This year, we will be below 10%, which is good. And we believe we can go to below 7% in 2030 based on the growth that we see, but also based on a number of the initiatives that we're taking to actually reduce cost and improve productivity. R&D, you see here, we made quite some investments in R&D. Again, that's our #1 priority to fund. I would almost say the higher it is, the better it is. Given the ROIC we make on these investments, we need to invest to stay ahead. It's critical, both in our equipment business, in our spares and service business. We allocate a lot of money to that -- to technology. And so far, it is delivering great returns and great growth. So as long as we do that, we should continue to do that. So we changed the target, as I mentioned already, to low teens in the years to come. And given all the opportunities that we see, all the inflections that are happening in, let's say, the coming decade, I firmly believe it's the right thing to do from a ROIC point of view.

Then this is a summary of the financial performance. You've seen the revenue growth from EUR 1.3 billion in 2020 to, again, EUR 2.9 billion in '24, a margin improvement of 25% to, let's say, 28%. And going forward, our target will be to stay north of 28% and from 2030 onwards north of 2030 -- north of 30%.

Tax rates, yes, we believe that we'll stabilize around the low 20s. Already -- last year, I should say, we have seen some impact from this global Pillar Two, the global minimum tax of 15%. It has impacted some of the incentives that we get in certain geographies to a top-up tax. That's just a new reality we will have to deal with. In addition, it is also dependent on the relative result development on a country-by-country basis because each country has a different corporate income tax rate. So depending on where we make the money, which is substance based, you will see a different impact to the global ETR, the global effective tax rate. Obviously, we monitor this properly. We manage this as good as we can. We see tax as a business partner. At the same time, tax, of course, is also contribution to society. So we pay our fair share of taxes based on substance. And we follow arm's length transfer pricing rules. We do all the right things, we believe and that results in a tax rate of low 20s that you can use in your models.

Working capital, we've done -- we had -- we used to have a guidance of 55 days to 75 days. We improved that slightly to 50 to 70 days. '24 is actually, I would say, exceptionally low. And why is that? You will have seen in our balance sheet that there's a huge amount of deferred revenue. And that is, of course, benefiting working capital. But that amount, I expect at least over time will come down somewhat. It's a large chunk related to China. In China, there is for a number of reasons that we can separately talk about, they -- let's say, when they put in orders, they only order what they need. They pay for more than what they need, and they keep the credits for spare parts or for a second tool or for other reasons to be able to order that by the time when they need these products. And in that time, there might be no budget for them but now they have the budgets and now they want to use these budgets. So there's a big chunk of deferred revenue in our balance sheet.

That's this -- the last point on the slide. It's called the higher contract liabilities, which basically means we have received the cash, but not yet fully, let's say, delivered on all performance obligations under the contract that we had with the customer. CapEx, a lot has happened here. You see quite a steep increase. So basically, the new guidance that we have is that in the years with infrastructure expansion, CapEx will be around EUR 150 million to EUR 250 million, somewhere in that range. And in the years without infrastructure expansion, between EUR 100 million to EUR 200 million. Of course, in the last few years, we had quite some infrastructure expansion to actually get ready for the growth that we see. We completed Singapore in '23. Korea, I think, Hichem, showed some nice photos. We completed that in this year, actually. The grand opening is by the end of this year. Phoenix, Scottsdale, we expect to finish that -- complete that in Q1 '27. We also still have plans for expansion in Europe. So that will also happen starting in most likely in '28.

And once that all is done, then we can have quite a few years without the infrastructure expansion, I believe, then we should go back to a level of EUR 100 million to EUR 200 million. So in the years with infrastructure expansion, it will be potentially EUR 50 million more. Cash flow, you see here the growth in the cash flow from year-to-year. We expect to be north of EUR 1 billion. So it's a minimum of EUR 1 billion that we target, again, based on the revenue and profitability targets that we have given and the working capital targets and the CapEx targets, so you can all model that, including tax. Of course, we need to grow as planned to be able to deliver this. We will continuously manage working capital very strict. We're really very much on top of that with our customers to make sure they pay on time with our inventory to make sure that we don't keep inventory that we don't need. And of course, also for CapEx, we try to manage tight CapEx budgets but we need to go through this phase of expansion, and we need to put all this infrastructure in place to enable the growth that we see ahead of us.

And then this is a summary of our 2030 targets. I think I covered all of them, so no need to repeat them. Then maybe just on a few, let's say, capacity and margin improvement initiatives, also gives some more comfort on the margin development but also on the CapEx, I think. In '23, as you know, we completed the second floor of the Singapore expansion, which was a big thing in terms of capacity growth, more or less doubling the capacity that we had in place. In '25, we expand further with Korea, which is a mix of manufacturing and R&D lab.

And then going forward, we believe that we have more than sufficient capacity to get our 2030 target and beyond even through, let's say, productivity initiatives, efficiency initiatives, but also through supply chain innovation and in particularly MIT, Merge-In-Transit, will contribute to that because what does merchant transit do? It simply means that we don't assemble all products that we ship to our customers in one of our sites. We only, let's say, produce, let's say, the chamber, which is also most sensitive from a competitive point of view in our sites. The hardware, the platform is basically manufactured at our contract manufacturer and tested and the chamber and the platform go straight to the customer. There, it's basically assembled, tested, installed and working.

Today, this all goes first to one of our sites. We unpack the platform. We bring it into our factory. We assemble the whole product. We test the whole product. We then have to basically disassemble it again, pack it again and ship to the customer. So with that, not all products, we cannot do it for all products but for the products that we can do it, it saves actually quite some, let's say, capacity that we have internally, which, of course, frees up capacity for the products for which we cannot do that. So we have more than enough internal capacity in place to deliver on our growth plans. Then some other initiatives. Actually, just 2 months ago, a little bit more than 2 months ago, was a big thing. We went live, a global big bank with a global -- a new global ERP system successfully, but also with a new global product life cycle management system. So we had 2 big banks on 1 day globally is a big thing but it's working well. So very good to see that since 2 months now.

And that, of course, provides and is a good foundation for subsequent initiatives to further drive improvement, to further drive real-time analytics and to do a number of things that we believe are important. We've talked quite a lot about platforms. So the drive towards common platforms has quite some benefits in terms of cost, in terms of inventory, in terms of lead times. Unfortunately, we cannot immediately switch every product to a new platform. That would be the best for us. But our dear customers cannot enable that. So it will be a gradual transition to the new platforms. But with that gradual transition, it will also gradually, of course, and structurally contribute to improve margins. I just talked about the manufacturing model, how MIT, but also further efficiency and productivity improvements that we can do in our sites will contribute further.

And all of that will structurally and gradually contributed 200 to 300 basis points in the margin, which is reflected in our guidance. So please don't add it to the guidance it is in. But anyway, there's a lot of work going on in the background to actually try to further improve margins. Capital allocation, no change, at least the capital allocation strategy is no change. There's one small change. The cash number that you see is around EUR 800 million. It used to be EUR 600 million. We just increased it a little bit in line with revenue growth and the business growth that we have. But in principle, all the priorities are still the same and the priority #1 remains investing in growth, R&D and CapEx, but also M&A. If we have opportunities, if we see opportunities like we've done in '22, we will definitely go for it. And priority 4 is still, if we have excess cash, we will return it to shareholders.

Just to show it, maybe also more visually, the purple bar is the free cash flow before allocating cash to R&D and before allocating cash to M&A and CapEx. So you see that the bulk of our money that we earn is reinvested into the business basically in all years. You see it here clearly presented, and that's a very deliberate choice to allocate that cash that we earn, back into the business. And in the bottom part, you see the dividends and the share buybacks, which, by the way, also have grown nicely. And then last but not least, sustainability. We have 5 pillars, innovation, people, planet, supply chain and governance. Innovation is in the product. It's part really integrated into our product development process now. We want to make our products more sustainable from a thermal point of view, from a chemistry point of view. And of course, we want to make our products safe. That's absolutely critical.

For people, we have a number of initiatives going on, one, making sure that we offer them a safe working environment but also we want to engage our peoples. We want to contribute to society and to the local communities in which we operate. We have a net zero target by 2035, as you know. We work closely together with our suppliers to also drive sustainability initiatives with them, one, because it improves, we believe, still planet earth but also it makes a hell of a lot of business sense. And I think you've clearly seen that in Jason's presentation. And of course, we try to apply the right ethical and proper governance to our company.

Under CSRD, we actually disclose a lot of, let's say, nonfinancial information. So what do we disclose? What we disclose is based on this, what you call double materiality assessment. So what is our impact as a company on people and planet. And on the other hand, what is the impact of potential climate change on the financials of our business. That's called this a double materiality assessment. Now we've done the analysis. We've reached out to multiple stakeholders, including investors. This came out of it. I'm not going to read through it in view of time but these are the key focus areas that we will report on going forward and actually already did in the year 2024.

Then some examples because the biggest challenge that we have is Scope 3. I mean Scope 1 and 2, you just saw a 85% reduction. There, we're moving in the right direction. Scope 3 is more complex. A big chunk will come from renewable electricity with our customers and with our suppliers. But in addition, we have quite some initiatives on sustainable chemistry, on energy efficiency and on greenhouse gas emission reduction. Again, it's fully integrated. I'm not going to take you through the examples here. We can do it separately, if you wish. But I think it's important to note that this is a key focus area. It's one of our strategic objectives, accelerate sustainability. And we focus more and more now on the product because also that is required from a Scope 3 point of view. And it makes a lot of business sense to do that. My key takeaways are clear.

Thank you very much for being here again. And I think I hand the floor back to Hichem now for the final wrap-up. Thank you.

Okay. I hope that we did not overload you too much today with the 164 slides that we showed to you, okay? But I think what you can gauge from today's presentation that we are very bullish about our industry. And even though there is like short-term uncertainty right now. I think the trend going into high-performance computing means that you need to have more performance. And with more performance, you need to have more ALD and more epitaxy. The industry is moving to 3D as we showed today. That means you need more deposition and we are a deposition company. And also, you've seen today the structure in the gate-all-around, in the CFET, into DRAM going from 6F square to 4F square to 3D DRAM. You've seen the structure becoming more and more complex. And with that complexity, you need more and more ALD. So we actually feel very good about our position in the market in ALD and was shown today and not only in ALD but on epitaxy and other parts of our industry.

You saw today the cool stuff that was presented by our colleague, Eric Shero, about where every monolayer matters. We actually can control Angstrom level in ALD. That shows really the prowess and the technology, our ALD technology that we have and that we're actually providing to our customer. To achieve this Angstrom-level control on the wafer, you need to have micron control on your equipment. And for that you need to have automation. And you saw that Jason Foster talked about the automation as part of our many outcome-based services. Service is becoming a very important part of our growth in the future because products and equipment is becoming more and more complex.

The other thing that we saw today that we're actually providing our customer with AI and ML opportunity on our equipment. Why we're doing this? Because for 2 things. AI/ML makes sense because it can speed up the innovation for our customer. It also makes sense because it allow them to keep the product availability and uptime much better because with AI and ML, you actually can pinpoint and find out any deviation in your product much faster and much easier so that you can fix it. We also have talked about sustainability and how sustainability is part of our DNA. And sustainability is very important because not only it's good for the environment but also makes really good business sense because it lowers the cost to the customer. And as deposition becomes a much higher intensity in semiconductor wafer equipment, really reducing deposition precursor flows intensity is -- chemical intensity is really very important because by reducing that, that you reduce the cost to your customer and that makes you also much more competitive.

We also talked about advanced packaging. We think that advanced packaging is a new growth area for us. And we are very -- we think that we are confident to really increase our served available market in this part of the technology to over 30% by 2030. We are very excited about the future. And I speak on behalf on all my colleagues here today that we are very excited about what's coming. So with that, I'm going to leave the floor right now to Victor to go to the Q&A session. Thank you very much.

Thank you, Hichem. Thank you. Now it's Q&A time, we need a couple of minutes to prepare the stage. So one moment, please. In the meantime, we will show our latest brand video to keep you entertained. On moment please.

[Presentation]

[Operator Instructions] First question. Tammy?

So firstly, Paul, may I ask you what has changed over the past 2 months since you've updated us on Q2, i.e., did you notice that foundry logic spending recovery was lower than expected? Or it was because of something incremental to what you heard in 2022 that made you to change your guidance? And also, what is the situation today, i.e., is it still not improving? Are you [ not ] going back to acceleration at all?

And the second question is for Hichem please. So you talked about all the TAM expansion story, which is amazing. And also at the same time, your market share remains at 55% and above level. If there is anything we should be aware from your competition strategy, i.e., your market share gaining pace kind of slow down than where you were before? Because over the past 10 years, you have been gaining market share purely because of you've been getting market share from the incremental new layers. Is that changing or the competition potential within the ALD market or IP market will be more competitive going forward?

Start with the '25 question. So what changed? So when we came out with the Q2 earnings, we, of course, have certain assumptions on Q3, Q4 and the remaining of the year. And based on where we are today, we've seen a number of changes. One is that, let's say, for leading-edge logic/foundry, we've -- as you know, there's a pretty mixed picture per customer. And let's say, for certain customers, we've seen actually things coming down even more than we anticipated. And for another customer which is still doing very well, is maybe, let's say, the acceleration we're planning is maybe slightly below what we were actually assuming at that moment in time. As I just mentioned as well, we also guided with Q2 earnings that bookings for leading-edge logic/foundry would increase in Q3 and Q4 compared to the low level that we saw in Q2. That's also happening. So that's still good, but maybe slightly lower than we thought, but still very good and very strong, but slightly different.

The other one, I think you referred to that, Tammy, is the power/wafer/analog. We have no plan for a recovery because although there were some encouraging signs but I think it was reasonable to assume that, let's say, the low that we had in the first half would improve somewhat. We don't see that happening now. Based on where we are today, we can, I think, almost conclude that, that's not happening in this year. This market will come back at a certain moment in time but we don't see it happening yet.

And thirdly, although we didn't mention that because it's not that big but still it contributes to silicon carbide, which is really very low, still low and it's likely to remain low also in '25 but was even lower than we already thought. And we had already put it pretty low, to be honest. So just quite a few changes. So nothing big, I would say, the structural trends that we talked about today are still in place, no change. And yes, that's where sure it is.

Okay. And to answer your question about why 55%. I think we say it's greater than 55% and can be any number greater than 55%. I think it's very -- it's something that we have talked about before. And to be honest with you, I mean, as you can see and I showed you today, we have gained share in ALD with the transition to gate-all-around and also we're going from the 2-nanometer to 1.4-nanometer technology node. So -- but we're not going to say exactly the number. I think I look into the number, maybe talk -- see what Tech Insight and so on talked about. I mean they have higher numbers that we have right now. But I think what we're going to say in the future, always greater than 55% and we're not going to change it if we go to 60% or 70%. We're just going to keep it greater than 55% because there's fluctuation from one quarter to the other and from one year to the other. But as long as we keep that, I think we're in good shape.

Jakob Bluestone, BNP Paribas Exane. Two questions as well. Firstly, on China, can you just help us understand what's embedded in your guidance? Are you worried about domestic suppliers taking market share and the sort of impact of restrictions? So just to sort of help us think what are you assuming longer term? And then just secondly, just staying with the market share. You talked about how for 2 nanometers, I think you said 50% of layers would be front end of the line and that would rise to 60%, 1.4 nanometers and that sort of play to your strengths given your stronger front end of the line. I wonder if you can maybe elaborate what are the dynamics between front and back end of the line? So are you actually taking share in the front end of the line and losing share in the back end of the line? Or any sort of color you can give us on how that breaks down?

Sorry, that's fine. I will take the China one. So what have we assumed -- so one that -- I'm not sure it's a trend but at least what we saw this year happening is, let's say, still a strong year in China but below last year. And that's actually a trend that we have continued in our assumptions. So we expect China to gradually come down. And at the same time, don't expect it to fall off cliff, let me say it like that. In addition, you're right, local competition is definitely getting stronger. We see steep growth in revenue of local competition. They're benefiting a lot from the vacuum that is being created by the export controls where we cannot participate. And it's for them the best way and the quickest way to learn because once you get a tool in the fab, that's how you learn.

That's how you get uptime improved. That's how you get your product further improved, et cetera, et cetera. So we see that also in China. There where we can compete head on, we still believe we have the better product and even a lower total cost of ownership given the performance of our tools that we have. But we also recognize that Chinese competition will get stronger. So -- and the assumptions that we have is that we see a gradual decline of China. That's basically the assumption. And then the other one was on share, back end of line versus front end of line.

Yes, I can talk about that. I think as we mentioned that when you go to 1.4 nanometer node, 60% of the ALD is actually in the front end of line. That's where we have most of our strength and that's where all the ALD was in back end of line, there is no ALD whatsoever right now. So there's no ALD being run in back end of line. I think the ALD that you hear about being in the back end of line, mostly right now, it's happening in the middle end of line in the [ M0 ]. But eventually, you go into the next generation, the 1.4 nanometer and 1.0 nanometer, then the [ metal 0 ] and [ metal 1 ] becomes also smaller, [ metal 2 ] becomes more so, and you see more and more of the metal deposition or molybdenum deposition going into the top levels. But right now, in logic and memory, there's no metal -- there's no ALD deposition in the back end of line.

Andrew Hayman from Independent Minds. I was just wondering how you are -- how do you see your relationship with ASMPT developing as you move into packaging? That's my first question. And then a second question, has there been a considerable learning curve as you move to outcome-based servicing? It's clear that the move to [ dry maintenance or dry servicing ] has been beneficial but has there been some areas or equipment where it's been more costly to keep those machines running to the desired level than you anticipated?

Yes, on ASMPT, so yes, ASMPT is a financial investment for us and it will stay a financial investment for us for now. You've seen the, I think, the pie that I think Hichem and also Vamsi has showed in their presentation, where we saw an overview of the total advanced packaging market. Then you also see that, let's say, bonding, hybrid bonding is not the biggest chunk of the whole addressable market over time. Actually, at this moment in time, PVD and [ ECP ], I believe, was largest. So today, our addressable market share is relatively small, only 15%. We want to double that. We have some ideas how to do that. And for that, we don't need ASMPT. So we hope, of course, that ASMPT will do very well but it's for us a financial investment. And for now, it will stay a financial investment.

I got the second one. The second one is on outcome-based services and have we gone through learnings? And are we potentially taking lower margins in certain areas as we go through? Yes, we have gone through learnings, right? When we approach a customer with outcomes-based services, many times, it's in the tool selection phase where we don't know what it takes to run that tool in high-volume production, right? So we do have to make commitments. And those commitments are really towards how do we expect that tool to run in a high-volume production fab environment. In many cases, it takes us some cycles of learning to get there to achieve better performance. So yes, we do have cycles of learning that we go through as we engage with outcome-based services with our customers.

Sandeep Deshpande at JPMorgan. Two questions from me. Firstly, on your guidance into 2027. Is this a top-down guidance based on WFE -- your WFE view in '27? Or have you done a bottom-up survey of your customers and based on how your -- they have given preindications on how the ramp-up will be. In particular, given that you've talked about these market share gains potentially in DRAM, will that be helping that in '27? And then secondly, the issue you had in '25 associated with the foundry, I mean, whether that foundry will be, or foundry/logic customer will be ordering in '27 or not has been taken into account. And my second question is on advanced packaging. You've talked about an opportunity in advanced packaging, but many of your peers in the semiconductor equipment space have already gone and staked up their positions in that market. Where is it that you're going to stake your position? Is it ALD in advanced packaging? Or is it epi, or where are you going to stake that position? And is this going to be a significant portion of your 2030 revenue?

Yes. So for 2027, the DRAM transitions I talked about are essentially [ D1, D2 ] onwards. So we do expect some initial investments on the [indiscernible], the inflections that I mentioned about it, but the -- most of the ramp is counted from '28 and beyond in that sense. So 2027, we see ramp of 1.4 nanometer happening and may not be -- so 2 nanometer will be a long node as well. So 1.4 nanometer will get started in 2027. So those are the 2 assumptions that -- but it's not going to be in very high ramp. It's going to be getting started out of the gate in 2027.

But from the foundry point of view, we don't see it whether it's going to go to customer A or B or C. So that's really agnostic to customer A, B and C because at that time, it's really the demand of the market that's really most important. And I think really demand is mostly important. And I think customers right now with the expansion they have in manufacturing, they will be able to get that demand independently of where that, which customer is more successful than the other one. And then the second question, you talked about advanced packaging. And I think in advanced packaging, yes, there are many a customers that are there. They have many competitors that are in those businesses. We think that we can offer something. I think we can disrupt some of that business.

First, okay, in PECVD, we have PECVD offering. And there is many, many new things that really needs to happen there. For example, you need to develop films that are -- that conduct electricity, conduct heat much better, so [indiscernible] film that conduct heat. So we're looking into new films, new chemistry, new precursor that actually can -- chemistry pretty much and we are let's say, we can -- we understand a little bit of the chemistry. So we're actually developing those films that can really dissipate the heat that's generated. And heat, as you guys know, in advanced packaging, heat is a big deal. There's also many films that where customer, you want to make sure that, okay, you can seal some of the better [indiscernible] in the bonding. Also, the films need to be compatible with CMP and so on. So really, customers actually are learning with us right now.

And there is lots of -- every day you hear, okay, this is a problem statement and so on and so forth. So they go to our competitor and also they go to us and we try to fix some of the solution. In some of them, we have been successful and we got the business from that point of view. Yes, we have ALD. So we're working on an ALD for advanced packaging, surface treatment that I talked about, that's one other area that we're working on. And as I mentioned, we would like to achieve greater than 30% of the TAM of that business. There are areas that we're also looking at. We're not saying that, okay, we're going to gain a huge market share in that part of the business. But I think if we do well, we should be able to -- if we execute well, I think we should be able to achieve good market penetration in advanced packaging.

All right. Eric, maybe switching to a question about the integrated solutions. I think one of your competitor has sort of thought about this for a couple of years. How big is it as a percentage of revenue now? And where do you see that end up? And what will be driving those type of integrated [indiscernible]

I can't really speak to the revenue but the integrated solutions are picking up steam, especially in the metallization space, where you have to have preclean also in the [ ASD ] space that Hichem spoke about. It's really important that everything be clustered together in those steps. So we have to pretreat the surface in order to enable [ ASD ] or metallization moly. Those are where I would say the most clustering is really critical. A lot of our -- a lot of technology does not require clustering, right? So it's -- but the XP8 is not just for clustering, it's also for throughput as well, high throughput processes. So parallel processing that don't necessarily have to be clustered. But I would say I can't really speak to the percentage but -- it's a small percentage, yes. But most important in the metallization space and areas like the deposition.

All right. Eric, well you said in the slide where operating margin would stay above 28% from '26 or stay above 28%. But given uncertainty next year, you start the year with a quite lower backlog than usual, what gives you confidence that, that's a lower threshold you can definitely be?

So you're right. So backlog, indeed, based on the guidance we've given, will come down in the coming 2 quarters. So the year will start soft. At the same time, given all the trends that we talked about, we do believe that somewhere in the year, things will pick up and improve. And based on that, we have a view, of course, on '26. That in combination with some of the initiatives we're taking on efficiency, on productivity, on platform and all the things that I mentioned, make it still possible to stay north of 28%. But you're right, we need a certain revenue level, of course, to be able to achieve that. We think that's still -- we're still on the money to be able to do that. But at the end of the day, it's still relatively early, too early to tell.

But yes, we know the year will start weak. That has been taken into account but we also expect that to improve in the course of '26. And it's really too early to tell how that will develop. You know everything that's going on with one specific customer, we'll see how that will develop. Power/wafer/analog is still early days to see how that will develop. Even China is a bit early days, although we think it will come down. But again, precisely how much is also difficult. So it's really hard to say precisely how much. But yes, we've made some assumptions that make sense, I believe. And based on that, we have said what we said.

Didier Scemama from Bank of America Securities. A couple of questions. Just wanted to go back to the booking things. So I think at the time of Q2, Paul, you said Q3 bookings will go up. And now I think what you said earlier is that leading-edge logic orders are up Q3, Q4. So basically, it means the rest is down. Is that fair? That's my first question. And then the second question, sorry, about bookings. We are financial analysts at the end of the day. For 2027 to see the snapback that you're sort of guiding for, given your lead times, would it be fair to say that your bookings have to materially accelerate probably at the latest Q2, Q3 next year? Is that the right ballpark?

Yes. So first on Q3. What we said at Q2 was that because Q2, bookings were relatively low compared to consensus. And of course, you want to know why. Now one of the reasons was because leading-edge logic/foundry was relatively low after 5 or 6 quarters of pretty strong bookings. At that point in time, we said leading-edge logic/foundry will come back in Q3 and Q4, will be strong again. That's still happening. We didn't say that the overall bookings would be strong because we also guided for a book-to-bill below 1 for Q3. Why is that? Because China is coming down a lot. So bookings in China will come down a lot, at least as the current view that we have in Q3 but even more so in Q4. So the net of everything we said was still not very strong bookings in Q3, just to clarify that.

Then on '27, yes, of course, at a certain moment in time, these bookings should come in. I'm not going to say it should be Q2 or Q3. We have the capacity in place, we showed that, so we can actually turn around the product pretty quickly. We can even deliver if needed, if we get a booking in the quarter. And if we have the material for which we can plan, we can even ship and invoice in the same quarter, so within 3 months. Most of it, what we do is, let's say, within 6 months, also because that's what we want, that gives us more time and better time to plan for material, for everything we need. But it's not impossible to do that.

And at the end of the day, it will all depend on customer forecast. We work with forecast, especially for the larger customers. There we get a pretty good indication what we want. That does not always one-on-one, let's say, convert into a PO, so into a booking that you see. But we already have more visibility, of course, in the pipeline than what actually you guys see in our order book. But you're right, it will have to come back. If it doesn't come back, it will get more complicated but we believe it will.

And just the final one, very simple one. The EUR 5.7 billion target '27, what sort of WFE is it based on?

EUR 155 billion.

And that's 2030 or that's, oh, sorry, sorry, the -- excuse me, the new '27 targets -- sorry, the 2030 and EUR 155 billion. So what's the '27 WFE assumption?

I think it's still the EUR 120 billion that we used to have. I don't think that we made any change there.

It's Adi Metuku from HSBC. Firstly, I just wanted to ask a bit about NAND. Your main competitor in ALD is talking a lot about share gains in NAND and you've spent very little time on NAND today. So I just wonder how do you see your market share in NAND, ALD? And any color you can provide around that market would be helpful. And I've got a follow-up.

Yes. So today, we focused on what we see as the areas of growth and potential for us going forward, depending on the end markets as well as where we see the huge intensity for areas of our strength going forward. So ALD moly is an opportunity that we are after. We are an ALD innovation company and moly is in different technologies. Moly is replacing ALD tungsten which was already in NAND. So any -- we are focused on logic applications as well as other memory spaces as well. So we are going to have ALD moly [ PTOR ] in the first-generation GAA. So that's a significant gain for us. So any moly win, [ PTOR ] win, we gain is a new business, new growth opportunity and new addition of SAM to us going forward.

So we are doing, in that sense, pretty well in the ALD molybdenum business. And we see also the ALD molybdenum continuing, and it's reflected in what we showed as the ALD layers going forward, especially in leading-edge logic/foundry. Most of the growth is still happening in front end of line, which is a place of strength for us and we are focused on looking everywhere but also looking at the same time maintaining share or gaining share going forward in all the space. So any gains we have in ALD molybdenum are an incremental new SAM for us and growth for us.

Got it. And maybe just as a follow-up. On epi, I noticed you didn't give any market share targets. I don't know if I missed it but previously, you had given a target of 30% in '27 and '30. So I just wonder if you have any targets to share.

Yes, I can give -- I will take that. We have -- of course, the target is to grow share. But for a number of reasons, we didn't want to be too explicit on the internal target that we have. So -- but the target is to further grow share in leading edge for epi business.

Francois-Xavier Bouvignies, UBS. My first question is on, I mean, this outperformance of WFE usual for '27 and 2030 is well supported by the layers A14, you mentioned it's going to be '27, all the DRAM opportunities in '28 and beyond. And is there any opportunities in -- on the layer count that you see in '26 as well? I mean do you still think you can outperform WFE also in '26 in that [indiscernible]? Because it seems that all the [ real ] incrementals are coming in '27 and beyond with all the things you described. And I was wondering if I'm missing something on '26 that could happen before.

I can take that question. I think if leading edge goes very well in 2026, based on what we showed today that we have good market share and increased market share in ALD, in the high-end logic part of the business, I think we should be able to continue growing and outgrow the market from that point of view. So it really depends on how the market will go in 2026. I think that there's still continuous bias from our customer for the 2-nanometer technology node. I think it's going to be still strong in 2026. We also see a pilot production starting to -- in the 1.4 nanometer technology node. And with that, we have good position in that technology. So we're confident as long as the high end goes well, I think we should be able to do very well.

And my second question is on 2030. Did you make any assumption on High NA adoption at all? I mean how it could impact your business if you move to double patterning to single? I guess it could have a very different impact as well. So is it because of the range [indiscernible] is High-NA? I guess it's important. So any impact on that?

So just to reset, 2030, we are looking at 1.4 nanometer ramp, heavy ramp and 1.0 initial production as well as memory 4F square ramp going on at that time with FinFET coming in at the same time in that year, not fully -- not full adoption but leading adopters will have that. So within all these assumptions, the -- what we showed about 1.4 nanometer front-end layers being higher than others as well, the impact of High-NA is in there implicitly, which is to say that it's not going to make that much of a difference for our assumptions and modeling in that sense.

It's Timm Schulze-Melander, Rothschild & Co Redburn. Maybe one question for Paul and then a follow-up for Hichem. Paul, on the 2027 margin guidance, you talked about some of the efficiency gains that are going to help the headwind from FX. But maybe could you be a little bit specific, which one of those are new and maybe which one of those are sort of the most sizable and kind of when can investors start to see them come through? And then I have a follow-up.

The operating margin or gross margin?

Ideally both.

Okay. So let's start with the gross margin then. So one is revenue growth, of course. There will be continued benefit from operating leverage. There's some fixed costs that we simply will come down as a percentage of revenue. MIT will start to contribute. We started this year already with some pilot products, so that will grow over time and that's actually quite a reasonable contribution. Platform is, I think, more gradual. It's something that we would like very much. And believe me, I would like it maybe the most but not all our customers want to, for good reasons at their end, immediately change. So that will go gradual but there will be some benefit of that as well.

Commercial negotiations, of course, with suppliers will continue because -- also with going to more common and more commonality in our platforms, you get better, of course, commercial negotiation positions, which for sure will help. And then on top of that, if you go to the operating margin, you have seen our target in 2030 47%. I mean that will continue, of course, from the 9-point something where we will be this year, [indiscernible]. So there you will see also further contribution in the operating margin. So that's -- yes.

Great. That's helpful. Obviously, I appreciate if you can give any color in the quarters when it does arrive. And as you look forward to 2027, 2030, I just want to talk -- think about the way that this group is organized and where you have your centers of excellence. We've talked a lot today about ALD. We've talked a bit about epi. And then we have an initiative in advanced packaging. Can you just talk about how those regions are going to be represented and particularly whether it's Europe, U.S.A., what you do in Asia and where most likely is your advanced packaging sort of center of excellence going to be, that would be really helpful.

I think that, as I mentioned earlier, as you can see, we're actually present everywhere. We are present in Europe and we're present in Asia and we present also in the U.S. And all of them are doing ALD for us, okay? So okay. And from that point of view, you have -- everybody has his own think on the ALD point of view. So when we say ALD, we're actually developing ALD worldwide. And that's really the strength. We take leverage of strength of each region to make our products better. The same thing is also happening in epi. I think epi is mainly between the U.S. and Europe, more than anything else.

I think for ALD, it's all over the world. I think for epitaxy, it's really working both in U.S. and in Europe. Right now, as far as the new application for advanced packaging. So since -- okay, we have ALD, so that's going to be all over the world. Since we have epi, that's between the U.S. and Europe. And we're looking into other opportunity and those opportunity, we haven't made a decision where they're going to be. But I think you're going to hear in the next few months some of that and we will make -- it's going to be very clear that where we're going to be then.

It's Janardan Menon from Jefferies. I just want to go back to your visibility on 2027. I'm just contrasting that with the, say, your point in 2023, where you were talking about 2025. And when I look at -- when you were obviously very bullish about the gate-all-around ramp, et cetera and your growth based on current expectations is probably going to be in the region of the sort of low 20% kind of range between '23 and '25. So you currently seem to be more bullish on 2027 than you were on 2025 at the same stage. So my question is, do you have so much visibility on 2025 -- I mean, on 2027 to be giving this kind of guidance? And secondly, what is the assumption you're making on China for 2027? Do you expect it to be flat or up or down at this point in time? And a follow-up is, one of your competitors has been talking about gaining moly metallization design wins in advanced logic as well. Just any comment on -- do you see any share losses on that front?

That's fine. So you have a better memory than I have, Janardan, because I don't know how upbeat we were in '23 about '25 and a little bit surprised here that we are more upbeat today than we were then. I mean, I think if you see where we are today and if we see how we will start '26, which will be relatively soft given the backlog development, it's definitely not a slam dunk, just to be clear. So if we sounded too upbeat, then let's bring that down a little bit. Having said that, there's quite a few trends that work in our favor. And we have modeled that very detailed. We have made assumptions around that. As we said a few times, at the end of the day, it will very much depend on end markets.

And so far, end markets, especially for leading-edge logic/foundry look actually quite good, quite promising. To us, it doesn't make a hell of a difference if customer A or customer B or customer C or customer D will be successful. As long as, let's say, there is enough visibility that these trends will continue, someone will very likely invest, or all of them because for some customers, it's more uncertain what will happen than for others, as you know. But that's -- especially if you look out 2 years, it's not super relevant for us, as long as, let's say, the macroeconomic uncertainty doesn't get -- I say that, does not pause investments, pause trends that we see. If that doesn't happen, we're actually -- yes, I think we will be in the money very likely. But there is, of course, this economic uncertainty going on. There's a lot of geopolitical stuff going on. There's whole tariff discussion is happening. To what extent will that impact the economy. I mentioned clearly on the slide, yes, we have not been able to take that into account because that's, that's really -- yes, we just don't know.

But if the trends that we see today, every day when you read -- when you open -- when you wake up and you open a newspaper, there's again, EUR 100 billion AI investment. So if these trends continue, yes, there is a good reason for us to stick to that target, given, again, the incremental SAM that we see for 1.4 nanometer. Then in '27, hopefully, we will also see the first products or orders coming in for 4 square, where we target to improve our share. So there's a number of things that can happen. But yes, at the end of the day, we don't have a crystal ball. But yes, we were not intending to sound more bullish than '25, let me -- than in '23, sorry. Let me say that upfront. And then the other one was on moly.

Yes. I think to ask the -- and then your second question is about the metallization and molybdenum in logic. I think the way I'll answer your question is that for us, metallization is a new market. We have been an ALD company, mainly in the dielectric, which is really in -- mainly in dielectric market and also in metal nitride and metal oxide. So metallization is a new thing for us. What I can tell you there is that if you look into the metallization part of moving from -- in ALD, anything we win in metallization for us is additional and we'll be really very excited about that. The other thing I can tell you that for what's happening in metallization is still a very small part of the whole ALD pie, as I talked about earlier.

I think metallization in the back end of line ALD will become more predominant in the 1-nanometer technology node and below. That's really where the size will increase. We are there. We're very excited that we have won some wins at our customer for the 2-nanometer win. I'm 100% sure that also our competitor have won some of those layers. But for us, it's new additional market. And, of course, for our competitor who have been in the metal market, it's really moving from tungsten to a different material. So as also mentioned to you earlier, we presented earlier that by myself and also by Vamsi that when we go from 2 nanometer to 1.4 nanometer, most of the ALD growth is happening in the front end of line. Most of the things really are happening there, many more ALD layer that's happening there than what's happening in the metallization. And that's why our growth of market share in the higher end 1.4-nanometer node is actually more better than at 10 nanometer because most of those ALD layer are actually happening in the front end of line. I hope that answered your question.

Andrew Gardiner from Citi. Just another one on the visibility you've got in terms of the market share gains, particularly as you look out to the next gate-all-around node. Hichem, you just mentioned that we'll start to see pilot line build-outs next year for 1.4 nanometer. Can you describe how well you are doing in terms of sort of the layers there? And I suppose specifically, what is the advantage that you've got in terms of your incumbency? Have you lost any layers that you had in the prior 2-nanometer node for the equivalent? And in terms of the gains you might have in the newer structures or newer layers for ALD, you're gaining more than your fair share, so we can look forward to that share gain?

Yes. So 1.4 nanometer, again, on a blended level, there are -- there's going to be more growth in front end of line blended across customers. There's going to be more growth in front end of line, as Hichem said multiple times compared to the middle of line and back end of line. And the current analysis that we have is that we have at least maintained our share going from 2 nanometers to 1.4 nanometers overall blended share for single-wafer ALD.

Mike Roeg from -- sorry, Michael Roeg, Degroof Petercam. [indiscernible] said that by 2030, everything stands or falls with the size and the mix of the wafer fab equipment market, $155 billion. Your most successful customer in leading-edge logic is TSMC. As for every new node they introduce, they're cannibalizing their older nodes in leading-edge logic. And you clearly see that the 7 nanometer utilization is below average. If there's a huge build-out in 2-nanometer, it will cannibalize the older fabs. Tools will become obsolete. And in the Q2 call, they clearly said, they're converting 7 nanometer to 5 nanometer, 5 nanometer to 3 nanometer, they will continue. So now my question is, how do you see the risk that a EUR 155 billion market in 2030 will be EUR 20 billion from converting existing tools, which leaves means only EUR 135 billion for new equipment sales?

I can take that. So -- without -- I'll get to your question but again, 2030 scenario, if you look back at our presentations and forecasts, it's essentially coming from an ALD logic/foundry focused company to now growth engines on ALD and epi as well as logic/foundry and leading-edge DRAM as well. So it's a more spread out in that sense of where we see the forecast coming from for 2030. But also coming back to your question, yes, we do the signals -- we do see the signals of what you just mentioned and that has been somewhat factored into our analysis. Of course, the reuse is a much bigger factor in DRAM than it used to be in logic/foundry but we do see that decisions are being made for at least 1 -- at least 2 nodes, if not more, going forward and that's part of our model.

I mean if I can also help in answering this question, I think we have demonstrated to you today that the device architecture is changing when we go from the 2 nanometer to the 1.4 nanometer to 1.0 nanometer. So -- and what does it mean? It means that you need new material and you need new equipment. So the reuse part really becomes not -- you cannot reuse many of those equipment from one generation. It's not as easy to use some of the equipment from one generation to the other because the complexity is there, which means you need to have new reactor. For example, plasma needs to be different -- different type of plasma. You need to get a certain capability, you need to have a new material. So that's a new precursor. And with the new precursor, you need to change the reactor.

So it's really not something that you can just repeat over and over from that point of view. What's important there is really the device structure is changing. I mean we showed you guys from 2 nanometer to 1.4 nanometer, you're going to have backside distribution. And then you're going to have more of these [indiscernible] players and there is also more of channel films on the top. There are many things that are really changing. And with that, you need to have more equipment.

Okay. So we really like to thank you all for joining us today in our third investor meeting. We thank you very much for being here and for your question. We are very confident. I think we hope that we really provided you with a summary of where we think the market is going and where we think that -- the role we can play. We feel confident about the future and it's really one of the greatest time to be in this industry. From someone who's been in the industry in -- for the last 31 years being in this industry, I can tell you what we've been developing right now in ALD is just amazing. I mean this Angstrom-level control that we are doing is -- I mean, it really blows my mind. For a person who's been like developing films at in -- the micron range, right now, we're developing film in the Angstrom range and we can control the thickness to 0.1 angstrom. If you ask me a few years ago, I'd said this is science fiction. But actually, it's reality.

And what I see right now, this industry is actually going to grow even further. I mean, right now, we think every monolayer matters. But what we're thinking right now about just 1 monolayer, which is actually 2D material. And we think -- we're actually thinking about the next, where you need to develop a new film for quantum computing. And for that, really, when we talk about quantum computing, we talk about -- it's really -- the problem in quantum computing is not quantum computing, it's really material. How can you make those materials that can really do the qubits at room temperature. So you need to -- there's lots of development from that point of view. But what I can tell you, we are very excited about this industry. It plays on our strength in material, in innovation.

And I hope that you have seen today that we also very -- we are very confident about the future to the 2030 and beyond. Thank you very much for your time.