Operator: Good afternoon. This is the Chorus Call Conference operator. Welcome, and thank you for joining the ASM Q3 2026 earnings call. As a reminder, all participants are in a listen-only mode. After the presentation, there will be an opportunity to ask questions. Should anyone need assistance during the conference call, they may signal an operator by pressing star and zero on their telephone. At this time, I would like to turn the conference over to Mr. Victor Bareño, Head of Investor Relations. Please go ahead, sir.
Victor Bareño: Thank you, operator. Good afternoon, everyone, and thank you for joining our Q3 2026 earnings call. With me today are our CEO, Hichem M'Saad, and our CFO, Paul Verhagen. ASM released its Q3 2026 results yesterday at 6:00 PM Central European Time. The press release is available on our website together with our latest investor presentation. Before we begin, I would like to remind you that today's conference call may contain forward-looking statements in addition to historical information. For further details on the risk factors relating to these forward-looking statements, please refer to our press releases and financial reports, which are available on our website. Please also note that during today's call, we will refer to profitability metrics primarily on an adjusted basis. Reconciliations to the reported numbers are included in the press release and the investor presentation.
Victor Bareño: With that, let me hand the call over to our CEO, Hichem M'Saad.
Hichem M'Saad: Thank you, Victor, and good afternoon, everyone. Thank you for joining our Q3 conference call. We will follow our usual agenda call. We start with a review of our Q3 financial results. I will then discuss the latest market development and our outlook, after which we will open the call for questions. Paul, over to you.
Paul Verhagen: Thank you, Hichem, and thanks everyone for joining our call today. Let me start with the financial performance. Revenue reached a new record of EUR 1.16 billion, at the upper end of our guided range at constant currencies. Compared to the same quarter last year, revenue increased by 44% at constant currencies, while sequentially, revenue is up 14%. Equipment revenue increased by 48% year-on-year at constant currencies, led by strong ALD demand. Service products also delivered another strong quarter, with revenue increasing by 31% year-on-year at constant currencies. This was supported by continued momentum in our outcome-based services, together with strong spares demand as customer fab utilization remained at high levels. In terms of customer segments, revenue was led by the advanced logic foundry segment.
Paul Verhagen: Sales related to 2 nm remained strong, while 1.4 nm made its first solid contribution as customers started investing in pilot line activity. 1.4 nm already represented a significant double-digit percentage of advanced logic foundry sales in the quarter. We also saw another healthy contribution from the 3 to 5 nm nodes. Mature logic foundry sales remained robust, mainly driven by the Chinese market. Memory increased strongly, both year-on-year and compared with Q2. The largest part of memory sales was related to high-performance DRAM for HBM applications. We have also seen increased memory demand in China, albeit from a low base. Finally, power analog and RF revenue remained stable at relatively modest level. For Q4, we expect revenue in this segment to increase significantly. Gross margin came in at 52% in Q3, broadly stable compared with 51.9% in the previous quarter.
Paul Verhagen: The continued favorable customer and product mix, together with the gradual benefits from our cost efficiency measures, supported the margin during the quarter. For the full year, we now project gross margin to be around 52%. The increase in SG&A expense versus the previous quarter was mainly explained by higher headcount and higher variable expenses related to increased activity levels and company performance. As a percentage of revenue, SG&A improved further to 7.6% compared with 7.9% in Q2, reflecting continued operating leverage and ongoing efficiency initiatives. For the full year, we now expect SG&A to be around 8% of revenue. Total company headcount increased by around 400, or approximately 8%, compared to the end of Q2, mainly to support the ramp-up in operations as well as continued expansion in R&D and our service organization.
Paul Verhagen: Gross R&D spending increased by 14% year-on-year to EUR 144 million as we continue to invest in our technology roadmap and future growth. Net R&D expenses increased more strongly, by 30%. The higher growth in net R&D compared with the gross spending was partly due to the lower capitalization of development expenses and also reflect a EUR 7 million impairment charge in Q3. Operating profit increased by 54% year-on-year at constant currencies with the operating margin improving further to 33.4% compared with 33.0% in Q2. If you look at the main movements below the operating line, financial results include the currency translation loss of EUR 22 million in Q3 2026, compared to a translation gain of EUR 22 million in the previous quarter, mainly explained by the weaker quarter and US dollar against the euro compared to the end of Q2.
Paul Verhagen: Our share of income from investments reflecting our stake of 24.6% in ASMPT amounted to EUR 15 million in Q3, up from EUR 9 million in the previous quarter. Let us 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 1.3 billion. Free cash flow remains strong at EUR 313 million in Q3. This was below the record EUR 355 million in Q2, but significantly higher than the EUR 139 million generated in Q3 last year. CapEx increased to EUR 55 million in Q3, up from EUR 38 million in the same period last year. We continue to expect full-year CapEx to exceed EUR 250 million, with the largest part related to our new Scottsdale site.
Paul Verhagen: Despite the strong sequential increase in revenue, working capital remains broadly stable due to improved days of working capital, which dropped from 50 days at the end of June to 43 days at the end of September. Looking into Q4, we expect working capital to increase due to revenue growth as well as higher days of inventory as we prepare for higher activity levels in 2027. During the quarter, we executed the EUR 150 million share buyback program, which was completed on 14 September. As most of you have probably seen in yesterday's press release, we also announced a new share buyback program of up to EUR 300 million. This reflects a strong cash position and financial performance, as well as our continued commitment to return excess cash to shareholders. In summary, Q3 was another strong quarter.
Paul Verhagen: We achieved record revenue and a very strong level of profitability while continuing to invest in innovation. With that, let me hand it back to Hichem.
Hichem M'Saad: Thank you, Paul. Q3 was another strong quarter for ASM. As Paul just outlined, revenue reached a new record, increasing 45% year-on-year and coming in at the upper end of our guidance. This was achieved against a backdrop of increasing pressure across the supply chain and reflects the tremendous effort of our teams to ramp output in close cooperation with our suppliers. Our investments in recent years have provided us with substantial manufacturing footprint to support further growth. In terms of manufacturing space, we have the capacity in place to support approximately three times our current sales level. Our focus today is therefore, first and foremost, on the broader supply chain. We are working closely with our suppliers to secure the capacity and critical components needed to meet rapidly growing customer demands in 2027 and beyond. We continue to see a very strong demand environment.
Hichem M'Saad: The continued build-out of AI infrastructure drives significant demand for additional semiconductor capacity, particularly at the leading edge. Expanding semiconductor capacity requires significant time, from fab construction through equipment installation, qualification, and production ramp-up. In response to sustained demand growth, customers are planning further ahead and working with us earlier in the capacity planning process than in prior cycles. This provides us with a good level of visibility into their requirement through 2027, and in certain areas into 2028. We see this as a positive indicator that customers are planning for sustained capacity needs, supporting a more durable growth outlook for the industry. Turning to advanced logic foundry, demand remains very strong. Investment in 2 nm continue at a high level as customers expand capacity and more advanced AI-related and other high-performance computing applications move to this node.
Hichem M'Saad: We are also seeing increased investment activity across the 3 nm to 5 nm nodes, supported by strong demand for products such as CPU for agentic AI, and base die applications for HBM stacks. At the same time, momentum at the 1.4 nm node continues to build. Q3 was the first quarter in which 1.4 nm made a solid contribution to our revenue, primarily reflecting investment in customer pilot lines. We expect this contribution to increase further in the coming quarters and to become a significant driver for our sales in 2027 as customers start investing in high volume manufacturing. There are several reasons why we are particularly excited about the opportunity at 1.4 nm. First, the opportunity for ASM expands substantially, both in terms of served available market and market share.
Hichem M'Saad: With the first generation of Gate-All-Around at 2 nm, the focus was primarily on successfully making the transition to a new transistor architecture. At 1.4 nm, customers are increasingly focused on extracting more performance from that architecture, driving demand for additional materials and ALD process steps. This plays directly into ASM's core strength in materials discovery. Epitaxy is another important growth driver. With the introduction of Gate-All-Around, nanosheet channel epi emerged as a significant new application, and we were very successful in capturing that opportunity. At 1.4 nm, we are building further from there, securing additional applications and increasing our share. Taken together, these developments are driving a significant increase in our served available market at 1.4 nm compared with 2 nm, while also supporting further gains in ASM's market share. Second, the wafer capacity opportunity at 1.4 nm looks highly attractive.
Hichem M'Saad: The substantial improvement in performance and power efficiency enabled by 1.4 nm will be particularly important for the next generation of AI logic devices. Customer interest is strong, with some already talking about 1.4 nm potentially becoming a larger node than 2 nm. Importantly, we see multiple customers actively working on this node, further expanding the scale of the 1.4 nm opportunity for ASM. Turning to memory, DRAM investment increased significantly in Q3. Capacity remains tight, particularly in high performance DRAM, and higher customer spending supported strong memory sales during the quarter. Looking beyond the current capacity investment, we see a substantial multi-year opportunity for ASM. DRAM technology roadmaps remain dynamic. While the industry transition to 4F² architecture may take longer than previously anticipated, development activity remains strong.
Hichem M'Saad: What remains unchanged is the need to continue improving the speed and power efficiency of each new generation of high performance DRAM. What we are seeing as a result is that some innovation previously expected to come with 4F² are being pulled forward to 6F². FinFET is a good example. Customers are now introducing FinFET in the CMOS peripheral circuitry at 6F² to improve performance. This drives additional ALD layers in the high-k metal gate stack, as well as creating opportunities in other applications. We are already engaged in a number of ALD applications for FinFET and other new innovations that enable these performance gains. We are also seeing epitaxy becoming increasingly important in DRAM. New epi applications are emerging as customers push device performance further, and we have already qualified additional applications for upcoming nodes.
Hichem M'Saad: While the precise DRAM roadmap continues to evolve, the underlying direction is very clear. Customers need more performance, and achieving that requires more advanced materials, more ALD, and more epitaxy. For ASM, this creates opportunity to expand our served available market and increase our market share. in 2026, our growth is still driven primarily by industry capacity additions. Beginning in 2027, our new innovations will start to have a greater impact in 2028 and beyond. Another area where we see attractive long-term growth opportunities is advanced packaging. As AI systems become more complex, advanced packaging is becoming increasingly important, creating new opportunities for our materials and deposition technologies across a number of applications. One particularly exciting area I would like to highlight this time is photonics, which is growing very rapidly. As the performance and bandwidth requirements of AI systems increase, photonics and co-packaged optics are becoming increasingly important.
Hichem M'Saad: We have established a leading position in epitaxy for photodetectors, positioning us well to participate in this growth. Aside from photonics, we are also developing a broader set of solutions across ALD, EPI, PECVD, and Preclean to capture the opportunities in advanced packaging. While we are starting from a relatively small base, we believe advanced packaging can become an important additional growth area over time. More broadly, we continue to look for opportunities to leverage our materials discovery and chemistry expertise to expand into new growth areas. Last month, we launched XP8 Vertos Flowable Carbon. It is the industry's first PECVD flowable carbon process technology and has already been deployed in high volume manufacturing. Developed around a novel ASM-designed molecule, Flowable Carbon addresses increasingly complex patterning challenges associated with advanced 3D logic and memory devices.
Hichem M'Saad: While conventional carbon films have delivered significant value to the semiconductor industry for more than two decades, they were not designed for today's increasingly challenging three-dimensional structures. Flowable Carbon enables carbon films to flow into higher aspect ratios and reentrant features while maintaining the low defectivity critical to carbon patterning applications. In addition, thanks to its unique flowability, the film naturally creates a more planarized and uniform surface across different pattern density and across the wafer, reducing process complexity and the need for downstream CMP steps. Flowable Carbon is a great example of how we can leverage our core capabilities to successfully expand into new growth areas and broaden our served available market over time. Let me finish with our outlook. As communicated in our earnings press release for Q4, at constant currencies, we expect revenue of EUR 1.23 billion, plus or minus 5%.
Hichem M'Saad: Supply chain conditions are increasingly tight, and we continue to work closely with our suppliers and customers to support the planned ramp in shipments. Based on our Q4 guidance, we expect full year revenue to grow by approximately 35% at constant currencies, with equipment revenue growing even faster at more than 40%. Looking further ahead, the opportunities I discussed today give us a high level of confidence in continuous strong growth in 2027. Advanced Logic Foundry is expected to be the key growth driver. Within this segment, 1.4 nm is expected to become a very strong driver for our sales as customers start investing in high volume manufacturing. While we also expect further growth in 2 nm. In memory, we expect strong growth supported by higher customer investment and our expanding ALD and epitaxy positions. We also expect the recovery in power analog wafer to strengthen further.
Hichem M'Saad: Against this backdrop, we project revenue to exceed EUR 5.7 billion in 2027 at constant currencies. This means we expect to achieve the revenue target set for 2030 at our 2025 Investor Day, three years earlier.
Victor Bareño: Thank you, Hichem. Let's now move to the Q&A. To accommodate as many callers as possible, please limit your questions to no more than two at a time. Operator, can we have the first question, please?
Operator: Yes, thank you. This is the Chorus Call conference operator. We will now begin the question and answer session. Anyone who wishes to ask a question may press star and one on their touchtone telephone. To remove yourself from the question queue, please press star and two. Please pick up the receiver when asking questions. Anyone who has a question may press star and one. At this time, the first question is from Jakob Bluestone, BNP Paribas.
Operator: Yes, thank you. This is the Chorus Call conference operator. We will now begin the question and answer session. Anyone who wishes to ask a question may press star and one on their touchtone telephone. To remove yourself from the question queue, please press star and two. Please pick up the receiver when asking questions. Anyone who has a question may press star and one. At this time, the first question is from Jakob Bluestone, BNP Paribas.
Jakob Bluestone: Hi. Good afternoon. Thanks for taking the questions. I've got two. Firstly, can you maybe just comment a little bit more on the degree of visibility you've got into 2027? So for example, how well covered is it out of your order book? The second question is, you've obviously hit your CMD revenue targets three years early. Just interested in, I appreciate there's obviously been a lot more capacity expansion than anticipated at the time. How should we think around the uplift in SAM per 100,000 wafer starts, for 1.4 and this sort of DRAM transition as well? Are those SAM uplifts in your view, still valid, or do you think those numbers have moved up?
Paul Verhagen: So let me first, this is Paul speaking. Let me first talk about the degree of visibility. What we see today is clearly a much better visibility than what we have seen at least in the prior years. That is also one of the reasons why we came out with an early guidance for next year. Based on discussions with customers, level of detail of the plans they share with us, timing of what they need by when, et cetera, it is clear that they are working very closely with us to ensure that they get what they need, and which can give us, of course, the confidence that the amount is real. We know their fabs, we know when they come on stream, et cetera, at least when they are planned to be come on stream. We can use that information again towards our supply base.
Paul Verhagen: Visibility is clearly much better. That does not mean that that is all already converted into POs. It is maybe slightly better than other years. Still, the level of visibility based on backlogs, based on POs is not a hell of a lot better, I would say. Also because specific configurations sometimes still can change last minute. Orders were really good, and usually the backlog at year end, we have mentioned that at one time, we disclose backlog. So early next year when we come to full numbers, you see that reflected in the backlog. On the SAM, Hichem already talked about it. We are very excited about 1.4. It becomes a very solid, or it has become a very solid contributor to revenue. This is going to significantly increase in 2027. It will be actually pretty big.
Paul Verhagen: We see growth in 2 nm in 2027 compared to 2026. But 1.4 will actually be the biggest contributor in that year. So already bigger than already a big number for 2 nm, which is good. In addition, also, 3 to 5 will have a healthy contribution, but at a much lower level than 2 nm and 1.4. The SAM that we communicated in the Investor Day is still valid, but we believe it is at the higher end of that range, what we see. So we guided for 450 to 500 million per day wafer starts. Based on everything we know today, we see today, we believe that is towards the higher end of that range. So all looking pretty good.
Jakob Bluestone: Very clear. Thank you.
Paul Verhagen: Thank you, Jakob.
Operator: The next question is from Tammy Qiu, Berenberg.
Tammy Qiu: Hi. Thank you for taking my question. The first one is, your decision of guiding 2027 is quite early. Is that because of the customer commitment has improved significantly in the quarter to make you make that decision to give us a precise number at this point?
Hichem M'Saad: Yeah, I will take that question. Yes, Tammy, we have a very good visibility, yes, with customer for 2027. The visibility became better the past quarter. You are absolutely right. And we feel that based on that visibility that we have at 2027, that we are able to provide with EUR 5.7 billion revenue in 2027.
Tammy Qiu: Okay. I just wonder how conservative you are on that number, because obviously your lead time is about six months. I guess a lot of moving parts potentially can happen in H2 2027. Have you based that 5.7 plus number on a very conservative estimation of maybe something will turn out to be not going on, et cetera? How conservative you are?
Hichem M'Saad: I think that if you look into how the industry is going to go in 2027, the growth is going to be over 30%. If you look into the number of 5.7 compared to our revenue, in 2026, that is about the same number. Okay? From that point of view, we already have always mentioned that we are going to grow at least at the level of the market, if not higher. The 5.7 actually mimics the wafer fab equipment increase in 2027. If you take further that our position in 1.4 nm is also very strong, both, in ALD and epitaxy, and the fact that the SAM has increased there. We think that EUR 5.7 billion is a very good number to go with.
Tammy Qiu: Thank you.
Operator: The next question is from François Bouvignies, UBS.
François Bouvignies: Thank you very much. My first question is on this FinFET on the memory side versus planar move. You mentioned, Hichem, that maybe 4F² could be later. My original idea was that FinFET for memory would be more 2028, 2029 timeframe. The fact that you have 4F² may be pushed out, do you see a pull forward of the 6F² FinFET into 2027 instead versus 2028? I thought that it would be, in terms of ALD layers, two or three times higher than the planar structure. Is that the right ballpark?
Hichem M'Saad: Yes, François, thank you for your question. Yes, definitely we see FinFET being deployed in 2027. We see that also the ALD layer in FinFET to be higher, 3X from that point of view. That really comes down to our confidence in the number that we project for our growth in 2027.
François Bouvignies: Thank you. My follow-up would be on these shortages. We hear more and more in the supply chain that some components are quite tight, and you seem to acknowledge that in the release. What area is particularly tight? Obviously you have a big growth project. I was wondering how much risky it is due to this components tightness, and what specifically is tight for you at this point?
Paul Verhagen: Yes, I will take that question, François. The supply chain is for sure challenging, to say the least, and we are on top of it. Also going into next year, that is definitely a continuous point of attention. There is certainly some risk. I would be lying if we would say there is no risk. We have people at a few critical suppliers. We are qualifying new suppliers, and we are taking a number of actions to manage it as good as we can. So far, we have been able, with pretty hard work, to manage it properly. It will be definitely at our radar as an increased risk going into 2027. We have quite a few actions ongoing to hopefully manage that. I can tell you the precise components, but we do not want to talk about that specifically for a number of reasons.
Paul Verhagen: I am not going to go into which components specifically. We see shortages. There are a few ones where we see a lot of stress, and that we are on top of, but we prefer not to go into detail which ones those are.
François Bouvignies: Okay. Thank you very much.
Hichem M'Saad: Thank you.
Operator: Next question is from Stéphane Houri, Oddo BHF.
Stéphane Houri: Yes, good afternoon. Actually, my question is, my first one is on 2028 because I know you made an effort to give us guidance for 2027. Given where the industry is and the plans with the new fabs in advanced logic, but also in memory, do you already have a first view or discussions with your client? Some of your peers, not really competitors, but like in the same environment, already talk about 2029 and 2030. So where does your visibility lead you at the moment?
Hichem M'Saad: Thank you very much for your question, Stéphane. To answer your question is that, we see good growth continuing beyond 2027 into 2028. With some of our leading edge customer, they shared with us their roadmap for 2029 and even 2030. Even on 1.0 nm node, we also have that kind of visibility right now. Saying this, we think that 2028 will be also a growth year. As you know that 1.4 nm HVM production will happen in 2028. That 2027 is only the pilot production from that point of view. The 1.4 nm, as was mentioned by many of our customers, is going to be a bigger node than the 2 nm node. Because the 1.4 nm has more players than the 2 nm node, we expect continuous strong sales in 1.4 nm node in 2028 and beyond.
Hichem M'Saad: Second, for us, the growth that we have experienced the past year and also in 2026 and 2025 in DRAM, for high performance DRAM, is being based mainly on capacity addition. By the end of 2027, and most importantly in 2028, some of the DRAM layers will use more ALD and EPI application, which we have won this year. So we expect also to see further growth in our DRAM market, in memory market in 2028 and beyond. The third factor is also in the area of advanced packaging. We are just starting there. We talked about photonics. We are very excited about photonics. We are using the photodetector using our epitaxy. We see new areas growth in also advanced packaging in PECVD. Most importantly, we see that we have more and more applications beyond our core EPI and ALD technologies.
Hichem M'Saad: Overall, what we feel is that, okay, that we have new growth area. Our served available market is expanding. We are gaining share in DRAM. So pretty much what would happen from one way or the other, that even if there is some economic effect and so on and so forth, I think the projection for us would be resilient and positive.
Stéphane Houri: Okay. Thank you. Very clear. Second question is around gross margin. You have reached, well, you say 52% for the year, which is clearly above the high end of the guidance that you have given. Given all the things that you are saying, you are also saying, if I am correct, that power is going to recover, mature nodes are going to be supportive again next year. Is there a reason why the gross margin would go down from the 52% level that you have now? Thank you.
Hichem M'Saad: I will answer this question this way. We would like to be able to continue to command a high profitability for the company. As you know, the gross margin also depends on product mixture, but also depends on customer mixture from that point of view. We are actually also having very strong program to continue improving our gross margin. Some of it right now, I can even mention that, okay, we have, in some cases, increased our pricing for the value that we are providing. We feel that we will be aggressive in maintaining the gross margin that we have from that point of view. On the other hand, I think the supply chain concerns, not concerns, but the fact that the supply chain is also tight. We see some also increase in some component pricing from that point of view.
Hichem M'Saad: The positive, you increase the pricing, you have more value, you have better efficiency that would tend to improve your gross margin. At the same time, component cost will go the other way around. Overall, we do believe that the trajectory for our gross margin will be positive going forward.
Stéphane Houri: Thank you very much.
Hichem M'Saad: Thank you.
Operator: The next question is from Didier Scemama, Bank of America.
Didier Scemama [Equity Research: Yes. Thank you. Good afternoon. I have got a couple of questions. Maybe first question for Hichem on the PECVD tool. I know you started to talk about that more prominently last quarter. Obviously, the CVD, which is bigger than PECVD, I get it. The TAM is three times the size of ALD. I just wondered with that sort of differentiated PECVD system, what sort of market share do you think you could get, and over what time frame? I have got a follow-up. Thank you.
Hichem M'Saad: Didier, for a person who spent 15 years in PECVD, I tell you that for ASM, this is really a very small market share than the other big competitor. PECVD is a huge business. We have been a small player in PECVD for the last 20 years, mainly having some locations at a couple of customers, but that was it. This PECVD film that we have, the Flowable Carbon, is a decision that we have made that our expertise and our materials discovery expertise, we would like to expand it beyond ALD, and PECVD is one of those area. With that, we have developed a new molecule, designed a new molecule that will provide these PECVD film, which we call Flowable Carbon. That is enabling much better patterning for 3D structure.
Hichem M'Saad: With it, we see a significant customer acceptance of this technology because it provides a much better yield of their devices. From that point of view, in this part of the market, this is a new market for us, new area for us in PECVD. We feel very good about where we stand from this point of view. On PECVD too, there are also application in advanced packaging. In that, we have some new products that we have developed in the PECVD for advanced packaging. We will be able to share with you in the next few quarters some of the wins that we have in this area when they happen.
Didier Scemama [Equity Research: Hey, congratulations and best of luck with PECVD. My second question is for Paul on OpEx. I think we saw a pretty big uptick in OpEx in the quarter, which is perhaps understandable given your growth rates on the top line. I am just wondering, given the guidance you have given for 2027, how should we think about OpEx growth next year?
Paul Verhagen: Thanks for the question, Didier. Indeed, you saw an uptake, but to a reasonable extent, at least in SG&A, that is variable expense based on the company performance and the growth. Still, it is an expense. In R&D, we had an impairment of EUR 7 million, so also that number is a little bit inflated. Adding it all up, you get the numbers that you saw. We added 400 across the board and not only in R&D and SG&A. Obviously, the biggest part was actually in operations and field service engineers. We added 400 people in the quarter to support the growth that we see. Also there, we are expanding. We also had some additional one-off, I say that, benefits for our employees, because it is happening almost across the industry.
Paul Verhagen: That is also important to properly compensate our people, and to make sure that we remain competitive in this industry where there is a war for talent going on. All these elements also play a role. Come to 2027, we guided for 2030, you might recall, R&D at low double digits, and SG&A based on EUR 5.7 billion revenue or more below 7%. That guidance is now also valid for 2027. R&D, we still target. If we can achieve it, I do not know, because we cannot even keep up with the growth, to be very honest, in R&D. We might lack a little bit there, but we target at least to be low double digits, and SG&A should come below 7% next year.
Didier Scemama [Equity Research: Very clear. Thank you very much.
Paul Verhagen: Thank you, Didier.
Operator: The next question is from Timm Schulze-Melander, Rothschild & Co Redburn.
Timm Schulze-Melander: Great. Thanks so much for taking my question. The first one is just trying to reconcile two things. I think you talked about a very strong book of client interest demand, supply chain tightness. I think you also mentioned that it hasn't converted to purchase orders, and I think that's something that's been echoed by some of your peers. Could you maybe just provide some color there? Is it the deposit rates that are being demanded? Why is that not flowing through? Then I had a follow-up, please.
Paul Verhagen: That is a good question, Timm. Why is it not followed through? I guess it is habit. I do not know. This is how we work with our customers. Typically, we see 3 to 6 months of backlog, and sometimes, depending on time of order, it can be a little bit higher or a little bit lower, but that is basically the trend that we have seen. That is how customers work with us. That is still the case. They are not converting what they need immediately into POs, partially also because we have quite some configurations. Configurations are typically decided later in this process. We know what kind of tool, we know what kind of platform, but we do not know the final configuration always that early. But what we do know is what I said in the beginning of this call.
Paul Verhagen: We have so much more granularity from customers compared to what we have seen in prior years, plus a much more committed tone, let us say it like that, from our customers. Based on that, and also based on the fact that their chip demand is still higher than what they can supply, that is where we base this visibility on. Of course, great visibility does not necessarily mean certainty. There is a difference between these two. But based on everything we know today, based on everything we see today, based on everything we hear from our customers, that has basically resulted in the disclosure for 2027.
Timm Schulze-Melander: Great. Just one quick follow-up. The commentary around the technology delivery has been really impressive for a number of quarters. You referenced epi wins, you have referenced Flowable Carbon. How come there has been a write-off in R&D? Maybe just give us some color about what that relates to. Thank you.
Paul Verhagen: We have, I do not know how many projects, but a hell of a lot. Really a lot. Tens and tens, if not 50s. You will always have, and actually we need to have, I would almost say, now and then a project where the win is not there or the customer finally restarts a certain development, but finally the benefits that we would anticipate do not materialize, so the customer will not move forward with a certain development. You need to have that, because otherwise you do not stretch the envelope well enough. This year, I think this is our first one. I am not done. Maybe had a small one, EUR 2 million, I am not sure if it was this year or last year.
Paul Verhagen: But given the amount of money we spent in R&D, and if you have this amount of impairment, I would almost applaud it, because we need to have that. So there will always be. We do not have projects, let us say, that require EUR 100 million development costs. We have projects that require EUR 10 million, EUR 15 million, EUR 5 million, EUR 25 million, but all much, much more. And sometimes, a project will not result in a winning proposition, and then you need to impair it. And that happened.
Timm Schulze-Melander: Very clear. Thank you.
Paul Verhagen: on some projects.
Timm Schulze-Melander: Very clear. Thank you.
Paul Verhagen: Thanks, Timm.
Operator: The next question is from Adithya Metuku, HSBC.
Adithya Metuku: Yeah. Good afternoon, guys. Thank you for letting me on. Two questions, please. Firstly, just coming back to the gross margins, Paul. You talked about an improving trajectory from here on for gross margins. If I looked at your etch and dep peers, they all recently updated their gross margin targets. They also gave targets at a similar timeframe, in 2025. They are talking about 300 to 400 BPS improvement in gross margins over the coming years. Would that be a reasonable assumption to use for you guys as well? Essentially, your gross margins can go into the mid-50s, maybe with a two, three-year view. Your peers face the same issues as you face in the industry around supply costs going up, et cetera, and yet they are talking about these things.
Adithya Metuku: Is there any reason why you would not see a similar level of increase, especially given you are on the leading edge where you could arguably say you have more pricing power than your peers who have more broad-based exposure to WFE? That is my first question, and I have got a follow-up.
Paul Verhagen: Yeah. Okay. Hichem already mentioned that, of course, the trajectory should. Our target is always to improve, let's say it like that. We will end this year with 52%. There will be pluses and minuses next year, also based what Hichem said, in terms of sales prices, cost, inflation, et cetera. I am not going to repeat that. There will be efficiencies in the supply chain that we are working on, in terms of, we have mentioned many times, this merge and transit initiatives that we are developing and that we are gradually rolling out in our products and platforms. We have efficiencies. There is also a positive structural benefit from operating leverage, although not a lot, but every year, 0-point something percent is added there given the growth of the company. It starts to add up. Could we move higher?
Paul Verhagen: I am not going to say yes, I am not going to say no. What I do want to say is that for next year, our starting point is where we are today. Of course, we will try to see how much we can improve. Then I would expect that during next year's Investor Day, which is very likely somewhere in the second half of the year, new guidance will be provided, whatever that will be. I think it is too early now to already give you a new range, but I am pretty sure that that will happen during Investor Day next year. For now,
Adithya Metuku: Understood
Paul Verhagen: You can starting point that Hichem already shared is that our trajectory should be good, and that is our target.
Adithya Metuku: Understood. Then, just a quick follow-up on the photodetector point you made earlier, Hichem. You talked about epitaxy wins. Can you talk a bit about what materials are you looking at? Is this gallium arsenide, indium phosphide or something else? Any color you can provide on that, and what are the metrics that are allowing you to win versus the incumbents in that space? Any color you can provide there would be great.
Hichem M'Saad: What I can tell you about the films that we are running is that it is group 4 element epitaxy. I am not going to say exactly what it is, but group 4. You can look at the table of elements to know what kind of elements are there. From that point of view, this is really where we stand. We have seen significant acceptance of our technology from our customer because when you are doing photodetector, you need to have a very good detection ability. For that, your epi film has to be super pure. Actually, we have some technical advantage in our technology whereby we can achieve very high-purity detector epi layer, which provide us with the leading share that we are having right now.
Adithya Metuku: Understood. I take it it is germanium, but I will not push you on that. Thank you.
Hichem M'Saad: Thank you.
Paul Verhagen: Thank you, Abhi.
Operator: The next question is from Sandeep Deshpande at JPMorgan.
Sandeep Deshpande: Hi. Thanks for letting me on. My first question is, you have now given 2027 growth, based on current guidance, as for this year is about 34%. This year, you are growing faster in the equipment market than in your spares and services, et cetera. Should we assume that sort of trend that the equipment growth will be better than the spares and services growth in 2027 as well? That would seem to indicate you continue to gain share in the WFE market, and that is probably because of the points you made earlier in the call. My second question is probably to Paul, which is to do with the overall backlog that you have. Have you been taking deposits for these orders that you are reporting into FY27? Has there been price increases in some of the tools that you are shipping into next year?
Sandeep Deshpande: Thank you.
Hichem M'Saad: Okay. Thank you very much for your question. We see a continued strong growth in our equipment sales in 2027, and we expect that growth in equipment sale in 2027 to exceed our growth in service products. Which really tells you that what we have been saying in the conference call, that we are gaining share in equipment, in our market share, and we are very excited about that. For the second question, I am going to let Paul answer that question.
Paul Verhagen: Yes. So on the backlog and whether or not we see deposits, we see actually for quite some time, deposits is what is accounted for in deferred revenue. These are invoiced revenue, but deferred. So invoice and paid, but deferred per the customer request because tools have not yet been delivered or spare parts have not yet been delivered or whatever the deliverable is against these credits. So for that part, a big chunk of that is paid for. Other than that, maybe a little bit, but no, nothing material, Sandeep.
Hichem M'Saad: Thank you.
Paul Verhagen: Except for the first part.
Hichem M'Saad: Yeah. Thank you.
Operator: The next question is from Ruben Devos, Kepler Cheuvreux.
Ruben Devos: Yeah. Thank you. I got two remaining. Just one on China. I think, in July, you talked about some ordering ahead of possible export controls, though not excessive. If we strip that out, what does underlying Chinese demand look like for 2027? Could you also talk a bit about the mix between mature logic, memory, and power, in China? I think you talked about not yet seeing aggressive participation in Chinese DRAM ramps, but has that changed at all in the meantime? Thank you.
Paul Verhagen: Yeah. China this year is actually very strong. We have seen very good growth in China, in particular in mature logic foundry, which is now actually gradually coming down from the height that we have seen, maybe in Q1 not a lot, but at least it is coming down a little bit, but still good. But that is fully compensated by growth in memory from a low base and also growth in power via analog.
Paul Verhagen: For next year, as you know, visibility in China is always a little bit more difficult there. We talked about good visibility, that is true, but in China, that is still different compared to some of the larger customers that we have. But still for China, let us say, it is still relatively good visibility. We still see continued good demand, strong demand, similar to this year, I would say. We do not expect a lot of growth there.
Paul Verhagen: It is anyhow too early to tell, but that is not what we see today. But it remains at a good, robust level. The growth that we will see next year and the growth that we talked about, that is all coming from advanced logic foundry and from, let us call it advanced memory for now. That is not so much the case in China. Then, of course, the uncertainty for China is the MATCH Act.
Paul Verhagen: Yeah, we have said before, it will have impact, but we are not going to speculate on how much. It is just impossible given the discussion that is going on and given potential changes that might still come in, and whether or not the bill will make it. But that is an uncertainty related to China. But if I forget that for a second, park that for a second, we expect still a good strong China next year.
Paul Verhagen: But based on everything we know today, not a lot of growth. Very similar, I think, to this year. That's the best that I can give you now.
Ruben Devos: Okay, great. Thanks for the color. Just a second one on NAND, actually. We're seeing investment pickup there as well. We've seen new fabs approved in Korea. I think some of your peers have included it in their growth outlooks for next year. I understood NAND is typically a smaller part of your business. But how relevant may it become again for you? And maybe could you talk about some of the relevant EPI or ALD layers here? Thank you.
Hichem M'Saad: Yeah. Hi, I can take this question. Yes, definitely we see increase in 3D NAND capacity and purchases for 2027. This is a significant increase in that, but for us, it's coming from a very small basis. Where we play in 3D NAND is mainly in ALD, mainly for layers like gap fill, for high aspect ratio layers, liners, and the patterning applications.
Ruben Devos: Okay, great. Thank you.
Hichem M'Saad: You are welcome.
Operator: The next question is from Lee Simpson, Morgan Stanley.
Lee Simpson: Great. Thanks for squeezing me in. Most of my question has been answered, but maybe if I could ask a clarification question. I think you were kind enough to say that 1.4 nms were about double digits of the advanced logic foundry business. Just wondered if you could clarify that. Then I think you said this will be bigger in dollar terms next year versus 2 nm. Am I right to assume those two clarifications? Thanks.
Paul Verhagen: Yeah. I think what we said on the double digit for advanced logic foundry for this year, that is correct. Relative for the advanced logic foundry revenue next year, and by the way, that was specifically for Q3, should not forget to add that. But it is a meaningful solid contribution already. Also the same we expect or even better for Q4. Next year, we expect a significant increase where I already mentioned that 2 nm is already strong and is expected to grow next year compared to this year. So 2027 versus 2026, there will be growth. But even with the growth in 2 nm, we still expect 1.4 to be stronger. So 1.4 is really contributing a lot next year, but already this year, it is a very solid contribution in H2 of the year.
Lee Simpson: Great. Very clear. Maybe just if I could ask a follow-up on China, following on from the earlier question. If indeed the dynamics here are for slower growth, let's say, relative to other parts of the business, does this introduce some mixed headwind risk on the margin next year?
Paul Verhagen: Everything else equal, yes. China typically is accretive, so less China or relatively less China, so relative to the total, would not be positive to margin. That is correct. But of course, we are not standing still. As I said already, we are basically making improvements on many fronts. That, of course, should also support margin, and hopefully offset and maybe even do better than whatever we lose because of relatively lower China sales. But yeah, everything else equal, if we would sell exactly the same mixes this year, with exactly the same prices, but just less China, yeah, margin would be impacted. It would be a little bit lower than. I am not going to say more than that. It would be a little bit lower. Yeah.
Lee Simpson: Got you. Very clear. Thank you.
Victor Bareño: Thank you, Lee. Operator, as we are running out of time, can we please move on to the last caller in the queue? We still have a few more in the queue, so if you did not have the chance to ask your question, reach out to us after the call, please.
Operator: Okay, so the last question is from Robert Sanders, Deutsche Bank.
Robert Sanders: Yeah, hi. Thanks for squeezing me in. Yeah, just coming back to the supply chain dynamics. When you look at what is constraining your most important sub-suppliers, like mass flow controllers, for example, when you look at what is scaling them up faster, is it the company's own mature chip supply, or is it other non-semiconductor factors? Do your customers get involved with sorting that out and prioritizing your demand over other OEMs? I have a follow-up. Thanks.
Paul Verhagen: I am not sure I fully understood the first part of the question, Robert. On the last part, I can say at least no, customers do not get involved in sorting out supply chain for us or with us. There it can be pretty straightforward. I am not sure if there was another part of your question.
Robert Sanders: When you look at what is constraining them, in the past, if you remember in 2021, it was not enough mature chips like microcontrollers, et cetera. Is that the case too, or is it something else?
Paul Verhagen: No, it is something else.
Robert Sanders: Okay. Then in terms of how to secure capacity as you ramp up with these sub-suppliers, are you prepared to pay down payments or even sign take-or-pay type agreements to secure capacity, or is that not something you are considering? Thanks.
Paul Verhagen: We will do what is needed to secure our supplies, and sometimes that might involve indeed some payments, but so far not a lot, to be very honest. But if that would be needed to secure supplies that we need, we for sure will seriously consider that. Yeah.
Robert Sanders: Okay, good. Thanks a lot.
Victor Bareño: Thanks, Rob. Okay. Thank you all for attending our call today, also on behalf of Hichem M'Saad and Paul Verhagen. Goodbye.
Operator: Ladies and gentlemen, thank you for joining. The conference is now over and you may disconnect your telephones.
