Afternoon, this is the Chorus C all conference operator. Welcome, and thank you for joining the ASM second quarter 2025 earnings call. As a reminder, all participants are in 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.
Thank you, Operator. Good afternoon and welcome, everyone, to our 2025 Q2 earnings call. I'm joined here today by our CEO, Hichem M'Saad, and our CFO, Paul Verhagen. ASM issued its second quarter 2025 results yesterday at 6:00 P.M. European time. The press release is accessible on our website, asm.com, together with our latest investor presentation. As a reminder, 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 our forward-looking statements, please refer to our company's press releases, reports, and financial statements, which are available on our website. Please note that the profitability measures mentioned in this call will be primarily based on adjusted non-IFRS figures. For the reported results, as well as the reconciliation between IFRS and adjusted results, please refer to the quarterly results press release.
I will now hand the call over to Hichem M'Saad, CEO of ASM.
Thank you, Victor, and thanks to everyone for attending our second quarter 2025 earnings call. Before we start, I'd like to remind everyone that we'll be hosting our Investor Day on September 23rd in London. We hope that many of you will join us in person in London or via the webcast. For today's call, we'll be following our standard agenda. Paul will begin with an overview of our second quarter financial results. Next, I will discuss the market trends and outlook, followed by the Q&A session. I will now turn it over to you, Paul.
Thank you, Hichem, and also thanks everyone for joining our call today. Our revenue in the second quarter of 2025 amounted to EUR 836 million, up 23% year-on-year at constant currency. Compared to the first quarter, sales increased by 7% at constant currency, which was above the top end of our guidance of 1%-6% constant currency increase. As you'll probably remember, with our Q1 results, we changed our guidance to constant currency based following the U.S. dollar weakening in recent periods and taking into account that the majority of our sales, 80% plus, is billed in U.S. dollars. Equipment sales increased 25% year-on-year at constant currency and were led by ALD, followed by Epi. Spares & Services sales were up by a strong 17% at constant currency.
Next to solid underlying growth in our outcome-based services, there was again some benefit from accelerated demand in China, albeit not as strong as in previous quarters. In terms of customer segments, revenue was led by foundry, followed by memory, and then logic. Logic foundry sales accounted for the majority of sales and were up both year-on-year and compared to Q1. Advanced logic foundry sales, for the largest part related to the 2-nanometer gate-all-around node, were up strongly to the year-ago period and slightly lower versus Q1. Memory sales decreased year-on-year but were at a similar level as in Q1. Sales in this segment continued to be predominantly driven by high-k ALD tools for high-performance DRAM in advanced HBM-related applications. 3D NAND sales were relatively stable compared to Q1 and lower year-on-year and remained the smaller part of our memory sales.
Sales in the power analog wafer segment continued to be under pressure and were down year-on-year and approximately similar to Q1. Gross margin in the second quarter came in at a very strong level of 51.8%, down from 53.4% in the first quarter and up from 49.8% in the second quarter of last year. The increase in gross margin was explained again by a favorable product and customer mix and including strong contribution from China and operational improvements. For the full year 2025, we still project the gross margin to be in the upper half of the target range of 46%-50%. This implies that the gross margin in the second half will be lower compared to the first half, with a more normalized sales mix amongst others in China. Similar to the statement we made last quarter, this excludes any impact from potential tariffs.
It is still unclear what the tariffs for our business will be, but we have prepared several scenarios to mitigate potential direct impacts, including flexibility to expand localized manufacturing in the U.S.. SG&A expenses were 8% lower year-on-year when adjusted for the one-off tax expense that was included last year. This decrease reflects our continued focus on cost control and discretionary spending. For the full year, SG&A is expected to be somewhat below prior year. Net R&D in Q2 increased by 18% year-on-year, reflecting ongoing expansion of the R&D programs. The faster increase in net R&D relative to gross R&D was due to higher amortization, as also discussed in previous quarters. For 2025, in line with earlier communications, we continue to expect that net R&D will be at the upper end of our target range, which is high single- digits to low double- digits as a percentage of sales.
Operating profit increased by roughly 40% year-on-year, adjusted for the aforementioned one-off in SG&A last year. This increase was the result of solid revenue growth, a higher gross margin, and strict cost controls on SG&A expense, while we continue to invest in growth and innovation. Below the operating line, financial results included a currency translation loss of EUR 60 million. This compares to a loss of EUR 40 million in the first quarter of 2025 and the gain of EUR 60 million in the second quarter of last year. As most of you know, we hold the largest part of our cash in U.S. dollars. The translation loss in Q2 was largely due to the meaningful depreciation of the dollar during the quarter. Let's move briefly to ASM PT.
Our share in income from investments, reflecting our stake of approximately 25% in ASM PT, amounted to EUR 4 million in the second quarter. It remained flat compared to the same period last year, reflecting the ongoing soft conditions in the backend equipment market. In addition, our net results include an impairment reversal gain of EUR 34 million. As a reminder, the first quarter of this year included an impairment of EUR 215 million of our stake in ASM PT, triggered by the reduced market valuation at that time. Following these impairments, the changes in the market value of ASM PT are included in our quarterly net results in case of further decline or until the impairment charge has been reversed. Let's now look at our order intake.
Our new orders amounted to EUR 702 million, a decrease of 10% compared to the first quarter and a decrease of 4% compared to the second quarter of last year at constant currency. Looking at the breakdown by customer segment, logic foundry was the largest segment, followed by memory and then power analog and wafer. Logic foundry orders were lower compared to the first quarter but up year-on-year. The sequential drop was explained by lower orders for 2-nanometer gate-all-around, and this is mainly reflected due to timing of orders. We believe the underlying demand trends continue to be healthy, and we expect orders in the leading-edge logic foundry segment to be up again in the next quarter in Q3. Mature logic foundry orders, mainly from Chinese customers, were again relatively strong in the quarter.
Memory orders increased compared to Q1 but decreased compared to the relatively elevated level in Q2 of last year. The largest part of orders in this segment was for advanced HBM-related DRAM applications. As we mentioned on earlier occasions, our sales and bookings were relatively high in 2024, especially in Q2 and Q3, which included some lumpy and relatively high memory bookings from China. Power analog and wafer bookings, including silicon carbide, remained at depressed levels in the second quarter. Turning now to the balance sheet, ASM's financial position continues to be solid. We ended the quarter with EUR 1 billion in cash, down from EUR 1.1 billion at the end of March. We generated a healthy free cash flow of EUR 125 million in the second quarter, up from EUR 103 million in Q2 of last year.
Days' working capital remains relatively low at 43 days, similar to the level at the end of March. CapEx amounted to EUR 44 million in the second quarter. We expect CapEx to further increase in the second half, for a large part related to the construction of a new facility in Arizona. For the full year, we expect CapEx to be north of our annual target. Range of EUR 100 million-EUR 180 million, probably somewhat in excess of EUR 200 million. During the second quarter, we paid the annual dividend to shareholders for a total amount of EUR 147 million, and we returned EUR 43 million through share buybacks. The EUR 150 million buyback program that started last April has been completed for 44% as of last week. With that, I'll turn the call back over to Hichem.
Thank you, Paul. Let's now continue with the review of the market trends. We delivered again solid results despite ongoing mixed trends across the different market segments. The overall market picture in the second quarter remained quite similar to the first quarter. The AI-driven segments, particularly gate- all-around technology and HBM DRAM, continued to show strength, while other areas experienced softer demand. Uncertainty continues to be relatively high because of tariffs, as well as ongoing geopolitical tensions, resulting in potential and direct impacts on the economy. We are closely monitoring development, but to date, business trends and customer conversation have not shown any significant changes. Let's now review the market segment in more detail, starting with the advanced logic foundry business. Revenue related to the gate- all-around technology was again robust in the second quarter.
Leading customers continued to invest in 2-nanometer capacity as they are about to ramp up this new node in high-volume manufacturing. Some of our customers have reported that in terms of initial demand, 2-nanometer is already outstripping the previous two nodes, and they expect 2-nanometer, including the sub-nodes that will be introduced in the next couple of years, to be a significant technology node. We have seen some further shift in the CapEx forecast among customers, but our view for the year has not changed. We still expect a strong increase in advanced logic foundry sales. Apart from an overall increase in advanced logic foundry spending, the gate-all-around node has also driven a significant increase in our serve-available market of $400 million U.S. dollars, as we discussed on earlier occasions.
On top of that, we have been able to maintain our number one ALD market share, and we have also delivered on our target to significantly expand our Epi market share in the transition from 3-nanometer to 2-nanometer. Following 2-nanometer, the next significant inflection will be the second generation of gate-all-around, the 1.4-nanometer node. We expect a further double-digit SAM increase in this transition. Complexity in the transistor, a key area of strength of ASM, continues to grow, requiring many new ALD layers in the next GAA nodes. One example is the use of multi-VT, which stands for multi-threshold voltage. VTs are generated by dipole layers deposited with ALD and are used to modulate the on-chip power and performance. The increasing complexity of next-gen nanosheet structures is driving an increase in VTs and thereby a growing number of ALD dipole layers and ALD patterning layers required for integration.
Other examples of new applications that we already previously discussed are the adoption of metal ALD and selective ALD in the next GAA nodes. Based on our very active R&D programs with all key customers, we expect that we can at least maintain our market share in ALD and Epi in the 1.4-nanometer node transition. We will, of course, talk more about the mid and long-term opportunities in our investor day next September. Next, the memory business. High-bandwidth memory-related DRAM continues to be the primary driver. ASM is strongly positioned with our high-k metal gate ALD technology, which is essential for enabling high-performance DRAM devices used in HBM stacks. Conditions in the other parts of the memory market remain more tempered. In the 3D NAND segment, we continue to see pockets of technology-driven demand, particularly for our ALD gap-fill solutions.
However, the overall contribution of 3D NAND remains modest compared to DRAM. Demand in 3D NAND tends to be closely tied to consumer markets such as PCs and smartphones, for which the recovery has so far been muted. As Paul just outlined, the year-over-year comparison is also impacted by memory sales in China last year. Even though our position in this segment is fairly small, the contribution of China memory in Q2 and Q3 of last year was relatively higher. We continue to expect memory sales to drop this year, and that's, of course, against a very strong level last year when memory sales jumped by more than 150%. As a percentage of equipment sales, we expect memory to decrease to somewhat less than 20% in 2025 versus 25% in 2024. Longer term, the memory segment is an important growth engine for ASM.
The DRAM roadmap is increasingly driven by the high performance, speed, and energy efficiency required by AI-related applications. The next nodes will see more and more logic-like technology, which closely aligns with our strength in ALD and Epi. We have strong R&D engagement in place with the leading memory customers, and we aim to further expand our share of wallet in these advanced nodes, particularly through increasing ALD layers as well as in nodes with new Epi applications. A significant inflection will be 4S squared or vertical channel transistor DRAM, which brings new 3D structure and increased process complexity requiring new ALD and Epi layers. In the power analog wafer segment, equipment demand remains depressed in the second quarter. This market has now been in a cyclical downturn for six quarters.
Some end markets are showing early signs of improvement, but this will most likely not be translated into new investment before 2026. We do not expect any meaningful sales recovery in the remainder of 2025. Silicon carbide is also experiencing a significant correction. We remain strategically well-positioned in this market, but demand is down by a double-digit percentage compared to last year. Looking at China, the sales contribution in the first half of 2025 held up well and was higher than in the second half of 2024, though still below the exceptionally high level in the first half of 2024. While some customers went into digestion mode, other customers made meaningful capacity investments. The power analog wafer segment remained soft, as just discussed. Memory sales in China declined year-over-year, as the prior year included incidental sales.
Apart from overall resilient demand in China, we also benefit from new customer wins. While ASM's sweet spot is in the advanced nodes, we invested in the past several years in innovative products for the mature nodes, as discussed on earlier occasions. This has contributed to many new customer engagements for ASM in China. Looking ahead, we now expect China equipment sales in 2025 as a percentage of total revenue to be around the top end of our guidance range of low to high 20s, though still below 2024 levels. This represents a slight improvement compared to our earlier outlook. For the second half of 2025, we expect our sales and, in particular, bookings in China to be lower than in the first half.
Before moving on to the outlook, I'd also like to highlight the continued solid performance of our spares and service business, with again double-digit growth in the second quarter. Since we started to invest in our outcome-based offering, our spares and services revenue has grown with a CAGR of almost 20% over the past five years. Apart from an increasing installed base, this growth has primarily been driven by new outcome-based services that create value for our customers by, for instance, lowering the cost of operating our tools and by increasing yield. In the remainder of 2025, spares and services will face a tougher comparison with the second half of last year, which, as most of you will probably remember, experienced accelerated demand in China. Adjusted for this, we expect ongoing underlying increases.
The structural growth potential of this business remains attractive as we further increase penetration with customers and as we continue to expand our portfolio with new innovative services. Let's now have a look at the guidance as outlined in our press release. For the second half of 2025, we expect revenue to be approximately similar to the level in the first half at constant currencies. For Q3 2025, we expect revenue to be flat to slightly lower in a range of 0% to - 5% at constant currency compared to Q2 2025. For Q3 2025, we expect advanced logic foundry booking to be higher than in Q2 2025 and China bookings to be lower, with the overall book-to-build ratio in Q3 projected to be below one.
Based on comparable sales in the second half, we expect revenue growth at constant currencies in full year 2025 to be around the midpoint of the guidance range of 10%-20%. The WFE market is still expected to grow slightly. As such, we remain on track to outperform WFE again this year. With that, we have finished our prepared remarks. We'll now proceed to the Q&A session. We'd like to ask you to please limit your questions to not more than two at a time so that as many callers as possible have a chance to ask a question. Operator, we are now ready for the first question.
Thank you. This is the Chorus Call c onference operator. We will now begin the question-and-answer session. Anyone who wishes to ask a question may press star and one on your touch-tone 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 François Bouvignies, UBS. Please go ahead.
Thank you very much. So my first question is, you know, we see your organic growth coming down for the full year. So H2 numbers versus H1 are coming down. China is coming down. Orders are also coming down. And with ASML last week, you know, that, you know, cannot guarantee growth in 2026 anymore. What does it mean for 2026, you know, this trend as, you know, numbers seem to come down in the second half? I am asking you because the consensus is, you know, before today at 14% growth year-over-year.
What can you tell us or help us and reassure the market around 2026 trajectory after all the moving parts?
Okay, thank you very much for the question. I think it is too early to provide guidance for 2026. I think if we look into the segments one by one, we are going to have a better idea. When you look into the data all around or the advanced logic foundry, 2026 will be a solid year, but really too early to be very specific. Some customers have already been talking about strong demand in 2-nanometer, and that is likely to continue adding capacity in 2026. For other customers, visibility is really more limited. In general, AI is really expected to be very strong in 2026.
The end markets that are more dependent on the consumer, such as smartphones and PCs, are more sensitive to the GDP development and any potential impact from tariffs. We definitely see that gate-all-around 2-nanometer being very, very strong, and customers are actually very busy working right now on the 1.4-nanometer technology node. We see also some buys in 2026 in the 1.4-nanometer node. If you look into memory, we actually expect memory in 2026 to continue healthy growth in HBM-related DRAM in 2026. For other commodity or conventional DRAM, it is really too early to tell. That would be very dependent on recovery in smartphones and PCs and also very dependent on the economy and the GDP. For power analog and wafer, the market now has been in downturn in six quarters. As you know, downturn starts in Q1 2025, and it is continuing for six quarters.
We see some really encouraging signs right now in the second half of 2025, which means that 2026 will be better in power wafer analog. The exact timing, we do not know, but we definitely see a recovery coming up in the power wafer analog in 2026. For silicon carbide, it is a little bit of a different story. Okay? We do not expect recovery anytime soon, but when that comes back up, we are actually very ready in silicon carbide. We have, as we mentioned, we have a new platform. We have made a significant improvement in cost of ownership for our customer and significant improvement in uptime and availability. We are really ready when the silicon carbide market will recover. If you look into China in 2026, it is really very tough to provide guidance in China.
It has been always very tough to provide any guidance for China. I think if we can compare to the very high spending levels we have seen in the last couple of years, in 2024 and 2023, and in the first half of 2025, I think we expect demand to normalize a bit in the coming years. That will be really a gradual process. We do not expect the demand in China to fall off the cliff in 2026, but we actually see it to normalize. Really, China also depends on export control, and that is really very, very, very difficult to predict. Overall, I think that we are really positive on gate-all-around in 2026. We are positive on HBM in 2026. We see the recovery in the power analog market from that point of view. I hope I was able to answer your questions.
Yeah, thank you, Hichem. It is much more than expected, so it is good. The second question is on, you know, you mentioned your market share, you know, in A14, you expect to be, you know, at least the same or above for ALD and Epi. Can you elaborate a bit more on that? I mean, where is it coming from, your confidence? I am asking because, I mean, obviously, everybody is seeing, you know, the opportunities in moly, Ru, VT tuning that you mentioned. Some of your U.S. competitors are a bit more vocal, saying that they might have market share increase in Logic ALD specifically, where it is, you know, it is your, you know, core business. Can you elaborate a bit more where your confidence is coming from on the market share sides? That would be very helpful.
I think it is really we have our internal intelligence and internal market share work that we have done looking into both the 2-nanometer and 1.4-nanometer technology node. We are very confident of our position in both generations. We have lots of, actually lots of engagement with our customer in ALD and Epitaxy, and we see our position to be very strong in both segments.
Thank you very much.
The next question is from Didier Scemama, Bank of America. Please go ahead.
Yes, good afternoon. Thanks for taking my question. I have got a couple. I just wanted to maybe follow up to François's question earlier on moly and on the A14 TAM. Appreciate you might want to discuss that more in detail at your CMD, but any chance you could help us quantify the opportunity?
For every 100,000 wafer start per month capacity addition, what would be the TAM for moly metallization layers? That would be my first question.
Yes, maybe let me start on moly because indeed you're right. We will very likely say more about this in the investor day. As you know, because we've said it a few times, we have some wins in moly in logic foundry and already have shipped some tools. More, let's say, precise info, detailed info, I think it would go too far at this stage. We're literally updating our numbers as we speak. We will give you more insight in, I think, two months from now, one and a half month from now, two months from now, 23rd of September.
Is it bigger than, you know, trends for ALD and Epi or Gate-All-Around, or how should we think?
Is that incremental, or is it like a meaningful opportunity?
What I mentioned before on molybdenum in logic application, it's a multi-generation shift and increase. Customers will introduce moly little by little as, especially, it's going to go with the lower level metal, like metal zero and metal one. As it shrinks, you go to 1.4, 1.0, you see that transitioning to higher and higher metal layers. Right now, it's really just the start of molybdenum incorporation in interconnect, and it's going to increase as we go to smaller and smaller technology nodes.
Got it. I'll wait for the CMD. For my follow-up, I just wanted to come back to the FX impact. If you could maybe help us understand a couple of things. First, what was the benefit from FX in the second quarter on gross margin, on OpEx?
Because you've talked about the direct impact on FX on the top line, but you haven't really quantified the benefits that you get on OpEx and COGS. Related to that, if I may make a suggestion, I appreciate you might want to guide going forward on constant currency. That's totally understandable. I think everyone would find it really helpful if you could actually give us a quantified guidance for Q3, Q4, or for 2025, whatever you like or whichever you want to pick, including currency, so that everybody's on the right page when it comes to numbers. Thank you.
Yeah. Maybe on the currency, on sales indeed, we have always disclosed that even before this recent, let's say, U.S. dollar, euro volatility, we have been disclosing sales at constant currency because we believe it's an important performance indicator, an important KPI. For that reason, always disclosed.
On the gross margin, actually, there's a negative impact from the U.S. dollar, euro development. Give you some insights. If, let's say, the. U.S. dollar, euro development over a full year would be, let's say, 5% lower than the year before, there's a few digits, 0 point and something, a few digits negative impact on the margin. Of course, the opposite will happen when the dollar or when the euro actually, or the dollar, as you say, would strengthen again. A disadvantage in our gross margin. Despite this unfavorable currency impact, we still have been able to actually come to a very good margin. I'm not sure if you said before, but also a large chunk of our, let's say, material purchased is in U.S. dollars. We have, from a, let's say, transactional point of view, a reasonably good natural hedge. It's not perfect, but it's reasonably good.
When you add to that the translation impact, of course, it becomes more significant. From a transactional point of view, it's the few digits that I just spoke about. If you would have, let's say, over a full year or a quarter compared to a quarter, it doesn't matter, but over a similar period, it would be a few digits impact. Down OpEx indeed, there is some benefit. I think in our press release, we disclosed the currency comparable change in OpEx, so you can see that. I don't know the numbers by heart, but measured in euros, there is some benefit, slightly more in R&D, because we have, compared to SG&A, more U.S. dollar-based cost in R&D, mainly because of our Phoenix site, of course, than in SG&A because our headquarters in Europe, so a larger chunk of our SG&A costs are in Europe.
Cost is a benefit, margin is negative, and revenue is negative with a weakening of the U.S. dollar.
Yeah, very clear. Thank you so much.
The next question is from Sandeep Deshpande, JP Morgan. Please go ahead.
Yeah, hi. Thanks for letting me on. My question is regarding your order intake. You had a stronger order intake in the first quarter. It has sequentially weakened into the second quarter, and the guidance is that you're guiding that you're going to have book to bill less than one in the third quarter. I mean, we're not sure how that relates to second quarter versus third quarter. Are you going to comment on how the orders will trend from the second quarter to the third quarter? Clearly, based on the guidance on revenue into the third quarter, it looks like it will be down from the first quarter still. So, why has this overall order intake suffered at this time, and looking that this is going to change by the end of the year in terms of the shipment of your order that you need be able to ship in terms of orders? I have a quick follow up.
Very good questions, Sandeep. So, in Q2, I think you've seen that the order intake was actually impacted by the lower gate-all-around. I think by now, it depends on the six maybe even seven quarter, at least six quarter strong in gate-all-around. This time, nothing more to reason to that. Order intake was lower, and we expect again a significant increase in Q3. So that's in itself a few timing and nothing else. Market demand for gate- all- around is very strong. As a result of that, of course, also you will see it reflected in our orders. In Q3, we see a different dynamic. There we see a meaningful decline in the orders in China. Actually, a very meaningful decline. As Hichem already said, visibility in China is low. We will still see how this will further develop for the rest of the year. The only, although we do not know, but the only thing that we think it is related to might be export controls where customers have just been accelerating, putting in orders due to concerns for export controls. If that is the case, we do not really know, but it is hard to understand otherwise the actually pretty material difference between first half and second half.
If I then move to Q4, based on what we see today, but that might still change, it is a little bit similar dynamics as what I just explained for Q3. We still expect gate-all-around again to be good, but China is still to be low. That might change, but at least today we do not see that. That does not necessarily mean that it might not come, because it was, yeah, pretty much front-end loaded. For the full year, China is still good. Also revenue. I think Hichem mentioned at the top end of our guided range of 30%, though below last year for the full year, but still good. For the years to come, we expect China to gradually normalize, but that will not happen in one go. That is most likely a gradual step-by-step decline. That is the color I can give you.
Thank you.
In terms of the timing of orders to shipments or revenue recognition, how should we look at this? I mean, your book to bill is less than one in Q2. It is less than one in Q3. Clearly, you are not disclosing Q4, but is this going to impact your revenue in Q4, or is it going to? Clearly, it is not because your overall revenue for the second half you have guided. Does this impact your revenue in 2026?
Yes. The conversion to revenue, we always say on average will be around six months, but we can also literally have an order and a delivery in the same quarter. So far, typically, customers order a little bit earlier, I guess, just to be sure that they get the product.
If needed, if we would get orders, let's say, in the beginning of the quarter, assuming we have the parts in stock, we can still deliver that at the end of the quarter. Of course, a higher order book is better than a lower order book. To that point, we fully agree, but it does not necessarily mean that revenue cannot still, let's say, increase through orders in the quarter and delivery in the quarter. It is definitely true. We've seen many examples of that as well.
Thank you so much.
The next question is from Stéphane Houri, ODDO BHF. Please go ahead.
Yes. Good afternoon. Thank you for taking the question. Actually, I have two. The first one is on the debate about the deposition intensity versus litho intensity from 3 to 2 nanometer and then from 2 nanometer to 1.4.
You said you are confident to keep your market share at least at 1.4. Can you maybe remind us how you see the transition to 1.4? Because from 3 to 2, you were talking about significant double-digit increase in terms of number of layers. How would it look like at 1.4 nanometer? I have a follow-up.
Like we have mentioned, I'll take this question. Like we have mentioned before, from node-to-node transition, we see double-digit increase in ALD intensity. When you move from 2- nanometer technology node to the 1.4- nanometer technology node, we see the same thing happening. There is no change in really the increased intensity of ALD layer throughout the different generation of nodes. For 1.4 nanometer, you know that, like we mentioned previously, we have been working with our customers.
We have engagement with all customers, leading customer logic and foundry on the 1.4- nanometer technology node. We see the new layers coming in as we have really predicted. I think there is no change in our forecast from that point of view.
Okay. Thank you very much. The follow-up is about 2027 and your guidance for, sorry, for $4 billion-$5 billion. I know it was not at constant currency. With the uncertainty that everybody sees for 2026, and even though you've given a quite detailed answer on 2026, still there is a doubt on the trend in 2026. Therefore, the question is, do you feel confident about the trajectory to the 2027 targets? Thank you.
Yeah, I think this is also a question for the investor day. We are doing our homework as we speak. Hichem indeed provides a lot of color for 2026, which might extend into 2027.
We will see that. Obviously, in absolute numbers, a stronger dollar is better than a weaker dollar for us. That is, of course, a given. We'll see if we need to, let's say, give you some more insight into an absolute number based on today's currency. Yeah, as you know, nobody will know what the dollar euro will be. Can be weaker, can be stronger. I mean, I have literally no idea. Otherwise, I would most likely not be here. To give you some more input, we could consider to at least give you a number of what the $4 billion-$5 billion range means in today's currency. That's a fair request, and it's easy to do for us. Of course, we'll give more color in two months from now.
Thank you very much.
The next question is from Nigel van Putten, Morgan Stanley. Please go ahead.
Hi. Hi. Good afternoon.
I want to follow up on China. Just before, you said, if I paraphrase you guys correct, China is not falling off a cliff, yet you see a meaningful reduction in orders into the third quarter and sort of a similar setup into the fourth quarter. At the same time, you're not entirely sure why that is. Just maybe the first question would be, connect the dots, like why would you know the fourth quarter order intake would be weak while you haven't really been able to pinpoint a direct source? And then I have a follow-up. Thanks.
Yeah. Thanks for the question, Nigel. What I already said is, and you're 100% correct, is that we don't know. China has very limited visibility. I started with that. Q4 might still change. Even Q3 might still change, to be honest, but that's what we see today.
If you ask me today, this is what we see, but things might still change. In China, we have seen, yeah, a very strong H1 intake. As a result, you would naturally also expect some weakening in the second half. What we see in the second half now is more than, let's say, a regular weakening. It's really a significant, meaningful decline, but that might still change. We have limited visibility and other than, yeah, concerns around export controls, we have no other explanation. What is happening as well in China, of course, is that because of export controls, the fact when that is created there where we are not allowed to deliver, of course, local Chinese players step in. You also see growth with the local Chinese players. Where we can compete, we still can win. Of course, there is a part where, yeah, the international players cannot compete.
Yeah, and that, of course, also does not really help. We'll see how Q4 will do. It's indeed too early to tell. I agree.
I mean, historically, you've been quite conservative about China. It's sort of, yeah. Great. Then my next question would be on sort of gate- all- around, basically the other way around. China might be weak, but. Does it make sense that you're not upgrading your gross margin target for the full year despite the strength in the first half? Would that imply that you see a meaningful acceleration in sort of revenue recognition for gate-all-around to the second half, considering your margins at these major customers might be not the same as China? Is that a fair approximation to make? Also, does that not set you up well for the first half of 2026, even though China doesn't?
Is that sort of a reason to be cautiously optimistic still, at least into the first two quarters of the next year?
Yes. There are a few things. One, as you know, China for sure plays a role. Less China, everything else equal, typically will put some pressure on the margin. The other uncertainty, of course, is currency. We also assume, let's assume if the dollar euro would weaken further, as I just explained, there will be some negative impact on margin. That could be another impact. Our product mix was actually very good and also driving margin. As you've heard from various comments, our product mix and our revenue in gate-all-around has been very strong. You could, let's say, derive from that that the margin we make there is also good.
China is, let's say, maybe a separate story in itself, but many products that we deliver, let's say, in this whole gate-all-around transition are actually giving us a good margin. Also, we are putting quite some, and that's already, of course, ongoing, but that will continue, quite some effort on efficiency and cost improvements that also will contribute somewhat. It is very likely, we already said it, at the very high end of the margin given where we are in the first half. You're right. We could have maybe upgraded a little bit, but there are a few small, let's say, dynamics, less China, currency maybe, that could still create some level of uncertainty, but it will be good for the full year. Given where we are after the first half, that's pretty safe we can say that.
Yeah. If I add something to what Paul has just mentioned.
Our gross margin, yes, it has been very good in the first half, Q1 and Q2, and also it's going to continue to be good. As you know, in Q2, in the second half, we're working at the dollar has weakened further from the first half. As Paul has just mentioned to you, the weakening of the dollar really reduces the gross margin. Even with that further decrease in the dollar in the second half, we're still very confident about our strong gross margin. I think the. Effort also that we have done the past few couple of years and we continue to do in improving our efficiency, both in manufacturing and the commonality efforts that we have done in our products really are starting to bear fruit little by little. We are very happy with that, and we continue to really focus on that going forward.
Thank you very much.
The next question is from Jakob Bluestone, BNP Paribas Exchange. Please go ahead.
Hi. Thanks for taking the question. Just to come back to the topic of China again, sorry. Can you maybe just clarify? I mean, you said that there might be issues around export restrictions. Can you maybe just clarify and have you actually seen any cancellations? And if you can maybe also just help us understand what share of your order book is China-related. Thank you.
Have you seen any what? Any escalations?
Cancellations.
Oh, no. No, we have not seen any cancellations. No, that is simple. And what was the part of the order book which is China? I do not know by heart. If I would know, I might most likely not tell you, but I do not even know it by heart.
There is still, I mean, as we said in the prepared remarks, that we expect also revenue to come down in H2 compared to H1, but that will be much less significant than what we currently see or expect in the order book, noting that we have limited visibility for the order book as we just explained in the previous question. There is a decline in revenue, but not as much as orders. That still means that revenue will still be reasonable. That also means that the part in the backlog is also still reasonable.
Understood. And then just secondly, just on the gate-all-around orders. I mean, you mentioned that there was some shift in CapEx forecasts for some of your customers. Was this also part of the reason for the soft orders, or as you say, is it just purely timing?
Yeah. I think that is, I will answer that question. It is purely timing. Really. It is not a softening whatsoever. Like we mentioned, I think gate-all-around is continuing to be very strong. The demand is super, super, super strong. We feel very, really happy with our position, and it is really timing. I think it is really growing, but you cannot grow in a straight line. A booking can go up and down, but still it is going up.
Okay. That is really the pressure. This is okay, yes. Thank you.
The next question is from Robert Sanders, Deutsche Bank. Please go ahead.
Yeah. Just coming back to the Molybdenum and the Ruthenium opportunity, I think a large part of that opportunity is to do with backside power delivery entering production and ramping.
It does seem like TSMC at least is being a bit more conservative on backside power delivery. Is that something that concerns you? It does appear that cost is a problem for a lot of the mobile customers, for example. Do you have any view on that? Thanks.
Yeah. I think that if you look into the backside power delivery, yeah, it's an added cost from that point of view. I agree, but it's really needed for high performance. We see the backside power distribution network to be used little by little, starting with the N2 technology node. When we get to the 1- nanometer node, I think the backside will be much more used in more and more devices. In the backside power distribution, you have many layers.
Some of them are also, yes, like molybdenum you talked about, but also they have other ALD layers like in spacer, like silicon nitride. You have tons of Epi layers like low temperature Epi, like silicides, like H-stop Epi, and like silicon and silicon germanium. Yes, I think backside is really good for us as a company, and we see backside being implemented little by little. By the 1- nanometer node, I think it's going to be a standard. More and more would be used in many more devices than what it is right now.
Thanks a lot.
The next question is from Tammy Qiu Berenberg. Please go ahead.
Hi. Thank you for taking my question. Firstly, on gate-all-around. Basically, I remember this is the second time you mentioned that there has been some capacity expansion or plan shift.
I'm just wondering at this point, have you always been seeing that maybe some customer lowered on expectation, but there will be someone else to pick up the order? Therefore, your expectation for GAA growth never changed in this year. Also, have you already seen, let's say, some customers give you really high indication of where they want to be, but it's actually unrealistic? Do you think that kind of double order or unrealistic order has been de-risked already into your numbers?
I'm not sure I fully got what was the did you get the full question, Victor, or?
I didn't really get the full question, Tammy, to be honest. I don't know if the line broke or that there was so what is what is?
Sure. Sure. I apologize. I apologize. First, my question was about the GAA capacity shift you talked about.
If I remember correctly, this is the second time that you mentioned that there has been some GAA capacity expectation shift. Your full year expectation on GAA never changed. Is that always the case that someone talked down number, then there will be someone talk up number, therefore your total never changed? Also, has your GAA already reflected or de-risked the unrealistic expectation or order from certain customers being removed?
I wish I could say yes to your first question, but that's definitely not the case. You're right. There have been shifts between customers in CapEx and gate-all-around compared to, let's say, what we thought maybe, I don't know, nine months ago. Where one customer actually has been higher than what we anticipated, and some customers are lower than what we anticipated. We talked about it.
We got many times the question, "Do you have impact from this particular customer?" And the answer has been yes. It is definitely not that that always will be a perfect match. In this particular case, while capacity is still being built up, it has turned out actually not too bad. Of course, if two customers at the same time would invest to their full potential, the build-up would most likely be even better than what you see today. At the end of the day, if you look medium to longer term, end market demand is most important. Then it doesn't matter too much to us if it's, yeah. If it's with one customer or two customers because, yeah, they will most likely not strongly exceed overall market demand.
Short term, of course, if market demand is really strong, then it could fairly be that if only one customer continues to invest because of good customer wins and because of good yields, that the build-up is, let's say, a little bit slower than when two customers would invest in parallel. That's, I think, the dynamics. We have looked at our own plans, and based on our own plans, we've seen one being higher than the other, and the other, yeah, coming down a little bit. There was a reasonable for this year, at least a reasonable answer. That's for sure. Not a guarantee that always happens.
Okay. That's very clear. Thank you. Also, secondly, when you talk about 2026, you said GAA will be solid. Does that mean there will be growth or it will be flat?
What we said is that the end market demands, we expect to be strong. Of course, we all know that there's one customer doing really well, so there is more confidence that there will be growth than with, let's say, some other customers there. What is very important, of course, and it's really too early to tell whether or not there will be growth, will depend on their customer wins and the development of their yields. We all read the same newspapers, let me say it like that. That's really too early to tell how that will develop. Let's say the customers who are successful in winning customers. They will likely continue to invest. Why? Because the end market demand is really good.
Okay. Thank you.
The next question is from Adithya Metuku, HSBC. Please go ahead.
Yeah. Good morning. Good afternoon, guys. Two questions, please.
Firstly, maybe just following on from the previous question. As you look out to 2026, you mentioned that with certain customers, you have low visibility around the gate-all-around capacity deployments. I just wondered, if this customer does end up changing their plans, do other customers have fab space and etc. to build up more capacity? Based on the discussions you're having today, based on the visibility you have today, or do you think they will need more lead time to prepare, and that means they can't necessarily make up for any changes at the other customer? Do you have any color based on the discussions you're having today? We've got a follow-up.
No, it's basically a repetition of the previous answer where, of course, I mean, again, over a medium to longer-term period, I mean, it's all the same, more or less, because your end market demand will determine the capacity investment. If you specifically ask 2026, yeah, there will be one customer continuing to invest in growth, but would they be able to absorb everything if the other two customers would, for whatever reason, slow down even more than what that specific customer today assumes? Yeah, that will be difficult short term to fully, let's say, absorb all of that, we would think. In the year thereafter, they would most likely do even more than what they currently think. Over a multiple-year period, I don't think it's too relevant.
In 2026, if suddenly two customers would invest less, then it's not very likely to assume that the other can fully absorb that. They might do something extra, maybe. That's at least how we see it. At the end of the day, we don't know. We don't have any insights into their overall plans. We only know what they tell us.
Understood. Just following up on China, you talked about China being a bit better than expected. I just wondered if you might be able to say a few more words on where exactly in China the upside surprise came from. Was it tied to logic foundry? Was it broad-based? Was it memory? Just any color there would be helpful.
It's been pure logic foundry. It's very simple. Power, wafer, analog is still very weak, so that's not growing.
Memory, as we said, is significantly lower than last year because we had some incidental sales last year. Memory overall, yeah, we don't really have a position in memory. It's mainly mature logic foundry.
Got it. Thank you.
The next question is from Janardan Menon Jefferies. Please go ahead.
Yeah. Hi. Thanks for squeezing me in. My questions are actually all on the memory market. If I just follow your comments from the beginning of the year on memory, at the beginning, you were sort of saying that memory could be flattish. You can't be sure. That's what you said in January. In April, you said it's. Likely to be slightly down. Now your guidance for less than 20% of memory revenue from 2025 suggests sort of a double-digit decline. I'm just wondering, where exactly is that weakness coming from?
Is it mainly driven by NAND flash, or is it mainly driven by commodity DDR5 kind of DRAM since you're saying that HBM is still strong?
There's a very short answer, John. It's China. We had these incidental sales, and we might have gotten them this year, but we didn't get them. That's, I think, the biggest. As we said in a few questions, visibility in China is very low. That's one big part of the question. The other one is that also conventional memory, what we call conventional, has not been very strong. Not a lot has happened there. These two have, let's say, gradually caused to gradually bring down the, how you say it, the outlook for memory. HBM actually is very much in line with what we thought, has been good, has been strong, no issue there.
In the past, you've said that in China, your DRAM exposure is on the low side. Would it be fair to say that this is more a NAND flash issue in China, or is it both on DRAM and NAND flash?
I cannot precisely remember because indeed, you're 100% correct. Our share in memory in China is very low. It's really very low for us, except for two quarters last year, where we got some incidental orders. I cannot for sure tell which segment it was within memory, but I know these were some incidental good, relatively high orders for us, at least. We've not seen them since then. We've not seen them in this year. We'll see what happens next year. So far, at least, we have not seen a repeat of that.
Last question for me is, when you ship a tool to, let's say, a non-Chinese DRAM maker, do you have visibility on whether that is going to HBM or DDR5, or do you just know whether it's going to a 1B or a 1C node? Then, depending on what the end use is, they would package it accordingly.
Yeah. I think we have a—I can answer the question. Yeah. We know that they go to the technology node, but we also have a very strong engagement with our top memory customers. We have visibility on where exactly they go, either they go to the DDR or to the HBM. So yeah, we have a visibility customer by customer because we engage with them on the different applications.
Understood. Thank you.
Operator, can we move to the final question, please?
The last question is from Timm Schulze- Melander, Rothschild & Co. Please go ahead.
Yeah. Hi. Good afternoon, guys. Thank you for taking my questions. I have two very quick ones, please, just to finish up. Both relate to China. Paul, you said when there are restrictions, you see local China players stepping in. Could you maybe just talk about what segments you're seeing that in? Is it ALD? Is it Epi? And then on China shipments that are taking place, could you maybe just share some color about what percentage of those China shipments are delivered to warehouse versus delivered to fab? Thank you.
On the first question, it's both ALD, let's say, more commodity or vehicle commodity layers, not the more complex, not the advanced ones. You see China absolutely improving, as well as on Epi. You might have seen the revenue numbers of NORA, who do both.
Yeah, they take benefit from the fact that we are not allowed to deliver certain tools for certain applications, which is extremely frustrating, to be honest, because it doesn't really support anybody if local Chinese players can deliver, but it just hurts us, and it hurts our peers. Anyhow, that's where we have to live with, I guess, for now. Then on the shipments, yeah, most of it goes straight into warehouse, into warehouse, into fabs. Some very temporarily in warehouse, but it's really easy to talk about a quarter or so for different reasons. No, we install it ourselves. We know precisely where they go, so it's mainly all fabs.
Yeah. That was very helpful. Thank you so much.
Yeah. Thank you.
Gentlemen, I turn the conference back to you for any closing remarks.
I would like to thank all of you guys today for attending our call.
On behalf of Paul and Victor, goodbye, and thank you very much.
Ladies and gentlemen, thank you for joining. The conference is now over, and you may disconnect your telephones.