ASM International NV

AEX:ASM

Earnings Call Transcript 2021-Q3

[Operator]

Good day, and thank you for standing by. Welcome to the ASM International Q3 2021 Earnings Call. [Operator Instructions] I must advise you that this conference is being recorded today. And I would now like to hand the conference over to your speaker today, Victor Bareño. Please go ahead.

[Victor Bareño, Director of Investor Relations]

Thank you, operator, and welcome, everyone. I'm joined here today by our CEO, Benjamin Loh; and our CFO, Paul Verhagen. ASMI issued its third quarter 2021 results yesterday evening at 6:00 p.m. Central European time. For those of you who have not yet seen the press release, it is available on our website, asm.com, along with our latest investor presentation.Please let me remind you that this conference call may contain information relating to ASM's future business and results in addition to historical information. For more information on risk factors related to such forward-looking statements, please refer to our company's press releases, reports and financial statements, which are available on our website.And with that, I'll turn the call over to Benjamin Loh, CEO of ASMI.

[Gek Lim Loh]

Thank you, Victor. And thanks to everyone for attending our third quarter 2021 results conference call. I hope you're all safe and well. First of all, I want to take the opportunity to thank our investors and analysts for your participation in our Investor Day last month. We were very happy that many of you took the time to join us to the webcast or in person in Amsterdam. The agenda for today's call is as follows. Paul will first review our third quarter financial results. I will then continue with a discussion of the market trends and the outlook, followed by our usual Q&A session. Paul, over to you.

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

Thank you, Benjamin, and thanks once more everyone for joining the call. So let's now go through the Q3 results. In the third quarter of 2021, our revenue increased to EUR 433 million, up 5% from the second quarter. Compared to the third quarter of last year, revenue increased 38%, both at constant currencies and on a reported basis. Revenue in the quarter was slightly above the top end of our guidance of EUR 400 million to EUR 430 million. Spares & services grew 7% at constant currencies. Our equipment revenue in the third quarter increased 48% in constant currency year-on-year and we offer the larger product driven by strong ALD sales.By industry segment, revenue in the third quarter was led by foundry, which reached a new quarterly record, followed by memory and then logic. Combined logic/foundry increased sequentially and continued to account for the largest part of our sales. Memory sales decreased sequentially, but we're still at a solid level and showed year-on-year growth, primarily driven by record quality sales in DRAM. Gross margin of 47.2% in the quarter was lower than last year's margin of 49.9%, which was impacted by an exceptionally strong mix and slightly down from 48.1% in the second quarter.SG&A expense increased by 26% year-on-year. Net R&D expenses on a reported basis increased by 11% compared to Q3 last year. Gross R&D, excluding IFRS, tax and impairment, increased by 20%. Below the operating line, results included a currency translation gain of EUR 13 million and this compares to a translation loss of EUR 2 million in the second quarter and a translation loss of EUR 14 million in the same quarter last year. As a reminder, we hold the largest part of our cash balances in U.S. dollar, and the currency translation differences are included in our financial results.Income taxes increased to EUR 25 million in Q3 compared to EUR 14 million last year. We still expect the ETR for this year to be high-teens percentage. Let's quickly say a few things over ASMPT. The normalized results from investments, which reflects our 25% share of the net earnings from ASMPT increased to EUR 28 million in the third quarter, up from EUR 19 million in the second quarter and up from EUR 6 million in the third quarter of last year. ASMPT reported sales of $802 million, up 20% compared to Q2 and up 70% from Q3 last year. Bookings amounted to $735 million in the quarter, down 22% sequentially, however, up 42% year-on-year.So now, I'm going back to ASMI again. So our new orders for ASMI, new orders in the third quarter were EUR 625 million, up 21% from the second quarter and more than doubling, up 106% from Q3 last year. And we already pre-announced on September 28 that we order intake would amount to more than EUR 600 million in Q3 compared to our earlier guidance of EUR 510 million to EUR 530 million. Please note also that we've decided to stop providing quarterly guidance for order intake as of next year. As this become clear in the last few quarters in which actual order intake was significantly higher than guided that is increasingly more challenging in the current environment to provide a meaningful outlook. We will obviously continue to report actual bookings as part of our quarterly results update.Now, looking at the breakdown in bookings by industry segment, logic was the largest segment in the third quarter, followed by foundry and then power/analog. Combined logic/foundry book has increased again to a quarterly record in Q3. Foundry booking has decreased slightly sequentially, while the logic segment surged to a new quarterly record. Power/analog bookings also increased to a new quarterly record high and were substantially ahead of the previous high of Q3 2028 (sic) [2018]. Memory bookings decreased compared to Q2 and were led by DRAM.Now, turning to the balance sheet. We ended the quarter with EUR 525 million in cash, which is up from the EUR 465 million in the previous quarter. The increase was driven by strong free cash flow of EUR 98 million, which includes EUR 15 million in dividends received from ASMPT and partially offset by cash spent on the share buybacks during the quarter. Free cash flow, excluding ASMPT dividend surged from EUR 47 million in Q2 to EUR 83 million in Q3. Next, the continued solid level of profitability, the increase in free cash flow is mainly driven by a more modest outflow for working capital of EUR 12 million in Q3 compared to an outflow of EUR 43 million in Q2.CapEx was EUR 17 million and we still expect CapEx for the full year between EUR 60 million to EUR 80 million. During the quarter, we spent EUR 45 million in share buybacks as part of the EUR 100 million buyback program that we started on July 28. And at the end of last week, this program has been completed for 70%. Then I'd like to conclude with the summary of the financial targets that we shared in our recent Investor Day. As we continue to drive growth through innovation, we target our revenue to increase to a range of EUR 2.8 billion to EUR 3.4 billion by 2025. This increase will be primarily driven by strong growth in our ALD and epi business in addition to a healthy growth in our spares and services and selective growth in our PECVD and vertical furnace product lines.We expect the gross margin to range from 46% to 50% over that period, depending on application mix and with cost efficiencies and operating leverage being the structural drivers. SG&A expenses will gradually decrease as percentage sales to high single-digit percentage by 2025. Net R&D, which is a key investment, is targeted at a high single-digit to low-teens percentage. We target operating margins of 26% to 31% in the '21 to '25 time frame, and this should enable us to generate a strong free cash flow. And as you most likely know, we kept our capital allocation policy unchanged. Investment in growth remains the key priority with excess cash returned to shareholders.And with that, I would like to hand back the call to Benjamin.

[Gek Lim Loh]

Thank you, Paul. Let's now look in more detail at the trends in our markets. It is clear that 2021 is a very strong year for our industry. On the back of robust end market demand and the parts of the market impacted by shortages, the global semiconductor market is now expected to grow by more than 25% in 2021. With continued strength in the second half, expectations for the WFE or wafer fab equipment market have also increased. We now expect WFE spending in U.S. dollar terms to grow by a mid to high-30s percentage this year, up from high-20s to low-30s still expected 3 months ago.Demand in the logic/foundry segment continues to be strong. Next to ongoing expansion of the existing leading edge node capacity, initial spending on the upcoming node also had a significant contribution to our logic/foundry bookings in the third quarter. Our customers expect the upcoming node be it 7-nanometer in leading logic or 3-nanometer in foundry to be an important node that will enable many new advanced products in, for instance, high-performance computing and 5G smartphones. We project a strong double-digit percentage increase in the number of ALD applications and layers going to this next nodes. And we believe we are well positioned to see further increases in our share of wallet with key customers.We expect advanced node spending in the logic/foundry sector to remain a solid driver for ASM going into 2022. Memory spending is also on track for a significant increase in 2021. A key driver for our memory business this year is the adoption of high-k metal gate ALD in DRAM. This is a key technology that enables greater power efficiency and improved performance of next-generation DRAM devices, specifically for big data applications. As we discussed in earlier calls, our solution has been selected by all leading DRAM manufacturers.In addition, our R&D engagements for other new applications in memory, both in DRAM and in 3D NAND continue to progress well, and we expect further increases in the contribution of memory to our sales when customers transition to the next nodes starting in '22 to '23. Of note, this quarter was also the record high order intake in our power/analog business, as just mentioned by Paul. The power/analog market continues to show a solid recovery following the negative COVID-19 impact last year. Our strong momentum in power/analog is also driven by the success of our new product introductions.For instance, we are seeing strong demand for our A400 DUO tool, our new high productivity vertical furnace for 200-millimeter applications, including from several new customers in China. If we then take a view, first view on 2022, it is still too early to provide forecast for next year, but we believe ASM is well positioned for continued solid performance in 2022. Based on our guidance for fourth quarter 2021, we expect to start 2022 with a record high order backlog.Next, I will provide an update on the supply chain and capacity. Regarding our own capacity, we believe we are in a good position. As we discussed on previous occasions, we completed our new manufacturing facility last year, and this provided us with sufficient capacity for now, even with much stronger than expected growth in demand in the recent periods. At our Investor Day last month, we also announced that we have started work or started the design work on the second manufacturing floor in our new facility, and we expect this second floor to be ready for production early 2023. This will further expand our capacity and will provide us with the flexibility to deliver on our 2025 growth plans.Looking at the supply chain, as we have already discussed in our call last July, conditions further tightened in the third quarter. This was especially impacted by the lockdown measures, and as a result, reduced factory outputs in Malaysia, which is an important link in the supply chains in our industry. The fact that we were able to slightly exceed the high-end of our revenue target, demonstrates, again, very strong execution on the part of our team in close cooperation with our supply chain partners and customers for which I want to thank them all. Recently, we have seen some improvements, but overall supply chain conditions are expected to remain tight in the fourth quarter. Meeting our customer requirements continues to be a key priority. Despite the supply chain limitations, we expect to achieve a higher sales level in the fourth quarter, in line with our earlier statement.Next, let's briefly recap the key messages of our Investor Day last month. Paul already talked about our financial targets. We aim to grow our revenue with a CAGR of 16% to 21% in the next 5 years with solid operating margins of 36% to 31%. This growth will be clearly ahead of the WFE market, which is expected to increase with a CAGR of approximately 10% over the same period. Important drivers will be our ALD and epi businesses, which are key technologies to enable the next-generation semiconductor device. We talk about the strength that set ASM apart, such as our network R&D model, our early customer engagements and our vast experience and know-how in ALD materials. We also gave you insights into our product differentiation, such as our broad offer of ALD solutions of unique reactor designs and our best in class filling performance and cost of ownership in both ALD and epi.We provided a deep dive into the inflections on the technology road map, increasing adoption of 3D structures and new materials, coupled with traditional scaling will drive many new applications in ALD, such as in high-k gate, VT tuning, EUV patterning, gap-fill, ALD metals and selective ALD. We expect ALD to remain one of the fastest growing segment of the WFE market with a CAGR of 16% to 20% till 2025. In ALD, we aim to maintain a leading market share in excess of 55% by 2025 based on the continued leadership in the logic/foundry space and an increase in our ALD memory share.Epi is another growth driver for ASM. Epi is a defining technology for the gate all around architecture in future logic/foundry devices. Combined with adoption in memory and structural growth in power/analog, we project the epi market to increase with a CAGR of 13% to 18% during the '21 to '25 period. In part supported by our new -- our recent new customer win in advanced CMOS, we target our epi market share to increase from approximately 15% last year to more than 30% by 2025. We also detailed the impact of gate-all-around, which will further increase the need for both epi and ALD. On top of meaningful increases in the upcoming technology nodes, we expect that gate-all-around will drive an increase of $1.2 billion in our addressable market in advanced logic/foundry by 2025.Further, contributing to growth for ASM is our spares & service business, where we are moving from a transactional model to an outcome-based model as well as selective growth in our vertical furnace and PECVD product lines. We also highlighted our increased focus on sustainability during the Investor Day. We showed many examples of steps we are taking to make our tools more efficient in terms of usage of energy and chemicals. We also announced our aim for net-zero emissions by 2035, including scope 1, 2 and 3. As an interim step, we target to source 100% of electricity for our global footprint from renewable sources by 2024, which will already drive a 90% reduction in our scope 1 and 2 emissions. Finally, let's have a look at the guidance that we have issued with our third quarter press release. For the fourth quarter, on a currency comparable level, we expect sales of EUR 470 million to EUR 500 million. Fourth quarter bookings on a currency comparable level are expected to be around EUR 600 million.With that, we have finished our introduction. Let us now move on to the Q&A.

[Victor Bareño, Director of Investor Relations]

We would like to ask you to please limit your questions to not more than 2 at a time, so that everyone has a chance to ask a question. Operator, we are ready for the first question.

[Operator]

[Operator Instructions] And the first question comes from the line of Stephane Houri.

[Stephane Houri, Research Analyst]

Actually, I have 2. The first one is about the supply chain constraint, because you -- back in Q2 you said that Q4 would be higher than Q3, but not specifying how much and it seems that it's clearly higher than Q3. So -- and at the same time, you're saying that there is still some constraints. So my question is to really understand if the supply chain constraints have gone a little bit lower or easier, let's say, and if there was no supply constraints, how much could you do on the quarterly run rate? That's the first question. And the second question is about gross margin, because the gross margin has indeed gone down, as you highlighted, but not so much. And at the same time, looking at the evaluation tool account in the balance sheet, it has come down by only EUR 10 million. So does it mean that you have not realized the disposal of those evaluation tools or does it mean that there are some that have come out and some that have come in, so some clarity about that would be nice.

[Gek Lim Loh]

So I will answer the question regarding the supply chain constraints and then maybe Paul can help with the question on the gross margins and the eval tools. So when we look at the supply chain, basically, as we have mentioned in the prepared remarks, during the quarter, we actually saw some tightening of the supply chain situation, primarily with our suppliers in Malaysia. And as we enter into the last part of the year, with the rapid, let's say, vaccination program that they are putting in, we are, I would say, cautiously optimistic that maybe as we exit 2021, things might be a little bit better in the sense that we'll -- there will be maybe more easing of restrictions, etc., as we enter into the new year. Now having said that, the -- one of the reasons why I think we are able to deliver also more in the fourth quarter or as we have guided despite the fact that the supply chain constraints are continuing is, as we have explained to you previously, we have taken measures to try to mitigate that. One of the measures, of course, is to try to only order as early as possible. That's one of the things that we have done as part of the COVID-19 learnings that we had from last year. And the other one, of course, is, again, as we have explained earlier, trying to qualify more suppliers. So to that respect, we are seeing some, I would say, initial returns on that, being able to lock in supplies or, let's say, orders, supplies earlier by giving orders earlier. And also, I think, even though we still need quite some time to qualify a lot of alternative suppliers that we would like to, but I think we are seeing some fruits of that. And that, coupled with, I would say, a slightly improving maybe situation as we go into the end of the year gives us the optimism that we should be able to deliver more than what we have done in the third quarter.

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

I'll take the question on margin, Stephane. On the margins outlook, first of all, I think that I'm pleased to say that we have a healthy margin at 47.2%, which as guided came indeed in slightly lower than what we had in Q2. In the eval tools, there have been quite a few movements. So as we already said, we expect in H2 a higher sales than in the first half, which indeed is happening. So we saw sales in Q3 to be higher than in Q2, having some impact on margin. 2, we also had new shipments into customer of new eval tools. So what you see is that the -- in the quarter movement of eval tools is very small, I think, it increased by EUR 1 million or so. It's a very small compared to year-end, indeed, increased by EUR 10 million. But there are pluses and minuses, of course, in movements through sales and through new eval tools. And maybe also -- could also already couldn't say that also in Q4, we expect, again, I think, even higher sales of eval tools again in Q3, which, of course, might impact the margin a little bit. But at the same time, I want to say that eval tools is, of course, one element that impacts margin, obviously, application mix and other elements also impact the margin. So it's always very difficult to give you precisely what the net impact of all that is. But what we've seen at least in Q3 as we guided somewhat lower margin amongst all because of the increased sales of eval tools. So I hope that is clear, if not...

[Stephane Houri, Research Analyst]

Yes, it is clear. Does it mean that Q4, you're guiding roughly flat gross margin? Would that make sense?

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

I don't think we guide the specific margin for Q4. What we -- what I can say at this time is that, that of course an increased level of eval tools will have an unfavorable impact, somewhat unfavorable impact on margin. That is only one element. There's, of course, other elements as well that will influence margin. And what I can say is, of course, the new rates that we have provided during our Investor Day that we do think that the margin for the year, but I can also say for Q4. At this time, will be somewhere between 46% to 50%.

[Operator]

Next question comes from the line of Marc Hesselink.

[Marc Hesselink, Research Analyst]

First question is actually on client order behavior and your visibility there. You already stated that you are not going to disclose that anymore. Is it -- did it change so much? Are clients completely different in the way they put in their orders? Do they do it more in bulky fashion or is it -- are there other things why you changed that? And then also related to that, because you have been building up such a large order book, when will that be converted into sales? Are those lead times much longer or is it just a temporary effect that we now see that build up and quite soon we should be more in the normal range again? Given the book-to-bill, I think, this year, 1.25. Yes, that's really high. When it will be more at the normal level?

[Gek Lim Loh]

I think maybe the easier way to answer would be to answer the second question first. So I think as quite a number of analysts and investors have correctly observed based on our fourth quarter guidance, we are probably going to exit 2021 with a record high order backlog. And of course, as we have also explained previously, one of the reasons why the order backlog is increasing is, let's say, due to the supply chain constraints. So if there were no, for example, supply chain constraints, we could probably deliver more. But it's something which is difficult to quantify because the supply chain constraints are actually quite, I would say, widespread. So we just have to work whatever we can to try to make sure that we get the supplies that we need. And as I've mentioned in the previous, let's say, question, we do this actually already from quite some time back by ordering earlier and also by trying to work on qualifying alternative suppliers. Now in terms of the orders that we are seeing, I think there is very strong demand. I think as we have mentioned in the prepared remarks, Paul has given some color that the combined logic/foundry order intake for the third quarter was a record quarterly high. The order intake for logic was again a quarterly record high. And finally, power/analog was also a quarterly record high. So we have seen very strong demand across multiple segments, and that is also leading up to the strong orders that we are getting. And of course, due to the supply chain constraints, this is manifested in the form of the order backlog.

[Marc Hesselink, Research Analyst]

But then still, really, why is it now then more difficult to forecast than it used to be?

[Gek Lim Loh]

I think we are seeing, let's say, probably more lumpiness in terms of orders coming. It has always been lumpy to be -- first of all, to clarify. Orders coming in from our customers has always been lumpy, because it really depends on, for example, the investment cycle. So it's not like they split whatever they're going to buy equally into 4 quarters and then they just give us the quarters now. It really is a timing on when is there investment. So it has already been lumpy, I would say, pre-supply chain constraints. But with longer lead times and I would say, an overall supply chain constraint in the industry, I think this has become more lumpy. And we have also shared that also in our second quarter earnings that we do see some customers placing orders a little bit earlier, because they want to make sure that they get the orders in front of us. And because of these reasons, we think that it's kind of difficult -- challenging for us to be able to provide with a proper, let's say, or accurate guidance as far as orders or order intake is concerned. As you know, we have like -- during the second quarter and the third quarter had to revise and give renewed guidance for our orders, because they came in much higher than what we expected. And we also think that going forward we will just provide you with the revenue targets or guidance. Now of course, we still have to report the bookings as part of our earnings cycle.

[Operator]

Next question comes from the line of Keagan Bryce.

[Keagan Bryce-Borthwick]

Just 2 from my side. The first on high-k metal gate DRAM. I know you've talked about wins across all 3 major players. One of these customers has obviously been quite public about the adoption. But I guess if I look at the market today, it still feels like high-k metal gate is still fairly niche. When do you expect to see a broader adoption of high-k metal gate for DRAM? And when you do, do you expect your market share, particularly in memory ALD, to meaningfully increase? And then my second question is perhaps a little bit longer term. You've done an operating margin to date of 29%, which puts you at the midpoint of your 2025 target. Now given you'll be close to doubling your sales base in '25, I'm surprised not to see more operational leverage come through. I'd always thought that you're lower margin versus, let's say, some of your bigger WFE peers was a scale problem. So surprised not to see sort of higher '25 targets. What other variables may be limiting sort of your operating margin improvement into '25?

[Gek Lim Loh]

I will take the first one and perhaps Paul can give you more insights into the operating margin question. So on high-k metal gate for DRAM, it has been, the solution has been adopted by all the 3 major DRAM manufacturers. And of course, as you correctly mentioned, it is used primarily to manufacture high-performance DRAM devices or, let's say, what we call high-performance DRAM. And it's a smaller part of the overall DRAM market, but we also see some signs, especially this year where we have had a fairly strong, I would say, business coming from high-k metal gate in DRAM, that the proportion of high-k metal gate DRAM is actually increasing. And how far this will go in terms of broader adoption, I think this is still left to be seen. But primarily usage of high-k metal gate DRAM is, of course, lower in terms of power efficiency and speed. And just to give an example, one of the applications that is driving that use is, of course, let's say, high-performance graphics, where speed is important. So that's one of them. But again we think that probably as the market find new applications for -- that requires faster speed, lower power and so on, we might see increasing adoption of high-k metal gate applications in DRAM, but that's more our opinion and we still have to see what happens at the end user market. Paul?

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

I think the question was on the mid-term guidance between operating margin, 26% to 30%. I think what's important is that -- let's turn to the gross margin and then the elements below gross margin -- that the operating leverage in the gross margin is not very big, although over the 4, 5-year period, there will be a structural improvement simply because of operating leverage. The year-on-year improvement is relatively small because the vast majority of our costs in our operation is variable because we have an assembly operation and no real manufacturing operations. So don't expect too much from that. But again, if you add it up 5 years in a row, yes, there is a benefit. Then on SG&A and R&D. This year, 2021, especially the R&D cost is actually lower as a percentage of sales than what it should be. We've grown very, very rapidly in this year. We have not caught up fully. We're investing in R&D simply because the growth was so rapid. And 2, it is not easy to find the right people. There is a war on talent going on, as you all know. So we will need to catch up. So compared to this year, R&D will actually go up and will be a negative in our EBIT margin. Compared to 2020, it will be similar or down somewhat. So compared to 2020, there will be an improvement, most likely. So compared to '20, you will see a benefit in our operating margin. Then we have SG&A, that's where we expect the most structural benefits from operating leverage and we've guided a high single-digit. We're now at double-digit in numbers. So there we will see a benefit. And if you add that all up, yes, more or less, you stay in this 26%, 30% range. Then, of course, from quarter-to-quarter and from year-to-year, there will be differences in application. As you see in every year, you've seen it every quarter, you've seen the margin in the beginning of this year which was in excess of 49% in Q1, now we are at 47.2%. So swings of 1%, 2%, sometimes even 3% is not completely abnormal simply due to application mix differences. But the structural drivers that we see, especially compared to this year, because you guys all look at the margin this year, take into account R&D as a percentage of sales is expected to increase somewhat, but will come down compared to 2020.

[Operator]

Next question comes from the line of Tammy Qiu.

[Tammy Qiu, Analyst]

So firstly, on the foundry customers and logic customers as well, so based on the tool you are shipping to the half today, will you be able to tell is that used for 4-nanometer or 3-nanometer for your foundry customer today? And will you be able to know that, therefore, next year, we may see a higher level of insertion because of their moving to 3-nanometer comparing to this year? And I have a follow-up as well.

[Gek Lim Loh]

I think what we are seeing, first of all, in terms of orders, we are seeing, of course, orders coming in for the additions -- capacity additions on the current node. But at the same time, we are seeing also the orders are coming in now for build-up of the next node. So in that respect, we are seeing both. And as we go on, we will be shipping not only, let's say, current node, let's say, capacity expansion tools, but we will be also shipping tools that will be used for the build-up of the next node as they go into a rise or pilot production, I would believe over the next year.

[Tammy Qiu, Analyst]

And just a follow-up on the DRAM high-k metal gate question. You mentioned that you still need to see what's the high-k metal gate penetration for DRAM over the next few years. And I'm just wondering, from your visibility today, do you see that as a straight light adoption curve going up from here or we may see a period of, let's say, pool from adoption perspective, because recording what happened at foundry and logic when they are adopting high-k metal gate, it was a big wave then it was sort of slow down a little bit before double pattern started to be using ALD. Are we going to see similar pattern or do you think it's more sort of in the smooth line adoption curve?

[Gek Lim Loh]

The honest answer, Tammy, is we hope it's a big wave. But we do see that there is -- in fact I would go to the point of saying that the number of tools that have been ordered and we actually have shipped and will be shipping, exceeded our forecast. So in that sense, we think that the adoption is actually higher. Is it going to be a straight-line or a hockey stake, I think that's going to be difficult for us to give any reasonable, let's say, feedback, because it's really very much dependent on our customers and they are customers who are the end users.

[Operator]

Next question comes from the line of Timm Schulze-Melander.

[Timm Nikolaus Schulze-Melander]

I had 3 very quick ones, if I may. The first one was on the order guide, which is down sequentially. So I just wanted to touch on whether there's something we should be taking away from that. Is that just a supply chain issue? Then the second question just to follow-up on the previous questions on the gross margin walk. Could you maybe talk about whether you've already seen any impact in the gross margin from cost inflation? And kind of what kind of year-on-year trends you're running with into 2022, please? And then the third question I had was on your service & spares profitability. Does it vary as much by application as the equipment business does? And can you maybe just give us some color as to how the margins in your service & spares business have sort of trended over the year.

[Gek Lim Loh]

I think I will probably answer the first one on the what you have alluded to as the lower order guidance and then perhaps Paul could give you, again, more color on the gross margin and maybe also the spares & service, let's say, profitability or gross margin level. I think in terms of orders that we have -- that we are seeing, we have guided the fourth quarter to be around EUR 600 million. And as I've explained in the earlier question to Marc from ING, we -- one of the reasons why we have only guided to, for example, one number today is we think that, that is going to be a fairly accurate number of, let's say, the orders that we will get. And again, we think that the demand is healthy, especially coming in from logic and foundry. But as we have also witnessed in the third quarter, the orders coming in for power/analog was also very strong. We think that if we look at a EUR 600 million order intake compared to maybe what we just delivered EUR 625 million there isn't really a lot of substantial difference and order intake or order demand continues to be at a very high level.

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

On the, let's first to the gross margin cost inflation question. So the answer is yes. We have seen impact of the cost inflation on logistics cost on some commodities. But at the same time, we've seen more benefits from order materials still. So net-net, you've still seen a positive. But if there would have been no cost inflation in, let's say, logistic cost, commodities, etc. the margin would have even been better than it is. But also important to understand that the net impact on our cost of goods take into account all the effort that has been done by our supply teams, supply chain teams is still a positive. So we have seen a small reduction in our gross margins. For next year, it's still a little bit too early to tell. Of course, we have a certain view, but I think it's too early because there's still quite some uncertainty to share that. In any case, as far as we see it now we do not expect a large impact of cost inflation. That's our current view. But of course, it's still early days, and we will have to see in the coming months, quarters, how this will further develop. But we do see opportunity still for certain parts in our build-up material, but we also clearly certain commodities and logistics costs still creeping up. And net-net, the impact should not be that big, but we'll update you in subsequent calls. Then on the margin of spares & services, yes, what we, of course, it's a very different dynamic than the equipment sales. But what I can say is that the margin that we generate is within a small range of the equipment margin. It's not very far away from that. It's sometimes very close, sometimes there's a few percentage point difference, but it's trends pretty much in line with the equipment margin.

[Operator]

Next question comes from the line of Didier Scemama.

[Didier Scemama]

Congrats on the quarter and guide. 2 questions, if I may. I think your question has been sort of tried to be asked earlier. If you look at your equipment revenues this year, you're not going to materially outperform the WFE market, which is sort of unexpected, I would say, given your exposure to probably one of the smartest and fastest growing segments of the market and we can see that in your bookings. And my question is the following. So do you expect to materially outperform WFE next year, number 1? And number 2, I'm going to put sort of pressure on you by asking, if I look at your order intake, is EUR 2.5 billion revenue 2022, a crazy number? And I've got a quick follow-up.

[Gek Lim Loh]

I think on -- when you look at what we think the WFE growth will be this year, as we have just, let's say, release of mid to high 30% and if we then look at our revenue, equipment revenue and if we look at what we have guided for the fourth quarter, we actually think that we are there. We probably are at the level of growing at the same or maybe even slightly above the WFE growth of mid to higher 30%. So that one we can say for 2021. But as far as 2022, I think it's still too early days. So we will probably have to really look into what is going to be the growth. And I think there's all kinds of projections as to what's going to happen in 2022, but it's still too early. On the EUR 2.5 billion revenue question, I can tell you, I would love that. But again something that is really too early for us to give any color or guidance. We will have, as we have always mentioned, we're going to end -- we're going to exit the year with a strong backlog. We think that the first half should be healthy for us. But visibility going into, for example, the second half is still very difficult. So we are not able to provide any color at this moment as far as 2022 is concerned.

[Didier Scemama]

And then on the gate-all-around buildup, I think some of your customers are -- have got like a sort of 2 to 3-year deployment schedule. Can you give us a sense of how much of your revenues in '21 has been sort of driven by the initial production build at one of your customers for gate-all-around? And how would you expect that to play out over the course of the next 2 to 3 years? How big can that be for you guys, maybe '22, '23?

[Gek Lim Loh]

As I explained earlier, I think the orders and the revenue that we are seeing at this moment is a mixture of, let's say, capacity additions at the current node, but also built up for the next nodes. And of course, when you take the 2 in comparison, high-volume capacity additions will be bigger rather than just build up for the next nodes. I think going forward into the next couple of years, we should see the built up progressing. And at the same time, I think the transition to gate-all-around should come at a fairly steady but gradual, let's say, pace.

[Didier Scemama]

Can I just ask a quick follow-up only one just for Paul. Coming back to the question on sort of margin leverage, I think we all understand the sort of variable cost structure you've got on COGS and I think you very clearly explained what's going on in R&D and SG&A. I just wondered, in SG&A, I appreciate it's where the leverage is coming from, but that's still going to go up in absolute dollars. And I appreciate you guys are growing like a train, so you need to spend more on SG&A. But is it more on the sales and marketing side or is it more in G&A simply because you need more controllers, more legal, more this sort of people because your customer base, your sales and marketing team, your customer is not growing, right, they aren't, it's a finite number of clients, I mean probably 10 clients, 15 clients at the most. I mean why would you need to spend on sales and marketing dramatically up from here?

[Paulus Antonius Henricus Verhagen, CFO & Member of Management Board]

And this is actually given the answer already yourself. This is indeed to support the growth that we're seeing and efforts seen. Also, I think in SG&A, the company has been running a little bit behind the fact simply because it has grown so rapidly and not always easy to get people. It's a little bit across the board because we have more contracts, there are more legal people. We need to improve processes. A certain process have worked very well, given the size of the company 2 years ago, but start now to be challenged more and stress more given the current size and especially given the future size. So we need to invest. We need to do something further in our IT infrastructure, which is self is good, but needs to be further improved to cope with the growth that we see. We have actually more than 10 customers. I do know that very few large ones, but we have a few more customers and also quite a few more contracts to deal with. So that definitely will need more sales and account people as well. So it is a little bit across the board. But having said that, obviously, over this 5-year period, it could be different again from 1 quarter to the other or from 1 year to the other where we might step up a little bit more, but then slowdown in the year directly, that will review from year-to-year and maybe even from half year to half year. But over time, of course, we will definitely not grow the SG&A expenses as much as the top line.

[Operator]

Next question comes from the line of Adithya Metuku.

[Adithya Metuku;Crédit Suisse;Director, Equity Research]

My questions have largely been answered, but I just had one question on the medium term. You, at your CMD noted that the 3D DRAM transition will happen in the mid-2020s and one of your peers immediately after you noted that it would be late into the 2020s. I just wondered if you could give us some color on the factors driving confidence in your forecast of the transition to 3D DRAM will indeed happen in mid-2025 -- in mid-2020s and not later. What are the factors driving that confidence? That would be super helpful.

[Gek Lim Loh]

So if we look at -- during our Investor Day, we brought up the topic, or let's say, the inflection, the potential inflection of 3D DRAM as a part of new growth opportunities actually for us for the longer term beyond 2025. And in the same context, we also spoke about, for example, second-generation gate-all-around, the transition, let's say, going from nanosheet to a forksheet in logic and foundry. And when we look at 3D DRAM, it's obviously still early. The timing of adoption will likely differ depending on the customer, but we actually expect the first adoption to be in the 2026 time frame. But the one thing that we wanted to deliver as a key message is that we are already in early engagement with leading DRAM customers. And based on these engagements, we think the 3D DRAM, when it comes and when it happens, it's going to be providing us with new opportunities. Maybe just to clarify a little bit that does not mean that our DRAM opportunities are just going to be dependent on the future inflection of 3D DRAM. As we have also mentioned a couple of times in the past, we have a pipeline of new ALD, epi applications that we have already been working on with customers. High-k metal gate was the first one that got adopted by all 3 leading DRAM manufacturers. And for the applications that we have in the pipeline, we do expect that some of them will be inserted or adopted and inserted during the 2022 to 2023 time frame, thereby helping us to grow our share in the DRAM segment. Maybe just to summarize, I think 3D DRAM is a very interesting long-term opportunity for us. But in the short to medium term, we have several new applications in the pipeline for next-generation of plan DRAM. And we expect this pipeline applications to actually be contributing to our growth in the DRAM business.

[Adithya Metuku;Crédit Suisse;Director, Equity Research]

Understood. And just as a follow-up to that, I mean, it sounds like some of the smaller guys because they can't afford EUV might end up going to 3D DRAM quicker than the bigger guys. In which case, if I look at your slide deck again from the CMD, the cost difference for a bit is so big that it feels like the bigger guys will not be able to continue with EUV shrinking for too long. While the smaller guys are providing bids at the much lower cost using 3D DRAM. Would you agree with that sentiment? So essentially, that will force the larger guys to move to 3D DRAM quicker than there was lot of work done.

[Gek Lim Loh]

That's difficult. I think we will need to wait until there is more clarity on the technology before we can make any kind of even speculation, because I think today, the technology is still at the discussion stage, and we probably will need some time before, we, for example, know what is the specs for 3D DRAM. I think that's still going to take some time. So today, I think it's difficult to make any kind of assessment along what you have just said.

[Operator]

Next question comes from the line of David O'Connor.

[David O'Connor, Analyst of IT Hardware and Semiconductors]

2 from my side, if I may. Maybe firstly, Benjamin, you mentioned the Q4 order intake of EUR 600 million that includes some benefit from supply constraints as the customers you mentioned are placing orders earlier. Just trying to quantify that. I mean is that a 10% impact on orders or is it more than that? Just trying to get a baseline of the true kind of order run rate as we exit 2021? And I have a follow-up.

[Gek Lim Loh]

I will be very upfront and honest. Actually, I do not know the answer. But by and large, when we talk about early ordering, I think it's -- of course we do see some of that. And some of our customers are ordering a little bit earlier in response to the supply chain constraints affecting the whole industry. But is it like a very significant amount or is it like, they are ordering months, months earlier? The answer is no. So I would say that there is some impact, but it's not that big.

[David O'Connor, Analyst of IT Hardware and Semiconductors]

And then maybe a longer term question. When you look at the future logic and foundry road maps, and in particular, you'd start to see more of the success of ALD and ALE steps happening under vacuum. Can you talk a bit about ASM's positioning to capture those steps versus the kind of larger etch peers out there who are also competing in the ALD markets? And what portion of ALD steps do these kind of vacuum steps represent, any kind of color around that how you're thinking about that would be helpful.

[Gek Lim Loh]

I think one of the things that you find with the increasing complexity as we get down to smaller nodes is that on some applications, the surface cleanliness is very critical. So before you can use ALD to deposit whatever materials that you would like to deposit, first of all, you have to make sure that the surface is very clean. And the reason for that is, in some areas, in some cases, the thickness of the ALD layers is what we call just 2 or 3 monolayers, 2 or 3 levels of atoms. And that's why the cleanliness is absolutely critical. Now we have, as we have shared during our Investor Day, developed what we call surface cleaning technologies and equipment that enables us to do that to make sure that before we do the deposition we are able to make the surface very clean. It's a form of what the broader industry likes to call ALE. But we have already developed those required -- or those technical capabilities. And we have actually been already delivering them to our customers. So in that respect, we are very, I would say, comfortable with our capabilities. But just to maybe be clear, the cleaning technology that we have made is in support of our ALD and Epitaxy business. We have no intention at this moment to, for example, go into an adjacent etch kind of technology. That's not our, let's say, intention. It is mainly whatever we do in terms of surface cleaning is just to support our deposition, let's say, products.

[Operator]

Thank you. That was the last question. And with this, I would like to hand back over to the CEO, Mr. Benjamin Loh for final remarks.

[Gek Lim Loh]

Thank you. I would like to thank you all for your attendance today. Also, on behalf of Paul and Victor, we look forward to meeting many of you soon again, either in person or virtually in one of the broker conferences or other investor events in which we will be participating. Thank you, again, stay safe, and goodbye.

[Operator]

That does conclude our conference for today. Thank you for participating. You may all disconnect.