Thank you all for standing by, ladies and gentlemen, and welcome to today's ASM International third quarter 2020 earnings conference call. Our presentation for today will be followed by a question and answer session. To ask a question over the phone, kindly press star one on your telephone. Please be advised the call is being recorded, and I would now like to hand the call over to your speaker, Mr. Victor Bareño.
Thank you, operator. Welcome, everyone. I'm joined here today by our CEO, Benjamin Loh, and our CFO, Peter van Bommel. ASMI issued its third quarter 2020 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. As always, we 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 the 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. With that, I'll turn the call over to Benjamin Loh, President and CEO of ASMI.
Thank you, Victor, and thanks to everyone for attending our third quarter 2020 results conference call. I hope you're all healthy and safe. Peter van Bommel will review our third quarter financial results, after which I will continue with a discussion of the market trends and the outlook. After that, we will have the Q&A session as usual. Now, before handing over to Peter, I want to take a moment to thank him for his outstanding contribution to the development of our company. As most of you have probably seen, on October 14, we announced Peter's decision to retire. He will stay on until the next AGM, and the supervisory board has started the search process for his successor. Peter joined ASMI in 2010 as CFO and a member of the management board.
Over the last 10 years, Peter has been instrumental in not only driving strong and structural improvements in our company's financial performance, but also in the overall strategic orientation of the company. To mention just a few of the highlights, since 2010, sales have grown at a compound annual growth rate of 16% and operating profit at a rate of 34% per year. In addition, during the last decade, ASMI returned more than EUR 1.8 billion in cash to our shareholders. Now with that, over to you, Peter.
Thank you for your kind words, Benjamin. Let me briefly comment on my decision to retire. This year, I turned 63, and after 10 years in my job as CFO of ASM, I decided that it's time for someone else to take over this role. As Benjamin just pointed out, the company is in a very good shape. I will stay on until next May to allow a smooth transition. I still expect to meet many of you, highly likely virtually, in the forthcoming periods, but I already want to take the opportunity to thank you all for your support and interest. It has been a great pleasure interacting with all of you during the road shows and other investor events over the last years. Back to business. Let's now review our financial results.
In the third quarter of 2020, our revenue amounted to EUR 315 million, up 16% from the second quarter of last year, and towards the higher end of our guidance of EUR 300 million-EUR 320 million. Compared to the second quarter, the revenue decreased 8%, half of which was due to negative currency effects. Services spares sales grew strongly, by 31% year-on-year, and represented 23% of the total sales. Our equipment sales increased by 12% year-on-year, and were again, primarily led by our ALD products. By industry segment, revenue in the third quarter was led by foundry, followed by memory, and then logic. Logic decreased sequentially, Foundry sales were up. The combined logic foundry sales continued to account for the largest part of sales. The memory segment was about stable compared to the second quarter, supported by a healthy level of DRAM sales.
Gross margin increased again to a record level of close to 50% in the third quarter, up from the already high 48.3% in the second quarter. The gross margin was again driven by an exceptionally strong sales mix. We started guiding the market on gross margin late 2012. At ACE, we informed the market that we targeted gross margins to be in line with the longer-term average within our industry. We also indicated at that time that there could be quarters where we would be slightly below or above this range of a low to mid-40s percentage. Since then, we have seen the margin in two quarters below that range and now for two quarters above that range. At this moment, we see no reason to deviate from the guidance that we first provided in 2012.
Having said that, we expect the gross margin in the forthcoming quarters to be more at the high end of the range. If we look at the operating expenses, SG&A decreased by 5% compared to the second quarter, which is explained by the one-off costs that were included in the second quarter. R&D expenses decreased by 6% compared to the second quarter. This is mainly caused by lower impairments, which dropped sequentially from EUR 5 million to EUR 2 million. The operating profit decreased slightly compared to the second quarter. That's mainly due to sequential decrease in sales. The operating margin percentage, however, further increased to 26.7% in the third quarter, which is again, a new record high for ASMI.
Below the operating line, results included a currency translation loss of EUR 14 million, which is mainly explained by the depreciation of the U.S. dollar compared to the end of the second quarter. This compares to a EUR 6 million currency loss in the second quarter and a currency gain of EUR 14 million in the year-ago period. As a reminder, we hold the largest part of our cash balances in U.S. Dollars. The currency translation differences are included in our results. The result from investment, which reflects our 25% share of the net earnings from ASMPT, decreased to EUR 6 million in the third quarter, down from EUR 11 million in the second quarter. ASMPT reported sales of EUR 551 million, down 1% compared to the second quarter and up 3% from the third quarter last year.
Their bookings amounted to EUR 583 million in the quarter, which is up 24% sequentially and up 12% year-on-year. Now turning back to ASMI's consolidated operations. ASMI's net earnings on a normalized base amounted to EUR 61 million in the third quarter. Our new orders in the third quarter were EUR 303 million, up modestly from the second quarter and up 4% year-on-year. Orders were slightly above our guidance of EUR 280 million-EUR 300 million, even with a negative currency impact of 4% compared to the previous quarter. Looking at the breakdown in bookings by industry segment, foundry represented again the largest segment in the third quarter, followed by memory, and then logic. Combined logic foundry bookings increased compared to the second quarter. While logic decreased sequentially, foundry bookings increased to a new quarterly record high. The memory bookings decreased compared to the second quarter.
NAND flash picked up somewhat, DRAM bookings dropped compared to the relatively strong level in the second quarter. Turning to the balance sheet. We ended the quarter with EUR 430 million in cash, basically stable compared to the EUR 432 million at the end of the previous quarter. Free cash flow in the quarter amounted to EUR 17 million. The continued strong level of profitability was partly offset by a cash outflow of EUR 26 million for working capital. Accounts receivable increased to a relatively high level at the end of the third quarter due to a distribution of sales, which was again heavily back-end loaded in the quarter. Inventories remained relatively stable following the increase in the second quarter, although the mix shifted from raw materials towards finished goods.
As we discussed in the last call, at the end of the second quarter, we had held higher raw materials for specific components in view of the bottlenecks in the supply chain in the second quarter. These raw materials were processed into finished goods during the quarter as the supply chain turned back to normal conditions in the third quarter. The decrease in raw materials in the course of the third quarter also led to a reduction in accounts payable. Assuming a continued normalized supply chain condition, we expect working capital to come down in the fourth quarter. In the third quarter, we have spent again EUR 19 million on CapEx, largely related to a new manufacturing facility in Singapore in combination with lab tools. During the quarter, we spent EUR 28 million on share repurchases.
As per October the 23rd, we have completed 40% of the current EUR 100 million share buyback program. With that, I hand the call back over to Benjamin.
Thank you, Peter. Let me first provide you with an update on the impact of COVID-19. Our focus remains the health and safety of our staff. During the third quarter, conditions in our supply chain and manufacturing operations largely normalized again. As just mentioned by Peter, our employees and suppliers delivered an excellent job. Thanks to their flexibility and relentless efforts, we have been able to continue supporting our customers in the best possible way. With the number of new COVID cases on the rise again in many parts of the world, we remain vigilant and continue to monitor the situation. Let's now look in more detail at the trends in our markets. Despite the impact from COVID-19 and the related slowdown in GDP, the overall semiconductor market has been resilient and is expected to show healthy growth this year.
Working and learning from home accelerated the digitalization trend and created healthy demands in segments such as PCs and servers. More recently, other parts that earlier had seen a negative impact from COVID-19, such as automotive, have also started to show some recovery. Next, zooming in on wafer fab equipment, spending also remained robust in the first nine months of the year. Looking at the market by segment, logic and foundry spending continues to be on track for a strong year. Investment in the most advanced 10 nm and below nodes continue to be the key driver in the logic foundry segment. Our customers have been stepping up their leading-edge manufacturing capacity in support of the wafer demand for multi-year growth drivers such as 5G, cloud computing, artificial intelligence, and autonomous driving.
5G, just to take one example, is expected to drive renewed growth in the smartphone market and new apps. Enhanced capabilities will lead to a higher semiconductor content as compared to the 4G smartphones. Demand in the coming years for faster and more energy-efficient chips to power these 5G smartphones is an important reason for customers in the foundry segment to invest in new leading-edge manufacturing capacity. As well known to all of you, investments in leading-edge logic and foundry have been a very positive factor for ASM. On the back of substantial share of wallet gains, we achieved in those most advanced nodes. Looking ahead to the upcoming nodes in logic foundry, we are strongly engaged in R&D, and we expect our served available market to grow again by a meaningful double-digit percentage. Looking at the memory market, end market trends were mixed in the third quarter.
Following healthy developments in the second quarter, pricing conditions softened somewhat due to inventory corrections in parts of the market. We confirm our earlier forecast that spending in the broader memory market in the second half will be higher than in the first half, mainly driven by an uptick in 3D NAND spending. Limited capacity over the last couple of years, combined with expected growth in end markets such as smartphones, brings the potential of a further improvement in supply-demand conditions. At this point, visibility in the memory market remains relatively limited. This year, we are having good traction in our DRAM business, driven by the high-k application wins that we discussed last quarter. Our ambition in memory remains to substantially increase our served available market over time as we further step up our customer engagements in new applications.
However, it is important to keep in mind that logic foundry represents the largest part of our sales, and therefore will remain the most important driver for our business in the forthcoming periods. A strong area of growth this year has been the Chinese market for the broader wafer fab equipment market and for ASM. We benefited from the investments we made in recent years to strengthen our position in this market and from the first meaningful investment by some of the domestic players in the more advanced nodes. Despite the uncertainty related to new U.S. export restrictions, we recorded again substantial year-on-year growth in our China sales in the third quarter. Another key driver for our sales has been the spares and service business. In the first nine months, we increased sales in this business line by 29%.
In part, growth in our spares and service were driven by temporary customer demand for buffer inventories in view of COVID-related supply chain uncertainties. A more structural driver has been the strong expansion of our install base in recent years. On top of that, we have been increasing our efforts to target additional opportunities, which is starting to bear fruit. To sum up, 2020 is shaping up to be a solid year again. For the full year, we estimate wafer fab equipment spending to increase by a high single-digit to low double-digit percentage. On the back of a continued healthy development in the second half, we increased our estimate for wafer fab equipment compared to last July. As also underscored by the guidance we provided for the fourth quarter, we believe we remain on track to outperform the overall wafer fab equipment market this year.
Against an uncertain economic backdrop, it is still too early to provide a quantified outlook for 2021, but we believe the fundamentals of our industry continue to look solid with ongoing healthy demand in logic and foundry. Now, let us look at the guidance for the fourth quarter we issued with our third quarter press release. For the fourth quarter, on a currency comparable level, we expect revenue of EUR 330 million-EUR 350 million. Fourth quarter bookings on a currency comparable level are expected to be in the range of EUR 340 million-EUR 360 million. Next, I want to share and update you on some of our strategic initiatives. In my first call with you in July, I talk about my initial observations that ASM is an attractive place with strong technologies and customer relationships, and well-positioned to benefit from solid long-term trends.
Now, three months later, I'm even more convinced of the strong opportunities ahead of us. ASM's existing markets offer excellent growth prospects. While ALD has already moved into the mainstream over the last years, we expect ALD to turn even more into a core technology that will help our industry stay on Moore's law. In the years to come, beyond 2021, we expect that ALD demand will substantially grow above the $1.5 billion market size that we have guided earlier. Increasing device complexity, new materials, and ever thinner films with higher required conformality mean that conventional deposition will run out of steam, and that ALD is going to take a bigger piece of the pie. Both in logic foundry and in memory, important technology inflections will drive substantial increase in ALD requirements in the medium terms.
To make sure that ASM is well-prepared to win the big part of all these new opportunities, we will keep investing in the growth of our company. For this, we remain further investing in R&D to develop the many new ALD applications that are on the industry's roadmap. Besides that, our CapEx investments in the forthcoming years will be more focused on lab equipment for customer demos and development of new applications. This will help us to further expand our serve available market and to keep our company at a high growth path in the medium term. EPI remains an attractive growth market, and we are strongly focused on adding new customers and driving our market share in the coming years. In vertical furnaces and PECVD, we continue our strategy to invest in targeted niche opportunities, which already drove decent additional top-line growth for our company in the recent years.
To conclude, prospects remain solid, and we are focused to ensure ASMI is going to benefit and stay on its growth path. Now, with that, we have finished our introduction. Let us now move on to the question and answers.
We'd like to ask you to please limit your questions to not more than two at a time, so that everyone has the possibility to ask a question. All right, operator. We are ready for the first question.
Thank you. We have the first question. It's from the line of Stéphane Houri from ODDO. You may ask your question.
Yes, hello. Thank you very much for taking my question. First question would be on Q4. There has been a lot of questions around the Intel issue, I would say. You were saying before that Q4 would be kind of flattish versus Q3, and now it looks that it's going to be a little bit better. What has changed during the quarter on the outside? Linked to that, can you update us on your view on the evolution of the situation with Intel that, according to what we know, remains a pretty important customer for you? Thank you very much.
First of all, Stéphane, thanks for calling in. The situation with the logic customer, we will not comment on the specifics, but I can maybe explain to you in more generic terms. One of the things that we have also explained in the last earnings call is that, between logic and foundry, if there's going to be a drop in logic production and the demand for advanced semiconductors continue, the capacity would probably have to be taken up by foundry. We are hearing of some noise in the market of some of these things that are happening. At the same time, I would say that because of the delay in the more advanced nodes, the logic customer is probably going to invest more in the existing high volume manufacturing node that they have. I think this was also confirmed in their latest earnings call.
When you look at all of that put together, as far as we are concerned, we don't see really a significant impact from, let's say, the announcement of the second quarter. We continue to see combined logic and foundry being very strong going into the fourth quarter.
Regarding your first question, Stéphane, we normally, as you know, don't guide two quarters ahead. Given the fact that there was so much uncertainty last time around the later part of the year, we decided to guide the market that the fourth quarter will be at least at the level of the third quarter, to make sure that everyone understood that the market was not falling apart in our view. In principle, there are not that many changes as compared to our view with regard to the fourth quarter as what we had a few months ago.
Okay. Thank you very much.
Thank you. The next question is from the line of Keagan Bryce from Barclays.
Hi, guys. Thanks for taking the question. One on gross margin. Could you give us a little more color as to what that mix was? Was it similar mix to last quarter? Is it more to do with a specific customer ordering for a specific node? Quickly on China, could you update us on where you stand in terms of shipping equipment to your largest Chinese foundry customer? On whether you are seeing any sort of demand pull-in from any of your customers in China. Thanks very much.
Okay, Keagan. Let me try to answer the first question with regard to gross margin. What we are seeing in our gross margin is that certain applications for certain customers have a higher margin or as an average, a lower margin. We had an extremely, I only can phrase it again, an extremely rich margin in the third quarter, even richer than the second quarter that we have seen. That is rather unique, to be honest. We expect, looking at what's going on in this moment with regard to the fourth quarter and looking also in the first quarters of next year, that these margins at this level are not sustainable.
As what I mentioned in the prepared notes already, we expect that we are moving more towards the high end of the range that we originally always have indicated, instead of that what we have seen in the past, where you have big shifts between low 40s and mid 40s. That's at least the view that we have for the next quarters to come.
Thank you, Peter. Keagan, on the China question, maybe to also elaborate a little bit. We are very encouraged to see that our initial efforts and investments in China is starting to pay off. This year, for the first time, we will see our China sales going into slightly more than 10% of our total revenue, which is a positive development. We have especially had strong, I would say, sales from China, in Q2 and Q3. We do expect that in Q4, the sales will come down a little bit lower. As far as the big foundry customer is concerned, at this moment, we are still able to ship a majority of our product line. We continue to be able to ship to them if they so want to buy.
At the same time, we understand the restrictions that are coming down from the U.S. government, and we comply with all the regulations. We strictly follow all the rules and the requirements that they have provided. Having said that, just to elaborate a little bit broader, when you look at China, our exposure to China is still small compared to our peers. As I've said, this year, for the first time, we will exceed 10% of our total revenue with sales coming in from China. Where else, for some of our peers, there is a significantly much higher level. I hope that gives you a better picture of our business in China.
Yep. Thanks very much, guys.
Thank you. Our next question is from the line of Adithya Metuku from Bank of America.
Yes. Good afternoon, gents. I had two questions. Firstly, I just wondered if you could give us some color around the levels of growth you saw with ALD in the third quarter and also with EPI. Secondly, I just wondered, as you look into 2021, if you could talk about the trends you expect by end market and also some color on the uplift you expect from the DRAM ALD layer win into 2021. Any initial color around that would be helpful. Thank you.
Okay. Adithya, thank you. On the first question, we, of course, see very strong growth in our ALD business. I think overall, when we look at the whole year 2020, we are going to see a fairly significant increase in the ALD business for us, and that continues to be the main driver for us, for our business. In EPI, we continue to be engaged with the main players in the markets to try to get customer engagements to qualify for the next nodes. That's ongoing and very encouraging. We have, as we have discussed or explained during the second quarter call, that we already have one customer that is already using our EPI tools for high volume manufacturing.
We think that as we go further into next year, hopefully when the customer starts to engage the more advanced nodes, we might see more, let's say, opportunities and our serve available market increase. In terms of trends for 2021, I think you were referring to specifically any kind of uplifts that we see in DRAM. I would just maybe provide you with a more generic picture. I think for 2021, it's still a little bit early, but we do think that logic and foundry will continue to remain healthy going into 2021. In terms of memory, I think there's been probably some pent-up demand in terms of investments. There's ongoing, I would say, announcements about investments in 3D NAND. As you probably saw today, KIOXIA, or one of the Japanese NAND manufacturers, announced a EUR 9 billion investment. I think the NAND market is healthy.
In DRAM, we think that it's more uncertain. It could be much more dependent on what is the demand pool due to the various, let's say, drivers. That's a little bit uncertain for us. Overall, I think it's still a little bit early for mapping out any firm trends for 2021. There's one thing I forgot about the EPI business for us. This year, for the power and analog market, as we all know, with automotive and industrial being severely affected by COVID-19, we are seeing lesser of a demand for our EPI tools in that area. Even though that part of the market, even though it's a very small part of the overall ASM, let's say, market, it does play a certain share in our overall EPI market. That is probably going to bring, I would say, some impact to us.
With the engagements that we have in terms of CMOS in memory, in logic and foundry, we hope that next year we will pick this up again and significantly increase.
Understood. Thank you.
Thank you. Next question. It is from the line of Tammy Qiu from Berenberg. You may ask your question.
Hi, guys. Thank you for taking my question. I would like to ask from technology perspective, looking into the foundry and logic pipeline for next few years, basically they are working on three nano or two nano or something even below that number. Do you know what is the adoption or insertion level of ALD comparing to where they were previously? Am I seeing incrementally ALD being used in more applications compared to previously, i.e., your TAM expansion is likely to accelerate going forward or would it slow down?
Good afternoon, Tammy. Thanks for the question. When we look at the logic foundry, which has been driving a lot of the sales and business growth for us as a company, every time when there is, let's say, a node advance, we generally see a double-digit percentage increase in the number of applications or layers that requires ALD. That's a very natural consequence of the scaling that is being done or that is required. Every time when there is a node change and things become smaller, circuits become more miniaturized, we see, in effect, a double-digit percentage increase in opportunities for us. As we have given or let's say explained in our prepared remarks, over the next couple of years, we see a significant increase in the overall ALD market size going forward, especially from 2021.
We are right now in the process of trying to look at and trying to quantify how big that will be. It's definitely going to be bigger than the previous size of $1.5 billion U.S. that we have guided several years ago.
Okay. That's cool. Also, what about competitors? Because we have been hearing that U.S. competitors try to basically aim in for the same market, and it has been a few years since they entered the market. What kind of progress do you see them making recently, given the TAM expansion is quite attractive?
It's probably one of the most, if not the most, attractive segment of the overall wafer fab equipment market. Competition is definitely going to come after and try to enter the market. I think so far what we have seen is that from our, let's say, first entry and our, let's say, penetration into a certain, let's say, applications where we have a dominant position, such as high-k metal gate and so on, we continue to be very strong in those areas because for most of the customers, for something that is proven and working, they don't like to change. We have a very strong, let's say, position in those kind of applications and layers. You are also seeing newer applications coming up, and that's where the competition tries to come in, and there is competition. We recognize that.
I think at the same time, because of our expertise, and because we have a long presence and continued efforts in both the materials and the chemistries that are required for ALD, I think we will continue to see ourselves being the market leader. If not, in fact, maybe winning more share as the market expands.
Okay, cool. Thank you.
Next question is from Nigel van Putten from Kempen & Co.
Hi. Thank you. Good afternoon. I'd like to follow up on the gross margin. The exceptional strength seems to coincide with exceptional strength in China. Would it be fair to say that growing your sales in China has been beneficial for the gross margin? That is my first question.
Nigel, the only answer that I can provide is a repetition of what I said earlier. Our gross margin is dependent on certain applications at certain customers. That's not country-dependent.
Okay. Yeah. We've discussed before, it's not too dependent. There's no difference between equipment and services. I'm just really trying to figure out, but I guess it's a mix of both.
It's a mix. There are certain, as you can imagine, and it's the color that I provided also in the earlier calls, I think. When you have a very unique solution, then the prices are mostly higher than when you replace an existing technology, where just the timing is there, that the customer is saying, okay, I can now make the choice to go from the older to the newer technology. There you see mostly big margin differences in the different products.
Got it. That's helpful. Thanks. A question, I think you just said you see especially strong ALD market expansion in 2021. Did I hear that correctly? What will be driving that into next year?
Nigel, I think, we continue to see logic foundry being strong, and at the same time, with memory inflections at the more advanced nodes, we think that the ALD will continue to be strong. What I was alluding to was that, especially from 2021 onwards, out into the several years after that, we think that the overall ALD market size is going to grow significantly from what we have guided earlier, which was several years ago, of EUR 1.5 billion. What we are trying to do now is to try and find a way where we could look at the future trends and all the requirements and try to quantify, if it's not going to be EUR 1.5 billion, what is it going to be? Hopefully, at some point in the near future, we'll be able to share this with all of you.
Sounds great. Thank you very much.
Our next question is from Rob Sanders from Deutsche Bank.
Yeah. Hi. First question would just be relating to the previous question, which is, your company's getting quite large now. There's a lot more sell-side analysts taking an interest in your stock. Do you not think it's a good time with new management to do a capital markets day and give a fuller explanation of your growth outlook? Relating to this, second question would be, if Gate-All-Around gets adopted at 2 nm by TSMC, as it seems, after Samsung adopts it from next year, what could be the impact on your served addressable market just from that alone? Have you had a chance to kind of understand how meaningful that would be? That's it from my side. Thanks.
Rob, thanks. On the first question, I think that's an easier answer. We are considering doing something, definitely not this year. If we do a capital markets day, it will of course be next year. If we decide to do it, I'm sure we will invite all of you to share with you our, let's say, strategy, et cetera. On the second question of Gate-All-Around, you are absolutely correct that the Korea site is looking at already insertion or switching over to Gate-All-Around at 3 nm, whereas the Taiwan site is looking at 2 nm. I think both sites are feverishly developing the process.
I think what we might actually see is that there's not going to be a lot of time difference when both of them launch the next nodes at Gate-All-Around. Probably, my guess is in the 2023 timeframe. If it can be earlier, it's even better. The reason for that is, when you move to Gate-All-Around, you will require definitely even more ALD applications. We are already in, I would say R&D and evaluation engagements with both customers. We have some idea, not 100%, but we have some idea what kind of layers will require ALD and also EPI. They will need more EPI layers, highly likely. We have, I would say, a pretty good idea that this is going to open up new opportunities for us, increase our serve available market, and especially in the ALD area, as we said.
It's just going to increase our market opportunities again and by a double-digit percentage compared to the previous nodes.
Okay. Thank you.
Thank you. The next question is from Marc Hesselink from ING.
Yes, good afternoon. Thanks. First question is actually on the NAND market. Clearly, there is quite some momentum. Can you explain what is your addressable market there now at the moment? Initially, you were not that active in 3D NAND, but you're clearly gaining more traction there. Also, how deep are you penetrated with all the NAND producers?
Good afternoon, Marc. On the topic of NAND, I think over the last two years or so, we have gradually but consistently increased our engagement. Much so that if we look at where we are today, we are engaged with all the NAND players in development work for the next nodes. When we look at where the market is today, it is still probably at a level where it doesn't play to where we have been doing or the work that we've been doing for the last two years. I would say that as we move into the more advanced nodes, hopefully next year or the year after, we will see definitely our serve available market and opportunities increase, and that's where we are targeting.
We are in a lot of developments with all the key players, and we know that at some point in time they will switch to the more advanced nodes, and that's when it will be I would not say significant, but highly interesting for us.
If you were to compare that to the opportunity in DRAM and the opportunity in logic foundry relative to each other?
I think over the next two years or so, I think logic and foundry would still be by far the biggest driver for us. Having said that, we do expect that the overall percentage of our business coming from DRAM and from NAND is going to increase from where we are today. It will play a more significant part in our overall and total business.
Okay, thank you. The second question is actually on the spare and service business. You already indicated that part of it is a little bit extra inventories at your clients for COVID-19, but partly also structurally. Is there a way that you can maybe split that out a bit to give us a feel on what, after that initial spike of COVID-19, what should be the normal run rate going forward?
The kind of buffering of spares inventory was actually mainly in the second quarter because I think there were a lot of concerns at that time about what's going to happen with the supply chain. We did see some of that. Now, it's probably, I would say, still a low single digit kind of percentage, so it's not really that significant, but we did see it. Now, structurally, we have, since probably the end of last year, beginning of this year, been looking at how do we improve and increase our spares and service business. One of the natural things is to look at, of course, the increased install base that we have and how do we service the install base.
At the same time, I think we are coming out now with more, I would say, innovative solutions on how to support our customers so that we can help them to kind of reduce their cost per wafer by providing better service, better, I would say, for example, solutions to help them to do that. I think we are seeing the beginning of that bearing fruit. I think going forward, we are going to probably continue to see our spares and service business growing, depending on how fast the customers would adopt our more, let's say, newer and innovative solutions.
It's clear. Thank you.
Thank you. The next question is from the line of Dominik Olszewski from Morgan Stanley.
Yes. Good afternoon, everyone. Just one, since we touched on a lot of topics. I guess as we're getting into Q1, you're expecting the ramp up gradually of the Singapore facility. Obviously, in the past, management had talked about the fact that equipment was sort of being manufactured suboptimally due to capacity and operational constraints. As this new fab is ramping up in Q1, should we expect any spike in activity, particularly front-loaded at the start of the year because of the new capacity availability that you have to play with? That's it. Thanks.
What we basically have said earlier is that indeed, it will lead to more efficiency in that organization. That will come in gradually because the plan that we now have is basically transferring late this year, early next year, the activities towards our new premises. What we stated also earlier, it will have some efficiency improvements, but it will not change the needle. It will not lead to a substantially higher gross margin as a consequence of that. It will make the way that we are dealing with our businesses much more fluid, Dominik. In that sense, you have to think about, as a consequence of that, quality issues, highly likely, will be much better, and the guarantee that we can give to deliver things on time. That will improve. There are a lot of quality improvement actions as a consequence of that also included.
Thank you.
Thank you. Our next question is from the line of Achal Sultania from Credit Suisse.
Yeah, hi. Good afternoon. Just maybe a question on gross margins again. If I think about the mix, obviously, we understand you had a very rich mix in the last couple of quarters. As we look into next year, you're talking about some of these new applications with DRAM, and even NAND customers. How should we think about why shouldn't this mix be more sustainable going into next year? A lot of the new products that you're launching probably should be carrying higher gross margins. We look at your evaluation tools on your balance sheet, and we've seen a big spike in that evaluation tool number, which again suggests that a lot of these tools are being trialed at your customer site.
I understand that some of that mix improvement may go away, but still, I'm just trying to understand why you shouldn't be able to retain a large part of that improvement as we go into 2021.
Yeah.
Thank you.
What I tried to guide also earlier is that, well, when you compare what we have done in the last years, that we were more in the mid part, low part or the mid part of the range of low to mid-40s, that we move more towards the higher end. There are a few reasons for that. I mentioned already the mix. You have to recognize also, Achal, that when we have new products, and we plan again to have a lot of new products, especially in the areas that you just mentioned, that will have an impact. The first tools that we are going to deliver, by definition, have always lower margins. That impact is also taking part of it.
I think that the last important one, I think, is that when you look over a longer period of time, you see that low to mid-40s are the normal ranges of gross margins in this industry. I said earlier, there will be periods where you might be above that level. I think on this moment, stating that we expect that our margin will be structurally above the levels that we earlier have indicated, will be giving you the wrong guidance.
Yeah. Okay. That's understandable, Peter. Thanks a lot. Maybe one last one on the buyback plan and shareholder return. Obviously, you've accelerated the buyback in the last few weeks, after being a bit slow at the start of the program. Is it fair to assume that you should be able to complete this buyback program well before November next year, which was the original timeline? How should we think about the stake at ASM Pacific? Is it something that you still think is core to the business, or are you willing to actually look at options around that stake?
It's highly likely that the share buyback program will be finished well before November next year. That's the answer on the first question. The second question, with regard to your second question, nothing has changed as compared to what we have said last time.
Yeah. Okay. Makes sense. Thank you, Peter.
We have another question again from Mr. Adithya Metuku.
Yes. Thank you, guys. I just had two follow-ups. Just firstly, looking at your peer group operating margins on a like-for-like basis, it feels like your margins, when adjusted for U.S. GAAP, are roughly around 10 percentage points below peers. I just wondered what you think the reason is for your margins being lower, and how should we think about the scope for improving profitability going forward? The second question is on CapEx. As you look into next year, could you give us some color on how we should think about CapEx? It looks like your capacity deployments were largely finished this year. Can we expect CapEx to go down maybe into the 35 to 40 million range next year? Is that a possibility? Thank you.
First of all, your first question about margins. I suppose that you talk about EBIT margins here, not about gross margins.
Yes. Your gross margins are broadly in line with peers, but your EBIT margins are lower.
Yeah. One of the major reasons for that is the size of our company. That's what we also explained in earlier calls. We plan to have gross margins. We make a distinction, of course, between R&D and SG&A. The R&D that we are reporting is around the 10%, but that's IFRS. When you compare that with the U.S. GAAP, what our competition is doing, then we are more at around the 14%-15%. That's also what our competition is doing. There, we don't have the major difference. The major difference is indeed in the SG&A cost, and it has to do a little bit of size. I don't want to argue with you if it's 10% or it's less, but there is a difference, and there size plays a role. We recognize that. The second thing is with regard to CapEx.
With CapEx, we have indicated towards the market that after the big investments that we have done in the new fabs, that we will, as an average, be back to the EUR 30 million-EUR 40 million, what we also have seen in the past years. That will be an average. What Benjamin already was referring to is that we are clearly looking on this moment, given the fact that there are so many opportunities popping up, especially in the new products that we see, that we might speed up there some CapEx investments in the next 18-24 months, and that are especially related to measurement equipment for our lab environment and eval tools. We need to serve all those new applications that are popping up, we might need to have somewhat more equipment that we put in our own lab environments.
As a consequence of that, our CapEx might remain on a relatively high level for an extended period of time. Again, that will be over time developed to the direction that I've mentioned earlier.
Understood. When you say high end of CapEx, are you thinking about the high end of EUR 30 million-EUR 40 million or higher than that?
Higher than that. Might be higher than that.
Understood. Very clear. Thank you.
Thank you, Adithya.
Thank you.
Operator, can we have the final question, please?
Yes. The next one, it's from Mr. Nigel van Putten again. Thank you.
Yeah, thanks a lot. Just a quick follow-up on foundry. Both strong in the order book as well as the next quarter sale. Just a question, is that sort of the big one or two spending more, or is that more a breadth with also Tier 2 foundry spending more?
Sorry. I think your question was, are the Tier 2 foundries also spending more? Am I correct, Nigel?
Yeah. What's driving the strength both in the order book as well as revenue into sort of the next two quarters? Is that sort of the big one spending more, or is that maybe some of these other foundries also picking up investment?
I think by and large, a large part of it is because the big ones are spending more, spending actually a lot more, and hopefully that continues. In their latest kind of earnings, they did not want to give a number for next year. I think the consensus of everybody is that in the Taiwanese foundry next year, they are even going to spend even more than what is already revised upwards this year. I think they will spend a lot more, and I think it's primarily driven by the demand that they are facing now from their customers. There's a lot of pull.
Would it be fair to say that as they spend more on the advanced node 5 nm, that your share of wallet will again increase year- on- year?
I think especially if they go into the next node. They are now in the current node where they are ramping up, and there's all kinds of guesses as to how far they will go in terms of capacity for this node. We are heavily engaged with them now. At the same time, we also heavily engage with them should they go into the next node. The next node, again, as we have always explained, it's going to see a double-digit % increase in terms of number of opportunities and layers for us. The faster they go into the next node, the bigger the opportunity is for us.
Great. Very clear. Thank you very much.
Thank you. There are no further questions at this time.
Okay. Thank you, everybody, and I would like to thank you all for your attendance today, also on behalf of Peter and Victor. We look forward to meeting with many of you in our upcoming investor roadshows and conferences. Once again, thank you. Stay safe, be healthy, and goodbye.
Thank you. That concludes our conference for today. You may all disconnect. Thank you all for participating.