Good day, welcome to the ASM International Q2 2016 earnings call. Today's conference is being recorded. At this time, I would like to turn the conference over to Mr. Han Westendorp. Please go ahead, sir.
Thank you, Elaine. ASM issued its 2016 second quarter results last evening. For those of you who have not seen the press release, it, along with our latest investor presentation, is accessible on our website, asm.com. We remind you that this conference call may contain information relating to ASM's future business or results, in addition to historical information. These forward-looking statements involve risks and uncertainties that could cause actual results to differ materially from those expressed or implied in such statements. For more information on the risk factors that could affect results, please refer to the company's press releases, reports, and financial statements, which are available on our website. With that, I will now turn the call over to Chuck del Prado, President and CEO.
Thank you, Han, thanks to everyone for attending our second quarter 2016 results conference call and for your continuing interest in ASM International. After a review of operations, Peter van Bommel, our CFO, will join me in answering any questions that you may have. Let's review our second quarter financial results. Revenue in the second quarter amounted to EUR 139 million, a decrease of 3% compared to EUR 142 million in the first quarter. Revenue decreased 31% compared to the second quarter of last year, which was a record revenue quarter with EUR 201 million. Revenue in the quarter was towards the high end of our guidance, which was a range between EUR 130 million and EUR 140 million. Our ALD business continued to be the key revenue driver. By customer segment, revenue in the quarter was led by foundry and logic.
The gross margin remained at the solid level of 43.8% in the second quarter, compared to 43.9% in the first quarter. The second quarter last year showed a margin of 45.1%. The quarter-on-quarter variations in the gross margin are mainly explained by mix differences. SG&A expenses decreased by 7% compared to the previous quarter. Reported R&D expenses increased by 6% compared to the first quarter. Amortization of capitalized development expenses amounted to EUR 4 million in the second quarter as compared to EUR 7 million during the first quarter. The total R&D expenditure, excluding amortization and impairment, but including capitalization of development expenses, amounted to EUR 25 million in the quarter, up 3% from the first quarter and up 5% compared to the second quarter of last year. The year-on-year growth in the total R&D expenditure is explained by an increase in customer requests for new applications and engagements, similar to last quarter's.
We generated operating income of €17 million in the second quarter, with an operating margin of 12%. This compares to operating income of €19 million and an operating margin of 13% in the first quarter. Financing results in the quarter was €10 million positive and mainly consisted of a €8 million translation profit. As a reminder, we hold a substantial part of our cash balances in US dollars, and the translational effects of currency changes are included in the financial results. Let's look at ASMPT for a moment. Results from investments, which reflects our 39% share of the net earnings from ASMPT, amounted to €16 million for the quarter, up from €6 million in the first quarter of this year.
The ASMPT results include a provision for restructuring of €9 million, of which €4 million is attributable to the ASMI results, and those restructuring costs are related to the relocation of manufacturing operations in China. In the second quarter of last year, results from investment was €20 million. These figures exclude the ongoing amortization charge, which amounted to €7 million in the second quarter. For the full year, this charge is expected to be approximately €27 million. In the second quarter, ASMPT sales increased quarter-on-quarter by 28% to HKD 3.7 billion. Compared to the second quarter of 2015, sales declined by 2%. ASMPT reported bookings of $561 million for the second quarter, an increase of 37% compared to the first quarter and an increase of 14% compared to the second quarter of last year. Let's now turn back to ASMI's consolidated operations.
ASMI's net earnings on a normalized basis amounted to €42 million in the second quarter, up from €12 million in the first quarter. While operating profit decreased compared to the first quarter, the net result showed improvement due to an €8 million translation gain on cash held in foreign currencies. In the first quarter, we recognized a translation loss of €11 million. Our new orders in the second quarter were €159 million, down 3% from €164 million in Q1. Orders were within our guidance of a range between €145 million and €165 million. Orders were mainly driven by our ALD business. Looking at our bookings, logic foundry represented the largest segment in the second quarter and showed a strong increase compared to the first quarter, driven by demand related to the 10 nanometer technology node. Orders in the logic sector decreased compared to the first quarter, while foundry orders strongly increased.
On the balance sheet and cash flow, at the end of June, the cash position decreased to €370 million, down from €416 million at the end of March. The decrease was mainly the result of €37 million in dividend payments and €25 million in share buybacks, partly offset by a €7 million dividend received from PT and the positive currency effects on the cash position. Net working capital stood at €130 million at the end of March, up from €113 million at the end of the first quarter. I think that's a mistake. Correction. Net working capital stood at €130 million at the end of June, up from €113 million at the end of the first quarter. Apology for that error. This increase was mainly caused by a higher accounts receivable position due to strong sales towards the end of the quarter.
The number of outstanding days of working capital measured against quarterly sales increased to 85 days at the end of the second quarter, compared to 71 days at the end of March. In the second quarter, we spent EUR 25 million to repurchase approximately 650,000 of our own shares as part of the EUR 100 million share buyback program that we announced last October. As of last week, we completed approximately 70% of the program. As the press release showed, up till the end of June, it was around 65% of the program. The number of outstanding basic shares decreased to approximately 60.7 million shares at the end of June, down from 61.3 million shares at the end of March and 61.9 million at the end of the second quarter of last year. Let's now look in a little bit more detail to our ALD business.
As mentioned earlier in this call, 10 nanometer investments in logic foundry strongly supported our bookings during the quarter. Single wafer ALD is a critical and enabling technology for logic and foundry customers to make the transition to 10 nanometer. Next to high-k metal gate applications, more precise and conformal deposition is needed for several other critical process steps to build these advanced infrastructures. As we have mentioned before, the shrink to 10 nanometer resulted in the qualification of our single wafer XP8 tool for low temperature ALD patterning applications in the logic foundry segment. In addition, we have developed new applications. As an example of such an innovative ALD application that we highlighted at our technology seminar earlier this month at SEMICON West, is the use of ALD metal oxide hardmask in the 10 nanometer node.
This is just one example where ALD technology supports our customers in addressing the challenges of ever smaller geometries in 3D device structures. Consequently, the number of ALD layers for 10 nanometer, for which we have been selected, has increased substantially compared to the 14, 16 nanometer technology node, which we believe positions us well for a significant increase in the share of wallet with key customers in the logic foundry segment. Longer term, the introduction of new materials and architectures such as gate-all-around will further increase the ALD opportunity we believe in the logic foundry sector. In the memory market, our ALD equipment has supported key customers in the ramp of several technology generations.
Multiple patterning continues to be a key enabler for customers in the DRAM sector in their move to the 1X technology node, especially with some of our most recent single wafer ALD process and hardware innovations. We are well positioned to serve these increased patterning requirements of our DRAM customers as soon as demand for capacity in this industry segment picks up. In NAND flash, customer investments are largely geared now towards 3D NAND. There is little need for multiple patterning-related single wafer ALD capacity, which up until last year was a solid driver for us in planar NAND. In the current 3D NAND devices that are in high volume, the number of new single wafer applications, non-patterning applications, has also been limited so far. With next device generations, we expect single wafer ALD to be required for a growing number of applications.
Our XP8 ALD tool has recently been qualified for several non-patterning related applications in 3D NAND. We expect that these applications will increasingly contribute to our revenue in 2017. At the same time, we are strongly focused on further expanding the number of applications in 3D NAND. Looking at the next few years, we remain confident that ALD continues to be an attractive growth market. Looking at the market environment as a whole, looking at the semiconductor end markets, conditions continue to be mixed. Worries about the global macroeconomic environment and Brexit persist. Softness in the PC and tablet markets and slowing growth in smartphones and continued oversupply in memory devices will result in a year, 2016, of 3% decline in the semiconductor industry.
On the positive side, the drop in pricing in certain segments seems to be bottoming out, some of the chip makers are becoming more positive about the rest of the year. With regard to the equipment market in which we operate, a key positive in 2016 wafer fab equipment spending outlook are technology transitions. In the logic foundry segment, the leading companies are investing in equipment for the 10 nanometer node. Another key wafer fab equipment spending driver in 2016 is the transition to 3D NAND memory. The NAND end market is strongly driven by solid-state drives, in particular for data centers. In contrast, spending for the DRAM sector, which was very strong in 2015, continues to be weak and is now not expected to improve throughout 2016.
The most recent forecast for 2016 wafer fab equipment from Gartner and VLSI Research reflect a roughly flat year, which compares to a low-to-mid single-digit decrease still expected earlier in the year. As far as the single wafer ALD market is concerned, 2015 has shown a strong growth. While we estimate this growth to continue, whereby the single wafer ALD equipment market will double in the period 2015 to 2018, 2019, as we believe, and as we shared already in the AGM presentation in May, we do expect the single wafer ALD market to show a double-digit decline in 2016 caused by a low demand in memory, which will not be offset by the strong increase in the logic foundry segment. Our single wafer ALD market share is expected to show an increase in 2016.
We expect the single wafer ALD market to strongly improve again in 2017. To provide these statements a little bit more color, in memory, the NAND market is shifting, as we discussed before, from planar NAND to 3D NAND, and the shift will significantly reduce the demand for additional multiple patterning related single wafer ALD capacity. A non-patterning related single wafer ALD demand is expected to gradually increase as next generation 3D NAND is going to ramp into high volume. In memory DRAM industry capacity demand was high in 2015 and has declined strongly in 2016. Based on current visibility, the demand for new DRAM related single wafer ALD capacity is believed to recover earliest in the course of 2017. In logic foundry, the transition to 10 nanometer technology node fuels substantially higher single wafer ALD equipment demand than in the previous nodes.
Based on that, now let's look at our short-term outlook that we also shared in our press release for Q3. We expect sales between EUR 135 million and EUR 150 million, while we expect an order intake of somewhere between EUR 100 million and EUR 130 million, both on a currency comparable level. We expect a stronger second half in terms of sales as compared to the first half of 2016. At this point, we are happy to answer any questions that you may have. We would like to ask you to please limit your questions to not more than two at a time so that everybody has a chance to ask a question. All right, Elaine, we're ready for the first question.
Ladies and gentlemen, if you would like to ask a question, please signal by pressing star one on your telephone keypad. Our first question today is from Peter Olofsen from Kepler Cheuvreux. Your line is open.
Good afternoon, gentlemen. First question on the H2 outlook. You gave an outlook, so I'm trying to understand what your expectations are for Q4. Do you basically expect to ship your end Q3 backlog, or do you also foresee some turns business for Q4? Then related to 3D NAND, from previous calls, I had the impression that we could already see some traction from you in the 3D NAND market later this year. But now it seems it's more 2017. Is that because the ramps by your customers have been pushed out or because it's not at the 48 and 64 spec devices, but only at a more advanced type of 3D NAND?
Peter, clear. On Q4, it's clear that if you combine our statement on our guidance on Q3 with our statement on the second half, because we now do expect a stronger second half in terms of revenue. You may have noticed, it's somewhat of a revision. We set a much stronger second half in Q1, and now we say a stronger second half. That is somewhat of a change compared to three months ago. At the same time, we still expect logic foundry to increase from the first half to the second half in terms of revenue. On memory, that's still to be defined whether memory is able to match the demand for the first half. That really depends on timing of further V-NAND demand, and that we will only get a better picture on that in the course of this quarter.
That is related to, this comes back to your other question, to your second question. If you look at the 3D NAND segment, then basically, most of the 3D NAND spending currently is done in the current mainstream HVM activities. We have said in earlier calls that we are not part of the current mainstream HVM activities in 3D NAND. The capacity demand for the new activities in 3D NAND, so the non-current mainstream HVM activities that we are engaged on, has basically been delayed by three to six months, a little bit depending on what customer you talk about.
Could you then be more precise? You say next generation 3D NAND, but I'm still struggling a little bit to understand what that.
Yeah. Next generation, it's the next generation for that specific customer. It is not necessarily, let's say, a new node in the industry because we are not allowed to talk about specific customers, but it is, let's say, the most advanced node for that specific customer. We are engaged with several customers, and for them, it's their most advanced 3D NAND node. What you do see is that depending on the customer, their ramp is three to six months delayed, and that impacts our outlook. There has not been a change in our, let's say, outlook for those customers as soon as they ramp. Because also we have quite a number of seeding machines for their, let's say, initial wafer starts that they have been building. Those tools, that install base has been building throughout this year so far, according to plan.
Okay. That's helpful. Maybe coming back on the comment on the outlook, just to understand it. If I look at the current backlog, that's around EUR 175 million, but that will decline in Q3, given that the order intake will be below the sales level. The end Q3 backlog, will that be a good indicator for your Q4 sales? Could you do better in terms of sales?
Let me try to answer that, Peter. When you look to the backlog that we have at the end of June, it's highly likely that whole backlog will be delivered in the course of 2016. That's normally the case. Moreover, when you look to the bookings that we will receive in the third quarter or so, much of those will be delivered in the course of this year. Last but not least, orders which will come in the course of the fourth quarter. A part of that will also benefit to the sales turnover for the whole of 2016. Only a part will be delivered in 2017. I think that gives you the color that you were looking for.
Yes, it is. Maybe one follow-up on gross margins. I have the impression that in terms of revenue mix, there will be a shift for more memory in H1 to more logic foundry in H2. Will that affect your gross margin?
No, I think that as far as gross margin is concerned, the same standard that we have set in all our calls in the last quarters, that the gross margin of different product or product lines doesn't show a big difference. There are certain applications for certain customers which have a lower margin or a higher margin than others. In general, the margins are more or less the same for most of the customers.
Okay, thank you.
Okay, you're welcome.
We will now take our next question from Mathias Santos Serra from Morgan Stanley.
Hi, gentlemen. One question more on memory. It's a bit similar to previous call question. What is demanded for NAND to drive the ALD market again? Is it simply migration to smaller nodes, or is it more building stacks on top of each other? What could be the potential timeframe for that? Second question, we see that ASML seem to have more and more traction on EUV. Is that something that could help NAND to go back to planar again and to benefit you guys? Thanks.
Okay, to start with your last question, we don't believe that changes in the schedule of EUV will influence our forecast on NAND revenue. To dive into your first question, to give you a little bit more color on the NAND market as it relates to our single wafer ALD business. We do think that already in 2016, the demand for 3D NAND related single wafer ALD will outgrow the demand for planar NAND related single wafer ALD. We don't think that any schedule changes for EUV or so will impact that. Our belief is that with next generation higher stack devices, that the complexity of these structures will increase and that as a result of that, the need for single wafer ALD will increase.
Of course, the speed with which 3D NAND will contribute to our sales will of course depend on the success of our customers to ramp the yield of upcoming versions of 3D NAND devices. It's not completely in our control. The best thing we can do is engage with them in the intense way we have been doing for the last couple of years. Then based on that, we trust that the payoff will come. Based on that, our current estimate is also that we do expect a strong double-digit % increase in single wafer ALD market size year-on-year, looking at the 3D NAND segment. Based on, let's say the relationships and the R&D engagements that we have been building with several 3D NAND customers over the last couple of years. I trust that answers your questions, Mathias.
Yes. One follow-up, though. Do you think planar NAND will come back? Is there some planar NAND, for example, in your long-term guidance for EUR 1.5 billion in 2018/2019?
No. Let's put it this way. Planar NAND is not instrumental in the estimates that we used to get to our estimate for the growth of the ALD market in the next three to four years. It has not played an instrumental role.
Thank you very much.
Okay, you're welcome.
We will now take our next question from Tammy Qiu, Berenberg. Your line is open.
Hi, Chuck and Pete. Thank you for taking my question. The first one is actually, I have a big picture question. As a smaller semi equipment company compared to your peers, your performance is definitely seems to be more volatile than them, and everyone has been seeing DRAM related weakness. Just wondering from a big picture perspective and long-term strategy perspective, do you think you are still better off as a smaller niche focused player instead of being a part of a large company? That's the first question. The second question, I have a follow-up on 3D NAND.
Okay, Tammy. Yeah. The first question is very simple. We don't think it is a factor at all from a fundamental point of view. This is just a matter of timing of specific parts of the industry, specific customers. You see that regardless of those dynamics, if you look at the P&L from a distance, it's still a very healthy P&L. Healthy growth margins, healthy bottom line, just a timing issue, We should not get too nervous about that.
Okay, thanks. The second question is about 3D NAND. Basically, all the other equipment makers have been reporting good 3D NAND momentum, some of them have been talking up, the 3D NAND capacity requirements for this year. You have mentioned that your version of 3D NAND is delayed by 3-6 months. I'm just wondering, is that because technically it's difficult to release that new version of design, or it is that the chip maker is actually realizing they can use ALD at a later stage instead of now?
I think, Tammy Qiu, it just has to do with two factors. First of all, again, we have been very consistent in our message on this for several quarters. It's not something that we just explain to you now because our numbers are somewhat worse maybe than the industry as a whole. We have always said that our engagement in 3D NAND, that's what I, after the fact, regret, we started three or four years ago with that engagement. Yeah, we were, until that time, mainly focused on Planar NAND, DRAM, and the whole logic foundry segment. The inception of new technologies takes 3-4 years from, let's say, the initial engagement with customers. As a result of that, we have just not been engaged in HVM with, let's say, the mainstream HVM activities at this moment in time in 2016.
We do expect that as, let's say, the engagement of our customers to their new nodes is going to HVM, that finally our engagement at least as we said, started three, four years ago, finally are going to pay off. In that respect, we are indeed delayed compared to some of our peers. As a result of that, we don't perform in 3D NAND in the same way as our peers at this moment in time.
Is that 3-6 months delay because of their design delay, or it's because-
No. To add that, we now are dependent on how quickly they can get their current designs to decent volumes based on the speed with which they can optimize their yield. That basically, the 3 to 6 months only has to do with that. It has nothing to do with our technology, the qualification of our technology. There are, as far as we know, no issues in that respect. It just has to do with their overall device yield, and as far as we know, that has nothing, not specifically to do with our ALD technology.
Okay, thank you. Last question is, basically, it is early August, what is the opportunity for ALD demand from foundry and logic market to be turned out to be better than expected, or DRAM market start to kind of recover slowly towards the end of the year. Do you see any chance for that to happen, or do you think this year is pretty much can be concluded as a double-digit down year?
I think the biggest impact, of course, is that DRAM. As you compare to, let's say, the outlook that we made, let's say a few months ago, the delay of DRAM from later in the second half, moving out to earliest, in the course of 2017, that has of course had a big impact. The delay of, let's say, HVM ramp of 3D NAND, as we discussed through several questions early on, has had an impact. At the same time, our logic foundry has been pleasantly strong. Very pleasantly strong.
Of course, what we are doing to optimize the result as much as we can for the year is of course to push the demand and the capacity for logic and foundry as best as we can and try to get any upside possible in terms of timing of 3D NAND ramping with our current engagement as much as possible, to benefit from that as much as possible. That's the only thing that could change our outlook in a favorable way. The DRAM part, we don't expect to change in any way. That's basically the color we can give Tammy. We are looking, of course, at any other opportunities longer term we may have in 3D NAND. There's so much happening there.
Through our engagements in ALD, we are looking at what opportunities longer term there are to engage in a stronger way from a company point of view in 3D NAND. That will not likely affect the bottom line for this year.
Okay. Thank you.
Okay. You are welcome, Tammy.
Our next question is from Chetan Udeshi from JP Morgan.
Hi, this is Sandeep Deshpande. Just a couple of questions, if I may. Firstly, on the memory market, this year you're seeing some impact on the transition to NAND, but with EUV happening, at the moment, DRAM is using multiple patterning, which is you're possibly using some of your ALD technologies. Will that have an impact in the future on when DRAM begins to switch in 2018, 2019 to EUV? Will that have an impact on your ALD demand? Secondly, with regard to market share in ALD, you're saying that despite the issues in 3D NAND and DRAM, your market share remains stable this year. One of your competitors who is also trying to break into the ALD market says that they are gaining share in some processes.
Can you just comment on overall how your position is at this point in terms of the new processes that are coming in which will possibly use ALD, and how you see your positioning in those processes using ALD? Thank you.
Yes. Okay. That's clear questions. Well, first, on the market share, let's start with the last question first. We have shared with you before that in logic foundry, the switch from 16/14 to 10, has, for us, been much more significant than the switch from 20 to 16/14. As a result of that, we have been gaining a substantially higher number of ALD layers at the 10 nanometer node, as a consequence, a higher share of wallet in the logic foundry segment, we believe. At the same time, there is also a mix effect because our market share in memory is strong, but somewhat lower than in logic foundry likely. For 2016, the memory part of ALD market is clearly dropping in a significant way.
Those two factors together make us believe that our market share in 2016, regardless from announcements on new products that some of our competitors are making, our market share likely will increase in 2016. Some of the introductions and/or commitments that competitors are making relate mainly to longer term future. Your question on EUV. Yeah, I would like to relate it to maybe our predictions on the ALD market as a whole, and our predictions that the market has a bright future towards 2018 and 2019. Yeah, in those projections, put it this way, the patterning part has a contribution, but in our current estimates, it's not the driving factor in those forecasts.
For us, whatever happens in EUV is not, in our opinion, going to change our view, our fundamental view that there is big growth opportunities for the single wafer ALD market in the next three to four years.
Thank you.
Does that answer your question?
Yes. Thank you.
Okay. You're welcome.
We will now take our next question from Jim Santinelli from RFSA.
Yeah, good afternoon. Thank you. Couple of questions. One, I wanted to dig a little deeper into your order guidance for the third quarter and understand exactly where the drivers of that are. In so far as, I understand your commentary around memory. Specifically, if we look at the logic and foundry market, specifically TSMC and Intel, they still have the majority of their budget to spend in the second half to get to full year guided CapEx. It looks to be in the region of a EUR 0.70 sequential increase in CapEx spend in the second half for both of those players to get to their guided number for the year. Against that backdrop, it's surprising to see your new order number down so strongly to the midpoint.
I'd like to get a little bit more color around that and why you're not seeing order flow from those two big customers. Secondly, I wanted to understand, in your commentary around 10 nanometer logic market share, whether within that you are including seven nanometers. Clearly we're seeing some name playing, I guess, within the industry around nodes. TSMC's seven Ns is effectively an extension of 10, and I want to understand whether, within your strong market share predictions for 10 nanometer, you are also including the seven nanometer for the foundry players, which is effectively an extension of 10 with little change in litho. Thank you.
Yeah. You said, first question, no order flow in logic foundry. That's not what we try to get across with you, because we definitely expect order flow in the second half of the year. You may not like the order guidance for Q3, the only thing I can say beyond the formal guidance that we gave through our press release is that we don't expect the logic foundry order flow to disappear in the second half. On the contrary, what we did say is that we do expect logic foundry revenues, sales, to be stronger in the second half than in the first half. Related to your seven nanometer, we absolutely have strong activities ongoing with the foundry segment on their seven-nanometer program. There are even tools already on the floor with our customers that are working on seven-nanometer programs.
Indeed, it's very likely that in case that segment would accelerate introduction of seven nanometer, that we could see further traction of that in 2017. That depends on the timing of our customers. The only thing we can do is to prepare ourselves with all the ALD solutions that we have available to accommodate our customers. And I think from that point of view, we are in very good shape.
Okay. Thank you. Maybe just to follow up on your first answer there around logic foundry orders. I don't think I was implying that you weren't going to see any logic foundry orders at all. I was just wondering why you weren't seeing or apparently not seeing a strong Orders for the third quarter should be reflecting CapEx spent over the second half for the logic and foundry sector. I understand DRAM is slightly down sequentially. I'm just trying to get a better feel for how half on half your order flow from logic foundry is looking.
What you also should take into account is that there is also logic foundry activity present in the current order book, in the backlog that has not been delivered yet. That's what you have to take into account. What you also have to take into account that activity level in logic foundry also differs quarter by quarter. It's not necessarily a linear pattern, or I can confirm you it is definitely not a linear pattern. Q1 looks different than Q2, and Q3 could look different than Q4. That's what you all have to take into account.
Sure. I guess, the implication for guidance for 2017, which is the ALD market overall, is going to see a strong increase into 2017 for all the reasons you've outlined. The implication there, I guess, is we should start to see order flow for that, obviously not in the third quarter, but probably in the fourth or first quarter.
Yeah.
Is that a fair indication?
I can follow your reasoning, we're just not used to give order guidance beyond one quarter.
Okay. Thank you.
Yes. You're welcome.
As a reminder, ladies and gentlemen, if you would like to ask a question, press star one. If you feel like your question has already been answered, you can remove yourself from the queue by pressing star two. We will now have our next question from Philip Schulte, Kempen.
Yes. Good afternoon, everybody. You are guiding for a strong improvement in the total ALD market in 2017. I was wondering what kind of visibility you actually have on that, and what your expectations, maybe also per segment are in that? What do you really need for that to really happen? I was also wondering whether you expect to maybe grow your market share further in that year with that strong growth. My second question is actually related to your reiterated expectations for the total market by 2018 and 2019. Why do you believe this development that you announced today is actually not impacting the timing of that doubling in those next couple of years?
All right. Okay. First of all, what visibility you have on 2017? Of course, it's the visibility of today and based on the relationships we have. To provide a little bit more color, our assumptions there are that logic foundry will continue to contribute in 2017 on further building at 10-nanometer capacity and also working on 7 nanometer. Especially in the foundry segment, preparing for 7 nanometer. We took that all into account in our assumptions for the single-wafer ALD market for 2017. We do assume that the DRAM market will recover to some extent in 2017. We do expect that the V-NAND, as we discussed earlier, that there is a double-digit % increase in the contribution of V-NAND to the single-wafer ALD segment also year-over-year. That has driven our statement on the 2017 market.
The second question is on the doubling of the market. What we can say there is that, at this moment in time, with the visibility we have now, we are still confident that the ALD market's potential to double, we are still confident that will happen. The key is that we believe in the structural growth drivers. Based on the applications that the different industry segments likely will need as they go to the next generations of their devices. What is difficult, of course, for us to predict, and this year provides evidence for that, is that the market performance in a specific year is more difficult to predict and dependent on specific timing, on customer investments. As we have seen indeed this year, for instance, with the timing of the DRAM investments that are pushed out completely of the year.
In that respect, are we 100% sure that timing-wise we are right on the statement, the amount of color that we try to give to the market on the ALD potential? No, we're not completely sure, but we think it gives a good feel to the market of what the potential of single-wafer ALD is compared to the baseline of 2015. That view, as of today, has not changed as a result of, let's say, the overnight press release.
Right. A quick follow-up, if I may, on the first one, actually, because if I hear you talk about the logic continuing to contribute strongly, that sounds to me like you may be able to maintain your market share under those assumptions in 2017.
I think that's too early. That's too early to say. We think there is a strong contribution of logic foundry, the combination, in 2017. Dynamics between logic and foundry can differ also for part of the year or by quarter. We feel confident that we are in a good position to benefit from the demand in that segment. What the contribution will be from some of our competitors in 2017 is difficult to predict in detail, but we feel confident, again, that based on the stronger relationship that we have built from 16, 14 to 10, that we are also in a good position to build on that towards seven. Based on that, translate a good traction in that part of the market into also a good performance in our opinion.
Okay. Thank you.
A strong contribution of that segment for our opinion.
Right. Cool. Thanks.
Our next question is from Marc Hesselink, ABN AMRO.
Yeah, thank you. First one, maybe some explanation in the, if I'm correct, and what you accept already is that your market share is better in logic and in memory. There's some competition. Is there a difference in memory on the DRAM side or the V-NAND side on where the competition is and where do you see them? Secondly, the second question is on your R&D cost. We've seen at the SEMICON West again, some new applications. Do you expect this to continue, that you need to invest in new applications? Is that something you're planning to speed up with the expected growth going into next year? Should we see something more similar that we see at the moment?
To start with your last question, you can see based on our R&D levels that we continue to invest in R&D, and that you should read that we continue to invest in staying and leading innovative company in this industry. We don't let ourselves get distracted by, let's say, these kind of quarterly deviations in our performance. We have a clear strategic roadmap from an R&D point of view, in our engagements with our customers, and we're holding on to that. Of course, our ambition is to even accelerate the time to market of new solutions. That's one. The other question is of competition in DRAM versus NAND. As we said, the single-wafer ALD NAND market is hardly there. DRAM has had significant volume in the past. That's where we experience most of the competition if you compare these two memory segments.
Okay then. Maybe then as a follow-up, in what you see when you're discussing with your clients going forward, do you think that competition will have also some traction in the NAND area, or do you think that it's more like what you're seeing right now in logic, that you are way ahead of the competition?
That's too early to tell, but I would be surprised if some of our competitors would not also be focused on the NAND segment. The only thing we use for our estimates is we try to estimate for ourselves the size of the single-wafer ALD market and the contribution to the whole single-wafer ALD market as a whole by year. We try to assess our bottom-up potential with the different customers year by year.
Okay, thanks.
We will learn more as we go. Okay? You're welcome.
Our next question is from Jonathan Rees, Apis Capital.
Yes, hello. Maybe two question on some other things. Maybe how you see a development going forward of your other business, especially epitaxy. If I remember, you also forecast a growth opportunity there going forward. For what do you see for the second half or for the coming two years in this space? Yesterday in the call of Besi showed a chart of VLSI that they expect assembly equipment market in the next two years to grow double digit. So far, you can maybe give us a light on this. Should we therefore expect even better years ahead for ASMPT? Hello?
For PT, I think we really have to focus on the way they reported themselves. What we like about their reporting is that their sales has gone up significantly, their bookings have gone up significantly, both strong double-digit numbers. Their position in quite a few segments of the market has improved. Their margin has improved. Of course, there is also some volatility in the margin, but there are a few good elements in their report. How that market will evolve going into next year, I think it's, for us as investor in them, too early to draw any conclusions. I think also for them, for themselves.
Only a short follow-on on this. You feel that they have maybe improved their positioning because they lost some market shares in the past, but you think they are on the way back?
I think in certain markets, we think they have improved their position. Maybe we can follow up on some of your questions through investor relations. I think for a great level of detail, I propose we refer you to their own press release. In some markets, in some of their products, we view that they really have improved their position. They also explain that in parts of the LED market, they have done a lot better. In parts of the CIS market, they have done a lot better. In general, their back-end equipment, they have been performing pretty well, both in terms of top line and in terms of margin development.
Okay.
I trust that provides some more color. Your other questions on the other product lines, I think in EPI in general. EPI this year so far has been developing favorably. At a much lower level than, of course, ALD, but it has been developing favorably. The vertical furnace has had a continued healthy, stable contribution to both top line and bottom line. We're very eager on our other product lines on PE, CVD, EPI, and furnace. We're very focused looking at the inflection points ongoing in several parts of the market to see how we can engage and benefit from that in the next couple of years. It's too early to share any more details in an investor conference call on that.
I can guarantee you that we are very focused on it, and as soon as we see any contribution to the P&L, we will share it with this audience.
Okay, thanks a lot.
Yeah. You're welcome.
We will now have our next question from Robert Sanders, Deutsche Bank.
Good afternoon, gentlemen. I just had a question, just to follow up really on 3D NAND. You flagged 3D NAND device yield issues as contributing to a delay. I was just wondering how many customers that relates to?
Yeah. The only thing I want to say, a few, but I don't want to say too much more.
Okay. Then just following up on 64-layer and beyond. Let's say you ship into a couple now, how many can you ship into out of the big five looking ahead to 2017, 2018, of the 3D NAND guys? Thanks.
How many can you have? Can you repeat your question?
Yeah. I was just saying, when you think about the big five 3D NAND guys that are out there, clearly your content or the amount of applications you're shipping into are fairly small today. When you look ahead to 64 layer and beyond, how many of those big five do you think you can get a reasonable contribution into? Thanks.
Our mission, of course, is to get engaged with all of them. We are engaged with all of them in one way or another. What we were talking about in terms of, let's say, short-term revenue contribution, it's a few. We are engaged with all of them, with all five of them in one way or another.
Okay, thanks.
Yeah. You're welcome.
Our next question is from Edwin de Jong, NIBC.
Good afternoon, gentlemen. A few questions left. To get an idea on the revenue segmentation of the first half, could you give some more color on what part was DRAM, what part was NAND, what part was logic foundry? Then maybe looking somewhat further into the future with the pushouts that we've seen in 2016 in DRAM orders, should we also expect that for 2017, you will be able to grow faster than the market? Finally, moving towards gate-all-around structure, which we were already mentioning earlier in the call. We're talking now about beyond 2019, I think, but what implications could that have for ALD in your view? At SEMICON West, we have also heard there were implications for EPI as well, probably in that area. Could you maybe have a little color on that as well? Those were the questions.
Can you repeat the second question one more time?
The second question was on 2017, with the pushouts that we've seen in DRAM for this year. Is it fair to assume that in 2017, you will be able to grow faster than the market?
Okay. All right. On logic foundry revenue distribution, again, we expect overall the second half to be stronger than the first half. We also expect, we also shared in the call that logic foundry, we expect second half to be stronger than the first half. That's an important driver for, let's say, the overall sales of the company to become stronger. I can say about the different segments. In our current forecast, we expect that to be the case for both individual segments also. For logic and foundry, for each of them also. That's on that question. You said on the market development, I don't want to go too much into whether we can. Of course, our ambition is to outgrow the market.
If the single wafer ALD market is going to grow in a significant way, and it's going to significantly outgrow the WFE market in 2017, there's a great opportunity for us to do so also. I don't want to run ahead and give financial guidance for 2017. We.
There's a fair chance on that.
We are aiming to do so. Again, if the single wafer ALD market will significantly outgrow the overall WFE market, we have a great chance to do so. On gate-all-around, that's earliest going to play a role as of, let's say, five nanometer. Again, earliest, depending on what our customers ultimately do. We do see indeed, also beyond ALD in areas like EPI, indeed, there's potential also for those product areas to grow further beyond the current market sizes. You're right on that.
All right.
Okay, you're welcome.
Our next question is from Jagvar Bajwa, Arete Research.
Hi, guys. Thanks for taking the questions. Could you just clarify how you see the catalyst for DRAM and NAND orders coming back? Is that 18 nanometer DRAM capacity being added, and is it 64 layer to 3D NAND? Hello?
Can you repeat one more time?
I just wanted to understand the drivers that will catalyze your DRAM and NAND orders. Is that 18 nanometer for DRAM market and 64 layers for 3D NAND market?
Oh, okay. Now I understand. You were not talking about end user drivers, but what nodes we have.
Orders. Yeah.
For DRAM, it could be 18. In principle, 1x node 18 is the next node, but it depends on what strategy our customers will choose. Whether they will choose 18 as the next node, or whether they, based on the delay in 16, will skip 18 and go to 16. It is definitely the 1x node. That will be the next node. Beyond 21, that has been the latest real meaningful volume node that we have been engaged in. That's for DRAM case. For NAND, it can be a mix. It can be a mix between 48 and 64.
Great. Maybe just on logic and foundry. When I look at the backlog in sales so far that you've had, is that quite concentrated among your customers, or is that quite broad adoption for 10 nanometers so far for your orders? If I look at the foundry market, roughly, we could expect that to maybe get to around 250k wafers at peak capacity. Is there any indication to kind of go by to how many tools you've shipped so far, and how much capacity maybe you are through that process node?
The 250k, for what segment were you talking about?
For foundry. Typically, we see around that size as a peak wafers node, 250k wafers a month. Can you give us any indication of how many tools you've shipped into the foundry segment so far? For 10 nanometer, that is.
Yeah, we're not used to giving the amount of tools, but our perception is that we are far away from the foundry segment reaching 250,000 wafer starts a month, as far as our intelligence can tell, for that specific node.
Right.
Yeah. That provides you more color. That's maybe helpful for you. Logic foundry, the amount of players. Well, it's clear that only a few players basically define the strong demand in that combined segment. It's only a handful-
Okay
that define that demand.
Can I just maybe sneak in one more? It's the 10 nanometer. When you look across those players, the small number of players, generally, are you finding a consistent level of content for ALD across those customers, or do you find that actually some of them are adopting ALD in more quantity than others? Thanks.
For foundry, your question was related to foundry?
For logic and foundry. We're seeing quite a few differences now between the process node at 10 nanometer, I guess, and the way that these designs are being done. Are you finding the amount of ALD used is quite consistent in terms of the number of layers? Or is there a kind of diverse amount?
Yeah, there's difference between the customers. There are different ways of implementation. Of course, there are similarities, but there are also clear differences. Also clear differences in the amount of layers.
Okay, thank you.
Yeah, you're welcome.
As there are no further questions, I would like to turn the call back to the speaker for any additional or closing remarks.
Okay. Well, I thank you. This was quite the dynamic quarterly release. Also, a healthy dynamic call. I trust it has provided you more color, and it has also contributed to the healthy credibility and trust in the company's performance. Please follow up on any remaining questions you may have through our investor relations channel. Looking forward to meeting quite a few of you on one of our next road shows. Thanks again for your attendance, also on behalf of Peter van Bommel and Han Westendorp. Thank you very much again.
That will now conclude today's conference call. Thank you for your participation. Ladies and gentlemen, you may now disconnect.