ASM issued its 2017 second quarter results last evening. For those of you who have not seen the press release, along with our latest investor presentation, it's 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. And with that, I will now turn the call over to Czech Dalprada, President and CEO.
Okay. Welcome everybody to the call today. Let's now review our second quarter conference call. Second quarter 2017 financial results. Sorry, I was still structuring my pages. Net sales in the second quarter amounted to EUR 202 million, a strong increase of 40% compared to the first quarter, and up 46% compared to the second quarter of 2016. Sales in the quarter were slightly above the top end of our guidance, which was a range of EUR 180 million-EUR 200 million. In terms of product lines, although all product lines showed higher sales, our ALD business was again the key driver. By industry segment, revenue in the quarter was led by memory, followed by foundry. Within memory, sales were predominantly related to 3D NAND.
We recorded a gross margin of 43.7% in the second quarter, relatively steady compared to the 43.5% in the first quarter and 43.8% in the second quarter last year. Note that we expect the gross margin to show more fluctuation during 2017 compared to what we have seen in 2016. Still is in the range of a low to mid-40s percentage as earlier indicated. This is caused by new product introductions and related initial costs in this year. SG&A expenses increased by 9% compared to the first quarter and were up 20% year-on-year. R&D expenses increased by 2% compared to the first quarter and by 14% year-on-year.
We generated operating income of EUR 38 million in the quarter with an operating margin of 19%, up from 10% in the first quarter and 12% in the second quarter of last year. Financing result in the second quarter was EUR 11 million negative, EUR 11 million negative, due to a currency translation loss of a similar size. In the first quarter, financing results included a translation loss of EUR 7 million and in the second quarter last year, a translation gain of EUR 8 million. Our net earnings increased from EUR 36 million to EUR 130 million in the second quarter and included a profit of EUR 84 million related to the sale of a 5% stake in ASMPT. Excluding the profit on the 5% sale and the PPA amortization, normalized net earnings amounted to EUR 56 million in the second quarter, up from EUR 42 million in both the first quarter and the year ago period.
Let's now take a closer look at ASMPT. In the first place, during the quarter, we reduced our stake in ASMPT from 39% to 34%. On April 24, we announced that we sold 20 million shares or a stake of approximately 5% through an accelerated bookbuild with gross proceeds of approximately EUR 245 million. Following the 13% stake sale in 2013, we have continued to regularly review our shareholding in ASMPT with a focus on long-term value creation. The most recent strategic review led to the conclusion last April that a further reduction of our shareholding of approximately 5% was justified. At the same time, our view remains unchanged that a significant stake in ASMPT is of strategic value at this stage of the company. In terms of operating results, ASMPT delivered another strong performance in the second quarter.
The first quarter result related to ASMPT of EUR 34 million included a one-off gain of EUR 10 million. Excluding this one-off, results from investments increased from EUR 24 million in Q1 to EUR 31 million in Q2, despite the dilution of our stake during the quarter. Results from investment excluding the ongoing PPA amortization charge, which was EUR 7 million in the second quarter and is expected to amount EUR 24 million for the full year 2017, assuming constant currencies. On a 100% basis and excluding one-offs, ASMPT's net earnings increased strongly by 38% to EUR 88 million in the second quarter. The second quarter, ASMPT's sales rose to $569 million, up 17% from the first quarter, and up 20% from the second quarter of last year. Sales in back-end equipment increased 22% year-on-year, while SMT Solutions increased by 19%.
ASMPT reported bookings of $661 million for the second quarter, an increase of 9% compared to the first quarter, and exceeding the forecast for stable quarter-on-quarter development. ASMPT's gross margin improved to 41% in the second quarter, with gross profit for the first half at a new record high. Let's now turn back to ASMI's consolidated operations. Our order intake in the second quarter remained at a solid level of EUR 206 million, up slightly from EUR 204 million in the first quarter, and up 29% from the level a year ago. Orders came in at the high end of our earlier guidance of between EUR 190 and EUR 210 million. The backlog remained at a solid level of EUR 210 million, which is 4% lower than the level at the end of March due to currency movements, but up 20% compared to June of last year.
Orders in the second quarter were primarily driven by our ALD business, but also included a decent and increased contribution from EPI. Moreover, we also received several orders for our PCD. Looking at the breakdown in bookings by industry segment, foundry was the largest segment, followed by memory. The foundry orders increased versus Q1 and were booked with multiple customers. This segment was driven by the continued ramp of 10 nanometer as well as early seven nanometer-related demand. Next to ALD, foundry orders also included multiple EPI tool orders. Within memory, bookings for 3D NAND again accounted for the largest part, while DRAM bookings remained at relatively low levels. Let's now look at our balance sheet and cash flow. At the end of June, the cash position stood at EUR 523 million, up from EUR 379 million at the end of March.
This increase was the balance of, on one end, the proceeds of the partial ASMPT stake and dividend received from ASMPT of in total EUR 264 million. On the other hand, a negative operating free cash flow of EUR 41 million during the quarter, and cash spent on share repurchases, and a paid dividend of in total EUR 82 million. Net working capital increased from EUR 139 million at the end of March to EUR 176 million at the end of the second quarter. The number of days of working capital measured against quarterly sales decreased from 86 at the end of the first quarter to 78 at the end of June. Working capital on a currency comparable basis increased with approximately EUR 50 million, which is mainly caused by a EUR 54 million accounts receivable increase due to the back-end loaded nature of sales in the second quarter.
Partly offset by improved profitability, the increase in working capital led to a negative free cash flow of EUR 41 million during the second quarter, compared to EUR 20 million positive during the first quarter. We paid EUR 42 million in dividends during the quarter. In the second quarter, we also spent EUR 40 million to repurchase close to 800,000 of our own shares. This is part of the buyback program that we announced October last year, and that we increased to EUR 100 million last March. As of last week, we completed approximately 85% of that buyback program. As already earlier indicated, we intend to use the proceeds of the partial sale of the ASMPT stake of approximately EUR 245 million for a new share buyback program. This program will start as soon as the current program has been completed. Our commitment to use excess cash for the benefit of shareholders remains unchanged.
We earlier announced the cancellation of 1.5 million of our treasury shares, which has been approved by the annual shareholder meeting last May, and will become effective by the end of this month. Let's now look more closely at the ALD market dynamics. Moore's Law remains a key driver for our customers as they are developing and introducing differentiated semiconductor devices at lower costs that are crucial enablers of new end market applications such as artificial intelligence, automotive, and IoT. As our customers transition to smaller geometries, new materials and complex device architectures, we expect that more and more deposition steps will require the precision and control, such as film conformality at low temperatures, offered by single-wafer ALD. At our recent analyst and investor technology seminar during the SEMICON West Conference, we highlighted several new ALD applications.
One new application in 3D-NAND is an ALD silicon oxide sidewall protection layer, which is deposited prior to the deposition of the metal source contact material. Another application in 3D-NAND is an ALD silicon oxide layer used to fill the select gate slit. Both applications are in high-volume manufacturing today. We remain confident that the ALD market will grow to reach a size of approximately $1.5 billion by 2021, as the number of ALD applications continues to increase. Based on our customer relationships that have strengthened substantially in recent years and our broad portfolio of ALD solutions and our continued focus on innovation, we believe ASM is in a good position to benefit from these long-term trends. In terms of industry segments of the ALD market, the outlook for the logic foundry segment continues to be healthy.
The transition to 10 nanometer already fueled a significant increase in the number of ALD applications. And with seven nanometer, we foresee a further increase in our addressable market. We expect logic foundry to remain the largest segment within the single-wafer ALD market this year. Looking at 3D NAND, the last few quarters marked the first meaningful deployment of single-wafer ALD for multiple applications in high-volume manufacturing. After we recorded in the first quarter, the first double-digit % sales contribution from 3D NAND, we have seen a further increase in the second quarter in 3D NAND sales, both in absolute terms and as a % of our total sales. The strength in 3D NAND that we have seen in recent periods is driven by the continued HVM ramp of the customers that we won last year.
We continue to project a strong year-on-year increase in the 3D NAND single-wafer ALD market for 2017. Longer term, we also expect 3D NAND to be a strong driver behind the overall single-wafer ALD market. With the transition to 96 layers and above, we expect ALD will be required for additional process steps to address the increasing complexity and high aspect ratios of these new generation devices. In DRAM, while overall investments in this segment are recovering, demand for new ALD tools has not yet picked up in a meaningful way. DRAM customers are still mostly focused on technology transition within existing fabs. Looking at ALD tools for multiple patterning specifically, we noticed that customers are still largely addressing their needs by reusing existing tools, which negatively affects demand for new tools.
For the full year 2017, we do not expect multiple patterning in DRAM to be a meaningful driver of the single-wafer ALD market. Next to ALD, we expect our epitaxy product line to have a meaningful contribution to ASM's overall top line this year. Earlier this month, also in connection with SEMICON West, we officially launched our new Intrepid ES tool. The benefit of this tool include, amongst others, substantially improved temperature control capabilities and a more efficient reactor clean cycle. As a consequence, the Intrepid ES offers both enhanced performance and cost of ownership reductions. With this tool, we target advanced applications in logic foundry and also in the memory market. With the Intrepid ES, we believe we have substantially increased our addressable market in epitaxy.
With our Q1 results last April, we announced that we won a first process tool of record selection for the Intrepid ES for an advanced epi application at the leading foundry. And that we booked the first high-volume manufacturing order intake. In the second quarter, we booked multiple orders as part of this selection. Let's now look at the broader market environment. In their most recent updates, market researchers like Gartner and VLSI Research increased their forecast for wafer fab equipment in 2017 to approximately 17%-18% positive year-on-year, compared to 8%-9% positive earlier this year. The reasons for the strengthening of the forecast are continued strong spending for logic foundry and 3D NAND, and an increased spending for DRAM following the low investment level in DRAM in 2016. Let's now look at our guidance. The company guidance as included in our overnight press release.
For Q3, we expect the sales level of between EUR 170 million and EUR 190 million on a currency comparable level, while for the second half of 2017 as a whole, we expect the sales level higher than in the first half of this year. After the very strong quarters in the past two quarters, we expect order intake in Q3 at a still healthy level of EUR 150 million-EUR 170 million, also on a currency comparable level. At this point, we are more than happy, Peter van Bommel, our CFO, and myself, to take any questions that you may have.
We'd 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, Lisa, we are ready for the first question.
Thank you. If you'd like to ask a question at this time, please press star one on your telephone keypad and please ensure the mute function on your telephone is switched off to allow your signal to reach our equipment. We will now take the first question from Peter Olofsen from Kepler Cheuvreux. Please go ahead.
Good afternoon, gentlemen. Just wanted to clarify, the comment on EPI, where you talk about several orders. These are several orders, but still from the same customer that you highlighted in Q1. That means we have yet to see a broadening of the customer base. Is that basically what you're saying?
That's correct.
Okay.
Short answer, Peter, but that's correct.
Okay. Then in terms of broadening of the customer base, apart from Logic Foundry, I think you're also mentioning opportunities in memory. What kind of horizon should we think of in terms of EPI gaining traction in memory? Is that something we might already see later this year? Or should we more think of 2018, 2019?
Yeah. Okay. To just repeat the bigger picture. It was a great opportunity for us, first of all, to remind that for the audience, refresh that for the audience, that this penetration really opened up for us, the EPI market beyond, let's say, the EUR 100 million market that we had been so focused on so far, which was the analog power part of the market. The total addressable market in EPI is around USD 600 million and is expected to grow in the coming years. A healthy CAGR is forecasted for the EPI market in the coming years. We invested for the last couple of years in several areas in the EPI space, of which this Foundry area was just one. We really indeed have now a first victory in that area with multiple orders.
We continue the seeding in other areas, which includes, indeed, the memory space. It's just too early to claim any victory. We have been working in those areas for a significant amount of time. So basically, for the last two years, we have been investing in that area, internally and in cooperation with customers. It's just too early to claim victory. We target to see results there, let's say, within the next 12 months. That's what we are targeting.
Okay. In memory, you're addressing both DRAM and 3D-NAND?
Yes. One area is maybe. We are targeting both areas, and certain applications are more, let's say, shorter term than other applications. We focus on the whole memory space. Yes.
Okay. My second question, I think for Peter, relates to the CapEx, which was a bit higher in Q2 than what we have seen in previous quarters. Is that just some quarterly volatility in there, or is it kind of indication that we are going to see higher levels there going forward? Related to that, can you confirm that given your outsourcing model, you are able to grow volumes in a meaningful way without major CapEx need?
Yeah. A few remarks there. Indeed, there is never a straight CapEx amount per quarter. We are improving and increasing our R&D facilities in the forthcoming periods. Hence, it's likely that our R&D of our CapEx will be a little bit higher than what we have seen in the past, in the next three to four quarters. All in all, in general, that remains the same. It's more related to the R&D area than to manufacturing. Given our outsource model, we think that we can deliver the requirements that the customers are asking for.
Okay, thank you.
You're welcome.
Thank you. We will now take the next question from Robert Sanders from Deutsche Bank. Please go ahead.
Yeah. Hi, good afternoon. Maybe a first question on EUV. Given that ASML is now saying that EUV will be used up for up to 14 layers, what does that do to your self-aligned opportunity, given that patterning is a big part of ALD? I have a couple of follow-ups. Thanks.
Okay. Yeah. On EUV, in general, we expect, let's say, deposition-related patterning to remain a solid part of our business and of the single-wafer ALD market in the coming years. Even with the introduction of EUV at seven and five nanometer, patterning complexity overall continues to increase and is our belief, besides the opportunities for, let's say, the EUV litho vendor, opportunities for the deposition players and a search for the single-wafer ALD market. Secondly, EUV patterning also requires more complex film stacks in order to overcome certain limitations or challenges that EUV introduction brings. That may also open up additional opportunities in the ALD space and a search for us. That's how we view it in general. Besides that, Rob, we would like also to emphasize that, patterning, as we said, stays a solid contributor to the market.
Versus 2015, the percentage contribution to the single-wafer ALD market has gone down in 2016 and is believed to further decline in 2017. In that way, let's say, the dependency on that part of the market has gone down, compared to, let's say, two years ago.
Got it. Just on the largest 3D NAND player. You've obviously had success with two of the largest, the top five. One of the big hopes was that you would get into the largest 3D NAND player for next year. It does look like they've now allocated vendors for that. They've done the plan of record. Certainly, your peers are talking about having already been selected. What are you thinking there about your penetration into that large account into 2018? Especially given that TEL is saying that their own ALD revenue is growing, actually doubling year-over-year, with their semi batch tool. Thanks.
Yeah. We're not allowed, of course, to talk about very specific customers. In general, we can say that we still target and count on expanding our presence in terms of number of customers in 3D NAND as the industry goes to, let's say, 96 stack. That schedule and those objectives have not changed.
Got it.
We think that the seeding efforts on that are ongoing in several areas, and we think that real HVM materialization will come in the course of 2018. Maybe some initial visibility will come towards the end of this year, but real material impact will be in the course of 2018 in our P&L.
Got it. Just lastly, on the Q4 guidance, the implied Q4 guidance is quite vague in the sense that you talk about growth half on half. Frankly, that could mean a wide range of numbers for the fourth quarter. It could be 170, 180, 190, 200. Can you just be a bit more specific about what kind of growth you're expecting half on half in the second half of 2017? Thanks.
Yeah. Of course, our formal guidance does not include Q4, because we normally only look one quarter out. As you know, there's no straight line in this industry when it comes to order intake on a quarterly basis. We just had two, three quarters of very strong bookings. Apart from the fact that we cannot comment specifically on the Q4 bookings level, we can say something about, let's say, the top line for the remainder of the year. As we included in our guidance, we expect the second half sales level to be higher than the first half. If there would only be a minimal difference of EUR one or two million, then we would not have included it in the guidance. That's one.
Secondly, we do expect to outgrow the WFE market based on, let's say, looking at the current forecast for the WFE market in 2017. If we look at our top-line projections for this year.
Got it. Thank you.
Okay. You're welcome.
Thank you. We will now take the next question from Jim Fontanelli from Arete. Please go ahead.
Thank you. Good afternoon. You kind of alluded to part of my first question on your last answer, which is just how you expect ASMI's top line to look relative to wider WFE spend over the next couple of years. You were talking about outgrowing it this year, but I guess, even assuming a EUR 190 million or EUR 200 million number for the fourth quarter, you'll be outgrowing it maybe by a couple of points this year. It would be interesting to hear your perspective on the degree of outperformance you might expect from ASMI or ALD versus wide WFE into 2018 and 2019. That would be my first question.
Yeah, Jim. That's a clear question, understandable question. 2016, of course, we underperformed the market. In 2017, you clearly see, first of all, that the single-wafer ALD market is expected to show a clear improvement after decline in 2016, which offered as, that's one dimension along which we can show a better performance this year. If you look at longer term, I think over a longer period of time, because I think your question is related to that, on average, we are of the opinion that in the coming years, the single-wafer ALD CAGR is expected to outgrow the WFE market. On average, let's say over a five-year period of time, that the potential of the single-wafer ALD market is really there, as we also alluded to further in our introduction. That's one area along which the company has potential to grow.
Second pillar is that we're also working hard and have been investing in technology, in customer relationships over the last few years to grow our served available market in the total addressable market for ALD in the coming years. That's another pillar along which we view that we have potential to grow. Thirdly, which first evidence is there, maybe on a somewhat smaller scale, but already in this year, is that we intend to grow in our other products like EPI and PECVD. By expanding also our served available market. Basically through focusing more on those products, we further expand our served available market in deposition. Not only in ALD, but also outside ALD in other areas where we so far have had more of a niche strategy.
That will go gradually, but every gradually increase immediately could have a meaningful impact to the top line of this company.
Okay, that's clear. I guess your point around the second pillar, which was increasing the TAM within ALD, that would imply some sort of market share relationship, right? You're effectively implying the potential to take back market share by addressing an increasing number of applications within ALD. Is that the sort of corollary of what you were talking about with your second pillar?
Yes. Of course, if you look at the increased share of wallet going for in Logic Foundry, for example, going from 1614 to 10. Basically, we have increased our served available market in Logic Foundry there, basically. We are aiming to further increase our served available market also beyond, let's say, the current nodes, not only in Logic Foundry, but also in the memory space. For example, what we are developing now in 3D NAND is basically increasing our served available market on ALD. The same we intend also to look in DRAM, what over time we can do beyond, let's say, the patterning space that we are engaged in at this moment in time. Will that have an impact immediately, short term on our P&L? No, but it will gradually become visible in our P&L. That's what we are working towards.
Okay. That's clear. Thanks. I guess secondly, maybe on a sort of shorter term thought process, but obviously you talked about the strategic review of the ASMPT stake and incremental 5% stake sale. Given PT's stock move overnight and obviously the reflection through to the ASMI share price this morning, how you view the ongoing PT stake, given that you're clearly showing improving operational performance within ASMI. You've talked about the opportunities lying in front of you, not just for ALD, but for the rest of the business. Yet, the day after you report numbers, your stock price is dominated by what's happening with PT in the back-end. How are you thinking about that, sort of, the strategic review option, against that backdrop?
Well, it would not be good if our strategic reviews would be led by market stock price movements in a 24-hour time space. That's not how we run our company. If you look at a little bit more detail to the PT results, then we think PT again delivered a strong quarter. If you look at the bookings after the strong Q1 order intake, they improved with another 9% to EUR 661 million in bookings against EUR 561 million in billings. That means a book-to-bill of 1.2 at the current billing levels, with expectations for further growth in Q3. Just to take one other indicator, gross margins, they are now at 41% level, while adjusted net income improved to 17% of sales.
If you look through the emotions of the day, and look at the numbers, the results of PT in a little bit more detail, then maybe, people will over time come to a different opinion. That's at least our view. Peter, you want to add something to that?
I think when you look to the PT performance, you all can actually read that, but also their bookings increased in the first half of this year as compared to the second half of last year with 14%. When you look to the billings, which Chuck already mentioned, we saw also an improvement quarter-on-quarter of 17%. Gradually, their gross margin is improving, and that's a function of all the things that have happened in the past period. The underlying quality of the results are improving further and further, while they have a strong position in most of the segments that they are playing. I think that is the long-term view that should be substantiated also, in our view about the performance of PT.
Great. Thank you. That's very clear. Maybe just very quickly, one clarification. Chuck, did you say that you were expecting to ship ALD units into the 96-layer HVM ramp next year? Did I hear that right?
Yeah. Just before I answer that, Jim, I just wanted to clarify that the way I answered your earlier question on ALD, I tried to provide a little bit more color to the audience on the, let's say the overall growth potential of the company. I was not talking specifically to ALD market share. As long as you really understood my answer well in that context. Secondly, the 96 stack, indeed, we are targeting to participate, in the 96 stack of beyond the current customer base that we have HVM business with at this moment in time. For 3D-NAND, let's say in the course of next year, probably by mid-next year or so. It depends a little bit on the timing of the customers, what timing they use to ramp their fab for that.
Okay. You'll start shipping units in at some point in the second half of next year to 96?
Again, that's what we are focused on. Again, they have not made their final decisions yet, to bring R&D partners to really freeze their processes for the areas that we are engaged on. That could happen in, let's say this, we do expect that they will start to do that in the second half of this year. We have no final yes yet, but if you look at the level of engagement, then we trust that we're on the right track.
Great. Thank you.
Yeah. You're welcome.
Thank you. We will now take the next question from Marc Hesselink from ABN AMRO. Please go ahead.
Yes, thank you. First question is on the order guidance for next quarter. Slightly at the lower level. Obviously, two very strong quarters in the first half. What is that slightly lower level? What's driving that? Is that simply digestion of the orders that we've seen in the first half of the year? Or is it also that in Logic Foundry, you've passed a little bit the peak of this cycle?
You're talking about bookings in general?
Yes.
Yeah. I think the way you really should view is that in this industry, and you know this industry well, again, like I said earlier, there's no straight line in terms of order intake every quarter. We really had around EUR 200 million quarters, Q1, Q2, also Q4 was a strong quarter in terms of bookings. Q3 is not that good, but not as good as Q1, Q2. We think it's still at a healthy level. Beyond Q3 and looking to, let's say, 2018 also, we believe that the company is well-positioned, that the underlying trends, as far as customers provide us visibility, is healthy, in the 3D NAND space, in the foundry space, separate from maybe some timing differences, one or two months pulling in or pushing a few things out. 3D NAND, foundry, healthy.
Logic spending pattern is maybe so far this year, not as high as maybe a lot of players expected at the beginning of the year. Still to be seen how that will develop towards the end of the year, going into next year. I think there only could be, likely there's more upside there than downside in the logic space as we view it now. DRAM. There can, to us, only be upside because the contribution there has been relatively low so far as we also elaborated on in the introduction.
Okay, that's clear.
You should not read too much in the number in Q3. We don't view it to be as a long-term trend. That's not the case at all at this moment.
Okay, that's clear. Thanks. The second one is on DRAM. As I can remember correctly, you said always that they're reusing some machines, and that's the reason why you don't see that there's no capacity being added there from your perspective. How long can that still continue? Is there a lot of room for them to reuse machines? When do you think that will end and that you see some strength coming into that DRAM space?
That's a great question. We would love it, we'd know the answer. Like we said, overall investments in DRAM have been recovering to some extent, across the board. If you look at specifically new ALD deposition tools, for patterning, then the demand is still modest because of the fact that most of the demand occurs in existing fabs. We're dealing with big memory customers that then find a way to reuse equipment from the DRAM or the NAND area. That has been providing them most of their additional capacity so far. As a result of that, for the remainder of the year, we don't expect DRAM to be a meaningful driver for the single-wafer ALD market. 2018, we don't know. It could change.
As I said earlier, based on Jim's question, we are ready for the patterning part of the DRAM market when it recovers. We're ready to engage with competitive products, in our view. Besides that, we're working hard also to look for ways to expand our served available market in the medium and long term in DRAM ALD, beyond patterning, and also outside ALD. That's for somewhat medium and longer term.
Okay, thank you.
Marc, does that answer your question?
Yes, it does. Thanks. Just one short one is on the OpEx. The SG&A or the cost there was burning a little bit higher than what consensus expected. Therefore a little bit less leverage on your additional revenues. Can you talk about what's driving that? Is that preparation for the future or?
Let's put it a little bit in perspective. Due to the higher sales, we had EUR 26 million more gross margin. Of that EUR 26 million, we used approximately less than EUR 3 million of OpEx. That's about the leverage. When you look to the cost increase in SG&A, approximately half of that amount is related to increases which are due to the fact that we have a higher activity level. Some costs are related to that activity level, and the other half is the further strengthening of the organization that we do in particular areas, especially to cope with the opportunities that we see and where Chuck already was referring to.
Okay, thanks.
Yeah. You're welcome.
Thank you. We will now take the next question from Tammy Qiu from Berenberg. Please go ahead.
Hi, guys. Thanks for taking my question. The first one is on the progress of 10 nanometer and 7 nanometer ramp. I remember you said that you have already seen early 7 nanometer demand. I'm just wondering, is 10 nanometer build already done from your perspective, like from equipment perspective? Is 7 nanometer still at early stage, there is more to come? If there is more to come, what is the rough timeline?
Yeah, Tammy. We see orders so far in 2017 for both nodes, 10 and 7. In the course of the year, of course, the level of 7-nanometer business increases. Whether the 10 nanometer really has stopped, I think that's too early to conclude. That's not what our customers have told us. If you look at the ratio 10, 7, then the ratio will shift more towards 7, towards the end of the year and going into next year. There's more to come definitely on 7 going into 2018. We're now carefully listening and waiting for guidance, let's say from the leading foundry customers on 7, on how soon we have to ramp for them. Let's say we don't have visibility going into 2018 yet on that, Tammy.
Okay, thanks.
No problem.
Really 7 nanometer is a 2018 story, and 10 nanometer has to support the order level from foundries for this year.
Well, again, we have already businesses in this year for 7. Their initial pilot capacity is being built already this year. What we do see is that there is pilot capacity invested in new applications. Also applications beyond, let's say the current 10 nanometer engagement. How quickly they will ramp from, let's say, pilot volume levels to really full flash volumes, that's still to be seen. That's still to be seen from our customer and probably also their supply chain.
Okay, thank you. The next question is about the demand level in foundry logic for 5 nanometer, i.e., beyond 7 nanometer. When you can get confirmation and say, okay, you are selected for certain application. Just to try to get a view about what are you doing and what is the progress for you to protect your market share in this market, or maybe even gain back market share?
Well, for 5 nanometers, that's ongoing now. I think in the next, let's say, 12 months already, a lot of directions are being defined by customers on 5 nanometer. We're now already talking about 5 nanometer engagements in having seeding systems at the customer to work on 5 nanometer. That's ongoing as we speak. That will likely only further develop in, let's say the coming quarters. It's ongoing now already.
Are you doing anything to protect your market share to make sure that you're not losing market share any further or maybe gain back market share?
Well, I think it's just we're focused on expanding our engagement. Of course, working on 5 is also on protecting your existing engagements. It's both. We're working on both areas. To expand into more applications, and at the same time make sure that we can provide continuity. Continuity can be from a cost point of view, but it can also be from a technology point of view. It depends a little bit on the application.
Okay, I see. Last question. On EPI tools, do you have a forecast on the total addressable market for EPI similar to the one you have on ALD?
Are you asking for the size?
For the size, yeah.
Yeah.
What is the future addressable market for you?
Well, basically, we said that we were engaged in around EUR 100 million addressable market. Now we are identifying opportunities in the broader, let's say EUR 600 million market. Are we immediately engaged in all the applications that make up that EUR 600 million? Of course not. We are developing and strengthening our platform and our reactors in such a way that we are a viable alternative in many more areas of that EUR 600 million market.
That is a market size today, or that's your forecast?
No, that's basically how Gartner estimates the market today. The CAGR for EPI for a number of years, it's depending on what firm you work with, but it's a small double-digit CAGR percentage that is forecasted for the EPI market. It's a little more than 10%. That, of course, changes. That's on average over a five-year period of time. It's definitely a growth area. It's perceived that EPI's applications are going to grow in both logic foundry and in the memory space in the coming years.
Okay. That's very helpful. Thank you, guys.
You're welcome.
Thank you. We will now take the next question from Johannes Ries from Apus Capital. Please go ahead.
Yes, good afternoon. Only one last question, a follow-up to the latest we had on epitaxy and PECVD. You mentioned the market size, you mentioned that you gradually can grow in this space to different application, and that could be meaningful for the company. Maybe in two years ahead or so could be the growth rate, because you mentioned even that the EPI market is growing around 10%, and you are maybe winning shares now and moving in new market segments. Could this area be growing in the same way as the ALD going forward, and how is the margin profile compared to ALD?
Yeah. Johannes, thank you for your question. That's just too early to tell. We now have the initial gains, but we have an established, strong competitor that will also fight for their business. We schedule to plan to gradually build a position there. We have a healthy ambition, and we also view, like how we introduced the tool at SEMICON West, the Intrepid ES tool. We presented there that we view the tool has some intrinsic capabilities that make it pretty competitive for a number of applications. We're going to build on that. It's to us to convince our customers that we are right with our view. Yeah, this market is now basically a one-vendor market, and I think it would be a win-win situation for the industry if we would be able to develop our position.
But again, it's for us to prove, and it's too early to share any projections. We just first want to do our homework, and then, step by step, we will share our results with the financial markets.
Yeah. Johannes, your question about the gross margin, is that in line with the other activities? There, the answer is, in time, we do not expect that the gross margins as well on the Intrepid, on PECVD, will not be different from the average that we see in our organization. In the initial phase, you can imagine that we have more cost. When you introduce a complete new product-
Yeah.
You have to go back to your supply chain, so they have initially higher cost. The customer comes back to you and wants to have some additional things, and especially since they are all new products for us, the initial phase, there, the margin is a little bit under pressure. That's one of the reasons why we have given that guidance also, that we expect a little bit more volatility, especially in the 2017 area where we are bringing those first products into the market.
Okay. Clear answer. Thanks a lot.
Thank you, Johannes.
Thank you. We will now take the next question from Charles from Natixis. Please go ahead.
Yes, good morning. I had a question on CapEx. I think you mentioned this year it will be probably higher, but I wonder, in terms of CapEx to sales, what are your expectations in the midterm basis? Thank you.
In the midterm basis, we expect that the CapEx will be slightly higher, so than what we have seen in the past. On the longer term, we expect that it will be back again on the numbers that we have initially mentioned, that's EUR 20 million-EUR 30 million on an annual basis. Again, it's a little bit dependent on the applications that we are bringing into the market, new applications that we're bringing into the market, because that might lead to some additional own equipment that we have to put into our lab environment.
Okay, understood. Thank you.
You're welcome.
Thank you. We will now take the next question from Edwin de Jong from NIBC. Please go ahead.
Good afternoon, gentlemen. A few questions left. Maybe I've missed it, but on the growth margins, for you, Peter, maybe, going into Q3, Q4, can you give a little bit more color on the moving parts there?
The biggest moving part is in principle, the mix. What we have indicated is that what I just answered on the question of Johannes is that in principle, when we have complete new products, when you have ALD products, there we have already a lot of experience with. There the impact initially on when you introduce new products on the gross margin is limited of scale. On the moment that you introduce complete new products like the Intrepid that we mentioned, like the inroads that we are making on this moment in PECVD.
Those products will have initially a lower margin because the supply chain is not used to making this sort of products. We are in aligning the details of the products with our customer base. As a consequence of that, you have initially higher cost. We expect that that is in a period of two to three quarters before such a product is going back into a more normal gross margin level.
The level of gross margins that we've seen in Q2, that's maybe also a good guidance to see for Q3 and Q4.
What we have indicated is that we expect somewhat more volatility, and the volatility is in line with the margin indications that we always have given for low to mid-40s. Again, it depends very strongly on how much of those complete new products that you are delivering in a certain quarter.
Okay. I can imagine. On PECVD, we were at SEMICON West and it sounded quite exciting, the opportunity that you have there also in 3D-NAND. Do you maybe have a little bit of feeling for how big that market could be in PECVD?
Well, PECVD is a very big market.
Yeah
In terms of size, there are quite a few established players there.
Different from the EPI market. There are a few U.S. companies, there are a few Korean companies. We have to carefully choose what customers and what applications, we would like to engage on. We think, our view is that there is room to improve, it will go step by step.
Not in the way it went with EPI at the end.
Could you repeat that one more time?
Not at the space that it went with EPI. The EPI, gained a lot of traction all of a sudden in a few quarters' time, and that's not the path that you would see for PECVD or?
No. Well, I think, again, we are just focusing on a number of new applications with specific customers and what you have seen now already, that we have an engagement in PECVD in 3D NAND. Yeah?
Yeah.
Which we did not have, let's say, last year. It's one application, one customer. We grow from there. The only thing I want to say is that the PECVD market is much bigger than the EPI market even.
Also the number of players there is also bigger. In both areas, we intend to grow gradually in a smart, focused way. It will not be a hockey stick effect. The good thing is that every opportunity we have to grow there, the impact to our top line, given the size of this company, could be meaningful.
Yeah. That's clear. Yeah.
All right.
Great. Thank you.
You're welcome.
Thank you. We will now take the next question from David O'Connor from Exane. Please go ahead.
Great. Good afternoon, gentlemen. Thanks for taking my question. Two from my side, if I may. Firstly, maybe on the strategic review, which prompted the ASM Pacific stake sale. Chuck, can you give me some idea of how often you intend to redo that exercise? Is this more a once every three-year type exercise or something a bit more dynamic? I have a follow-up on China. Thanks.
Well, the first one we're not going to do certainly based on today's stock price movement. That's the first remark. We just do this in a structural way. We review our strategic portfolio and our strategic direction of the company multiple times a year, among management. Also regularly, we review it with our board, and that's what we will continue to do. Again, we did that of course, on our way to. We did it earlier in the year. We again did it on our way to the AGM, because of course, that subject was on the agenda of the AGM. It's not that as a result of today, that we immediately would have to revision. Our exercise of, let's say, one or two months ago.
All those elements are, of course, always taken into account in the regular strategic reviews that we will do in the future. That's it. The strategic importance, the strategic value of PT to ASMI, as we expressed at the AGM, that position today has not changed.
Okay. Thank you for that. As the follow-up on China, can you talk a bit about the opportunity in China among particularly the new domestic memory players? How are you developing the relationship with them? When do you expect some orders? Are you seeding tools? Any color around that opportunity with the domestic memory players in China, and timing will be helpful. Thanks.
Yeah. A few short answers. The revenue in China this year, although, let's say still at a low base compared to, let's say the other regions in the world, is significantly higher than last year. We expect next year it to significantly increase compared to this year. The number of engagements, the number of, let's say, interactions with customers, grow every month. If we look at, let's say, how our activity level in China has developed over the last six months. As a result of that, or basically also ahead of that, because it's a little bit chicken and the egg. On the one hand, we decided to strengthen our infrastructure in China, significantly. That has resulted in more contacts, but also the intensity level has triggered us also to do that even more actively.
Very helpful. Thank you.
In general, the activity level in China is increasing, but again, from a low base, because historically in China, a lot of investments were done in older technology nodes, which did not fit well with the majority of our product portfolio. More and more, we are aligning our portfolio with the needs of the China market. That also, of course, what helps is that in time, they need more advanced technology.
We still have people in the queue, but we are running out of time, so we have to conclude the conference call. For any remaining questions, please contact us after the earnings call.
Okay, Victor, that means that we now conclude the call?
Yes.
Okay.
Operator?
Okay. Before the operator concludes the call, sorry if any of your individual questions have not been answered by now, also on behalf of Peter. Of course, feel free to contact us after this call with any remaining questions you may have through Victor. Remains me to say thank you very much for your interest today, and let's stay in touch.