Good day, welcome to the ASM International Q3 2017 earnings call. Today's conference is being recorded. At this time, I would like to turn the conference over to Victor Bareño, Director of Investor Relations. Please go ahead, sir.
Thank you, Sergey. ASMI issued its 2017 third 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. For more information on risk factors related to such forward-looking statements, please refer to the company's press releases, reports, and financial statements, which are available on our website. With that, I will turn the call over to Chuck del Prado, President and CEO.
Thank you, Victor. Thanks to everyone for attending our third quarter 2017 results conference call and for your continuing interest in the company. After a review of operations, Peter van Bommel, our CFO, will join me, of course, in answering any question that you may have. Let's start with a review of our third quarter financial results. Performance in the third quarter was in line with our expectations. Revenue in the third quarter decreased by 8% compared to the second quarter and increased by 28% from the third quarter of last year. At a level of EUR 185 million, the third quarter revenue was towards the higher end of our forecasted range, which was between EUR 170 million and EUR 190 million. Reported sales in the third quarter were impacted by a 5% negative currency effect, both Q-on-Q and year-on-year.
The revenue stream was led by our ALD business, at some distance, followed by epi sales, which increased strongly compared to the previous quarter and included multiple tools of our new Intrepid epitaxy platform. Revenue in the third quarter also included multiple PECVD tools related to the new application win that we announced with our second quarter results. By industry segment, revenue was again led by foundry, followed by memory, then logic. Foundry remained at a strong level compared to the second quarter and included an increasing portion of 7-nanometer related investments. Within memory, sales were predominantly related to 3D NAND investments, which continued to contribute a solid double-digit percentage of our total equipment revenue. The gross margin decreased to 40% compared to 43.7% in the second quarter, and 44.2% in the third quarter of last year.
As already flagged with our second quarter results, gross margin in the third quarter was negatively impacted by new product introductions and related initial costs. The impact on the gross margin of these newly introduced products amounted to 5 percentage points or EUR 9 million. We expect that the gross margin will continue to be impacted by these initial costs in the next few quarters before normalizing again in the course of next year. Looking at the operating expenses, SG&A increased by 3% from the second quarter, while R&D decreased by 10%. The operating result amounted to EUR 26 million, down from EUR 38 million in the second quarter, and up from EUR 17 million in the third quarter of last year. The financing result in the third quarter was EUR 8 million negative due to a currency translation loss of a similar size.
Translation losses amounted to EUR 11 million in the second quarter and EUR 3 million in the third quarter last year. Our net earnings decreased to EUR 42 million in the third quarter from EUR 132 million in the second quarter. As a reminder, net earnings in Q2, of course, included a profit of EUR 84 million related to the sale of a 5% stake in ASMPT in that quarter. On a normalized basis, excluding the profit on the stake sale and the PPA amortization, our net earnings amounted to EUR 48 million in the third quarter, compared to EUR 56 million in Q2 and EUR 40 million in Q3 of last year. Let's now have a closer look at ASMPT.
Results from investments, which reflect our share of approximately 34% in the net earnings from ASMPT, increased to EUR 32 million in Q3, up from EUR 31 million in Q2, and EUR 27 million in the year-ago period. These figures are on a normalized basis and exclude the book profit on the ASMPT stake sale and the ongoing amortization charge. ASMPT had another solid quarter. Sales in Q3 increased to a quarterly record of $654 million, up 16% from the second quarter, and up 22% from the third quarter last year. SMT Solutions, the Surface Mount Technology Solutions in particular, had a strong quarter with a sales increase of 54% year-on-year. ASMPT reported bookings of $574 million for the third quarter, a decrease of 13% compared to the second quarter, but 32% higher than in the third quarter last year.
For the first nine months, ASMPT reported 24% growth in sales and a doubling in adjusted net profits. Now turning back to ASMI consolidated operations. New orders of EUR 160 million in the third quarter were within the range of between EUR 150 million and EUR 170 million that we had guided for. Orders decreased by 22% from the second quarter, but were up 31% compared to Q3 of last year. Orders were for the largest parts driven by our ALD business, but also included, again, multiple EPI and PECVD tools. Bookings were led by memory, closely followed by foundry, and then logic. Within memory, 3D NAND continued to account for the majority of the order intake, while DRAM remained at relatively low levels. Now turning to the balance sheet and cash flow. At the end of September, the cash position stood at EUR 525 million, essentially unchanged compared to the end of June.
We generated positive operating free cash flow of EUR 20 million during the quarter and received EUR 18 million in dividends from ASMPT. We spent almost EUR 33 million on share repurchases during the quarter. Net working capital decreased from EUR 176 million at the end of June to EUR 163 million as per the end of September. Outstanding days of working capital remained approximately stable at 79 days. At the end of August, we completed the EUR 100 million share buyback program that we announced in October of 2016. Under this program, we repurchased in total 2 million shares at an average price of EUR 50 per share. Furthermore, during the third quarter, the earlier announced cancellation of 1.5 million of our treasury shares became effective. On September 22nd, we announced the start of a new EUR 250 million share buyback program.
When we announced the 5% ASMPT stake sale last April of this year, we said that we would use the proceeds of approximately EUR 245 million for a new share buyback program. The share buyback authorization that was granted by the shareholder meeting in May runs until November 2018, but we strive to complete the full program well before that time. We intend to cancel the shares that we repurchase as part of this program. Far, as of October 27, we have completed 22% of this new program. Our commitment to use excess cash for the benefit of our shareholders remains unchanged. Let's now have a closer look at the trends that we see in our markets. 2017 is shaping up to be a very strong year for our industry.
Total semiconductor industry sales are forecasted to increase by almost 20% this year, which compares to an average of 2% growth in the last six years. A key driver behind this semiconductor growth is the memory segment, with strong volume and price increases in both NAND flash and DRAM. VLSI Research and Gartner now expect equipment market growth in the high 20s, up from the high teens, that was still expected last year as a forecast for this year, with most of the growth driven by memory spending. For 2018, these market watchers expect a further low- to mid-single-digit increase in WFE, in the wafer fab equipment, with, again, memory as the main growth driver. For the single-wafer ALD market, we continue to see a clear improvement year on year in 2017.
Looking at the different industry segments in more detail, we see that the logic foundry part of the single-wafer ALD market continues to be healthy this year and will probably remain stable compared to the strong level that we saw in 2016. Following the success that we had in the logic foundry segment last year, we expect this segment to remain the largest segment for us in terms of sales contribution in the full year of 2017. Next, to continued investments in 10 nanometer, we have seen the first early investments in seven nanometer this year, which are likely to further increase in next year, in 2018. We are seeing a further expansion of the single-wafer ALD 10, the total available market, in the transition from 10 to seven nanometer in logic foundry.
In the next years, this sector continues to be a key contributor to the long-term growth of the single-wafer ALD market. 3D NAND is clearly the key driver this year for the growth of the single-wafer ALD market, and also for the growth in our business. Demand in the third quarter continues to be solid. From a lower base in 2016, the 3D NAND single-wafer ALD market is experiencing strong growth in 2017 and is this year, for the first time, expected to account for a meaningful double-digit % of the single-wafer ALD market. We expect the number of single-wafer ALD applications in 3D NAND to further increase with the next generation higher stack devices.
In DRAM, while investments in the broader market have started to pick up, new tool demand for single-wafer ALD patterning continued to be weak in the quarter, in line with our previous indications. Once customers start to invest in new fab capacity for DRAM, we expect tool demand for new ALD patterning to increase. At the same time, we are very focused to, over time, expand our scope in the single-wafer ALD DRAM market beyond applications in multiple patterning. Momentum in our EPI business is increasing. As mentioned, EPI had an increased revenue contribution in the third quarter and included multiple Intrepid epitaxy systems under the first customer win that we reported earlier in the year. These tools are part of the leading-edge foundry capacity being ramped in HVM by this customer. We continue to expect a strong increase in our total EPI sales in 2017.
This first win has been strategic for us in expanding our addressable market in EPI. We remain focused on further building out our position in EPI market in the coming years. In short, in 2017, we expect to benefit from strong market conditions with a clear recovery in the ALD market, and also a strong increased contribution from our EPI business. Looking at next year, we aim to outgrow the wafer fab equipment market. The long-term outlook remains positive. We continue to expect the single-wafer ALD market to grow to approximately $1.5 billion by 2021, and we believe our company remains well-positioned to benefit from important technology inflections in the coming years. Now let's look at our Q4 guidance as we communicated in our press release.
For Q4, on a currency comparable level, we expect sales of between EUR 190 million and EUR 210 million and an order intake between EUR 170 million and EUR 190 million. This is the end of our introduction. At this point, we are more than happy to answer 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, Sergey, we are ready for the first question.
Thank you. Ladies and gentlemen, if you wish to ask a question at this time, please press star one on your telephone keypad. If you wish to cancel your request, please press star two. Please make sure the mute function on your phone is switched off before asking a question. Our first question comes from Pieter Olofsen of Kepler Cheuvreux. Please go ahead.
Good afternoon, gentlemen. Two questions related to memory. Maybe first on 3D NAND. You were somewhat late to the 3D NAND game, so there was opportunity there to broaden your customer base and make some inroads there. Could you maybe shed some light on the progress you're making there at the moment? My second question relates to DRAM. The contribution to sales and orders in Q3 was fairly limited. Do you already see that changing into Q4? Or is the demand there still held back by the reuse of installed ALD machines?
Okay, Pieter. Thank you. First of all, on 3D NAND. I think we are pretty happy with the strength of the 3D NAND segment for our company in 2017. If we look at the contribution, it has contributed to more than one product line. It has contributed to both ALD and PECVD. Both product lines have benefited from our ability to support the ramp of the key followers in this market segment in 2017. If you look at the ALD segment, then currently, if we look at the amount of applications that are out there in single-wafer ALD for the 3D NAND segment, our estimate is that we address a little bit more than 50% of the single-wafer ALD addressable market now. We are very eager to increase that share of the addressable market in 3D NAND going towards higher stacks.
Beyond that, as we shared with you before, we're very focused for the leader in this industry to address their needs in a 96 stack. For the followers, it's the nodes that are applicable to them going into next year. Again, going into next year, we expect to further expand our business in 3D NAND across multiple product lines and with more applications. As we said before, we would like to expand the number of customers, especially as if we can close on the so far, let's say, successful R&D progress that we have made in qualifying for the 96 stack in this industry.
Just to clarify, have these selections for next year already been made?
No, they have not been made. I think towards the end of this year, early Q1. In Q1, I think those decisions will be finally made or will translate into, let's say, high volume purchase orders. We make our assessment based on how well, let's say, the R&D programs are going. From that point of view, we can only say that we are on schedule.
Okay.
No final decisions have been communicated by the customer. Okay? That's not specifically to us, but I think that's more generally their timing in freezing all their processes or most of their processes, in preparation of the new node. On DRAM. Your question on DRAM, is there any improvement expected for Q4? The short answer is no. We see some business activity in the P&L this year. There is DRAM capacity expansion in there. Not, let's say, with the momentum that we are used to from the past. In that DRAM segment, our assessment has not changed compared to earlier calls that most investments, or basically all of the investments so far, from the big customers has been done in existing fabs. That's mostly done through tech migration and migrating equipment from older nodes to newer nodes in existing fabs.
A lot of reuse. That has been especially applicable to patterning-related applications. The good thing is that we do see early indications now that going into 2018, that meaningful investments by customers may be done in new fabs, in the newly, more recently built fabs. That could open up more significant investments in patterning-related applications by then relevant technology nodes. We're preparing for that. That's one. Secondly, we also have continued our investments in R&D, for non-patterning related applications in DRAM. We started those investments already, let's say almost two years ago. We are working hard for that also to start to pay off, let's say, towards the second half of 2018 and going into 2019. That's because those R&D evaluations are ongoing now already for those final selections toward the end of next year.
That's another pillar that we are working on to really expand our penetration in the DRAM space, with regards to ALD.
Okay. Thank you.
Okay, you're welcome, Pieter.
Our next question comes from François Meunier of Morgan Stanley. Please go ahead.
Yes, thank you very much for taking the question. Actually, this epitaxy tool you're ramping up at a foundry guide, do you know for what type of end market this is used? How big the opportunity could be with that customer by 2020, 2021? What is roughly the opportunity you think it could grow to? Just to compare with ALD and your $1.5 billion market guidance. Thank you very much.
Yeah, François. Okay. Thank you for your question. Thanks for being on the call. I think our assessment is that it is an application in that specific foundry segment for the N7 mainstream, N7 node. Further details, even if we would know, we would absolutely not be allowed to share because that's customer confidential information that we would never share publicly with the market. It's a mainstream as far as we know, a mainstream CMOS N7 application. Of course, in N7, there are more EPI applications that are likely being used. We are not the only supplier of EPI applications in that node. It has been for us, the first entrance in the mainstream EPI space. It offers us a lot of opportunity to prove our capabilities.
In that respect, François, we view it as an expansion basically of our served available market in the EPI market as a whole. If you look at the EPI market as a whole, being, let's say $600 million-$700 million US dollar market, until recently, we only engaged in the analog power part of the market, which was maybe up to $100 million. This opens up the big other part of the market because the rest of the market, the biggest chunk of the market and the biggest amount of money is spent in the mainstream CMOS space. That is our intention to get out of the $100 million US dollar market, or not to get out of it, but to expand our presence into the much bigger chunk. Yeah.
Our aim is to expand the number of customers and to expand the number of applications in the coming years. Of course, there's strong competition from the incumbent player, but at the same time, I think the industry could benefit also a lot from having an alternative in an EPI market, which is a market with a very tough technology and a very high entry barrier. I trust that answers your question, François.
Thank you.
Yep.
Go ahead, please.
Sorry. I was asking, what is your unique selling point in that market if it is really competitive? Is it because you've got a technology which is one generation ahead of the others, or is it just because you don't need as high a gross margin than your competitor?
Yeah, I think it's a combination. Of course, I think the initial drive from the industry likely is to, if there is a capable second supplier to bring in a second supplier, viable supplier, a long-term viable supplier in this industry. That likely has been so far a big incentive for the industry to bring us in. Ultimately, you should ask our customers, not us. That has opened up an opportunity for us. That as a starting point, is a great opportunity for us to also, in working with customers, now getting the opportunity to work with them, to explore also where our tools could be differentiated compared to the existing tools. Personally, I don't think there's reason for us to exclude the possibility that we may have unique selling points in future applications. Yeah.
That is for us to prove in the near future. If you would ask us, is that a realistic opportunity? The answer is yes. We're working hard on that. Again, the jury is still out. The team is working hard to prove that in the coming years with multiple applications and across multiple customers.
That's great. Thank you very much.
You're welcome.
Our next question comes from Jim Fontanelli of Arete Research. Please go ahead. Apologies. Our next question comes from Marc Hesselink of ABN AMRO. Please go ahead.
Yes. Thank you. My question is about your comment that you expect to outgrow the industry in 2018. If I'm correct, I think see four points to that. It's the higher ALD intensity in the next logic foundry nodes, so the 7-nanometer. Now you're going to 3D NAND 96 stacks, that you're gaining market share there. Epitaxy and DRAM, a new application going into the second half of 2018. Can you indicate what will be the biggest driver of your outperformance of the market?
I think it's a good question, Marc. It's maybe a little too early, since we're only in October, and you know this industry, a lot can change. The color that we can provide on 2018 is that our current assumption is in making those statements on our, let's say targeted performance in 2018 is twofold. First of all, we do expect that the logic foundry market, will stay healthy, going into next year. Of course, there likely will be a mix in spending from 10 to 7-nanometer. We trust that going towards 7-nanometer, that also our let's say, level of penetration with single-wafer ALD can grow, going from 10 to 7-nanometer. That's one element that we take into account. Of course, what is still not completely clear is, again, the spending will shift towards 7-nanometer.
What still has to be assessed is how much reuse from 10 to 7 will be applicable. We made certain assumptions there. 3D NAND spending is expected across the board based on the visibility we have now from our customers, is expected to remain strong. As we spoke to earlier in the introduction, we expect the use of single-wafer ALD to further increase, going towards next device generations such as 96 stacks for the leader and other advanced nodes for the followers. We expect those additional spendings for those new nodes to really to start kicking, from, let's say, mid of the year, especially in the second half of the year. We're also focused on increasing the number of customers in that industry segment. On DRAM.
We have assumed, like we also touched on earlier with Pieter Olofsen's question, we do anticipate that the DRAM multiple patterning landscape, we assume that is going to improve in the course of 2018 based on real DRAM patterning spending in new fabs. Indeed, on top of that, towards the end of the year, we trust that also initial investments into non-patterning related DRAM applications will start to pay off. On top of that, and last, but certainly not least, we expect the momentum in our new product business at the PECVD to really contribute further into next year. We do expect the contribution of both product lines to increase in a meaningful way year-on-year, in 2018 compared to 2017.
Last but not least, from a regional point of view, we also think China will start coming from a low base, but also we expect that to increase significantly. Again, coming from a low base, but it would not surprise us if the revenue would double in that region next year. I trust that provides you some color, Marc.
Yes, it does. Thanks. The other question I have is more on the ALD market in general. We have this $1.5 billion market target that is already from a little bit ago. It seems that the market has been a bit stronger the last quarters. Also, we've seen that EUV is getting maybe better than people expected earlier. Can you talk about the impact for your longer-term market outlook?
What you mean is, what the impact could be specifically from EUV from a patterning point of view?
It's actually two parts in the question. One is the market in general seems to be stronger than when you provided that $ 1.5 billion target. The second thing is what is the impact of a more successful EUV for that market target?
Okay. As you know, in this industry, demands go in peaks and they go in waves. Our 2017 has been, so far, a very strong year. As a result of that, our forecast for 2020, 2021 has not changed. It has further provided evidence that the single-wafer ALD market has a lot of potential. It goes in waves, so our long-term forecast has not changed. What we do think is that if you look at the time frame 2017 to 2021, that time frame of four years, that the single-wafer ALD market has a good opportunity to outgrow the WFE market during that time frame. We continue to believe that. That's on your question on the ALD market. On the EUV element, we think it has two elements.
One element is that patterning in general, going from 2015 to 2016, the contribution of the patterning market to the single-wafer ALD market has gone down significantly, not only in absolute terms, but also in percentage contribution to the single-wafer ALD market. Going forward, at least for the next few years, we don't think the percentage contribution of patterning to the single-wafer ALD market will change that much. It will not come back to the percentage levels that it had in 2015. The single-wafer ALD growth in the coming years, this year and in the coming years, more comes from, or basically mostly comes from non-patterning. The market has become less dependent on patterning as we speak now. In the past, it was mainly memory that drove it. At this moment in time, more patterning demand this year comes from logic foundry than from memory.
That gives you some color on how that market changed. That's one element. The other element, EUV, as a result of that, not that concerned of EUV. EUV indeed will likely, when it ultimately comes, it will replace some layers, some patterning layers, but as said, in a landscape that has changed according to, as I just explained. Secondly, the positive thing that EUV can bring is that it allows the industry to continue Moore's Law in an easier way. That opens up when Moore's Law continues. It likely will further increase the demand for deposition technologies that can accommodate the smaller geometries of a continued Moore's Law. That's where single-wafer ALD can bring the solution. In that respect, I think it offers opportunities to this market, let's say, beyond a couple of years from now.
Thank you.
Okay, you're welcome.
Our next question comes from Jim Fontanelli of Arete Research. Please go ahead.
Yeah. Hi, afternoon. Thank you. Just following up on your PECVD comments in the press release, could you maybe give us an idea of how to think about sizing that opportunity, both for the second half of this year and more particularly into 2018? Just specifically, what are the new applications? I guess you're referring to TEOS for 3D NAND, but I just wanted to confirm that.
Again, PECVD for us will stay, for the foreseeable future, we will stay a niche player. The market has become, by now, close to a $3 billion market. We're exploring how we can expand our position still within that niche strategy to cherry-pick some good segments where we view that we have a differentiating position from a cost point of view or from a technology point of view. Then grab those opportunities. Again, 3D NAND is an example, is an important example where we explore our opportunities. So it will not be changed into, let's say, a broad-based across-the-board strategy, but a smart, expanding niche strategy. Where, as you know, let's say small increases in market share in a $3 billion market can be meaningful increases to our top line, given the size of this company at this moment in time.
That's what we are focused on. In that respect, we do expect that, we do foresee that we have an opportunity in 2018 to year-on-year, in a meaningful way, grow the revenue in that product line.
Meaningful is what? Is double-digit million incremental revenue? How do we think about that in more concrete terms?
Double-digit million, yes.
Okay. Maybe just to drop that down to Epi, could you give us an idea in the third quarter, whether your new Intrepid Epi business was into double-digit EUR millions or was that single-digit EUR millions of revenue?
The total Epi business in Q3 was absolutely in the double-digit EUR number.
Incremental for Intrepid, the new volume shipments of Intrepid you mentioned in multiple tools.
Well, the color I can give is that, let's say, until recently, until early, let's say this year, our mainstream business was the analog power business. Our position there, our competitive position is still the same as it was before. Over 2017, let's say