Good day, ladies and gentlemen, and welcome to the ASM International Q4 2016 Earnings Call. Today's conference is being recorded. At this time, I would like to turn the conference over to Mr. Victor Bareno. Please go ahead, sir.
Thank you, Alex. ASMI issued its 2016 Q4 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, adem.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 Chuck Del Prado, President and CEO. Chuck?
Yes. Thanks, Victor. Okay. Welcome to everybody here on the call today. Looking at our company's financial performance, 2016 was a mixed year.
Despite strong gains in the Logic and Foundry segment, the single wafer ALD market showed a double digit drop caused by weaker demand from memory customers. This caused our revenue to drop by 11% in 2016. In this environment, we kept our gross margin stable and supported by continued strong financial position, we have announced a stable proposed dividend and an increase in the size of the current share buyback program. We finished the year with a better than expected Q4. Looking ahead to 2017, we believe our company is in good shape and we expect the single wafer ALD market to return to growth again.
I would like to thank our employees around the world for their continued dedication and hard work during the year of 2016. So let's now review our 4th quarter and full year 2016 financial results, starting with the 4th quarter. Net sales in the 4th quarter came in at 170 €3,000,000 up 20% from the 3rd quarter and up 19% compared to the Q4 of 2015. Sales in the quarter beat the top end of our guidance, which was a range between €150,000,000 170,000,000 In terms of product lines, the key driver was again our ALD business. By market segment, 4th quarter revenues was led by foundry customers.
The gross margin increased to 44.9% compared to 44.2% in the 3rd quarter. SG and A expenses increased by 8% to the previous quarter last quarter and were roughly flat compared to the year ago period. R and D expenses were almost stable compared to the 3rd quarter and dropped by 35% compared to the Q4 of 2015, which included €60,000,000 in impairment charges. We did not record any impairment charges in last year. Operating income increased strongly to €30,000,000 in the 4th quarter, up from €17,000,000 in the 3rd quarter and €4,000,000 in the 4th quarter of 2015.
Financing results in the quarter was €19,000,000 positive and mainly consisted of a translation gain of a similar size. As most listeners on the call are probably aware of, we keep a substantial part of the cash position in U. S. Dollar, and the related translation differences are included in the financing result. The financing result included a €3,000,000 translation loss in the 3rd quarter and a gain of 6 €1,000,000 in the Q4 of 2015.
Results from investments, which reflects our share of approximately 39% of the net earnings from ASMPT, amounted to €18,000,000 for the quarter, down from €27,000,000 in the 3rd quarter and up from 2,000,000 in the Q4 of 2015. These figures exclude the ongoing PPA amortization charge, which amounted to €7,000,000 in the Q4. And for 2017, this amortization charge is projected to be €27,000,000 In the 4th quarter, ASMPT sales amounted to US455 $1,000,000 down 16% from the 3rd quarter. Compared to the Q4 of 2015, sales increased by 20%, which was better than to a high single digit increase that ASMPT had guided for. ASMPT reported strong bookings of US471 $1,000,000 for the Q4, up 24% year on year and up 8% from the 3rd quarter, which was also better than the sequential decrease that ASMPT had expected.
Now turning back to ASMI's consolidated operations. ASMI's net earnings on a normalized basis amounted to €69,000,000 in the 4th quarter, up from €40,000,000 in the 3rd quarter and €18,000,000 in the 4th quarter of 2015. Our new orders in the 4th quarter amounted to €177,000,000 up 44% from the 3rd quarter and up 31% from the Q4 of 2015. And also clearly higher than the range of between €130,000,000 €160,000,000 that we had guided for. The order intake in the Q4 is a new record high for our front end activities and ALD was the key driver behind the order intake in the quarter, but we also recorded a decent level of booking in some of the other product lines.
Looking at the customer segments, equipment orders in the quarter were led by the foundry segment and followed by memory, in particular, 3 d NAND. So now let's discuss the full year results. Our net sales in 2016 decreased by 11%, reflecting the contraction of the single wafer ALD market as a whole during the year. Despite the drop in sales, ALD continued to represent clearly more than half of our equipment revenue. And for the full year, sales were led by the Foundry segment, followed by logic and memory.
Gross margin was 44.2 percent in 2016, marginally up from 44.1% in 2015. The fact that we were able to keep gross margins steady despite the pressure on sales reflects the impact from the effort we undertook in the last several years to further improve the efficiency and flexibility of our manufacturing and supply chain operations. As also discussed at earlier occasions, these measures included new outsourcing initiatives and a stronger focus on sourcing of complete subassemblies. Over time, we have further reduced the fixed cost part of our total cost of goods sold. Our target remains to run operations at a gross margin percentage level of lowtomid40s, bearing a market downturn for which the target is a high 30s level.
On a quarterly basis, gross margin continue to be impacted by factors such as sales mix and utilization. Operating expenses remained under control during the year. SG and A expenses dropped by 5% and increased as a percentage of sales from 14% in 2015 to 15% last year. Total R and D expenses increased by 2% and as a percentage of sales from 13% to 15%. Excluding the impairment charges in 2016 and excluding R and D capitalization and amortization, R and D rose by 10% in 2016.
This was in response to a further increase in customer requests for new applications and new engagements. Operating profit dropped from €111,000,000 to €82,000,000 And the operating margin remained in the double digit area at 14% compared to almost 17% in 2015. Results from investments on a normalized basis increased from €43,000,000 to €68,000,000 in last year. And the total sales as reported by ASMPT increased by 10% to US1.8 billion dollars in 2016. Sales of the back end equipment business increased by 20 3% in 2016.
Apart from a recovery in the overall back end market, ASMPT did well in specific market segments such as equipment for CMOS image sensors and LED. Sales of SMT Surface Mount Technology Solutions dropped by 4% for the full year, but returned to growth in the second half of the year. Asia MPT increased gross margins from approximately 36% to 37.6% in last year, And this was the balance of a strong 4% improvement for back end equipment, offset by a 2 percentage points decrease in SMT. Looking at our consolidated numbers again. Normalized net earnings decreased by 12% to €163,000,000 And on a per share basis, normalized net earnings decreased by 8% to €2.65 So looking at our balance sheet and cash flow.
At the end of December, cash amounted to 3 €78,000,000 up from €363,000,000 at the end of September. This increase is mainly explained by positive currency effects with free cash flow of about €10,000,000 in the 4th quarter, offset by cash used for share buybacks. At the end of December, net working capital stood at €155,000,000 up from €127,000,000 at the end of the 3rd quarter and €114,000,000 at the end of 2015. The number of outstanding days of working capital measured against quarterly sales increased to 81 days at the end of December, up from 79 days last September and 69 79 days last September and 69 days at the end of December of 2015. For the full year 2016, free cash flow dropped to approximately €20,000,000 down from €104,000,000 in 2015.
Next to lower profitability, the drop in cash flow was a result of an increase in working capital, as I just mentioned, which I was explained by the back half weighted character of sales in the Q4 of last year as well as a rise in inventory in anticipation of increased shipments in the forthcoming quarters. Quality and the underlying level of working capital remained healthy. During the year, we used a total of €140,000,000 in cash for dividends and share buybacks, up from €116,000,000 in 2015. This reflects our continued commitment to use excess cash for the benefit of our shareholders. We spent €13,000,000 during the Q4 to repurchase approximately 250 1,000 of our own shares.
During the Q4, we completed the previous 2015, 2016 share buyback program, and we started the new program that we announced last October. As of last week, we have spent approximately €20,000,000 of this program. Yesterday, we announced an increase in the current share buyback program from €50,000,000 to €100,000,000 which is enabled by our continued strong financial position. The program will run as earlier communicated until ultimately November of this year. In addition, we have announced that we will propose a dividend of €0.70 per share, which is stable compared to last year.
We also announced in yesterday's press release that we intend to cancel 1,500,000 Treasury shares. Let's now have a more detailed look at our ALD business. After strong growth in the 2013 to 2015 period, the single wafer ALD market went through a double digit contraction in 2016. Weakness in memory was only partially offset by strong growth in the LogicFoundry segment. In 2017, we expect condition to improve and the ALD market to show a clear recovery compared to last year.
Longer term, we continue to expect healthy growth for ALD. We forecast the single wafer ALD market to reach a size of US1.5 billion dollars by the 2020 2021 time frame. We earlier expected the market to reach the US1.5 billion dollars level by 2019. Our updated view is mainly explained by lower expectations for the DRAM segment. Still, our updated forecast implies a solid double digit growth path for the coming years.
General drivers are the introduction of new materials and complex chip architectures such as advanced FinFET technology that are crucial enablers of Moore's Law. Displaced to the strengths of ALD has more highly precise and conformal film deposition is required for the critical process steps in leading edge semiconductor manufacturing. If we look at logicfoundry, we substantially increased our share of wallet in this segment during 2016. As these customers started to ramp the 10 nanometer technology node. Single wafer ALD is a key enabling technology for the 10 nanometer transition.
Compared to the previous 14, 16 nanometer node, the number of single wafer ALD layers has increased significantly. Next to our existing Hi ks metal gate applications, the precision and film core formality of ALD is required for several new critical process steps to build these advanced devices. In addition, ALD for multiple patterning, an area where we already build a leading position in the memory market, is also the key enabler of the most advanced logicfoundry nodes. We believe we have strategically strengthened our our position with leading customers in the logicfoundry segment during the year. In Foundry, our revenue increased significantly north of 50% year on year to a new record high and well above previous levels.
For 2017, we see continued healthy demand in logic foundry ALD supported by spending on the 10 nanometer node and some early investments in 7 nanometer. Looking over a multiyear period in the transition, so looking over a multiyear period time frame, in the transition from 14, 16 nanometer to 10 nanometer and 7 nanometer node, we expect the single wafer ALD served available market in logicfoundry to more than double in total. In this context, I would also like to highlight the award that we received from TSMC last week. As one of 5 equipment companies, we have been awarded with TSMC's Supplier Excellence Award, the first time for us specifically for the performance and support of our ALD equipment and technology during 2016. We are very honored to have received this prestigious award from TSMC.
In the memory market, the single wafer ALD market last year saw declines in both DRAM and the NAND flash segment. In 2015, memory and DRAM in particular was still a very important driver for growth in the ALD market. Memory still represented our largest customer segment in that year, in 2015, but was overtaken by logicfoundry in last year. And in 2016, overall spending dropped substantially in the DRAM sector. In the course of the year, it became clear that the investments in new ALD tools for the transition to the 1x DRAM node were pushed out, which led to a weaker than expected development of the overall single wafer ALD market.
ALD continues to be an enabling technology for DRAM customers to transition to next nodes. Compared to the 25 nanometer generation devices manufactured at the 18 nanometer node and below require a higher number of ALD based patterning layers. We believe ASM remains well positioned to support key customers in their transition to the next node. At the same time, however, overall spending in DRAM is expected to remain relatively moderate compared to the strong level in 2015. In 2017, customers are more focused on technology transitions within existing fabs, which means a substantial amount of reuse of existing tools.
As a result, looking at 2017, we expect the DRAM part of the single wave ALD market to show a modest recovery compared to 2016. As far as the NAND flash sector is concerned, 2016 was, as expected, a transition year for the LD market. The mix of spending shifted almost completely from planar NAND to 3 d NAND in the course of the year. In planar NAND, we had a strong exposure with ALD for multiple patterning, which is, however, not an important requirement in 3 d NAND. This meant that ALD demand for this application, which was still a healthy driver for our business in 2015, largely disappeared in last year.
And at the same time, the need for new nonpathaling single wafer ALD applications was limited in this phase of the in that phase of the 3 d NAND transition. 2016 was important, however, as we booked the 1st significant process tool of record selections, PTYR selections, followed by multiple tool orders for 3 d NAND in the Q3 and again in the Q4 of last year. As the customers where we have been selected ramp high volume manufacturing this year, we expect a significant increase in our 3 d NAND related sales as a result in this year. Taking a longer term view, we see 3 d NAND as an important driver behind the expected double digit growth of the overall single wafer ALD market in the coming years as the industry transitions to more complex higher stack 3 d NAND devices. In short, we remain confident about the growth potential of the ALD market.
As our customers move along Moore's Law, further miniaturization, introduction of new device architectures and new materials will further increase the need for superior position and conformality offered by ALD. As a leader in ALD, our company continues to be well positioned. So let's now look at the broader market environment. The wafer fab equipment market had a solid year in 20 16, growing by almost 10%, with increases in logicfoundry and particularly 3 d NAND offsetting declines in DRAM. The difference was the weaker development in the ALD market, as just discussed, is explained by the 3 d NAND spending, which strongly boosted WFE, but so far in 2016 has not yet been a strong driver for the single wafer ALD market.
Our industry has entered 2017 in good shape. Market watchers such as VLSI and Gartner expect again a mid to high single digit increase in WFE wafer fab equipment spending in 2017. While the increase in semiconductor end markets is expected to accelerate from 2% growth in last year to 6% growth this year. So let's now look at our guidance as included in our press release. We continue to expect a clear improvement in the single wafer ALD market in 2017.
First of all, demand in logicfoundry segment is expected to remain healthy. Secondly, we believe the 3 d NAND contribution to the single wafer ALD market will show a strong increase in 2017 and leading that leading to increased orders for ASMI in the 3 d NAND space. And thirdly, the DRAM segment is expected to show a modest recovery this year. So based on all of that, we project a year on year sales increase for the first half of twenty seventeen, whereby we expect a sales level of between €135,000,000 and €145,000,000 for Q1 and between €160,000,000 €200,000,000 for Q2, both on a currency comparable level. The order intake in Q1 is expected to remain healthy at the level of somewhere between €170,000,000 and €190,000,000 also on a currency comparable level.
And finally, on the sales guidance for Q2, you may have noted that the range that we provided for sales in Q2 is a little bit larger than normal. And this is explained by 2 factors. First of all, the rapid increase in demand that we have recently experienced has resulted in slightly longer lead times for selected products, and this may impact the specific timing of the sales. And in addition to that, also bookings of certain Japanese customers are increasing, and these typically take a couple of months longer to be recorded as sales compared to bookings from other geographies. At this time, we are very happy to take your questions.
We'd like to ask you to please limit your questions to not more than 2 at a time, so that everybody has a chance to ask a question. Okay. Alex, we are ready for the first question.
Thank you. We will take our first question from Nigel Van Putten of ING. Please go ahead. Your line is open.
Thanks, guys. I have two questions then. First off, on the phasing again, 2017, you changed the DRAM wording and you've just explained why. But competitors, I think, one at least has said that they expect 2017 to be more front end loaded and also yes, your order book is quite strong coming into the first half. So would it be fair to assume that that could be the case this year?
Or do you see ongoing strength and perhaps some acceleration into the second half? And I'll take my second question directly as well. It's on the competitive positioning. It seems to be a strong foundrylogic year again 2017. So how do you think your relative market share will perform?
Will you be able to gain a bit of share again? Or is it less mix dependent and you see others making inroads? Thank you.
Okay. So on your so basically, you're asking a little bit what outlook we see for the full year or for the second half compared to the first half. That's how I understood your first question. Well, we think current momentum for the company is solid. As we said, Q4 sales and particularly order intake were better than expected.
Order intake in Q1 will again be strong. So we should end Q1 with a pretty strong backlog. And the overall outlook for sales in the first half looks pretty healthy, we believe. Same time, it's a little too early to provide guidance for the second half. Yes, as far as the full year is concerned, yes, we have provided guidance to you for the single wafer ALD market to show a clear recovery in 2017.
So that already should give you some indication on the year as a whole. And together with that, we also shared that we previously expected recovery in DRAM to be more concentrated in the second half of the year, but we now expect, yes, a modest recovery for the full year as a whole. And we already see, yes, some modest improvement in the first half, but we do not have visibility that on, let's say, DRAM, it will be different in the second half. That's just too early to tell. Just too early to tell.
On LogicFoundry, we do see that it will be very healthy for the full year. That's what we with visibility we have now. Certainly, it will be healthy for the first half. And but so far, we have no indications that the second half would be any different. And for 3 d NAND, yes, we expect a strong improvement throughout the year, year on year.
And so that's very positive. So I trust that provides you some insight on the outlook for the year. Again, the most important thing is clear recovery of the single wafer ALD markets for SHEIN year on year. And we are building our strengths on that. So market share development, yes, well, I think our logicfoundry, we shared before that our share of wallet in logicfoundry has increased significantly going from the 16, 14 nanometer node to the 10 nanometer node.
And we don't yes, we expect that our market share position we foresee that our market share position this year will stay strong in the logicfoundry segment. DRAM, yes, well, based on the only modest recovery, I don't think the spending there will impact our overall market share in the single wave for AOD market that much because the absolute spending is not, let's say, instrumental as part of the total based on current visibility. And yes, in the 3 d NAND market, again, we will year on year, our revenue development will increase significantly. And but of course, also some competitors are also benefiting from that. And so and of course, our market share is likely a little higher in logicfoundry than in 3 d NAND.
So as a result of that, maybe our percentage market share may go down slightly in the year as a whole if you take all these items together. But again, that is regardless of the company being in good shape for the year as we see it now. I trust that answers your question.
Yes, very much. Thank you.
Okay.
We will take our next question from Matthias Santos Silva of Morgan Stanley. Please go ahead. Your line is open.
Hello. Thank you for taking my question. So I only have one question. It's on DRAM. If you could please clarify the recent headwinds.
Is it purely cyclical? Or could factors like reuse, for example, continue in the coming years, making it more structural? Just if you can give some clarification on that. Thanks.
Yes, the recent headwinds, yes, the recent headwinds, I think it's just a continuation of what we have seen already for the last one and a half year. And yes, our impression is that in general, let's say, the underlying fundamentals in the DRAM segment have improved, let's say, inventory levels, pricing of, let's say, on a device DRAM device level have improved. At the same time, you see that the big DRAM players have given priority to invest their new invest in their new fabs primarily in 3 d NAND. That's currently the prioritization that they have chosen, to use that capacity in new fabs only for 3 d NAT. And let's say, work on the new technology nodes in DRAM through technology migration in existing fabs.
And that automatically stimulates, yes, reuse, reshuffling of equipment from older nodes to newer nodes over investing in new equipment. Those dynamics might change over time as a result of a different demand pattern in DRAM itself and maybe also as a result of different dynamics in the 3 d NAND market. If the 3 d NAND market at some point in time slows its pace of growth and you get more a little bit more overcapacity there or so, that might also trigger customers to make other choices. But again, we cannot speak on behalf of our customers. You would ask you should ask them, but that is that's the visibility we have at this point in time.
But I would like to emphasize, we have not there is not any we don't have any insight that our competitive position has changed in the DRAM space. This is just more dynamics that are happening at our customers that we have little influence on.
We will take our next question from Peter Olofsen of Kepler Cheuvreux. Please go ahead. Your line is open.
Yes, good afternoon. Two questions from my side. First one, I've noticed that the ASMPT share price is now back to the 2013 level at which you reduced your stake. At the same time, your front end business still looks relatively undervalued. So has there been any change in your thinking about the stake in ASMPT and how to unlock the value of the front end business?
And then maybe question for Peter on the tax rate. Can you explain why the tax rate was so low in 2016? And what should we expect there for 2017?
Okay. Yes. First of all, on the PT stake, we're, of course, very happy shareholders that the share prices of PT has developed in that way. So that's first observation. And secondly, the only thing we can say is that there are no concrete plants at this moment in time to do anything with the PT sec.
Yes. With regards to the tax rate, Peter, I'm more than happy to give you there some color. That remains low. It was also low in 2015. And as you might know that we have restructured some of our activities a few years ago.
And as a consequence of that, we are now better able to make use of the NOLs, the deferred tax losses that we have in particular areas. And as well in 2015 as in 2016, that's had led to substantially lower ETR than what you normally would have expected. Going forward, I think that it gradually will increase, but that also in the 2 years to come that our tax rate or ETR will not be above the 10%.
Okay. Thank
Our next question comes from Robert Sanders of Deutsche Bank.
Maybe if you could just talk a bit more about the change in your view on the long term ALD market. Do you have a different view on the amount of leading edge capacity in the long term? Or is there a different view on ALD adoption in a particular segment? And then second question would just be around 10 nanometer yields. It does seem like a lot of your growth now is tied to the 10 nanometer ramp at TSMC.
So given that there's a lot of press around yields and some products already being delayed, I was just wondering how sensitive is your revenue outlook to potentially weaker yields this year? Thank you.
Okay. First of all, the long term model, is there do you have a different view on, let's say, the long term potential of the single wafer LD market? Well, the short answer is no. Absolutely not. The only adjustment that we made, as we tried to explain in the introduction, is that we yes, that the only change is that we project a more modest recovery in 2017 of the DRAM market.
And we adjusted mainly the main adjustment we made was adjust the model for that. And so as a result of that, we project now that the growth to a €1,500,000,000 level is now forecasted in the 2020, 2021 time frame. But that still means that there is a double digit CAGR associated with the growth of the single wafer ALD market. So that should illustrate to you that the fundamental potential of this market has not changed at all. Okay?
So then on logic on the foundry, 10 nanometer, so you you said if your revenue outlook is very yield sensitive. Well, I think well, based on the visibility we have now, our forecast is pretty is not changing, has not been changing a lot recently for the foundry segment. And it's for the year, there's a mix of 10 nanometer and early 7 nanometer investments. And of course, the best for our customers would be that the yield is good. If the yield is less good, then of course, it could incur a higher demand.
But in general, I think our foundry our very respected foundry customers have scheduled for a very solid yield as they have been performing to high yields throughout their existence. So I don't at this moment in time, we don't at least we don't foresee any negative impact from yield results, again, based on current visibility. Okay? I trust that answers that part of the question.
Yes, that does. Thank you.
Okay. You're welcome.
We will take our next question from Jim Fontanelli of Arete. Please go ahead. Your line is open.
Thanks very much. Hi, good afternoon. Just a couple of questions. One is, could you help us maybe size the NAND opportunity relative to DRAM this year? I know you obviously don't give absolute numbers, but just in terms of relative size of the NAND opportunity versus DRAM this year, could you help us think about that?
Well,
there is let's put it this way, that contribution to our top line is going to be instrumental this year. It's going to be a significant double digit percentage of revenue. And if you look at DRAM and the current projections for DRAM are not in that space. So our current projections for DRAM in the first half are so yes, the same applies to the first half, And we see no reason currently that, that will be different in the second half of the year.
Okay. So sorry, just to clarify, are you saying that at the moment, it looks like NAND overall for the year will be larger than DRAM in terms of revenue contribution?
Absolutely. Absolutely. Great. For sure. And I also think that it's also interesting to say that the NAND we currently project the NAND revenue to be the 3 d NAND revenue for this year to be bigger than the planar NAND revenue that we had in 2015.
Interesting. And if I could, just a way you could help us understand the relative contribution of 48 layer and or 64 layer within that within the overall 3 d NAND pie?
Could you repeat that one more time?
Yes. Within the overall revenue pie for 3 d NAND, is 48 layer making a greater contribution than 64? Or is it mainly driven by the ramp of 64 through the second half of the year that's driving overall net revenues for you?
Yes, yes, that's a good question. Well, it's really a mix, and that's also because I think customer the customers also especially the newcomers are also still I think it's still to be determined how much volume will really be made in 48 and how much will really be made in 64. But we are in as of 48, 48 technology with the customers that we are engaged in. So from that point of view, I think towards if more volume would go to 64 at the customers that we are engaged in, it would not negatively influence our outlook.
Great. And maybe secondly, just a follow-up on prior question. You were clear in terms of the thought process around the PT stake currently. But in terms of how you see the underlying value of the front end business, clearly, exiting out the rising value of the PT stake, the front end business hasn't really seen its value its implied value move over the last couple of years despite pretty strong backdrop for overall SPE and a broader rerating for the sector. So if you could just update maybe your thought process on value around the front end business and if there's anything there that can release further value beyond just continuing to deliver good operational results.
Yes, yes, yes. Yes, well, Ali, I fully understand your reasoning there. And again, we are also somewhat surprised that there is not a stronger recognition in the market for the front end results based on the track record of the company. In terms of, let's say, implied value, we're also a little surprised. And we still trust that continued strong operational performance will unleash at some point in time the right valuation to the front end.
But at the same time, of course, on a continuous basis, we as management, Peter and I and we involve others in that, we do the analysis also based on the implied value, based on the valuation of the development of the PT share. And I can just guarantee you, again, ultimately, we also are aiming for driving the most optimal long term value for all stakeholders in the company. And we should take all these elements into account. So I only can say finally, I we clearly hear what you say, and we continuously analyze and review.
We will take our next question from Tammy Kwei of Berenberg.
Hello. Hi. Thank you for taking my question. So you explained already your 3 gs Now opportunity that you're already in the 48 layer design with your customer. I'm just wondering how should we think about the route from 48 to 64 and 228?
Is your TAM basically expand exponentially? Or is your TAM actually growing at a lower pace compared to the layer increase?
Well, I don't know if I fully could you provide a little bit more color on your question, Tammy? I really want to understand what you mean.
Yes. So for example, your ALD opportunity from 64 layer three d NAND compared to 128 layer, is your total addressable market opportunity doubled or is less than that?
Well, we trusted 2 elements that are part of our 3 d NAND strategy. First of all, we want to increase the number of customers that we are engaged with because currently we're not engaged with all of the customers, yes? We started strongly. We are now strongly engaged with, let's say, the newcomers. And but we want to complete the whole scope of customers.
We want to broaden our scope, especially as of 96 Tech, then I think which will become important in 2018. Then we want to further increase the scope of customers that we are engaged with in HVM because we are, of course, engaged with all of them in R and D. And secondly, we do indeed our belief is that we can increase our the level of engagement with them as we go to higher stakes. So we are driving our 3 d NAND business across both angles. And based on that, we think that percentage wise, the relative growth of 3 d NAND within the single wafer ALD market will be significant in the coming years.
The percentage growth percentage of growth of that segment as part of the total single wafer ALD market. Logic foundry is expected to be stay the biggest. In the past, to provide you a little bit more color, in the past, it was DRAM. DRAM was the number one area. But since 2016, logicfoundry is clearly the biggest.
And with the current insights, we foresee that logicfoundry will stay the logicfoundry combined will stay the largest segment for a number of years. But 3 d NAND will grow in a percentage wise rapidly. And as a result of that, in absolute terms, we'll also grow to a decent size in the next few years.
So will actually your ALD addressable market or ALD steps, which requires your single wafer tool, doubles if the number of 3 NAND double, like the 3 NAND layer doubles and your ALD steps, would that double as well?
Well, I think it's very well possible that compared to, let's say, yes, compared to last year, I think it will for sure double the single wafer contribution of 3 d NAND. And I think in the yes, it will again, we are fine tuning our projections. It depends also on how soon certain applications from us become really POR for HVM, but a high double digit percentage growth also looking at the level of the 3 d NAND single wafer ALD market this year is also very likely to have significant high double digit percentage growth compared to the level that we will reach this year.
Okay.
Yes. And this year the level we will reach, the single wafer ALD market this year is, in my opinion, already significantly higher than the planar NAND single wafer ALD market in 2015.
Okay. I see.
Yes. And our and the contribution to ASM of 3 d NAND in 2017 will also surpass the contribution of planar NAND it had to ASM in 2015.
Okay. That's clear. And also, I'm just wondering what is the upside from a gross margin perspective? In ALD. So, where is the upside from your gross margin perspective?
Are we likely to see some kind of higher margin products when you actually move to more segment?
No. I think that's tell me that the gross margin will develop as what we have indicated also in the previous years and in previous calls, we expect that the gross margin to fluctuate between the low to mid-40s as what we have seen also in the past years.
Okay. Okay. Thank you.
We will take our next question from Edmond de Jong of NIBC Markets. Please go ahead. Your line is open.
Good afternoon, gentlemen. Most of my questions have been answered, but maybe a few left. On the other projects, the Epi and the furnaces, You were mentioning in your introduction that you saw some significant part in the order intake of Q1 was also related to those products. Could you give some more color on that? And maybe on sales in Q4 and what you think of 2017?
That's my first question. And then the second question also in the introduction, you mentioned on the inventory stage, you had some inventory ready for shipment in the coming quarters. Must I think of this as that this equipment has been sold already? Or is it more speculative? Or how should we see this?
Those are the two remaining questions.
Okay, Edwin. Okay. So for your first question, so yes, indeed on so we had some yes, some decent contribution in terms of bookings in Q4 in the epi and PCVD space in Q4. And so really healthy contribution. And if you look at 2017, yes, we do count on the contribution of Happy MP CBD to go up in certain areas.
But again, it's we're working hard on that, and it's just too early for us to really quantify it. But we do think that the contribution will also go up in this healthy industry climate. And the first indication of that was in the Q4 bookings that we saw.
And I'm very happy that's still the power devices?
Yes. It's a little bit of mix of it's a little bit of mix of different markets. And when the momentum would develop further, we can provide the market with a little bit more color. But it's a little bit of a mixed bag on Epi, which is good. We're not betting on one specific market.
But most of it was indeed in the mainstream. But again, we have been investing over the last few years in both PECVD and in epi in several new areas. And yes, we trust that at some point in time, it's going to be back, a payback. And it would be great if some of that would become visible in 2017. And as soon as that has materialized, we will definitely share that also in this forum.
That will be very good.
Yes. Very good. Yes. And then the second question
And your question about the inventories, I think it's a mixed bag in the inventories. On the one hand, we have seen that the finished goods inventories were coming down. So they are on a very low level as we speak today. So that's not speculative at all. They're highly likely they will going to increase in the forthcoming period a little bit because that has to do what Chuck already mentioned about the Japanese customers.
So we recognize there the sales turnover in a little bit later stage than what we do with other customers. And as a consequence of that, we will have some finished goods on our balance sheet increasing in the forthcoming period. The rest is related to raw materials and work in progress, which
is, of
course, at this moment increasing simply because of the fact that we have all those new orders coming in. And Chuck already explained the rationale why that not immediately turns into sales turnover already in the Q1. That has also to do with those Japanese customers. So that's a little bit of it. So it's not speculative at all.
It's simply the inventory levels that we need for ramping up our activities in line with the order book development.
Okay, clear.
We will take a follow on question from Peter Olofsen of Kepler Cheuvreux. Please go ahead. Your line is open.
Yes. I wanted to come back on the earlier discussion on the 3 d NAND opportunity. If you take a certain number of layers, say 64 layers, is there a difference in the ALD opportunity for you between charge trap structures and floating cage structures? And when it comes to emerging types of memory like 3 d exploit, Is that an area that might already contribute to your sales this year?
Yes. Could you repeat your second question, Peter?
Yes. Whether emerging memory technologies like 3dXPOINT might already provide some contribution to your sales in 2017?
Okay. Okay. Okay. So yes, on the let's start with the latter one. On Xpoint, on, let's say, new memories in general because Xpoint is maybe specific to certain customers, but let's talk about new memories.
It could be crosswind, it could be our ERAM, it could be all kinds of new memories beyond DRAM and NAND. Yes, there will be some contribution this year, but we so far have projected yes, just a little bit. And that has to do not again, we have been investing in this area for a number of years. We have positioned ourselves in a certain way. It just depends on when the customer will launch will be able to launch a real product to their end markets.
So the investments that are being made into our equipment are really to support their R and D, but not high volume related yet. Again, that could change going into 2018 based on some customer projections, but it would pleasantly surprise us if customers would ramp towards the end of this year. And that's current visibility is that, that is not the case yet. So then on 64, as you know that most customers are on the chargeback path in the 3 d NAND space. So of course, without becoming too specific on individual customers, it should be clear that most of our efforts, again, all both areas are important to us because all customers are important to us.
But of course, most of our efforts currently are in the charge up area because that's the biggest part of the market. And yes, so that's I trust that answers your question there.
But theoretically, there would not be a major difference being the use of ALD in these two segments with these two technologies?
Yes. Well, I think, again, I don't want to get into too much detail on an application level. But just as a matter of fact, if the majority of the customers are in the charge up area, then by definition, the served amenable market there is much bigger than in the other area.
Okay. And maybe one thing to clarify on what Peter earlier mentioned on Japan and the revenue recognition. The fact that it takes a bit longer to recognize the sales has had to do with the fact that it's Japan or has it to do with maybe new client that you're not able to recognize on shipments? What's the reason it takes longer?
Yes. There are 2 different sort of revenue recognition, as you might be aware or so from other semiconductor equipment companies, you mostly we recognize the revenue on the moment that we ship. Then we recognize the biggest part of it. And only the little big bit is recognized on the moment that the product is installed and released by the customer. And there are certain customers, and especially is that Japan related, where the release only takes place on the moment that where the revenue only will be recognized on the moment that the final release takes place.
And that's the difference. So it's more a general practice in certain parts of the world than anything else.
So it has nothing to do with customers that we did not engage with before.
Okay. Yes, that's helpful. Thank you.
We will take a follow on question from Jim Fontanelli of Arete.
I may have missed it earlier, but
could you give us
a thought process on how to think about OpEx for the remainder of the year, particularly R and D and how you see that scaling?
Yes. I'm more than happy to do that. The color that we have given previously and that is not that has not changed, I think. We expect that R and D gradually will increase. So we expect that to be as an average over the longer period of time in the low to mid teens.
While SG and A in general is expected to have to remain relatively stable. There will be some slight increases, but it will not increase in line with when we would have a 10% sales increase, we would not see an SG and A increase of that level. So I expect that only with very small steps would increase in time. And that's in principle what you have seen also in the previous years when you look to the last 4 or 5 years, then you see R and D gradually increasing with the activity level, so in that low mid teens as an average. And SG and A has remained with the exception of currency movements relatively stable in time.
Thank you.
That will conclude
Okay. Well, thank you everybody for your participation today and for the amount of questions that you were able to raise in which gave us an opportunity to further clarify. So thank you very much, and let's stay in touch. And any follow-up questions you may have, please get in touch with us through Victor Beranjo. Thank you again, and have a nice day.
Bye.
Thank you. That will conclude today's conference call. Thank you for your participation,