ASM International NV

AEX:ASM

Earnings Call Transcript 2019-Q2

[Operator]

Good afternoon, ladies and gentlemen, and thank you for standing by. Welcome to today's ASM International Q2 2019 Earnings Call [Operator Instructions] I must advise you that this conference is being recorded today, Wednesday, 24 of July 2019. I would like to hand the conference over to your speaker today, Victor Bareño. Please go ahead, sir.

[Victor Bareño, Director of Investor Relations]

Thank you, Maria. ASMI issued its second quarter 2019 results last evening at 6:00 Central European Time. For those of you who have not yet seen the press release, it along with our latest investor presentation, is accessible on our website, asm.com. As always, we remind you that this conference call may contain information relating to ASM's future business and results in addition to historical information. For more information on the risk factors related to such forward-looking statements, please refer to our company's press releases, reports and financial statements, which are available on our website. And with that, I'll turn the call over to Chuck Del Prado, President and CEO of ASMI. Chuck?

[Charles Dean del Prado]

Thank you, Victor, and thanks to everyone for attending our second quarter 2019 results conference call and for your continued interest in ASM International. Before I start with review of the second quarter results, Peter van Bommel, our CFO, will first discuss the financial impact of the patent litigation settlement, a settlement that we announced recently. So Peter, please go ahead.

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

Thank you, Chuck. On July 1, we announced that we reached a settlement agreement with Kokusai Electric to resolve our lawsuits regarding the use and infringement of patents. The agreement includes cross-license agreement for period until July 2021, 2 years from now and a payment of an amount of USD 115 million by Kokusai to ASM. This amount EUR 103 million has been included in our Q2 bookings and sales, which amounted to EUR 373 million and EUR 363 million, respectively. The EUR 103 million also dropped to gross profit, resulting in a gross margin of 59% and to operating results, which increased to EUR 150 million. In our earnings release, we showed the financial numbers both including and excluding the settlement. The legal costs related to the patent litigation cases increased in Q2 as compared to Q1. These legal costs, as in the last several quarters, have been included in our SG&A expenses. We expect that the legal expenses in second half will be at a much lower level than in the first half of 2019. The settlement agreement is also impacting our tax rates. We started the year with tax compensation possibilities, also called NOLs or net operating losses, for approximately EUR 82 million in the Netherlands. EUR 25 million of these NOLs have already been recognized in our net results in the past years. Due to the income from the patent litigation settlement, combined with the profitability from normal operations, we fully utilize these NOLs in the course of the second quarter. As a consequence, part of the income in the Netherlands was taxed at a regular rate of 25%. This is the main reason why tax expenses increased in the second quarter. At earlier occasions, we guided that the structural tax rate would increase to a mid- to high teens percentage, once we would have fully utilized the net operating losses. As we have now reached that point in the Netherlands, the tax rate going forward will be in the range of mid- to high teens. Lastly, it's important to note that the settlement did not have a cash impact in the second quarter. Half of the EUR 103 million cash payments will be, and in the meantime, has been received in the third quarter, and the other half is expected to be received late in the fourth quarter or early in the next quarter -- in the first quarter of 2020.

[Charles Dean del Prado]

. Okay. Thank you, Peter. Let's now review our second quarter 2019 financial results. As Peter just discussed the financial impact from the patent litigation settlement, I will focus on the results, excluding this impact. Sales in Q2 reached a new quarterly high of EUR 260 million, which was up 5% compared to the first quarter and ahead of our guidance, which was between EUR 230 million and EUR 250 million. Equipment sales increased 5% compared to Q1, spares and services accounted for 20% of total sales in the second quarter, and we're approximately flat sequentially. For the first half of 2019, equipment sales were up 44% year-on-year, and spares and service sales up 20%. In terms of product lines, our ALD business continued to be the main driver in the second quarter, but the other product lines also have again solid contribution. Looking at the first half of the year, all of our product lines showed healthy sales growth, both sequentially and year-on-year. By industry segment, sales were led by logic, followed by foundry and then memory. Combined logic/foundry sales were approximately similar to the record high level in the first quarter and continued to represent the largest part of our sales, with logic sales up sequentially and foundry down sequentially. Memory sales increased slightly compared to Q1. The gross margin improved to 42.8% in the second quarter up from 41.3% in Q1. This increase is explained by favorable sales mix during the quarter. As a reminder, our gross margin can fluctuate from quarter-to-quarter due to variations in the sales mix. In addition, we continue to have a solid pipeline of new products. The related costs, such as for evaluation tools, can impact the gross margin from quarter-to-quarter. SG&A expenses increased by 19% compared to the first quarter. This increase is explained by; first of all, rise in legal cost for the patent litigation; second, an increase in variable expenses related to the higher activity level; and thirdly, as a result of investments in the strengthening of our organization and business process. R&D expenses increased by 9% sequentially and 20% year-on-year, driven by increased spending on new product development. Operating income in the second quarter was EUR 47 million, stable compared to Q1.Financing results in the second quarter included the currency translation loss of EUR 5 million compared to a gain of EUR 4 million in the first quarter. As a reminder, we hold the substantial part of our cash balances in U.S. dollars, and we include the translation effect in our -- we include the translation effects in our financing results. Net earnings on a normalized basis and including the patent litigation settlement amounted to EUR 125 million, up from EUR 53 million in the first quarter. Let's now briefly look at ASMPT. Normalized results from investments, which reflect our share of approximately 25% in the net earnings from ASMPT. Those normalized results from investments decreased to EUR 2 million for the quarter, down from EUR 3 million in the first quarter. In the second quarter, ASMPT sales were USD 461 million, down 1% from the first quarter and down 31% year-on-year. Bookings were down 17% year-on-year but increased strongly by 31% compared to Q1 of this year.So now turning back to ASMI's consolidated operations. Our order intake in the second quarter amounted to EUR 270 million, excluding the settlement, and as such, exceeded our guidance, which was a range between EUR 240 million and EUR 260 million. Orders increased by 15%, 1-5 percent compared to Q1.Looking at the breakdown in equipment bookings by industry segment, logic was the largest segment in Q2, followed by memory and then foundry. Combined logic/foundry orders decreased compared to Q1, but were still at very healthy levels. Logic bookings increased compared to Q1 and were primarily driven by 10-nanometer related demand and the early tools for 7-nanometer. Foundry orders decreased somewhat compared to the record high level in Q1 and primarily reflected further investments in the 5-nanometer node. Memory orders in the second quarter increased compared to low level in Q1, mainly driven by DRAM. The increased DRAM bookings in the quarter were largely related to specific customer demand, and in our view, not indicative of a broad-based recovery and spending in this segment. Looking at the bookings by product line, while ALD was again our largest product line, we also experienced healthy demand in, for instance, our PECVD and LP business in the quarter. Looking at our balance sheet and cash flow, at the end of June, the cash position increased to EUR 382 million, up from a EUR 312 million at the end of March. And as Peter just explained, this does not include the cash receipts related to the patent litigation settlement. The increase in the cash position was mainly the result of a strong free cash flow, and to a lesser extent, EUR 16 million in dividends from ASMPT and partly offset by the EUR 49 million dividend, the equivalent of EUR 1 per share that we pay to ASMI shareholders. Free cash flow increased to EUR 105 million in Q2 to EUR 126 million in the first half year. In the second quarter, free cash flow was driven by a strong decrease in working capital next to continued solid profitability. Excluding the EUR 103 million in receivables from Kokusai related to the litigation settlement, working capital decreased by approximately EUR 70 million compared to Q1. This decrease was primarily driven by an underlying reduction in accounts receivable as well as customer prepayments received during the quarter. In addition, the measures that we earlier implemented to structurally reduce inventories are also very improved. Days of working capital, excluding the effect of the patent litigation settlement, dropped to 49 days at the end of June, down from 77 days at the end of Q1. Capital expenditure dropped in the second quarter, but this was due to the timing of investments. Our guidance that CapEx will remain at the higher level in full -- in the full year, full year of 2019 is unchanged as we are investing in our new facility in Singapore, which is well on track for completion in the first half of next year. With the Q2 report, we've also announced a new EUR 100 million share buyback program that will start as soon as feasible. This reflects the solid increase in our cash position and our unchanged position -- policy, our unchanged policy to use excess cash for the benefit of our shareholders. Furthermore, on July 23, the cancellation of 5 million treasury shares has become effective, reducing the issued share count by almost 9%. Let's now look more closely at the dynamics in our key markets. Against the backdrop of economic uncertainty and global trade issues, the semiconductor end-market is expected to decline by a high single-digit percentage this year. Looking at the wafer fab equipment, or WFE market, average expectations for WFE spending now show a year-on-year decrease of high teens percentage in the full year of 2019 compared to a decrease of mid- to high teens still expected 3 months ago. This is the balance of a further weakening of the outlook for memory spending, which is expected to be down substantially this year, while spending in advanced logic/foundry is expected to increase in 2019, supported by solid demand for the most advanced nodes.Looking at memory market, wafer fab equipment spending is slow due to continued oversupply in both the 3D NAND and the DRAM segment. While the visibility remains limited in memory, our view about the market is unchanged that spending in the broader memory market will not show any material recovery in the second half of this year. As a consequence, for the full year of 2019, memory spending is still expected to be down significantly with most of the spending on technology transitions. For ASM, memory sales in the second half are likely to be somewhat higher as compared to the first half as we already indicated with the Q1 result. As mentioned, this increase is driven by customer-specific demand in the DRAM segment and not so much an indicator of a broader recovery. We believe we are on track to meaningfully outperform the WFE market this year. In the first half of 2019, our equipment sales increased by more than 30% in U.S. dollars term. This compares with high teens percentage decrease year-on-year for the WFE market in the first half. Logic and foundry are the key drivers behind our outperformance. While our memory sales were down year-on-year in the first half by a double-digit percentage, basically in line with the broader market, this was more than offset by very strong growth in our sales in the logic and foundry segments, as customers stepped up spending on the advanced nodes. As highlighted in the last few quarters, logic and foundry account for the largest part of our business. In addition, we have strongly increased our share of wallet with logic and foundry customers in their most advanced nodes as we are engaged in a substantially higher number of layers. This is driving our strong current performance. In advanced logic, demand remained strong in the second quarter. Equipment spending was focused on additional capacity for 10-nanometer HVM and initial 7-nanometer investments. The transition to the 10-nanometer node in advanced logic has driven for us a substantial increase in the number of ALD layers as compared to the 14-nanometer node. In foundry, we continued to experience healthy demand in the second quarter primarily related to 5-nanometer investments, where we have successfully expanded our position, resulting in substantial share of wallet gained in this segment. We expect the strength that we have seen in the logic and foundry market in the first half to continue in the second half of 2019. Finally, looking at the analog segment, as we already flagged last quarter, we anticipate that demands in the analog segment will be lower in the second half of this year, following healthy sales levels in the last several quarters. In terms of product lines, ALD continues to be a solid driver for our company. The long-term outlook remains strong. The current most advanced nodes, 10-nanometer in logic and 5-nanometer in foundry, have been a major inflection in terms of ALD needs, driven by further miniaturization, new materials and by new more complex device architecture that are on the industry's roadmap, the need for additional ALD applications at future nodes will only further increase. This will support continued healthy growth in these segments of ALD market over the longer term. Our focus in the memory segments of ALD market remains the expansion of our served available market, or so-called SAM. We continue to invest in broadening our portfolio of ALD applications for future DRAM and 3D NAND device technology. In 3D NAND, for instance, as the industry moves to higher stacks of the 96 layers, 128 layers and beyond, the increasing device complexity and high aspect ratio structures will stimulate the need for higher number of single-wafer ALD applications. We are targeting to increase our SAM and our share of the memory market step by step as customers' transition to next-generation devices over the next years.Next to ALD, epi has now become a second growth engine for ASM. The prospects for a structural growth of the epitaxy market are healthy, as next-generation devices in the logic/foundry market are expected to require a steadily rising number of epi steps. At the same time, we remain focused on further increasing our market share by broadening our engagement in the advanced CMOS segment. An important step in this context is the introduction of the Previum that we announced in Semicon West in San Francisco earlier this month. Previum is an integrated epi pre-clean module, integrated as a module in our intrepid epi product. Pre-clean steps can be critical to enable high-quality epi film growth in advanced devices at 5-nanometer and smaller. The integrated Previum modules, therefore, strengthens our offering in the epi market. In short, ASM is well positioned both in short-term as we are on track to meaningfully outperform WFE this year and in the long-term, as ALD and epi are expected to be the key growth segments in the deposition market. Now let's close by looking at our Q3 closest introduction by looking at our Q3 guidance as included in our press release overnight. So what we stated there was for Q3, on a currency comparable level, we expect sales of between EUR 250 million and EUR 270 million, while bookings on a currency comparable level are expected to be also in the range of between EUR 250 million and EUR 270 million. And as stated earlier for 2019 general expectations are now that the WFE market will decline with the high teens percentage, while expectations for the memory segment have further weakened. The healthy demand in logic/foundries -- foundry segments in the first half is expected to continue in the second half of this year. Based upon this current market view, we maintain our expectation to meaningfully outperform the WFE market in 2019. At this point, Peter and I are happy to answer any questions that you may have

[Victor Bareño, Director of Investor Relations]

[Operator Instructions] Okay, operator, we are now ready to for the first question.

[Operator]

The first question is coming from the line of Quang Lee.

[Quang Tung Le, Research Analyst]

Can I just firstly confirm with you that your Q2 cash balance, does it include half of the litigation cost -- settlement or not at all?

[Charles Dean del Prado]

Not at all.

[Quang Tung Le, Research Analyst]

Not at all, all right. And if I can have a follow-up regarding your new orders, it was led by logic and followed by memory and then foundry was last. Do you see orders declining from your largest foundry customer or is it just quarterly fluctuation?

[Charles Dean del Prado]

We definitely see it as quarterly variation.

[Quang Tung Le, Research Analyst]

This is quarterly variation. I got it.

[Operator]

The next question is coming from the line of Sandeep Deshpande.

[Sandeep Sudhir Deshpande, Research Analyst]

Could you just comment on your epi wins, and where and what kind of a process you have won in epi and how? Because you talked about epi being the next big driver of sales for the company. I mean ALD is a very considerable driver, so is epi going to get to that scale? Or what kind of wins are we talking about on the epi side? Secondly, in terms of the process transitions. We have these logic process transitions, which are clearly helping your orders very much because of the use of ALD in the 7-nanometer/5-nanometer node, yes, the EV-based nodes, for instance, do you expect in the other logic customers when they start ramping by the end of next year in the 2021 that they will also be using a lot more ALD in the process?

[Charles Dean del Prado]

Okay. Yes. Thank you for your question, Sandeep. So first on epi, what process, well, as we have shared over the last 2 years in this conference call, we clearly said that we started, of course, with a strong baseline in analog power in both 200 and 300-nanometer. And as you may recall, we started to penetrate into mainstream CMOS high volume early 2017. That was the start at N7 leading -- N7 node, the leading foundry. And we further expanded and we shared it with this audience at that time that, that was the result of a conscious strategic decision that we took in 2015 to get back to become a meaningful player in mainstream CMOS besides our healthy position in analog power because, of course, mainstream CMOS is the biggest part of the epi market, and the epi market is expected to have a high single-digit CAGR for the coming years. And we also sense very, very clearly that the market would love to have a viable second player in mainstream CMOS because it's just healthier ecosystem. So we got that opportunity at a leading foundry N7. And we further expanded our number of layers to into 5-nanometer. And we are very eager not that it doesn't stop there, we have quite some R&D ongoing in many ways, first of all, to further expand our engagement with existing customers, but also to expand into more customers. And of course, logic/foundry is by far the biggest segment in, let's say, mainstream CMOS epi, but we also don't close our eyes for memory, we also have R&D engagements already ongoing for quite some time in memory. And we are learning a lot from that. So that is -- so as a result of that, we have been able to develop epi as a clear second-growth engine for the company. And now for 2 years in a row, the epi product line has significantly grown in their absolute contribution to the company, and also we foresee this year, year-on-year that epi will meaningfully grow in its absolute contribution to the company, Sandeep. So that's on epi. Then on, yes, logic, of course, we are -- I think the best way to see it is that the success that we have -- are experiencing now in the transition in advanced logic from 14-nanometer to 10-nanometer. The significant amount of layer increases that we have -- are experiencing now so basically, increasing share of wallet. The same we are experiencing in foundry going from 7 to 5. And that is, of course, a great baseline to work from in expanding our presence also at other players in the logic foundry arena. Are we immediately at the same level in terms of layers? No, we are not. But of course, also the volume at those players is so much lower at least currently that in terms of market share, the big players that we are engaged with today, determine mostly our position. But we are very, very eager if you look also at the developments in advanced logic/foundry towards gate all around, we are all over, the needs in the market in that -- for that market space a few years from now.

[Sandeep Sudhir Deshpande, Research Analyst]

And then -- I mean one last question on the memory market, I mean what are you hearing from your memory customers? I mean are you getting any indications of when they may come back? Or is this still too far away, they are still waiting to see a stabilization in their market before they start talking to equipment suppliers such as yourself?

[Charles Dean del Prado]

Yes, it's a fair question, Sandeep. And yes, we don't see -- we in our model, we have not -- we are not forecasting a meaningful recovery of equipment demand in 2019, let's say, beyond the demand that we have in our forecast today. And as we shared that, we have a specific customer demand in DRAM that is very, very healthy. So as a result of that, our memory contribution is not bad this year. But in overall recovery, in that segment of the market, we don't foresee before 2020 from an equipment point of view.

[Operator]

The next question is coming from the line of Stephane Houri.

[Stephane Houri, Research Analyst]

I have two questions. The first one is to come back on the strong investment cycle, which is currently driven by logic and foundry. So I know it's a bit early to talk about 2020, but how do you see self-evolving and then there is the question mark also of memory, but at least if you could give us some visibility on the logic/foundry discussions that you have with your customers? And second question is about gross margin. It seems to be improving a bit, it was at stuck at 41%, now it's 42.8%. I know there is volatility on the quarterly basis, but can we expect this positive uptrend to continue through the out of the -- the rest of the year, sorry?

[Charles Dean del Prado]

Okay. So first of all on 2020, yes, it's, of course, early -- too early to provide real forecast on 2020, but we can try to provide you some, let's say, general color on what we sense with the market today. And it is that, of course, we sense that from an end-product point of view, there is a lot ongoing in terms of 5G data centers, artificial intelligence that is -- let's say, those are healthy underlying end-markets that could fuel the logic and foundry segments over the medium and long-term. If you -- so that's, of course, good to know. Looking at the spending itself at this moment in time, in advanced logic, we have no visibility at this moment that 10-nanometer investment, for example, in 2020 would immediately dry up, and on the contrary, we -- at this moment in time, we see a healthy climate in advanced logic. And in addition to node 10-nanometer investments drying up going into 2020, we see also a strong ambition in advanced logic to continue on the path towards 7-nanometers. So from that expect, it would also no surprises that we will face healthy investments -- R&D investments for 7-nanometer in 2020, but that's the best we can say at this moment in time. Again, we don't show -- we don't see advanced logic falling off a cliff at this moment in time, absolutely not. But how strong it will be throughout the whole year, that's a little bit too early to tell. On foundry, yes, we are seeing as we, of course, have also shared, during introduction, a very healthy climate in the first half of the year that we also see, yes, that climate not deteriorating in the second half based on the visibility we have today and yes -- and if you then look at 5-nanometer, the 5-nanometer investments that the customer is now, let's say, is basically preparing for, is by far not the Lth capacity level that is planned for 5-nanometer. So also in that respect, if the macroeconomic environment stays healthy, 5-nanometer investments in 2020 are very likely to happen, and we are excellently positioned to benefit from that as you have seen on -- at the numbers in our guidance so far. So then on gross margin, Peter?

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

Yes. Yes, we always have stated we expect the gross margin to be in the low to mid-'40s. We also have provided some guidance in the last quarters that we expect due to the introduction of a lot of new products that the gross margin would fluctuate a little bit more in 2019, maybe to recall that one important reasons that we have a lot of evaluation tools that we are placing at our customers. And evaluation tools, you only place at the moment that you think that you have higher volumes there, but we have also some smaller customers on the certain moment where we are introducing new applications, and that's the first tool that we placed there is something -- sometimes coming with a discount. So that's -- and those costs are also reported by us under the introductory costs or under the evaluation costs. Last but not least, what is also important to mention here is on the moment that you bring your first products to a new customer or new application to an existing customer, then most likely the introductory costs or the installation costs and related R&D cost might be a little bit higher than what you see when you do this for the second, third or the fourth time. And that -- especially due to the strong growth that we're seeing in 2019, the impact to the gross margin is a little bit more. So I think that 2019 still will show somewhat more volatility than what we expect for later years.

[Operator]

The next question is coming from the line of Nigel Van Putten.

[Nigel Van Putten, Analyst]

Could you please provide some color regarding the DRAM order and potentially also a rough indication of the size of it?

[Charles Dean del Prado]

Yes, the size of it, it's customer-specific, project-specific. We have been working with this customer for multiple years. And yes, we just are benefiting from fact that year-on-year, this application range gets more attention and more demands. So as a result of that, we have seen, yes, strong bookings momentum in the first half, which will become visible in terms of billings in the second half of the year.

[Nigel Van Putten, Analyst]

Maybe take a quick clarification, just an example of how you can expand your service to addressable market at DRAM customers or is this...

[Charles Dean del Prado]

So it's tens of millions. So that's the -- yes, that's color from our size point of view that we can give. And Nigel, your second question was on services, please continue.

[Nigel Van Putten, Analyst]

Also related to the same DRAM order. Is it related to the SAM expansion? Or is this -- what was the reasoning for the customer to place the order, I mean to the extent that you can provide us a detail?

[Charles Dean del Prado]

No, it's just an expansion of the capacity. We were engaged in this application already. And they just specifically expanded in this space, and we were very well positioned. So we were more than anybody else able to benefit from that.

[Nigel Van Putten, Analyst]

Appreciated. Maybe one quick follow-up.

[Charles Dean del Prado]

Yes, sure.

[Nigel Van Putten, Analyst]

With regards to the comments Peter made on the share buyback or the reason behind it, you obviously haven't received the cash yet at least when you reported you hadn't. So should we assume that as the cash from Kokusai settlement comes in that there is additional room for share buybacks or other capital returns to shareholders?

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

Nigel, you know what our process is. On the moment that we have an amount which is substantially above the EUR 300 million then we look how we can fulfill our promise to our shareholders that we will use excess cash one way or another for the benefit of our shareholders. Since the end of the second quarter, we had a substantial higher cash amount, we desired to do that. And we will look at the same sort of situation. And on the moment that we are gained and after having done this above that EUR 300 million and then we will look at the sort of solutions that we will have, which might benefit our shareholders most.

[Operator]

The next question is coming from the line of Marc Hesselink.

[Marc Hesselink, Research Analyst]

The first question is actually on the new epitaxy module. So the first move away from the positioning into edge. Is this very specific to what you can do in epi? Or is this also something that might be a first step to do other edge applications? And my second question is on your capacity. You're doing clearly very well and you're, at the moment, investing in new capacity, and what kind of point of utilization are you today? Can you grow from the current base from without the Singapore expansion? Or do you really need that to grow to the next level?

[Charles Dean del Prado]

Okay. So on -- yes, on the edge part, the way you should see it is that in epi especially when you go to more advanced nodes than pretreatment of the wafer ahead of the critical epi processes is becoming more and more important. So you should view this announcement in that context. As we go to smaller nodes and the amount of epi steps are expected to increase, especially in logic/foundry then the pretreatment becomes more important. So to make sure that we really are successful in increasing our served available market in epi in the years to come, we put this on our roadmap, and this is introduction is part of that. And in general, beyond epi, you may see it also that also pre- and post-treatments before and after deposition also in the ALD space, in other areas of deposition becoming more important and we are paying a lot of attention to that. But in epi, it's very specifically critical and that's why we made this separate announcement, which is very valuable to our customers in viewing us as a viable partner also on their future roadmaps.

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

Yes, with regard to the capacity, I mean you can imagine that with the growth that we've seen in the -- especially in the equipment in the first half of the year that we are stretching our capacity on this moment to the max. There will be enormous relief on the moment that the new building in Singapore will be up and running, but that will cause some time. But I think so far, we have been able to handle that, and we are also taking some additional actions to ensure that we will not run into capacity issues in the next year to come. And that's the -- that year is necessary to overcome that because we expect Singapore operation be ready by mid next year.

[Operator]

Next question is coming from the line of Robert Sanders.

[Robert Duncan Cobban Sanders, Director]

I guess first question would just be on the applied Kokusai deal. You traditionally haven't been in the batch area of ALD. Do you anticipate that market to grow as fast a single-wafer ALD and -- or do you think that single-wafer ALD will basically ultimately grow faster than that? And then I've got a couple of follow-ups.

[Charles Dean del Prado]

So you're talking specifically not to the batch market as a whole, but the batch ALD part and compare that with single-wafer ALD, right? That's what your question is. So yes, our analysis so far, but of course that could evolve over time is that the batch ALD market is a very healthy one and it will stay a very healthy one for years to come. But our estimate is that the single-wafer ALD market will grow faster and also today is bigger than batch ALD markets. But it is a factor that it will stay a factor, the batch ALD market for a number of years, but the dynamics we -- I just explained.

[Robert Duncan Cobban Sanders, Director]

Got it. And in terms of next year, it sounds like you're saying that in terms of your 2 largest customers, they're only about 1/3 of the way through those big ramps where there is 10-nano or 5-nano, respectively. If so is it not a quite likely that you will continue to grow next year, given that trend? Or is there some pull forward or lumpiness that we should be worried about in terms of thinking about the comparables next year?

[Charles Dean del Prado]

Yes, I would love -- we would love to answer that question also if -- of course, but we cannot at this moment in time. We talk to the climate a little bit on one of the earlier questions on the 2020 climate for logic/foundry and there's not too much to add to that. Again, at this moment in time, we see pretty healthy climate in logic and foundry, and we don't foresee the 10 and 7-nanometer demand in the logic and the 5-nanometer demand in foundry to, let's say, fall off a cliff in 2020. That's not what we hear so far based on the visibility we have. But how it will develop in the course of next year, we don't know. The only thing you and I know is that the end-markets. So many attractive end-markets that are gaining a lot of traction at this moment in time, artificial intelligence, 5G, many things that we could elaborate on for much longer, but we don't have the time for that. So exactly on a quarter-by-quarter base, it's difficult. The -- but the overall climate for our business, especially in advanced nodes, is in our view a -- perspective for us is good over the longer term.

[Operator]

The next question is coming from the line of Tammy Qiu.

[Tammy Qiu, Analyst]

So basically you mentioned that your addressable market and step wins within next 2 generations foundry and logic, it's already known. Can you actually comment about your addressable market expansion opportunity you have with your -- within your single-wafer ALD market. Because I remember you used to say by the time of 2021, your addressable market will be x. Is that actually different today because the sum of locations you're guessing is actually bigger than you thought or can you comment on that?

[Charles Dean del Prado]

No, I think in -- I think our view is still the same. Let's be clear, we are -- basically, we put together a strategy a number of years ago to address to increase our served available market across for the whole -- within the whole deposition, Tam, yes. So we're not -- not only was in ALD, but also in epi and also in a more modest way on the 2 other product lines. So we're really looking for how we can address our served available market across-the-board with of course ALD and epi the biggest opportunity as well. We talked to epi a little bit ago, zooming in on single-wafer ALD. Yes, we are -- we have been successful now as you can see in increasing our served available market in a meaningful way, going from 7 to 5 in foundry and from 14 to 10 in advanced logic. As a result of that, our share of wallet, those customers is increasing significantly and it will not stop there. So we are very eager as we go in foundry from 5 to 3 and in logic from 10 to 7. We foresee we're working on R&D programs that if we are successful will enable us to further increase our layer of participation in that transition. And in memory, as we spoke to earlier, it will go more gradually. It will really go step by step that we will in DRAM and 3D NAND, node by node, increase our presence, yes. But as a result of all of this, we see big potential for this company.

[Tammy Qiu, Analyst]

Okay, I see. And you covered batch-tool impact. What about mini-batch tool because Tokyo Electron has been talking a lot about mini-batch tool and commenting that they are actually a very sort of popular solution for the memory makers.

[Charles Dean del Prado]

Yes, maybe that's, in general, are included, Tammy, that's in general included by us and also by the industry watchers in the single-wafer segment.

[Tammy Qiu, Analyst]

Okay. Do you see going forward Tokyo Electron's mini-batch because the efficiency is actually higher, is it likely to be firstly adopted by memory guys before single wafer, the real single wafer is being adopted?

[Charles Dean del Prado]

Yes. The distinction between a mini -- it's really dependent on what kind of product they are -- we are talking about. But in general, again, I don't want to talk too much on specific competitors, but Tokyo Electron, for example, has a strong presence in batch. So really they're real batch and with specific mini-batch products in the single-wafer space. And those products have been there already for quite some time. And in specific segments, they have been very, very successful. And we are very much aware from that -- of that, Tammy. And we're developing our strategy based on our own strengths and based on whatever move -- moves our competition makes in that space, but I -- so you mentioning their presence in mini-batch is not new to us.

[Operator]

The next question is coming from the line of Achal Sultania.

[Achal Sultania, Director]

Just one question on gross margins. Like obviously, we have seen a number of moving parts in the last few quarters, whether it's your mix within your customer, your Singapore operations being ramped up, can you help us understand like what could -- what are the biggest moving parts going forward in your gross margins? Is it just the mix within the customers or by foundry, by logic? Or is it some incremental investments that are needed further to ramp these new addressable markets that you're targeting? Just trying to help us understand whether like what could be the biggest delta for gross margin going forward? And this is more like longer-term question, not just for the second half but general in 2020 and 2021.

[Charles Dean del Prado]

Besides what I already we have mentioned that so the introduction of new products, the introductory extra costs which are, by the way, not only in the manufacturing operations, they are more in the installation so in our service organization. The biggest swing factor is, of course, the certain -- application for certain customers. So it's not customer-driven specific, it's not memory or logic or foundry-related, but there are certain applications, which have a lower or an higher margin than the average as you can -- so what we have said on earlier calls, all our product lines are showing healthy gross margins, but within the product lines, we have products which have higher or lower gross margins. And in a lot of cases, the effect of that in total is relatively small, but there could be quarters where you have biggest swing factors leading to a slightly higher average gross margin than in other periods or to a slightly lower gross margin. I think that's the biggest swing factor besides the introductory cost that I just have mentioned.

[Operator]

Next question is coming from the line of Michael Roeg.

[Michael Roeg, Equity Analyst]

I have a question about your current run rate. Judging by Q1, Q2 and your outlook for Q3, your run rate is roughly EUR 1 billion in sales, which is quite a lot more than last year. How would you say that, that growth is currently broken down if you would have to distinguish between existing clients buying the same products as last year, and on the other hand, all the new innovations, the new products and the clients using new applications for existing product? So can you give us accrued breakdown?

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

Yes, that's of course information which is rather difficult to provide. I mean in general, we introduced a lot of new applications just what we already have stated earlier. So the percentage of sales coming from completely new applications is especially in 2019 very, very high. What we have already stated in the earlier calls is also that we are working very hard to increase also our position with new customers in, as an example, China. So that our strategy is also becoming visible as we speak. But it's a lot of the applications, to make that mix, you need to have a good definition of what you mean with that. But I think that in general, Michael, what we can say about it is that a lot of the sales turnover that we have seen in 2019 is coming from rather new applications.

[Charles Dean del Prado]

Yes. And I -- and to add to what Peter said, I think as we said during our introduction that the transition of the nodes in logic/foundry has been instrumental, also in the year-on-year growth. So that's -- a lot of that is, of course, new applications and not only ALD as we spoke to also epi. And then there is also a pretty healthy contribution from the 2 other product lines, which is not all advanced stuff in a way that those applications were, let's say, formed in the last 2 years or so. But they are just steady footprints at existing customers where the volume is increasing now. And we're basically harvesting from basically the strong coverage at those customers for multiple years. That's a little bit of color, Michael, I hope that's good enough for now. And of course, you always -- possible to look for some further context on that with Victor outside...

[Michael Roeg, Equity Analyst]

No, that's very helpful, especially the comment that is quite sizeable all the innovations. Maybe a follow-up on that. I noticed in the balance sheet that the amount of evaluation tools remains rather high, which is also very promising going forward. I assume that, for instance, the evaluation tools of 1 and 2 years ago in your balance sheet are now driving that particular growth this year, is that correct to assume?

[Charles Dean del Prado]

It's an element of it. So it's one of the elements, is what I mentioned earlier, we also have a lot of new customers there we place -- which are placing orders of tools which we have already developed a few years ago which we have already been installed with other customers. And yes, that's also driving our growth.

[Michael Roeg, Equity Analyst]

Okay. Good. And the evaluation tools that you have today, I assume part of it is with existing customers but there's also potentially new customers within that framework?

[Charles Dean del Prado]

As you know in this industry, we basically are doing business with all the players in the top 10. So most of the new evaluations are with the big spenders because top 3 is spending more than 50% of the CapEx today, top 10 is spending more than 70%, 75% of the CapEx today. So that's where most of your seeding tools should go to. But of course, you see here and there some dynamics of new areas, where new customers pop up, for example, China, where big investments are being made. So selectively, we also make investments there, but it's more -- it's mostly existing customers, new applications in ALD, in epi, and sometimes in other areas.

[Petrus Antonius Maria van Bommel, CFO & Member of the Management Board]

And Michael, because there we tested the first tools and the -- for the first applications that has been done and then a new customer would pop up or a second customer, which is able to do that. And we're trying to do that, as I mentioned earlier, that we're giving sort of a discount on that first product that we're going to deliver. So to ensure that our evaluation tools are not at the highest level. That's further clarification on the color that I tried to give that we discussed [ with those parties ] earlier in this call.

[Operator]

The last question is coming from the line of David O'Connor.

[David O'Connor, Analyst of IT Hardware and Semiconductors]

And for squeezing in, maybe just one, Chuck, on the capacity and your current run rate. Can you continue to grow sales into 2020 if you're currently at max capacity and Singapore doesn't come online until the middle to late 2020?

[Charles Dean del Prado]

Yes, we can do that. We have limited capacity in Korea in our own facilities limited, but we have found ways to expand the capacity. So capacity should not be the limitation in that.

[David O'Connor, Analyst of IT Hardware and Semiconductors]

Okay. Got it. And then for Singapore, when exactly is that ready for tools coming off the line in 2020?

[Charles Dean del Prado]

In -- we expect that the building will be delivered to us in the course of the first half. So I think that at early second half of next year that we can start high-volume manufacturing in Singapore.

[David O'Connor, Analyst of IT Hardware and Semiconductors]

Okay. Great. And maybe if you can comment also on the lead times, how they are there currently trending?

[Charles Dean del Prado]

Lead times for orders you mean?

[David O'Connor, Analyst of IT Hardware and Semiconductors]

Lead times for orders, that's correct.

[Charles Dean del Prado]

Yes. In general, you see that they are 3 to 4 months, except for Korea. That has been the general practice throughout the years, but more recently, you see that sometimes in the logic/foundry space, also orders are placed a little bit with a shorter lead times.

[Operator]

There are no further questions. Please continue.

[Charles Dean del Prado]

Okay. I think we're out of any further questions from the audience. So on behalf of Peter van Bommel, Victor Bareño and myself, I would like to thank you all for attending today's conference call. And of course, feel free to contact us on any remaining questions that you may have in the coming days or weeks, and let's stay in touch, of course. Thanks very much for your interest in the company, and let's stay in touch. Thank you, again. Bye-bye.

[Operator]

That does conclude our conference for today. Thank you for participating. You may all disconnect.