ASML Holding NV Management Discusses Q4 2013 Results - Earnings Call Transcript

January 22, 2014

ASML Holding NV (NASDAQ:ASML) Q4 2013 Earnings Call January 22, 2014 9:00 AM ET

Executives

Craig DeYoung - Vice President of Investor Relations - ASML Tempe
Peter Th. F. M. Wennink - Chief Executive Officer, President and Member of Management Board
Wolfgang U. Nickl - Chief Financial Officer and Executive Vice President

Analysts

Sandeep S. Deshpande - JP Morgan Chase & Co, Research Division
Timothy M. Arcuri - Cowen and Company, LLC, Research Division
Francois Meunier - Morgan Stanley, Research Division
John W. Pitzer - Crédit Suisse AG, Research Division
Stephane Houri - Natixis S.A., Research Division
Srinivasan Sundararajan - Summit Research Partners, LLC
Gareth Jenkins - UBS Investment Bank, Research Division
Mehdi Hosseini - Susquehanna Financial Group, LLLP, Research Division
Jerome Ramel - Exane BNP Paribas, Research Division
Mahesh Sanganeria - RBC Capital Markets, LLC, Research Division
Sumant Wahi - Redburn Partners LLP, Research Division
Weston Twigg - Pacific Crest Securities, Inc., Research Division
Simon F. Schafer - Goldman Sachs Group Inc., Research Division

Operator

Ladies and gentlemen, thank you for standing by. Welcome to the ASML Fourth Quarter and Annual 2013 Conference Call on January 22, 2014. [Operator Instructions] I would now like to turn the conference over to Mr. Craig DeYoung. Please go ahead, sir.

Craig DeYoung

Thank you, Sarla, and good afternoon and good morning, ladies and gentlemen. This is Craig DeYoung, Vice President of Investor Relations at ASML.

Joining me today from our headquarters in Veldhoven in Netherlands, is ASML's CEO Peter Wennink; and for his very first time, ASML's new CFO, Wolfgang Nickl. Welcome, Wolfgang.

The subject to today's call is ASML's fourth quarter and annual 2013 results. This call is also being broadcast live over the Internet at www.asml.com, and a replay of the call will be available on our website for approximately 90 days.

Before we begin, I'd like to caution listeners that comments made by management during this conference call will include forward-looking statements within the meaning of the federal securities law. These forward-looking statements involve material risks and uncertainties. For a discussion of risk factors, I encourage you to review the Safe Harbor statement contained in today's press release and presentation, found on our website asml.com and in our annual report on Form 20-F and other documents as filed with the Securities and Exchange Commission.

As a reminder, the length of the call will be 60 minutes. Now I'd like to turn the call over to Peter Wennink for a brief introduction. Peter?

Peter Th. F. M. Wennink

Thank you, Craig. Good afternoon, good morning, ladies and gentlemen, and thank you for attending our fourth quarter 2013 results conference call.

Before we begin the Q&A session, I'd like to introduce and welcome Wolfgang to his first ASML quarterly results conference call. Wolfgang joined ASML in December of last year from Western Digital, headquartered in Irvine, California, where he spent 18 years in various operations and financial roles, ending his time there as CFO. I'll save further introductory remarks for later, as many of you will get to meet Wolfgang for our upcoming post-results roadshows later this weekend next.

Now, Wolfgang and I would like to provide an overview and some commentary on our fourth quarter results and provide our view of the coming quarters. Wolfgang will start with a review of our Q4 financial performance with added comments on our short-term outlook, and I will complete the introduction with some further comments on our future business outlook and update you on our emerging and EUV product programs. Wolfgang, if you will?

Wolfgang U. Nickl

Thank you, Peter, and welcome, everyone. It's a pleasure for me to be here today and talk to you about our results and our guidance.

Q4 revenue came in very much as guided. Revenue was driven again by foundry shipments, plus an uptick in IDM shipments, which together produced 65% of total quarterly system sales.

Sequentially, memory system sales were up strongly as well, contributing the remaining 35%, driven by NAND.

During the quarter, we recognized revenue for our first NXE:3300 EUV system. Overall, last quarter of unit shipment were more balanced between ArF immersion and KrF than the prior quarter, indicating the addition of wafer capacity. As a result, the ASP of EUR 25.7 million for all systems sold in Q4 was lower than Q3 at EUR 28.2 million.

Service and field option sales again grew significantly to EUR 407 million from EUR 359 million in Q3. This includes Cymer service sales, as well as higher application sales.

Q4 net bookings and booked ASPs remained at Q3 levels of EUR 1.45 billion for 52 systems booked with ASPs at EUR 28 million. These bookings exclude EUV.

Memory bookings were present at almost 60% of total bookings value. Our system order backlog at the end of Q4 grew further to almost EUR 2 billion, excluding EUV. The key of change in our backlog profile appears at year end, driven by the increased memory bookings, leaving memory backlog at 55% of total versus 35% in the prior quarter.

For Q1, we expect total sales to be about EUR 1.4 billion with a gross margin of around 42%. Q1's margin is slightly negatively impacted by the product mix relative to Q4 2013, which is expected to reverse again in the coming quarters.

As discussed last quarter, we expect that Q1 will include revenue recognition for 1 EUV system, which currently is dilutive to overall gross margin. Excluding EUV, Q1 gross margin would be around 44%.

In 2014, we expect to deliver the remaining 8 EUV systems on order. In Q1, we expect to ship 1 EUV system. As with all new technology introductions, revenue recognition of EUV system is currently taking place after installation and customer acceptance. This revenue recognition method will continue until we demonstrate predicable installation timing.

Upon tool acceptance, about EUR 60 million will be recognized per system with about EUR 10 million per system being deferred until each system receives light source upgrade. We currently expect to recognize revenue for 1 system in Q1, 1 system in Q2, and 6 systems in the second half of the calendar year 2014.

As mentioned, we again saw a meaningful increase in our service and field option sales revenue to EUR 407 million in Q4. This was driven by a significant quarterly increase in performance-enhancing field option sales. We expect that quarterly service and field options revenues throughout 2014 will leverage about EUR 350 million per quarter.

Combined R&D for the first quarter will be about EUR 280 million. Other income with contributions from participants of the Customer Co-Investment Program will be about EUR 20 million.

Throughout 2014, we will walk quarterly R&D down with a targeted quarterly combined R&D spend of about EUR 250 million exiting 2014.

Combined SG&A is expected at about EUR 85 million for Q1.

We will again propose to increase our dividend by 15% to EUR 0.61 per ordinary share. During 2013, we bought back 4.6 million shares for a total of EUR 300 million, representing 30% of the EUR 1 billion buyback program announced last year for 2013 and 2014 combined.

ASML has now returned more than EUR 4.5 billion in dividends and share buybacks since 2006.

As our current business status, memory customers have somewhat aggressively moved to the lithography systems order front with bookings in Q4 of EUR 850 million, a significant increase over Q3, which was higher than Q2, leaving us with the memory backlog of almost EUR 1.1 billion. This backlog appears to be supported by tight mobile DRAM capacity and a steady growth path and stable market outlook in NAND Flash.

Foundry customers are evaluating 20-nanometer demand and are preparing for 16 and 14-nanometer node capacity additions later this year, and the MPU sector continues to add 14-nanometer capacity.

Our current business visibility along with a high level of business certainty provided by our current backlog and understanding of near-term demand, allows us to reiterate our revenue guidance of around EUR 3 billion for the first half of 2014, excluding EUV. This creates a solid basis for the remainder of the year.

Now with that, I would like to hand it back over to Peter.

Peter Th. F. M. Wennink

Thank you, Wolfgang. As Wolfgang highlighted, we've seen a significant shift in our backlog in favor of the memory sector, which will help drive our first half revenues.

This year, we will see the continued node transitions in NAND Flash and DRAM, as well as the addition of NAND Flash capacity, including an initial phase of 3D NAND manufacturing capacity build. Based upon customer interactions and research data, we expect low to mid 40% bid demand growth for NAND, and between 20% and 30% demand growth in DRAM, driven largely by the mobile application space.

First half, we'll see the foundry logic sector taking some time to absorb the rapid 20-nanometer capacity build that we witnessed in the second half 2013, before we expected 16- to 14-nanometer node ramp will start in the second half of this year. Like Wolfgang said, the 14-nanometer microprocessor manufacturing ramp started late 2013 and will continue throughout 2014.

The industry has entered a very interesting time as the challenges of device designed and performance have in turn, created significant manufacturing and cost challenges in pursuit of continued cost-effective strength.

In past conference calls, we have discussed our company strategy to address these issues, which is based upon execution around a very significant R&D investment for technology leadership. I think it's important to repeat 4 key areas of our strategy in the context of today's environment.

First of all, it's the continued significant performance improvement of our current, dry and immersion scanner architectures, which enabled the most cost-effective leading-edge imaging solutions, which will help our customers to address the imaging complexity involved in multipass patterning strategies at the current and the future of those nodes.

Secondly, the buildup of our application products or our Holistic Litho portfolio, enabling and differentiating further our dry and immersion products in support of sub-20-nanometer process control challenges, many of them related to the aforementioned multipass patterning strategies.

Thirdly, the introduction of EUV, the next-generation litho technology, the delivery of which will enable the continuation of Moore's Law, by providing large cost reductions, device power savings and further device performance improvements.

And fourth, the creation of a strong and highly competent R&D infrastructure to react to any relevant request from the industry, such as the potential 450 millimeter wafer size insertion.

The execution of this strategy enables us to support any short and long-term lithography need of the industry. This becomes particularly evident when we look at the challenges of our customers to address the 10-nanometer logic node transition. In order to control risk, customers may decide to initially focus on the use of our most advanced NXT:1970 immersion scanner, enabling execution of their complexed multi-patterning strategies supported by the complete suite of Holistic Litho products for maximum process optimization, thereby making this technology the current process of record.

As mentioned previously, the progress we have made in the development of leading edge immersion technology has culminated into our new NXT:1970 immersion system. This machine combines world leading performance of 250 wafers per hour with overlay achievements of less than 2 nanometer, which, when matched with a suite of application products, enables aggressive multi-patterning strategies, which our leading-edge logic device customers can safely choose to plan their sub-20-nanometer designs.

The NXT:1970 has seen widespread and quick adoption in recognition of its value, and we shipped 5 tools in the 3-month period post-introduction and have built a 14-system backlog with a value just over EUR 700 million. The value of this system to our customers is translated into a very strong average selling prices, currently and slightly above EUR 50 million while maintaining very healthy margins.

The adoption of our application products, again key to the performance at the 20-nanometer and below, is expected to grow again in 2014, likely to a level between EUR 500 million and EUR 600 million. While it is clear that the performance of our newest immersion systems help our logic customers, we also see significant demand for these systems from our memory customers. The complexity of 3D NAND also requires best available overlay and focus capability, and will furthermore require a significant number of immersion systems, which we are seeing now as 3D NAND chip manufacturing volume start to grow.

We are very satisfied with the acceptance of our NXT:1970 as the most advanced lithography solution in the industry and as a viable choice for the 10- nanometer logic node. However, we realized, together with our customers, that the continuation of further shrink, therefore, Moore's Law, ultimately and inevitably depends on the introduction of EUV. We are therefore working very closely with selected customers on the industrialization of our EUV technology, which we will very likely adopt when their economic targets for initial insertion of these more cost-effective solutions are actually met, thereby optimally addressing the economic and cost challenges of this node.

Regarding EUV, we're happy to report good progress. First 3 of 11 NXE:3300B scanners have been shipped. In the last 2 weeks, the fourth -- a fourth system has been signed off and is in the process of shipping, while 7 additional systems are in state of integration.

In the fourth quarter, progress has been shown. First, source code to all systems, enabling the increase of output power by 40%, showing up to 50 wafer per hour capability in our factory.

Second, source of time and cost of ownership improvements through institute cleaning technology of the EUV mirror in the stores. Third, defect management through the development and use of chemicals on the photo maps.

Despite our continued good progress, we have currently not yet reached a stage of industrialization, whereby our customers can confidently assign their critical layers for their most advanced future nodes to EUV. However, the progress that we see gives us full confidence that we will reach that stage before our customers put these nodes into full volume production. We therefore remain fully engaged in EUV insertion planning discussions with our customers, some of which at this point, have already qualified EUV for imaging at the 10-nanometer foundry logic node. Going forward, with the benefits of EUV clearly understood, insertion decisions will be based on the economics of EUV, driven by productivity, stability and cost of ownership improvements, in turn driving layer-by-layer insertion when these improvements become available.

With that, we would be happy to take your questions.

Craig DeYoung

Thank you, Peter. Ladies and gentlemen, the operator will instruct you momentarily on the protocol for the Q&A session. [Operator Instructions] Now Sarla, if you could give your instructions and then we'll have the first question, please.

Question-and-Answer Session

Operator

[Operator Instructions] Our first question comes from Sandeep Deshpande from JPMorgan.

Sandeep S. Deshpande - JP Morgan Chase & Co, Research Division

Peter, if I could ask, I mean there has been some confusion about what TSMC said about EUV insertion at the 10-nanometer node, and the use of and the potential economic gain to you or lack of from immersion [ph] at that node? Could you make a comment on?

Peter Th. F. M. Wennink

Yes. I think TSMC didn't say that much different than what I just said. It is clear that TSMC has found out with the introduction of our deep UV systems that the safest route for a 10-nanometer node is using the maximum capability of our NXT:1970 with the full suite of the Holistic Lithography products, which actually gives them a kind of a safe route into 10-nanometer. Having said that, it is a route with a very high intensity of litho tools, making this not a very cheap node. This is also why we, at the same time, work very closely with TSMC on the introduction of EUV. We actually have seen some very encouraging results, and that working on that particular no node will lead to a situation where when we meet the economic targets that TSMC has set for the introduction of EUV, they will immediately insert it, and this is where we are. And I think this is also what TSMC or Mark [ph] said at the call. He said the process of record is deep UV because we know what to do, but we need EUV because it is more cost effective when our economic targets are met and those are not targets, which are the 125 wafers per hour. Those are significantly lower targets because also TSMC needs to learn. And that is where we are. So it's a 2-track road, whereby deep UV is currently the safe route, not the cheapest route because the future is with EUV.

Sandeep S. Deshpande - JP Morgan Chase & Co, Research Division

And then Peter, maybe you have -- based on what you have said today, you haven't really changed what you said after your Q3, that there is this Plan B and that there are these 2 routes to the 10-nanometer node. Are you giving any roadmap on the EUV and saying that you are on schedule? Is that all that you're saying or is that something more to it as such really?

Peter Th. F. M. Wennink

You're absolutely correct. I think, we are not saying that much different than what we said 3 months ago. And what we are confirming is that the 70 wafers per hour, that is what we see happening this year and 125 next year, but I think that the difference between now and between the third quarter conference call is that we've made good progress on some important elements that drive the reliability and the stability of all the tools, which has to do with -- like I said in the introductory comments, have to do with the cleanliness of the EUV mirror and also which has to do with the development of pellicles for the photomask. So those are infrastructural issues and our industrialization issues that actually provide and approve of a good progress that we have made since the third quarter.

Operator

Our next question comes from Timothy Arcuri from Cowen and Company.

Timothy M. Arcuri - Cowen and Company, LLC, Research Division

Peter, I just wanted to zero in a little bit on what your EUV shipment plan is for this year. Last call, we were thinking that there'll be sort of 12 to 14 tools shipped this year. It sounds like there's going to be a few less than that. Yet, you also sound like you're a little more constructive on there being a round of follow-on orders for EUV. So can you help me bridge those 2?

Peter Th. F. M. Wennink

Yes. Let's first focus on the 11 orders that we have on the book and 3300 systems because I think that is what we said also in the introductory statements of Wolfgang and myself. The 11 tools on the 3300, the 3300 will be followed by the 3350. The 3300 are 11 tools, one of which will go to Research Institute and we will see -- not in the sales level but will be in the R&D line. So that's 10 systems left, 3 of those systems were shipped in the fourth quarter, one of them led to revenue recognition, so it leaves us with 9 systems. Now the way that we currently look at the shipment schedule, and as Wolfgang said, installation and customer acceptance happen after the shipment, we currently see or plan that 1 of those 9 systems will potentially move into 2015. So that's why we are left with 8 systems. So 8 systems we will recognize out of the 3300. Now having said that, we also started the integration of the first of 3350 systems. Now based on -- so we are preparing for 3350 shipments without having specific orders because we are confident that we will reach levels of productivity and reliability that will make customers take those systems. So depending on the progress that we are making over the next few quarters, it could very well be that we would target a shipment or a few shipments of 3350 also towards the end of the year. That would not lead to revenue recognition because that probably is going to be in 2015, but we are preparing to next to the 3300 be able to start shipping 3350s. And then revenue, we will probably see in Q1 of 2015. Does that explain a bit what you're asking, Tim?

Timothy M. Arcuri - Cowen and Company, LLC, Research Division

It does, it does, Peter. And then just a follow-on on R&D. R&D, I didn't think was going to go too much higher than 250 or 260 a quarter and yet now it's creeping up to 280. I know you're trying to bring it down throughout the rest of the year, but is the creep all related to the Cymer acquisition or is there something else happening there?

Peter Th. F. M. Wennink

Wolfgang, do you want to answer?

Wolfgang U. Nickl

Yes. I mean, it's just accumulation of a lot of investments that we're making, not only in the EUV including the Cymer, we're going to source there, but also in Holistic Lithography and imaging, quite frankly. We are happy to invest that money right now in order to solidify the roadmap and make sure that we deliver the value, but we also have a clear path to work it again down to the 250 level by the end of the calendar year, which by the way still means that we are investing about EUR 1 billion run rate into this business from an R&D perspective. It's a very conscious decision.

Operator

Our next question comes from Francois Meunier from Morgan Stanley.

Francois Meunier - Morgan Stanley, Research Division

The first one would be about your comment in the press release about the capability to expose up to 50 wafers per hour on the EUV systems. If you could explain really what it means, capability, is it because it's still a bit service, you call this throughput? That's my first question.

Peter Th. F. M. Wennink

Well, this is not a theoretical throughput. It's actually driven by the fact that we have the source power to go up to 50 wafers per hour. As a -- just as a side comment, we would like to now talk about wafers per hour because our customers are now talking about economic metrics and are driven by productivity and stability, and just the driver of the productivity is the source power. So 50 wafers per hour equates to roughly 70 Watts. That is what we have shown here for an extended period of time. That equates to up to 50 wafers per hour. So this is why we said it. And going forward, we would like to talk about wafers per hour because that's the metric that our customers are using.

Francois Meunier - Morgan Stanley, Research Division

Okay, that's very clear. But does it mean that the laser is in -- is at the Cymer premises, does the rest of the equipment is in Veldhoven?

Peter Th. F. M. Wennink

No, no. the lasers are also here. It must have been quite a time since you have been here because then you could have seen that we have here in the EUV base, it's all lasers and scanners. So it's all connected. And we do by the way R&D on both sides of the ocean, on the West Coast and here in Veldhoven, but they are all connected.

Francois Meunier - Morgan Stanley, Research Division

Okay. Now in terms of the learning curve with EUV, given that there is some slight delays again with the progress, basically the learning curve could take longer. Do you agree with the fact that if it's not used very much at 10 nanometer, I'm saying if, if it's only introduced at a later stage at 10 nanometer, then it will take longer to learn how to reach higher margins with EUV, so basically for potential peak earnings in 2019 for 7 nanometer?

Peter Th. F. M. Wennink

I've always said and it's true that the cost reductions is a matter of the learning curve. So whether the learning curve starts in year 1 or in year 2, we actually need about 24 months after volume, that way [ph] reduction that has not changed. I think from a learning curve perspective, there's not much difference. So we need the first 2 years after volume introduction to actually get the margins up to the corporate average, which is still the plan. So I don't see any -- not much difference there. And the learning curve is not only for costs, I think a big part of the learning curve is with the customers. Actually, if we would ship an EUV tool that could do 100 wafers per hour today, where the infrastructure of the customer side is they would probably have to throw away 80% of those wafers, because they're all infrastructure and their process are not able to deal with that productivity. That's why customers can settle for a lower productivity at insertion of a 10-nanometer node. Too much productivity is not good either because they need their process development in the factory also.

Operator

Our next question comes from John Pitzer from Crédit Suisse.

John W. Pitzer - Crédit Suisse AG, Research Division

Peter, I guess relative to the improvements that you guys are driving on double patterning, can you help quantify from here how much improvement is left? And is there -- I know you're trying to get away from wafers per hour, but is there a throughput number on EUV, where independent of double-patterning improvements, it still makes sense to move to EUV?

Peter Th. F. M. Wennink

That's a very good question. In fact, we have set, we want 70 wafers per hour this year, which is still our target and we're going for it and also 125 in 2015. Having said that, at -- for the insertion at the 10-nanometer node, the first phase of it -- customers actually settle for a lot less than 70 wafers per hour. Customers would settle for 30 or 40 wafers at 80% availability, much larger than 80 wafers at 40% availability. So it is what I said earlier in the -- as an answer to the previous question, it's not only us that needs to go through this learning curve. Too many wafer cycles were in a process that is not stable. In the customer factory, it's not good either. So customers had 30 to 40 wafers but give me a reasonable stability, yes? So this is what we are focusing on. So the productivity in the first phase, which will be 6 to 12 months, is not that important. Now once you start ramping -- and don't forget the ramp of 10 nanometer is 2016, 2017, yes? Then they need that 70 or that's 100 wafers per hour. So you need that initial learning period, not only for us but also for the customer. So you will very likely see insertion point, whether productivity is not used at 70 wafer, it is used at a lower level, but they want stability. Is that clear, John?

John W. Pitzer - Crédit Suisse AG, Research Division

Peter, as my follow-up -- that's helpful. As my follow-up, if memory serves me correct, this time last year, you actually gave a full year revenue number for the company or target. And today, you're sitting here giving us only a first half. Independent of EUV adoption, is there something different in the CapEx environment this year that's giving you less visibility or less comfort, giving us the full year revenue number?

Peter Th. F. M. Wennink

The main difference is there's a new CEO. But it's also because the visibility is what it is. And I feel that they quantitatively very comfortable about 2014, because that the 3 core segments of the business, logic, memory and microprocessors, are all in a very healthy state. Memory, we talked about in the introductory comments. Microprocessors is now starting to ramp new nodes at the end of last year and is fully ramping in this year. And logic is as strong in 2014 as it was in 2013. So qualitatively, I'm very confident and with the quantitative guidance of the EUR 3 billion, excluding EUV -- if include the EUV, it's between EUR 3.1 billion and EUR 3.2 billion. If you combine the 2, I feel very confident about a solid 2014. But things that I cannot see full clearly, although I do have a gut feel, I'm not going to comment on quantitatively.

Operator

Our next question comes from Stephane Houri from Natixis.

Stephane Houri - Natixis S.A., Research Division

I have a question about lithography intensity because when you will be switching -- when the industry will be switching from 20- to 10-nanometer In the past you used to say that from 28- to 20-nanometer, the litho intensity was growing by 1.7 to 1.9. Do you have the same kind of metric for the 20- to 10-nanometer transition?

Peter Th. F. M. Wennink

It's a bit difficult to answer at this moment in time. What we are seeing because we -- it is a bit different customer per customer. We're still compiling that information and I don't feel comfortable enough to give you that particular, very precise answer. It might be next quarter or conference call, I might be able to actually do that. It will go up, that's for certain. With the number of layers go up and also the -- from the double patterning process will go to multi-patterning process. We know for the fact that for short-term in all layers quadruple patterning is actually needed. So to give you a precise number is we're still working on it and I'd like to give you it, but I'm not going to speculate or to guess what is going to be significant. So the litho intensity for the 10-nanometer node will definitely go up.

Stephane Houri - Natixis S.A., Research Division

Okay. So I will wait for next quarter. Maybe a follow-up, you talked about the R&D increase. What can we say about SG&A, because they have increased since you acquired Cymer? Is there any cost-cutting measure there?

Wolfgang U. Nickl

I would assume for now that it's relatively stable for the time being.

Stephane Houri - Natixis S.A., Research Division

Okay. On the EUR 90 million level?

Wolfgang U. Nickl

EUR 85 million is what we said, yes.

Operator

Our next question comes from Srini Sundar from Summit Research.

Srinivasan Sundararajan - Summit Research Partners, LLC

Just to -- could give -- please give some color on what was involved in the acceptance of EUV tool in the Q4 quarter?

Peter Th. F. M. Wennink

What was involved in the meeting is -- basically this meeting...

Srinivasan Sundararajan - Summit Research Partners, LLC

Meeting, what were the criteria that the customers felt was sufficient for them to...

Peter Th. F. M. Wennink

There were technical criteria and the most important one was that the customer ran wafers what we call OPC wafers with the -- let's say with the target features of the mask being printed on the wafer, giving the imaging results confirming the 10-nanometer device requirements. And that's what actually happens. What -- it was actually asked for is that I want a machine that an integrated machine with an EUV source and an EUV 3300 scanner with the maximum optical capability that will give us a 10-nanometer feature on the wafer using a 10-nanometer OPC wafer. And that's what it is. And then they run a patch of those wafers, and that need to be batches where everything was fine and which was fine and was passed. That's why we say the qualification also in the introductory comments on the imaging results were 10-nanometer as actually being done. We are qualified from an imaging point of view on 10. And that's when the customers say I have a tool, with which I can make my 10-nanometer device. And then I will insert that tool when the economics, i.e., the stability of the tool, availability of the tool are actually met. And that will I think happen both over the next foreseeable future. Because we worked hard on it, and we are very close with our customers. We've made good progress and we know what to do. So I think the acceptance of that first tool showing those OPC wafers is also -- was a big thing.

Srinivasan Sundararajan - Summit Research Partners, LLC

Just a short follow-up, what percentage of memory orders were based on 3D NAND -- wafer 3D NAND?

Peter Th. F. M. Wennink

We don't give that kind of detail what is different on 3D NAND? But NAND shipments, as we all know, there's one major outfit out there in the world that is taking tools in Asia and that is forward NAND. So we're not going to be specific on this. So you can easily deduce how many systems and tools were taken by that particular customer, which I'm not going to do. So...

Operator

Our next question comes from Gareth Jenkins from UBS.

Gareth Jenkins - UBS Investment Bank, Research Division

I guess we've spoken a lot about logic incision and 10-nanometers, can you talk about EUV and DRAM 18-nanometers? And what the progress there is and what your expectations are with those of 1 to 2 layers in this 2015 timeframe?

Peter Th. F. M. Wennink

Yes. On DRAM, we actually -- the 19-nanometer node that's what going to be the customers are looking at, which is probably going to be done with deep UV and double patterning. It's the 16-nanometer DRAM node, which is the node where we have to -- where we actually have to use EUVs. So with our DRAM customers, this is the node that we're talking about the EUV -- about the EUV insertion. And that actually means that we will have to see some level of interaction with our customers on EUV in the course of this year to do the first development steps with a 16-nanometer DRAM node.

Gareth Jenkins - UBS Investment Bank, Research Division

And as a follow-up, completely unrelated, Peter or Wolfgang indeed. The gross margins on the 1970, I'm just wondering if you could help us quantify, obviously, it's a very helpful color on the ASP and the backlog. Could you just help us with the gross margins on -- about tool? I presume given the throughput at the decent level as well.

Peter Th. F. M. Wennink

The only thing I can say is that there's gross margins are higher than the -- currently than the corporate average, because of the higher percentage of application products that are in there. And the application products are software and hardware related. It's YieldStar combined with the software control products, which carry good gross margins. So the 1970 is not only a tool that has the highest average selling prices, but very healthy margins. The only indication I can give use as you can understand, I'm not going to give in the public for all the margin profile on the tools specific, which of course a lot of customers would have like and especially the purchasing departments, but it's really you have to look at the margin development of the company. I think Wolfgang made it clear in the press conference this morning, when you look at the fourth quarter gross margin and you take the EUV impact out and the let's say non-cash accounting adjustments for Cymer, then you will see that actually the fourth quarter was above our last peak gross margin profile. So if you do the math then Wolfgang help you or one of the IR guys can help you go through those numbers. But that gives you an indication of what it means for the company.

Operator

Our next question comes from Mehdi Hosseini from SIG.

Mehdi Hosseini - Susquehanna Financial Group, LLLP, Research Division

Peter, when you look at what your foundry customers are doing, can you help us understand what the -- if you were to have the 70 wafers per hour this year and the 120 next year, what is the number of layers at 10-nanometer that your foundry customers are going to adopt EUV for, based on 70 wafers per hour this year?

Peter Th. F. M. Wennink

70 wafers per hour this year. Well, like I said earlier, it's -- if it's 70 wafers per hour, the question is where they're going to use 70 wafers per hour. I think they are going to use the tool as much as they can. And I think they will be looking at insertion points, where the economics are met for the layers that they are targeting at. Now, per customer that is different, that it can vary from 2 to 4, a 4 maximum, 2 I think is the most reasonable assumption. And we have the 70 wafers per hour capability, but it means that if they do their learning, they will probably start at a lower level because they have to go through their own process and own learning also. So what we are really looking at is, I would say 2 on average, could be 3. It's been not all the same, maximum of 4, but that's I would say max. So if you want to go average, 2.

Mehdi Hosseini - Susquehanna Financial Group, LLLP, Research Division

Okay. And then your foundry customers are also talking about escalating costs and especially in the absence of EUV. Would there be a scenario in 2016, where the key customers come back to you and say, look, you got to lower the immersion ASP. Is there a scenario where you will actually be pressured by immersion ASP, as a way to alleviate escalating costs in absence of EUV? And again, this is just one scenario of many.

Peter Th. F. M. Wennink

Well, if you say in the absence of EUV, I don't think that first that is not realistic scenario, because in 2016 there will be EUV. But if you say suppose there's not going to be EUV, then there's -- it's very hard to imagine that there are going to be any other tricks up their sleeves to go from 10 to 7. If they do, it's going to be extremely difficult from a cost point of view to keep using deep UV. So I think that -- what you're basically then saying is that Moore's Law will slow down. Now -- so what we are seeing today, I think is more realistic. Yes, it's EUR 50 million-plus price, but it does provide the value to do 10-nanometer shrink. So obviously -- and the customers make those calculations better than we do, that they know what their litho spend is going to be, they know what their production spend is going to be, and they must be convinced that the value that they get out of the shrink to 10-nanometer is significantly higher than that added cost, otherwise they would not do it. So it's obvious that there is a cost increase. And this is the reason why they are working so closely together with us on EUV. Because ultimately you could say that 1970 multi-pass patterning, Great Bridge solution for the 10-nanometer node, there is a kind of a process of record. But once we made the EUV targets, yes, the economic insertion targets, they will go EUV.

Operator

Our next question comes from Jerome Ramel from Exane BNP Paribas.

Jerome Ramel - Exane BNP Paribas, Research Division

They were the -- a few months ago, you had a vision of EUV market in 2016 potentially being at 60 units. Why don't you describe today with on average maybe 2-layered [indiscernible] EUV. What is the more realistic scenario for 2016? And a follow-up will be; if that the tuned EUV market it being, whatever, 30 or 42, in absolute term that are turn what is the impact on your addressable market?

Peter Th. F. M. Wennink

Yes. we said 2015, 12 to 15 units, where with some upside, if we would see a faster adoption, if it would meet the economic targets of our customers faster. We have the capability of doubling that in 2016 and doubling that again in 2017. And our scenarios, our simulations there is that we run based on the customer roadmaps lead us to believe, also after discussions with the customers, that we are going to use that capacity if we meet those economic targets. So 70 wafers per hour this year, and 125 next year. And we can meet those economic targets in terms of availability and in terms of the stability of the tool, I think we all were going to use our capacity.

Jerome Ramel - Exane BNP Paribas, Research Division

Okay. But realistically, what could be the number of total units you targets for EUV in 2015?

Peter Th. F. M. Wennink

I said 12 to 15 units in 2015, doubling that in 2016, doubling that in 2017. I think that's what we said last quarter also and that is still the case.

Operator

Our next question comes from Mahesh Sanganeria from RBC Capital Markets.

Mahesh Sanganeria - RBC Capital Markets, LLC, Research Division

Here is the -- I have a question on your booking composition. It's a little surprising. Your memory booking is pretty close to the peak and second half of 2010, and foundry bookings are much lower and I would've expected that it will be, considering that 20-nanometer capital intensity at foundries are so high. So maybe on both positive and negative surprise on memory from my perspective, if you have any comments, that will be helpful.

Peter Th. F. M. Wennink

Yes, I think memory in the bookings is indeed high, but I think it is understandable. If we look at especially the DRAM situation, where when we look at the installed capacity and the reduction of that capacity due to -- when I talk about the wafer out capacity in the DRAM industry, it has come down with DRAM prices going up. And the reason why its total capacity went down is because there were conversions from DRAM into NAND in 2012 to early 2013. And also the complexity of the DRAM chips has gone up with additional ArF immersion layers, which also means that your wafers outgo actually down. No new DRAM fab is build. DRAM prices going up to very healthy levels. It's not a surprise that this comes back. And at the background, NAND being pretty stable, they are all going to the next node. So when you look at the value, it's a bit deceiving because it's 3 to 4 years ago, while if you look at the units that's have been booked and the average selling price is different between 2010 and 2013, '14 is quite different also. So from a unit point of view, it's not a crazy number, but that's for memory. For foundry, foundry was a bit lower in the fourth quarter. And if you look at the backlog, the end-user is only about 13%. So it means foundry still has to come and this is exactly what is the case. We had one foundry investing pretty significantly in 2013, ramping before the others, which the others are -- now will follow, simply haven't met than all translated it into the paper that is called an order, but those will come. There's no doubt.

Mahesh Sanganeria - RBC Capital Markets, LLC, Research Division

Okay, that's very helpful. And one follow-up question on EUV shipments. Are you going to ship all 4 with upgradable to MOPA PrePulse in field? Or are your -- the later shipments will go with mobile proposals? Is there a difference in shipping to different customers for the 3300s?

Peter Th. F. M. Wennink

Yes, there is a difference. So some take no more and with an upgrade in the field, others want to wait until MOPA PrePulse is actually there. So there's a difference between customers. And as you can understand because of the limited number of customers, I'm not going to give you the details, who does what, but yes, there is a difference.

Mahesh Sanganeria - RBC Capital Markets, LLC, Research Division

And so that's the driving that you have less revenue recognition in the first half and there is a -- it's for shipments and then there's a gap and then once that MOPA PrePulse has to come available than shipment more? Okay.

Peter Th. F. M. Wennink

Yes. That's one of the reasons. Yes.

Operator

Next question comes from Sumant Wahi from Redburn Partners.