Ladies and gentlemen, thank you for standing by. Welcome to the ASML second quarter results conference call on July 15th, 2015. Throughout today's introduction, all participants will be in the listen-only mode. After ASML's introduction, there will be an opportunity to ask questions. If the participant has difficulty hearing the conference, please press star zero for operator assistance. I would now like to turn the conference over to Mr. Craig DeYoung. Please go ahead, sir.
Thank you, Aaron. Good morning and good afternoon, ladies and gentlemen. This is Craig DeYoung, Vice President of Investor Relations at ASML. I'd like to inform you that today we're conducting our conference call from two locations. We're participating here in San Francisco at the SEMICON West show, and with me is Peter Wennink. In our headquarters in the Netherlands, we'll be joined or are joined by Wolfgang Nickl, ASML's CFO. I mention this just in case there is a technical problem. When one of the calls drop, we'll get back to you as soon as is technically possible, although we don't anticipate any problems. As a reminder, the subject of today's call is ASML's 2015 second quarter results. The length of the call will be 60 minutes as usual.
This call is also being broadcast live over the internet at www.asml.com. A replay 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 the conference call will include forward-looking statements within the meaning of the federal securities laws. 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 materials found on our website and in our annual report on Form 20-F and other documents as filed by the Securities and Exchange Commission. Now with that, I'd like to turn the call over to Peter for a brief introduction.
Thank you, Craig. Good morning, good afternoon, ladies and gentlemen. Thank you for joining us for our second quarter 2015 results conference call. Before we begin the question and answer session, Wolfgang and I would like to provide an overview and some commentary on the recent quarter and provide you our view of the coming quarters. Wolfgang will start with a review of the second quarter financial performance with some added comments on our short-term outlook. I will complete the introduction with some further comments on the current general business environment and our future business outlook. Wolfgang, if you will.
Thank you, Peter, and welcome everyone. For Q2, our net sales came in at EUR 1.65 billion with system sales again nicely balanced between memory and logic. Memory represented 47% and logic represented 53% of system sales. Our system sales included one EUV tool. Service and field option sales came in at a very healthy EUR 520 million, driven by strong demand for field options and upgrades. Gross margin for the quarter was 45.6%, above our guidance and impacted by revenue recognition of one EUV tool, which shipped in Q1. R&D expenses came in at EUR 267 million and SG&A expenses came in at EUR 88 million. Both slightly above our guidance, driven by accelerated product development and infrastructure projects. Our effective tax rate for the quarter was approximately 11% of pre-tax income. Turning to the balance sheet.
Quarter-over-quarter cash equivalents, and short-term investments dropped to EUR 2.52 billion from EUR 2.84 billion at the end of the prior quarter, in part due to a EUR 302 million dividend paid in Q2. We also paid EUR 166 million during the quarter for repurchased shares. Regarding the order book, our Q2 non-EUV system bookings came in at about EUR 1 billion. This was better than we expected in both memory and logic. With the additional booking of six EUV systems, our total system bookings during the quarter were EUR 1.5 billion. We finished the quarter with a strong overall backlog of just over EUR 3 billion. As it relates to our EUV backlog and our backlog in general, I do want to point out a change in our definition driven by this quarter of EUV orders, which allows us to make these orders visible to investors.
Our prior definition included a timeframe of requested system deliveries within one year. That time limit has been removed as the current delivery lead time on new EUV systems is greater than one year. With that, I would like to turn to our expectations and guidance for Q3 and share our view on the balance of 2015. We end the Q3 with a strong and nicely balanced system backlog. Adding continued strong service and field option sales at the level of Q2, we expect Q3 revenue to be between EUR 1.5 and EUR 1.6 billion. Q3 will not include any revenue recognition for EUV systems. Based on expected customer and product mix, we expect gross margin for Q3 to be around 45%. R&D expenses for the third quarter will be about EUR 275 million. SG&A is expected at about EUR 90 million.
R&D and SG&A are both up slightly from the Q2 levels, driven by investments in future technology and continued effects from a strong US dollar. As to the rest of 2015, we expect less tapering in our memory business in H2 than we anticipated last quarter. This is driven by continued capacity additions in two new DRAM fabs, and by some additional capacity needs in the one existing volume 3D NAND fab. We continue to see a stable logic business throughout 2015 in support of current FinFET node DRAMs and next node developments. Our service and field option sales exceeded the EUR 500 million mark in Q2, and we expect that level to continue throughout H2 due to increased adoption of our Holistic Lithography products and the purchase of system node enhancement packages supporting customer node migrations.
We expect to see Q4 revenues holding up at the Q3 level with some upside opportunity possible due to logic's need for additional advanced node development tools. This means that we are on track to another record sales year for 2015. Peter will talk more about EUV shortly, but I would like to make a few points regarding 2015 EUV shipments. As most listeners are aware, we continue to show great progress in improving on key EUV performance metrics related to productivity and availability, with unchanged targets of improving both further throughout this year. We are delighted about the receipt of purchase orders for 6 EUV production-ready tools during last quarter. Of the 6 tools, 2 tools are scheduled to be delivered this year, and the remainder will ship from next year on. This is an indication of EUV's continued progress along the new technology adoption curve.
For 2015, we now see the opportunity to ship 5 NXE:3350 production-ready systems, supported by 4 system orders and with close customer interaction on the fifth system continuing. Also, we plan to ship 1 additional NXE:3300 before year-end. With the 1 NXE:3300 that we shipped earlier this year, we therefore now forecast a total number of 7 EUV shipments for the year. 1 of the 2 remaining prepaid NXE:3300 is planned to be upgraded to an NXE:3350 or higher, and for the other system, discussions with the customer have not been conclusive as of today. We expect the EUV revenue for the year to be limited to the 1 NXE:3300 that we recognized during Q2. Revenue recognition for the remaining shipments this year will depend on the achievement of certain performance milestones and attribution of overall revenue to the system itself and related services.
Revenue recognition will vary from customer to customer based on the specific terms and conditions in the respective agreements. We will provide our assumptions for timing of revenue recognition after we have shipped incremental EUV systems. Once the performance of EUV technology is matured, we expect to recognize revenue in the same fashion as we do in our DUV business. Finally, as an update on our capital return program, our proposal to increase our dividend by 15% was accepted by our shareholders at our annual meeting in April, resulting in a payment of EUR 302 million in dividends in Q2. In addition, relative to our 2015, 2016 buyback program announced in Q1, we now executed total repurchases for the first half of 2015 of EUR 285 million at an average price of EUR 96 per share. Now with that, I'd like to turn the call back over to Peter.
Thank you, Wolfgang. As Wolfgang highlighted last quarter, we had an expectation that our strong second half of last year would continue into the first half of this year. We can now say that the strength in the first half of this year is expected to continue through the second half, leading us to a record revenue year in 2015, as Wolfgang mentioned. This is clearly supported by our strong service field options and upgrade business. Our customers appear to be preparing for continued growth in their business. As we see, for example, the two new DRAM fabs that Wolfgang mentioned continue to install capacity. In addition, we've all heard announcements of the 2 new NAND fabs likely to begin taking manufacturing equipment in 2016 for volume build of Vertical NAND.
On top of this, we see two new foundry fabs beginning to take tools in the second half this year in support of advanced FinFET process node ramps and early development of next node logic devices. Before I move on to EUV, I'd like to share some highlights of our Deep UV and holistic lithography programs. Firstly, with respect to our Deep UV program, we will begin shipping this year our NXT:1980i immersion product, which is capable of 30% improved overlay accuracy versus our prior offering, along with the world's first 275 wafer per hour immersion capability. This will help our customers to deal with the increasing cost of complexity relating to multiple patterning strategies. Next to these platform improvements, we are increasingly focusing on the availability performance of our systems.
I'm glad to report that this year, the average availability of our worldwide installed base of more than 300 NXT tools increased to above 96%. Again, an important driver of affordability. Secondly, I'd like to highlight that our holistic lithography products now have a 100% attach rate on newly sold immersion systems. Our integrated metrology systems, which we call YieldStar, is now broadly accepted with more than 250 systems in the field. Our holistic lithography concept of providing imaging, measuring, and modeling capability allows unique support for controlling today's and tomorrow's most challenging advanced patterning processes across all industry sectors. Thirdly, we feel that we are strongly positioned with our wide range of field options and upgrade products.
These products allow for extraction of maximum performance from the install base of our lithography tools, and extension of that performance over multiple nodes allows for a more efficient capital use of our lithography equipment. As witnessed this past quarter, demand for this product continues to grow, with combined service, field options, and upgrades crossing the EUR 500 million mark last quarter for the first time ever, as Wolfgang mentioned. It's expected to continue to stay around this level for the next few quarters. Finally, on EUV. As most of you are aware, we continue to demonstrate real progress against our targets in system throughput and system availability. These are the key metrics of performance that drive new lithography technology adoption once imaging and overlay performance are demonstrated and accepted.
As mentioned on many prior occasions, our focus in 2015 is on improving EUV stability and availability with continued steady progress on productivity. As it relates to productivity, during the first quarter, we have demonstrated 1,022 wafers exposed in a 24-hour period at one customer site with an 80-watt configuration. Our target is to be able to repeat this at several customers and at several sites. Upgrade of systems in the field to 80 watts has virtually been completed in the second quarter. This gives us further confidence that the customer targets are realistic and should be met. In addition, we have demonstrated 130 watts dose-controlled source power in our facilities, enabling further productivity improvements. On system availability, which is targeted at 70% by year-end, we've been running at a 55% average on customer-installed systems.
With current upgrades, multiple sites are now achieving average availability of greater than 70% for one week, with one customer achieving a 70% average over a four-week run. Excellent progress on this important front. With the first two orders for our fourth generation NXE:3350B production tools announced late last year and entered into our backlog, this past quarter we not only signed a volume purchase agreement for a minimum of 15 tools with a U.S. customer, but have also taken a six-tool purchase order against that agreement, which has now also entered our backlog. As Wolfgang mentioned, also discussions are continuing with multiple customers on the exact timing of their requirements for more EUV systems. Clearly the EUV adoption curve will be driven by the continued progress we are making on these key EUV performance metrics.
Question of how many EUV orders we will see during the next few quarters can only be answered by relating the aforementioned progress to each customer's specific roadmap, adoption drivers, and risk appetite. These are, of course, different per customer, but generally we can say that our logic customers are most aggressive given the complexities of multiple patterning strategies, followed by DRAM and ultimately NAND customers. In any case, in order to deal with this eventual and inevitable demand, we recently opened our new EUV factory, which will enable an output of 24 systems by 2017 and potential further output capacity growth to 60 systems in the years to follow. With that, we would be happy to take your questions.
Thanks, Peter, and thanks, Wolfgang, as well. Ladies and gentlemen, the operator will instruct you momentarily on the protocol for the Q&A session. Beforehand, I'd like to ask, as always, that you kindly limit yourself to one question with one short follow-up if necessary. This will allow us to get to as many callers as possible. Aaron, could we have your instructions and the first question, please?
Of course. Ladies and gentlemen, at this time we'll begin the question and answer session. If you have a question, please press star one to register for a question and star two to withdraw your question from the queue. Your questions will be answered in the order they are received. If you're using speaker equipment today, please lift your handset before making selections. One moment please for the first question. The first question comes from Sandeep Deshpande. Please state your company name followed by your question, sir.
Yeah. Hi, Sandeep Deshpande, JPMorgan. If I may ask on EUV, in terms of what you're discussing with your customers, you've given some indication of the U.S. customers' orders, which are going to ship into next year. Do you have any sort of visibility about the other major customers and how EUV shipments would go into 2016? Clearly, you're going to have capacity now to do that given the opening of your new EUV factory. Secondly, my question is on OpEx. You've had some impact to your OpEx from the euro-dollar exchange rate. Is there any further impact to the OpEx from the euro-dollar exchange rate, or is the current OpEx reflecting what the current euro-dollar exchange rate is? Thank you.
Sandeep, I will answer the first question. Wolfgang will take the second one. Your question, do we have visibility on the EUV shipments in 2016? Well, part of that visibility is pretty certain because we've received the orders, and we know what to do, and I think they're currently planning on the 3300s that will be upgraded to 3350 performance that have not been shipped yet. That could be part of the 2016 shipment pattern also. I would say, like I said in my introductory statements, the question of how many EUV orders we will see during the next few quarters, you should really see in the context of the progress we are making on the availability and the productivity, vis-à-vis what customers' specific roadmaps are and what their risk appetite is and their adoption drivers.
I think we've made good progress on availability and on productivity, which is evidenced by our current performance after the 80-watt upgrade. I would say, if you would have a conservative view of 2016, I would say, you should pencil in the same number of EUV systems that we've seen in 2015, which is a combination of the 3300s and the 3350s. If you're a bit more positive, and that will have to show over the next two quarters, as in the second half of this year. That could run up, and that could run up to approximately 10 systems. It's anywhere in between what we ship today and 10 units. That is what our best estimate is today. Don't forget that the real ramp is in logic first, which is basically 2018 production for our customers, which means 2017 shipment.
The real shipment ramp will be in 2017. 2016 will be, you could say, an extension of the 2015 development work that our customers are currently doing.
Peter, just to clarify, you're talking about shipments and not recognition, correct? When you say that 7 to 10 or whatever, that is shipments and not recognition.
Yes. What is important is that those tools get shipped. Once they're shipped, and it's a EUR 100 million tool, customers will use them. Usage is extremely important in the learning curve and in the adoption of EUV technology. Shipment is very important. I would say the order of priority is to ship them first, then get the cash in, which is part of the orders, then we do rev rec, which is, you could say, a discussion we will have with the accounting community. Wolfgang can go into that into more detail. It just means that you need to go through the list of agreed performance criteria, and you need to tick every box. In that order, we need to ship first. That's the most important for learning curve and for the adoption, and then we get the cash.
Revenue recognition will be with the auditors.
Thank you.
Sandeep, I'll go into the OpEx question. If you look at where we started the year in R&D around the EUR 260 level, then in SG&A, in the low EUR 80s or so, now we're guiding to EUR 275 and EUR 90 respectively. It's indeed true that the majority of that jump comes from FX effects. If you look at R&D alone, I think we have about 25% or so of our spending in the U.S., with exchange rate going from the 130s to 110. That probably accounted for two-thirds of that jump from EUR 260 to EUR 275. The balance is really in accelerating investments in future technology before we get some of the wind down of investment in existing technologies. SG&A, very similar, an FX impact, of course. We are investing in certain areas, for instance, in business development teams for field options and upgrade sales.
We are completely focused on the 2020 model that we published last November, we plan on being at about 13% of revenue in R&D and about 4% in revenue for SG&A. I think we are in good shape. I want to mention on the FX effect that, of course, as a company, we also have positive effects from FX because we're selling field options and also services, and particular services abroad in local currency. That's about 10% of our revenue in total. There we have a positive impact on our P&L as well. Net, there is probably one percentage point on the net income that we were impacted with from FX. I hope that clarifies it, Sandeep.
Thank you very much.
You're welcome.
The next question comes from Mr. Kolsaat. Please state your name and your question. Go ahead, sir.
Yeah. Thanks for taking my question. It's Kai Kolsaat at Merrill Lynch. I had two. The first one was on EUV. Peter, just to make sure I understand correctly, you're saying that demand and orders will depend essentially on the performance over the next six to nine months. My understanding is the availability is already moving up to 70% now with what is essentially an old tool, and the 3350 will have, I guess, material improvements in availability in source power. My question is, by the time these tools ship, why would the major logic customers, when they see those improvements compared to the 3300 that they're using today, not order those tools? Is there anything really that you think the 3350 will not be able to deliver, let's say, in the first 12 months of shipments that your main customers need?
That would be the first question. The second one would be for Wolfgang, just on the balance sheet. From what I can tell, it's been not a great cash flow quarter because obviously the dividend's been paid, but you're still above that EUR 2.5 billion threshold, I think, that you've mentioned to us. Is it still the right for us to assume that any cash that you generate essentially from here onwards will find its way back to the equity holders, either through buybacks or the divvy? Thank you.
Yep.
Okay. Yeah. Perhaps the shortest answer is probably on the second question, why don't you do that?
Yeah, I'll get started, Kai. Yes, the assumption is right. We haven't changed anything on our returns policy. We have established that minimum growth cash balance that we think that we need to run the business. It's around EUR 2.5 billion. Last quarter, indeed, the cash balance went down by some EUR 300 million. Like you mentioned, it was purely a function of share buybacks and dividends. There was a nice free cash flow in the quarter as well. The policy remains intact. We look first at dividends on excess cash over the EUR 2.5 billion and our goal there is a stable or preferably growing dividend. Then the balance we used for share buybacks. We have a EUR 1 billion, thereabout, program announced in Q1, and we have executed EUR 285 million against this program in the first two quarters. Same policy as we had before. No change.
Good. On your question, basically, say with all the progress that we have made in the introduction of the NXE:3350B, why wouldn't customers order those tools? That is a very good question, and I think I would ask the same question. Why wouldn't they order those tools? Because we are making very good progress. I think the answer is, we've had, over the past couple of years, we had many expectations and promises on the EUV introduction. We've, as a company now, basically over the last one and a half to two years said, we're going to tell you, and to give you EUV guidance on what we actually see. What we are seeing is a significant improvement, I would say, in availability and in productivity. Your assumption is why wouldn't they come? It's a very fair one.
On the NXE:3350B, we still need to ship the NXE:3350B. We have the expectation that the NXE:3350B will perform better than what we have in the field today because you're correct, it will be higher productivity, will be better availability with in-situ cleaning in there. This is indeed our expectation. What we do not want is to give you all our expectations. We would like to guide you on what we actually see. Like I said, if you want to be conservative and think that progress is going to be slow, you say, well, perhaps 2016 is going to be an extension of what we saw in 2015. If you're a bit more positive, you would pencil in higher numbers for next year. If you have higher numbers for next year, you also need to book the orders.
Given where we are today, and you know me a bit, I'm more an optimistic person. We are not here yet to give you specific numbers on orders and when orders will come and when we will ship.
Great. Thank you.
The next question comes from Muse. Please state your company name, followed by your question.
Yeah. Hi, this is C.J. Muse with Evercore ISI. I guess first question, you talked about potential upside in Q4 led by logic. Curious if you could provide a little more color there as to what the main driver and magnitude could look like. I guess as a follow-on to that, as you see the 1980 layer in, presumably beginning in Q4 and then throughout calendar 2016 and beyond, how should we think about the uplift for ASPs?
Yeah. On the potential upside in the logic, that's really driven by the fact that when you look at the announced node-to-node transitions, we make it specific and say that's the 10 nanometer node transition. What we have witnessed is also true on the 14 and on 20. We see an extended period that customers take for development and for qualification of those nodes as it relates to the timing of the ramp. It just takes longer to go to the ramp of that node, which effectively means if the ramp of that node is a given, then you need more development time, brings the shipment of the early development tools back in time. You need them earlier. This is where we see the upside.
When we talk to customers and they talk about their newest node, they just take more time to do the development and the integration work. That means that is an upside that we are seeing for the second half of the year. It's really advanced technology transitions, earlier acceptance and earlier delivery of those tools because they need more time.
I can cover-
Sorry, you want to cover the ASP, Wolfgang?
I can cover the ASP. I can't give you an exact number, but you should assume, C.J., that the ASP is up. You see it in our press release, for instance, the tool, the 1980 provides for a 30% improvement in overlay and all the better throughput of 10% to 275. That's significant value to the customer, and as like we always do, we share that value, and so you should assume that tool has a higher ASP, a few EUR million than the 1970, yeah.
Great. I guess as a quick follow-up on EUV, I know there's a lot of timing uncertainty, and you guys are more focused on shipments and the ramp in 2018. Wanted to get your view on what would be sort of a low-end number and a high-end number in terms of revenue recognition on the EUV side in calendar 2016.
Well, I'll take that, it's Wolfgang. I won't be able to give you a number today. I can explain to you a little bit about the complexity and what we are going to do in the future to provide you some guidance. Like we said, we expect to ship seven systems this year. Of that, only one we'll recognize. You have the opportunity to get some revenue recognition in 2016 of these tools. We will ship other systems like, for instance, the upgraded 3300, and you can expect to get some revenue recognition there as well. As it relates to the 3350, as we launch this tool, the revenue recognition will actually depend on a few factors. It will, of course, depend on shipment acceptance, but it will also depend on certain performance milestones.
We have the revenue, and we got to attribute it to the system and related services. It's fairly complicated, and it's actually then made more complicated that every customer agreement looks different from each other. It's very difficult if you can't go into the detail on all these schedules and milestones to give you a forecast. As you can imagine, they're also not in cast in stone from a timing perspective. What we're going to do is we're going to give you an indication as we start shipping these tools, what the timing and the EUR amount of the revenue recognition will be. The first one of these could go as early as this quarter.
You could expect us in Q4 to talk about this particular tool, how the revenue recognition will go, as we ship four tools in the fourth quarter and January, I'm certain that we'll give you some specific guidance, can't do it today. I'd also like to mention, this is a temporary situation. Ultimately, the objective is, of course, as the technology has matured, that we go and recognize revenue in the same fashion as we do it for our EUV system. Sorry to not be specific about the numbers, that's kind of the complexity behind it, you will see us giving you more specific guidance as these tools actually ship.
Makes sense. Thank you.
You're welcome.
The next question comes from Mr. Gareth Jenkins. Please state your company and follow by your question.
Yeah. Thanks. It's Gareth Jenkins, UBS. A couple if I could. I just wanted a bit of clarification, Peter, on something that you've talked about. Just on the 10 nanometer insertion for EUV, presumably the design window is closing for that in terms of insertion. I just wonder whether you need further, or whether you expect further orders for EUV to support 10 nanometer insertion, or should we just now expect a kind of full insertion at seven nanometers across the three main logic customers? Then I have a follow-up.
It is clear that I think the EUV insertion is now focusing on seven. However, things are a bit fluid in the sense that when you talk about 10 nanometer, you listen carefully to what customers say about the 10 nanometer node. There have also been clear public statements by customers about phases in that 10 nanometer node. You have a first phase and a second phase, and they will have node nomenclature that are changing. From our point of view, when you talk about a 10 nanometer node, it is a certain pitch or a half-pitch, and that there could be several phases.
In this particular case, and given the complexities of 10 nanometer production using very complex multiple patterning strategies, I would never rule out that with progress of EUV, the way it is today, and even if we can accelerate that progress, that there's a customer that says, "Listen, in the last phase of a 10 nanometer node, I'm going to use EUV for two reasons. One, it might be beneficial to reduce complexity at that moment in time, and you have to learn. You have to learn for 7 ." I'm not ruling this out. It's going to be a function of the progress that we are making over the next one or two quarters.
Just on that, Peter. Presumably the half-pitch shrink you're talking about maybe something like eight nanometers. Maybe EUV's not inserted at 10, but something like a half-pitch at eight. Is that the sort of thinking?
The half-pitch stays the same, whether it's 10 or X, whatever that node's going to be, nine, eight, seven, eight, and eight and a half, 10 minus, 10 X. Whatever customers are going to think of a node nomenclature, which effectively is the second phase of the 10 nanometer node. The pitch and the half-pitch is going to be the same. Like I said, this is going to be a function of progress that we are making, the complexities and the yields that customers are experiencing in using DUV multiple patterning strategies. Those two things together will determine a potential, if you were to say, second-phase introduction. I'm not going to say that that's a slam dunk. With seven nanometers, I think it's clear.
We see very clear announcements from customers that say, "We're going to use EUV at seven nanometer." It is really the performance and their complexities that we are engaging or running into that are going to determine a potential second-phase introduction. That is for learning also.
Great. That's very clear. Just on slightly shorter-term question, you mentioned that you're seeing earlier development in H2 2016 than maybe previously. Does that mean that you're actually seeing some pull-in, effectively, some pull-in business from H1 2016 to H2 2015 that we might have expected a bit later? I'm not asking for H1 2016 guidance, but can you help us out with the sort of phasing through into next year?
When you look at the potential upside that we're talking about, that's really like I explained earlier. That is because of the development and the integration time for the next nodes, for instance, the 10 nanometer node, is more complex and takes more time. You see, as compared to the original assessment of when they would need the early development tools, which would be the first half of 2016, we see a potential of that being pulled into the fourth quarter of 2015. That is correct. That's technology transition. With respect to capacity, like we mentioned in the introductory statements, both in DRAM, NAND, and logic, there are several new fabs out there that will take tools, and they will take tools for the most advanced nodes, and they will definitely ramp the first phase. That first phase will happen.
That's why we are positive, or we feel comfortable about the second half of this year. How that will translate into 2016 ramp and the speed of that ramp, we don't know yet. The fact that those are new fabs, that they have to be ramping the first line, gives us a lot of confidence over the next couple of quarters.
Thanks.
The next question comes from Mr. Arcuri. Please state your company name, followed by your question.
Cowen, thank you. I guess my first question is, Wolfgang, I want to make sure that the EUV systems that you're putting in the backlog this quarter, these are the 125 wafer per hour spec tools. The ASP we should think about is like EUR 110 million, and that's about the ASP that we should think about for all the EUV systems that you book from here on in. Is that right?
Colin, the backlog includes some of the 3350s, there is also some of the next generation 3400 included there as well. We said before that the 3350s list price is somewhere in the mid EUR 90s, 3400 would be over EUR 110. You got to be a little bit careful when you look at it in the backlog because, what you got to do is, you got to take the ASP and split it over several elements of things that you deliver. There may be a little bit that gets allocated to a service that you deliver or an installation or a warranty. You got to be a little bit careful with the backlog ASP. I can tell you for the six tools that we took last quarter, the backlog ASP is somewhere in the low EUR 90s.
Low EUR 90s. Okay, great. I guess I also had a question on Holistic. Peter, you said that the attach is now up to 100%, I'm assuming you're talking even in memory. I guess my question is, of the EUR 1.9 billion in service you're going to do this year, how much of that is going to be Holistic and YieldStar? What's the right opportunity for that going forward? I guess I'm thinking about what gross margins are going to be on your immersion systems, it looks like in June, if I assume zero gross margin, that the non-EUV gross margin is sort of 47.5%. It looks like maybe if Holistic is that strong, that you could get to 50% maybe next year or the year after. Thanks.
I think on the margin question, Wolfgang, are you going to take that? I think on the Holistic business, yeah, it's 100% attach rate. That is definitely true for advanced logic, but now also increasingly for advanced memory and especially DRAM. Also there you see multiple patterning strategies. You see multiple patterning spacer strategies, which basically also need the control software that we can provide together with the scanner to make sure that customers can control their process, and they can make a yield. Yes, you see it now also in memory, and that attach rate will probably expand over time to the entire product base and to, well, I would say across all industry sectors. Wolfgang, you want to take the margin?
Yep. As it relates to the gross margin, it's important to know, first of all, when we report revenue, we report in systems, and we report in second category field options and services. For the holistic lithography, it depends on whether the customer orders it with the original system, the option, then it gets reported in the system. When it's later on ordered as an upgrade option, then it will be in our field options and services. When you hear us talking about holistic lithography achieving over EUR 500 million last year and us aiming for EUR 1 billion by 2017, you'll actually find that in both buckets. I think that's important to know. As it relates to the gross margin, we have said that before. There is a strong software content in that business, and there is some hardware in there with the YieldStar metrology tool.
The blended margin is somewhere in the mid 70%. As we approach EUR 1 billion in 2017 and as holistic lithography outgrows the rest of the business, yes, it's going to be accretive to the gross margin over time. I don't think I have an exact number to share with on what it could be in which scenario, but in general, it's accretive.
Okay. Thanks so much.
You're welcome.
The next question comes from Mr. Mehdi Hosseini. Please state your company name, followed by your question.
Thanks for taking my questions. Mehdi Hosseini, Susquehanna International. My first question goes back to some clarification. You are increasing the age of backlog for EUV system. Now it seems like going to extend more than 12 months. It seems like you also have incremental confidence on seeing a double-digit EUV shipment by 2017. You have a capacity of 24 systems per year by then. In that context, can you help us understand what is your base assumption for the mix of these systems that are going to be shippable in 2017? What I mean by mix, either on the logic side, you can help us seven versus 10, and would there be any DRAM that would be included?
The fact that your U.S.-based IDM customer has stepped up and placed an order also makes us wonder if they have revisited insertion at 10, or is it still going to be at seven nanometer? I do have a follow-up.
Well, on the last question, we're not going to be specific on any customer. That's unfortunate, Mehdi, but I don't think it is really appropriate in this call. On the split in 2017, logic DRAM and the potential. Well, like I said in my introductory statement, the majority will be logic. However, if you look at the DRAM roadmap, in 2018, we will be in the mid 1X DRAM nodes, which is EUV territory. We are indeed seeing potential for a 2017 shipment, which also includes advanced DRAM. The majority will be logic. That's going to be, if you take 24 units, it's going to be significantly over 50% is going to be logic, which is microprocessors and foundry. I would say it's a minority. If you want to get a number, it could be 75%-80% logic and the rest DRAM.
Would you expect majority to be 7 nanometer logic, or would there be a mix of-
Yeah, I think it's going to be 7 nanometer. Yeah, largely, yes.
Okay. My second question has to do with your core business. The bookings in Q1 declined 26% on a sequential basis, and Q2 is down low single digit. Is this the kind of a base booking that would be required for technology investment? If there is any capacity coming in later on, it would help with any kind of improvement in booking? Is that a fair way of thinking about your current booking?
I think we probably sound like a broken record in this sense, but we're going to repeat how bookings are being administered in the company. We only have a handful of customers. Actually, we plan our production and our shipments to customer based on an agreement that we have with the operations executives of the customers and looking at their ramp schedule. Then we agree with them that we have a certain production capability there to do the shipments. Then the orders are almost an administrative follow-on part of the process. Just to give you an example, with a few customers, we have volume purchase agreements that span a node.
Basically, you have a whole node agreement, which is a lot more units that are under the volume purchase agreement. They will just follow on with orders based on the lead time that we give them, and they will be very disciplined on giving us the orders, taking into account the lead time. With other customers, we have annual volume purchase agreements. With other customers, we have, let's say, volume purchase agreements which span the period of two years. When you have an annual volume purchase agreement and you negotiate that for three, four, five, six months and it comes in, the orders will follow because then you only have six months left. Then you get an input of orders, as you could say, is a big lump of those orders in one quarter.
What I'm trying to say is, don't put too much emphasis on the order intake. Over a longer period of time, you are correct, it needs to reflect our future business. A quarterly order intake is not indicative of our business going forward and might be almost deceiving if you understand how the ordering goes in our business. More important is, I think, our guidance on what we believe that we ship. That is based on the agreements which we have with operational executives of our customer.
Got it. Very helpful. Thank you.
The next question comes from Mr. Andrew Gardiner. Please state your company name followed by your question. Go ahead, sir.
Thank you. Good afternoon. It's Andrew Gardiner with Barclays. Peter, you do seem sort of more optimistic now than you were earlier in the year. It seems as though despite sort of broader concerns about the end markets and the cycle, investment's clearly continuing. Your optimism, though, seems more around capacity additions as we come to the balance of the year rather than a very high level of a sort of technology race. I think we can see this in the tool mix as well. Looking at 2Q, you had a higher level of KrF shipments, for example. As we then move into next year, it seems as though you're talking much more about technology-driven year, particularly when thinking about logic, beginning the initial investment at 10 nanometer and that requiring a high-end immersion.
I'm just wondering, is this a fair characterization or is it perhaps too much of a generalization? What are your expectations of your, or those of your customers in terms of capital intensity for next year? I'm just wondering why wouldn't your product mix move more towards the higher end, in 2016 relative to where we are in 2015, if indeed that sort of technical complexity increases?
Yeah, I think this is generally, the ASPs are going up as you've seen it, that is just a reflection of the increased need for tools that can help our customers address the complexities of the next nodes. Whether it's on improved overlay, higher productivity, it was an earlier question, it will translate into a higher ASP. I think that is definitely the case. Now, to your point, yes, I think it's too much of a generalization and to try to split the shipment patterns between capacity and technology. When we look at the capacity additions, they are not in existing fabs, they're in new fabs. Which all of those new fabs are addressing the most advanced nodes in those sectors. That in fact is technology. Yeah?
I would definitely say that the first phase ramp of those fabs is a technology insertion, but they have to add capacity because it is basically a new fab. For this year, this is what's happening. We have two DRAM fabs taking those tools. We have logic fab taking those tools. That is driving our business in the second half, and that will happen. If you ask 2016, it's too early to give you guidance on 2016. Customers generally, when they have went through the first phase, they just Where they are, what the end markets do, and they will time the remainder of that ramp to fill up the fab. That you could argue is then capacity demand. I would say this is how we actually look at it, because 14, 16 nanometer logic is now ramping. We are now shipping.
How much of that capacity will be filled in 2016? We don't know yet. The same is for DRAM. I would say it's probably just the other way around a bit. What we're currently seeing is more technology capacity additions, and whereby 2016 will be driven by the situation in the end demand. That is for us, still opaque. We don't know that, and I think our customers don't know that either. We'll just have to wait and see.
Understood. Thanks for the clarification.
The next question comes from Mr. François Monnier. Please state your company name, followed by your question.
Okay. Thanks for taking my question. The first question is about the new machine, which is about to ship, the 3350. How is it going to work? You're going to ship the machine, unpack the machine, plug it in, put the floppy disk, switch it on. What shall we expect from this machine? Is it going to start at like 500 wafer per day, 800, 1,000? What's the next data point on this new machine? Thank you.
Basically, the 3300 that we now have has an upgrade package, which brings us to 80 watts.
The 3350 ships with a new drive laser, which brings you to 125 watt. That is a productivity increase from 80 to 125 watt. Which you could say is almost the linear impact on the productivity. Maybe 500 good wafers per day on the 3300, 80 watt, then you can calculate, you don't even need a floppy disk for that. You can calculate on a piece of paper what that means for the productivity. As you know, we have a target of 1,000 good wafers per day next year. We need that productivity of the 3350. It's not only productivity, it's also availability. We have a 70% target for this year. But the 3350 with all the new, I would say, options on it. Which includes better stability of the droplet generator, which includes in-situ cleaning, actually brings the targeted availability to 85%.
If you go from 80 to 125, which gives you more productivity, you get an availability that goes from 70% to 85%, it gives you more productivity. You can just calculate that the 1,000 good wafers per day is a very reasonable and good target for 2016.
Right. Very good. Now, a more longer term question. There's been quite a bit of news flow recently about China and even a rumor of a Chinese company wanting to buy Micron this week. It's pretty clear that the Chinese government wants to get more involvement in the semiconductor industry. Do you expect China in general to become a bigger customer of yours going forward?
Yeah, I think, you just mentioned it. What you just said could have been my answer. Yes, it's clear that there's more focus in China on building semiconductor capabilities, both in logic and also I think there have been statements clearly that they want to be somehow involved in advanced memory. Yes, I think China is a bigger market. Don't exaggerate it either. They're still, when you think about logic, as compared to the leading edge logic makers, they are two generations behind. Moore's Law is not slowing down. As a matter of fact, you could argue at 10 nm, it is actually being pulled in. Yes, there is a lot of attention. There's more attention. I spent more time in China than I did the years before. We have more interaction with our customers. 20 nm is now ramping in China. There's our 20 nm logic.
There's definitely more activity, but like I said, the generations are still two generations behind. In terms of leading-edge memory, there's a lot of leading-edge memory in China, but that's in NAND and in DRAM, and that's not in Chinese ownership, that's in Korean ownership. Yeah, China is moving, but don't think it's going to be a blowout. It's going to be gradual growth over the next five to 10 years.
All right. Thank you very much, guys. Thank you.
Thank you, François.
We have five minutes left. The next question comes from Mr. Amit Harchandani. Please state your company name, followed by your question.
Thanks. Amit Harchandani from Citigroup. Good morning, and good afternoon. Two questions, if I may. The first question relates to your field option and service sales. You see an elevated level in this quarter, and I'm wondering to what degree is this an outcome of increased reuse by customers as they move, say, within a node or across nodes. How should we think about service and field option sales going forward versus the EUR 2.5 billion-EUR 3 billion target that you have in your 2020 model? The second question regards to EUV. We've talked about throughput and productivity. Matching and overlay seem to be under at acceptable levels. Could you give us an update on the defectivity aspect and where is that trending versus customer expectation? Thank you.
Okay. I'll take the EUV defectivity question and the growth drivers for field option and service. I think Wolfgang can then talk about the 2020 target.
Yep
which basically, I think you're referring to our model that we showed at the Investor Day last year. On the EUV defectivity, this is currently not a, well, that's not entirely true. There was a defectivity worry, which largely had to do with mask-induced defectivities. I think with the introduction of the EUV pellicle, which we talked about last quarter, basically those worries are gone. It's now making EUV pellicle, the mountable and demountable pellicle, bringing that to an industrial state, and that will take care of that. Clearly there are resist-induced defects. There are process-induced defects. This is why customers need those development tools. This is why they are running hundreds and hundreds of good wafers per day today to figure that out. We do not get the feedback from our customers that says that is a showstopper.
This is what I would say, it's the normal development activities that customers have to do before they can go into a production ramp. We don't have those numbers. Even if we would have those numbers, I would not give them to you because this is a public call. It is customer privy information. It's not a major concern, and especially since we have now a mask solution. On the drivers for the field option and service growth, very clearly, it's the install base that grows.
It's the drive of the customers to look at their install base and say, "How much CapEx have we still on the balance sheet, and how can we reuse that, or how can we use it more efficiently by either getting more wafers out, doing productivity upgrades, doing overlay and focus improvements so that they can use those machines for the next nodes?" That is the big driver, and it includes upgrades. Very advanced upgrades that includes a lens swap, a new optical system could be over EUR 20 million, EUR 20 million-EUR 25 million. Those are drivers for the field option and service sales. I would call it install base management. That is what customers are focusing on to have a better capital efficiency. Wolfgang, you want to take the target number?
Yeah. First of all, also on the reuse, yes, reuse is enabled by us through these packages, and therefore, we have also modeled that in our 2020 revenue numbers and know the assumptions have not changed as it relates to reuse in our field options and services. The EUR 2.5 billion-EUR 3 billion that you refer to in the different scenarios, it will depend a little bit on, like I explained earlier on an earlier question, on whether customers order holistic options right with the system or whether they do options later on. Again, at EUR 400 million in the first, EUR 520 million in the second, and that being stable, you'll be somewhere over EUR 1.9 billion. It's still quite a bit of growth left to get to the EUR 3 billion level. We feel very comfortable about it. It's good business for us.
It's approximately at the average of the gross margin of our systems business. We believe that's still a good target range for now.
Thanks. Thanks, Wolfgang.
You're welcome, Amit.
Okay.
Ladies and gentlemen, with that, our 60-minute call time is expired. For those participants that were unable to ask a question and have a need to do that, please feel free to contact the investor relations department with your question, and we'll get back to you as soon as we can. With that, on behalf of ASML's Board of Management, I'd like to thank you all for joining us today on the call. Operator, if you could formally conclude the call, I'd appreciate it. Thank you very much.
Thank you. Ladies and gentlemen, this concludes the ASML second quarter 2015 results conference call. Thank you for participating. You may now disconnect your line.