艾司摩爾 (ASML) 2014 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by. Welcome to the ASML fourth-quarter and 2014 annual results conference call on January 21, 2015. (Operator Instructions).

I would now like to turn the conference over to Mr. Craig DeYoung. Please go ahead, sir.

Thank you, Eileen, 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 here in Veldhoven, The Netherlands, is ASML's CEO, Peter Wennink, and ASML's CFO, Wolfgang Nickl.

As a reminder, the subject of today's call is ASML's fourth-quarter and 2014 annual 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 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, found on our website, and in ASML's annual report, 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, and now I would like to turn the call over to Peter Wennink.

Thank you, Craig. Good afternoon, good morning, ladies and gentlemen, and thank you for attending our fourth-quarter and annual 2014 results conference call.

Before we begin the quarter -- the Q&A session, Wolfgang and I would like to provide you with an overview and some commentary on the fourth quarter and provide our view of the coming quarters.

Wolfgang will start with a review of our Q4 and 2014 financial performance, with some added comments on our short-term outlook, and 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.

I would first like to highlight some of our financial accomplishments for 2014. Year on year, we grew our revenue by 12% to a record level of EUR5.86 billion. 2014 gross margin was up almost 3 percentage points to 44.3%, while our earnings per share increased by 16% to EUR2.74 per share.

Last year, we paid a EUR0.61 per share dividend, worth EUR268 million, and used an additional EUR700 million to repurchase 10 million shares, returning a total of EUR968 million to shareholders throughout 2014.

So all in all, a very successful year for ASML.

Looking to Q4, our net sales came in at EUR1.49 billion, which was ahead of our guidance. Memory accounted for 65% of our system sales in the quarter and also drove the upside in system sales. As anticipated, we recognized revenue for one EUV system during the quarter and service and field option sales came in at very healthy level of EUR409 million.

Gross margin for the quarter was solid at 44%, compared to our guidance of approximately 43%. Higher gross margin was driven both by product mix and higher volume.

R&D expenses came in slightly above our guidance at EUR268 million, affected by both the strong US dollar and a one-off settlement with a farmout partner related to our suspended 450-millimeter program.

SG&A expenses came in as expected at EUR79 million.

Turning to the balance sheet, quarter over quarter we grew our cash, cash equivalents, and short-term investments balance to EUR2.75 billion. Regarding the order book, with the NXE:3350 we started to include EUV orders in the bookings and backlog for the first time.

Overall, Q4 systems bookings were strong again at almost EUR1.4 billion, including two NXE:3350 systems.

An anticipated shift in bookings strength toward logic occurred, which represented 73% of bookings. This was largely driven by foundry. Although memory bookings dropped to 27% of total this quarter and Q4 memory sales were strong, as mentioned, the memory sector backlog remains high at 43% of total.

As a result of the strong foundry bookings, our backlog grew to almost EUR2.8 billion, including EUV.

With that, I would like to turn to our expectations for Q1 2015 and share our initial view on the first half of 2015.

Our strong logic bookings last quarter are now giving us overall a nicely balanced backlog going into 2015. We expect that strong sales to the memory sector and strength in our service and field option sales will continue in the first half and that sales to the logic segment will increase from the second half 2014 to the first half 2015.

For the first quarter of 2015, we expect total revenues of around EUR1.6 billion. We do not expect any EUV revenue recognition during the quarter. As mentioned earlier, we expect strength in service and field options revenues to continue and we expect revenues in this portion of our business to be about EUR400 million in Q1 and to strengthen throughout the year.

We expect gross margin for Q1 to be around 47%, driven again by sales volume, by the expected rich product mix, and continuing strong system performance enhancing option sales. R&D expenses for the first quarter will be at about EUR260 million. SG&A is expected at about EUR83 million.

Peter will talk more about EUV shortly, but I would like to make a few points regarding EUV shipments and revenue recognition for 2015. As most listeners are aware, we have shown excellent progress in improving on key performance metrics related to productivity and reliability in 2014.

Also, as announced last quarter, one foundry customer has ordered two production EUV tools for the eventual insertion in 10-nanometer high-volume manufacturing. The first one of these systems is planned to ship to the customer midyear.

It is highly likely that the demonstration of all key performance metrics will not be confirmed until 2016. The accounting rules under these circumstances are not allowing us to recognize any revenue until all performance metrics are met; hence, we do not expect to recognize revenue on the two TSMC NXE:3350s in 2015.

We are at various stages of negotiation for the other four NXE:3350 tools, as well as the three customer-requested NXE:3300 to 3350 upgrades, which are in our 2015 production plans. The final negotiations on these systems will determine shipment dates, payment terms, and revenue recognition.

Finally, we have now previously mentioned strong financial position cash flow prospects. We will continue to return our excess cash to shareholders through dividends and share buybacks. We intend to increase our dividend per ordinary share for 2014, again by 15%, to EUR0.70 per share and accordingly will submit a proposal for authorization to the 2015 annual general shareholder meeting, which will take place on April 22, 2015.

We also intend to continue to purchase our own shares during 2015 and 2016, including 3.3 million shares to cover employee stock and stock option plans, as well as up to EUR750 million worth of shares intended to be canceled.

The buyback program will start tomorrow, and at current share price, these intended repurchases represent a total value of approximately EUR1 billion.

Now with that good news, I would like to turn the call back over to Peter.

Thank you, Wolfgang.

So as Wolfgang highlighted, we had a very successful year in 2014. We had record sales, record annual gross margins leading to significant EPS growth. Alongside our 2014 financial performance, we can also claim significant successes in product development, new product introduction, and product adoption.

As Wolfgang also highlighted, we expect our strong second-half 2014 sales to continue into the first half of 2015, driven by continued substantial memory spend and an increased spend in the logic sector, which is all supported by our current backlog.

Today, we see that foundry is essentially working on three to four nodes at the same time. The 28-nanometer node, the current high-volume node, is still not at its full 100% capacity as new demand for this node keeps growing. Therefore, we expect that our two shipments in 2015 will be used for capacity additions at this node in order to meet the growing demand.

At 20 nanometers, the first chips are on the market, so that node has moved into volume at some foundries. The 16- and 14-nanometer nodes are been qualified for production and there is process development happening at the 10-nanometer node.

As we mentioned in earlier calls, the foundry industry effectively sees the most aggressive ramp of new nodes ever in their quest for faster, more energy-efficient and cost-effective shrinks for mobile and communications applications.

In memory, industry analysts are expecting both NAND and DRAM markets demands to continue growing at a similar rate to 2014, which means a little under 30% for DRAM and about 35% for NAND. New DRAM fabs are planned and some are already being equipped to meet demand, compensating for some lost capacity in DRAM due to device shrink, node complexity, and compensating for growing device size with movement from PC to mobile DRAM.

Next to the initial introduction of vertical NAND, NAND fabs are continuing their planar node shrinks and are expected to add greater capacity to meet forecasted bit demand.

As to our product development and product introduction successes, we have ramped our NXT:1970C immersion product, launched late in 2013, faster than any new product introduction before, with over 50 shipments in 2014.

Our TWINSCAN immersion system set new productivity records at over 1.5 million 300-millimeter wafers imaged in a 12-month period, creating huge value for our customers. And we have had significant success in customer adoption of our new YieldStar petrology products, with an installed base of all YieldStar products touching on 200 units.

Regarding EUV, we commented in great detail at our investor day in November, so I won't repeat myself too much here. But as a reminder, we have met the 500 wafer per day target that our customers set for us in 2014 as we have now demonstrated these productivity levels at multiple customers over multiple days. We also demonstrated source power at 100 watts with 96% die yield, which is a key factor to get to the productivity levels that our customers need for volume production.

Our 2015 productivity target remains at 1,000 wafers per day, and importantly for our customers and for our EUV program, we received the first two orders for our fourth-generation NXE:3350 EUV tools, the first of which is planned for shipment in the middle of this year.

Our EUV product focus for 2015 is on improving the stability and availability of the machines in the field and to continue the encouraging progress that we have made with EUV productivity so far. We will furthermore concentrate on shipping the first six NXE:3350s so that our customers can continue their process integration planning and to begin to introduce these tools into their production environments.

Before I end my introduction, I would like to comment on some recent events regarding our customer core investment partners' equity holdings in ASML. All co-investment partners are approaching the end of their ASML equity lockup period whereby they can begin to sell all or parts of their holdings, if they wish to do so. Each company will act in their own best interest and are generally not obliged to inform us of their decisions.

I would like to make it clear, however, that whatever they decide in this regard, it in no way represents a change of support for ASML or ASML and there is no impact on the co-investment R&D contributions committed by each partner company.

Now with that, we would be happy to take your questions. I would like to turn back to Craig.

Thank you, Peter, and thanks, Wolfgang, as well.

Ladies and gentlemen, the operator will instruct you momentarily on the protocol for the Q&A session, but beforehand, as I always do, I would like to ask you to kindly limit your questions to one with one short follow-up, if necessary. This, again, will allow us to get as many callers in as possible.

Now, Eileen, could we have your instructions and then the first question, please?

(Operator Instructions). Sandeep Deshpande.

This is Sandeep Deshpande from JPMorgan. Thanks for letting me on. My first question is regarding your EUV shipments. You have talked about those two EUV shipments to TSMC coming through in terms of recognition in the following year. How is your negotiations on the other tools going in? How do you see -- do you expect some of those other tools will also be recognized next year, as well as how do you see the recognition going on the tools associated with the upgrade from the 3300 to the 3350?

Secondly, my question is on the 16- and 14-nanometer node. You had strong ramp-up in the foundries in the fourth quarter. Do you see that all those 16/14-nanometer orders have come in or is there a tail to that? Thank you.

Sandeep, Wolfgang, I will take the first part on the revenue recognition. As both of us mentioned, our prime focus here right now is to focus on shipping the tools so that customers can make progress on 10-nanometer and mid-node insertion and then preparation for 7 nanometers.

Just to ground us, we shipped three tools in 2013, we shipped four tools in 2014, and we have currently a plan to ship up to 10 tools in 2015.

As it relates to revenue recognition, it's different from the mature technologies where we recognize at shipment. And if you look at revenue recognition for this year, you've got to almost look at it in a couple of different slices.

First of all, there was one 3300 that remained to be shipped and recognized, and that system, I can tell you, has shipped by now and we will recognize that, I guess, in Q2.

Then secondly, there are six tools of which 3350s. We got an order for two of them. We shipped them starting midyear, and like I mentioned in my remarks, the revenue recognition rules, they require us to meet all performance milestones in order to recognize any revenue, and we forecast that to happen in 2015.

For the remaining four 3350s, the negotiations are ongoing with several customers, and it will depend on that outcome and those terms of conditions when we ship, when we get paid, and then when we recognize the revenue. The same applies actually for the three systems that we are upgrading for customers. Those negotiations are also ongoing, and again, we would have to see final terms and conditions before we can tell you when exactly we will recognize them.

Then before answering the second question, I will answer the second question. I would just like to add, Sandeep, if we ship the tools somewhere midyear -- let's say early Q3, then we take three to four months to ship and to install the tools, so before the customer can start to do the first qualifications and we start meeting the performance metrics, we will be easily in the beginning of 2016. It's just the time that it takes.

On the second question, the 16- to 14-nanometer node ramp have orders coming. Well, you could actually look at the backlog and the order intake and look at the geographical distribution of our order backlog, how that changed during the fourth quarter, and it won't be specific on the customers. But you can deduce from that that's a -- some of the logic orders still need to come from other parts of the world. I don't think everything's in.

C.J. Muse, Evercore ISI.

Thank you for taking my question. I guess, first question, in terms of your backlog you are now at roughly three-year high, and historically you have provided guidance out a little bit more than a quarter, so curious what's changing this time? Is it really a factor on the foundry side in terms of uncertainty of timing of shipments and/or yields or is there something else going on?

Yes, I think we -- during our introductory call, I think we were pretty positive on the actually qualitative guidance that we are giving based on the backlog, because you are right, the backlog is high. It's well distributed amongst memory and amongst logic.

For the first quarter, we have much better indications of the, let's say, time in which we can ship and which we can recognize, because we now talk about the general -- the regular NXT immersion tools. But for the second quarter, there could be timing differences towards the end of the quarter, let's say in the June/July timeframe, where tools might be pulled in a bit or went from one month to the other month.

Now if you have four tools, it's EUR200 million. Those things are still fluid, so it's not a matter of that we have any doubts on the total -- the amount. There is just a matter of when do we see the cutoff of those very expensive tools that are now over EUR50 million a pop. It's more that than anything else.

Sure, makes sense, and I guess if I could ask a quick follow-up. In terms of field services, I think your installed base on the immersion front is roughly 224 tools, give or take. Curious in terms of YieldStar upgrade, what penetration are you there today and what kind of followthrough could we see in 2015, 2016, as well as, I guess, the upgrades on the EUV side, to help us understand how to model the spares and service part of your business?

Yes, I think on the last question, how do you model it, I think we will guide you. When you talk about upgrades, it's a very wide variety of upgrades.

You refer to YieldStar. That's one upgrade possibility, but another upgrade possibility, which is actually happening in -- we are signing POs for this as we speak, are upgrades for entire tools, for instance from a 1950 to a 1960, from a 1960 to a 1970, a 1970 to a 1980.

There are four generations upgrade possibilities there, which is, of course, an upgrade. Next to the application products that we are bringing out, which can be, let's say, brought back to earlier scanner models. So the whole upgrade program has a very wide variety of different products. That means it's very difficult to guide you on all those different products.

What we will do, and I think Wolfgang said it, that we start the year with about EUR400 million in sales, and I think, Wolfgang, this will grow throughout the year with, I would say, a significant double-digit percentage towards the end of the year.

This is clearly a growth engine. You might remember that on the applications business, the holistic litho business, we said we have a target of reaching EUR1 billion by 2017. I think we're just over halfway, so there is still something to go, but there's not that many years left, so the compound annual growth rates are pretty steep and that's only for the holistic litho, and that does not include, for instance, the system upgrades that I talked about.

All in all, it's a very strong source of growth for us with good profitability.

Great, thanks so much.

Kai Korschelt, Merrill Lynch.

Thanks for taking my question. I had two, if that's okay. The first one was really on memory. It looks like the order intake has already moderated a bit, but we have also seen a new fab announcement, I think, in December from Micron, Toshiba, and some of these people. Presumably those haven't ramped yet for you, so I am just wondering what is your maybe 12- to 18-month view? I know it's difficult to predict, but do you think that memory cycle we have just seen has legs? That's my first question.

The second one's, then, on the 10-nanometer foundry ramp. Just roughly when do you expect that to materialize? And also, could you remind us whether the initial rollout, which presumably will be without EUV, will that be triple patterning and what increase in the lithography content could we expect to offer? Maybe you end up double patterning 20-nanometer or 14/16-nanometer nodes. Thank you.

Okay, let me just make some notes. On your first question, does memory has legs? I think if you look at all memory customers, DRAM and NAND, and you go through all the major customers, every customer is expanding their capacity. Every customer has either a new fab or an extension to a fab.

There were major fabs announced in Korea for both major players. Everybody knows that in Japan, there is a major effectively extension ongoing. Micron announced one recently.

Every major memory maker is looking at adding capacity and that capacity will not be up in two months' time, so it will take time to build those factories and it will take time to fill those factories up. That will not happen in two quarters. It will be stretched out over a much longer period and it just reflects, I would think, the discussions we have with our customers as not only that we have with them privately, but also I think how they talk about their business going forward.

I think we internally also mentioned this amongst the senior managers here that we haven't seen our memory customers as, let's say, positive about the future applications of memory products in a very long, long while. It has to do with the fact that memory starts to take many, many forms, which will be used in many, many different applications, and that's what we see happening. So I think it's a very broad base and I don't think it is something that will easily go away.

The 10-nanometer foundry ramp, when we see that happening, that will be end of 2016, going to 2017 and 2018. That's when it will happen. Definitely there will be higher litho content. We always see this at 20 nanometer. I think node on node, we see about a 35% to 40%, 45% litho intensity increase because of double and triple patterning.

Now that all looks nice and definitely for the nodes that we are currently seeing, but we all know that there is, from a cost point of view, only one solution going forward and that's ultimately EUV. That's about 40%, so on average.

Thank you.

Mahesh Sanganeria, RBC Capital Markets.

You just mentioned something about the 450 program. Can you provide a little bit more details? And what happens to the customer co-investment? I think Intel invested significantly on the 450, and so what happened to that agreement if there is a change in the 450 program?

This is Wolfgang, so it won't change the total contributions to ASML from that particular customer. It's just being allocated to other engineering projects, EUV and other projects.

In terms of the 450, I think it's publicly acknowledged that that program is past and we will see when that product -- project comes back online when the industry is converging as that one is a solution. (multiple speakers).

The settlement that I mentioned was a relatively small settlement. That was a clean-up from one of our farmout partners in our R&D line.

Okay. The second question on the foundry orders, you talked about the different technology node, the different phases. Can you give us an approximate distribution of how do you see the full-year 28 nanometer, then 20 and 16/14, how will your shipment distribution will look like? Is it primarily 16/14 or more towards 20?

I asked the same question when we looked at the order backlog and the order intake, and that in fact it's across those three categories. It's -- we are seeing capacity additions on the 28. We see some capacity additions on the 20, and as you know from a lithography perspective, the litho requirements for 20 and 16 and 14 are virtually the same.

We are definitely also seeing 16 and 14 coming out of -- they're currently in qualification, so they will ramp in the course of the year, which, of course, the installed capacity for 14/16 is still limited, so that's where definitely capacity needs to go. And by the end of the year, we are also shipping for 10 nanometer. I mean, 10-nanometer initial risk production will be on immersion only and the tool sets that we are developing today together with our customers to be able to deal with that node also require shipments towards the end of the year.

Actually, it's a difficult question to answer because you will see capacity additions at every new node. Now if you would ask me would there be an emphasis on one of them, I would say 16/14. That's where probably the majority will go.

That's very helpful. Thanks.

Gareth Jenkins, UBS.

Thank you and happy new year to you gentlemen. Just one quick question on R&D. Should we expect R&D through the course of 2015 to head lower from the EUR260 million level in Q1? And then, I have a follow-up, please.

Gareth, the R&D, the total R&D for this year is in total euro terms going to be lower than it was in 2014. We are shooting for this EUR1 billion run rate level for the midterm.

Now the US dollar exchange rate throws us a tad bit of a curveball, so it could be in the EUR260 million range, too, but we are getting down and we're getting into that neighborhood. We will control it at that level for several quarters, and then at our capital markets day, we said that for the longer term we really will drop it as a percent of our revenue, and we're right now operating at about 18%, and with our EUR10 billion ambition by 2020, we should settle it in somewhere in the 13% range or so. It remains very well controlled.

That's great, thanks. Just a somewhat different question on Chinese -- the Chinese market and what you are seeing there. You signed a large frame agreement pre-Christmas. I just wondered what you are seeing in China and whether you'd expect any of the demand to be fulfilled with the re-use of tools. Thank you.

I think the Chinese market after its years of silence has actually come back. Yes, we had this announcement pre-Christmas on a deal that we signed with a Chinese foundry.

Generally speaking, there is a lot more activity there. It is also focused on, let's say, what I would call an advanced node. It's the 32-nanometer and 28-nanometer node. Now what I said earlier on that 28-nanometer node is a very large node where we constantly see devices being brought onto the 28 node level.

It seems to continue making it a very large node, and I think the Chinese foundries will benefit from that also, or at least that is what they are anticipating.

So that market has come back to life, definitely. It is not only used tools. The Chinese market is not a market where you only sell secondhand tools. These are new projects which require new NXTs, and let's be honest, 28 nanometer is a pretty advanced node. So, yes, it's -- after years of silence, it's back.

Thanks.

Timothy Arcuri, Cowen and Company.

A couple questions. First of all, Peter, can you give us an update on the forecast you gave us last quarter. You said that the 20/16/14 wafers, you said there would be about 175,000 worth of outs by the middle of this year. Can you update the number and also maybe extend the forecast, given how much the orders have grown in December quarter from foundry? Can you update us where you think that will be exiting 2015?

I think -- there is no reason to change what I said earlier, and I may remind you of what I said earlier when somebody asked the question, where does this foundry demand go to? What kind of node? And it's basically four nodes. It's 28, it's even 20. It's 16/14, with I think the emphasis on 16/14, and on 10 even.

Yes, it does sound like a lot, but I mentioned it in my introductory call, also. We're in probably the most aggressive ramp period that we have seen for a very, very long time in the logic industry, but no reason for us to change our statement of last quarter of about 175,000 for that 20/16/14 node, but the outliers here are 28 and 10.

Got it, okay. Then two more questions. One, I just wanted to make sure that your comment, that I actually heard it right. If your equity partners do decide to sell their stake, that has no impact at all on the NRE funding agreement? Is that correct?

You understood it correctly.

Okay, thank you. Then just last thing, did I -- am I hearing some change in your EUV revenue recognition policy or is it just that it's now going to take longer for you to get the sign-offs from the customer and for you to hit the metrics in the field, or has there physically been a change to the revenue recognition policy on the EUV initiative?

No, Tim, there is no change to the revenue recognition policy.

To give you just a bit of background, because I think you are referring to the 3300, the 3300s were sold with the idea to go to the development sites of our customers, and there were hardly any performance criteria in there, except that you should be able to image a wafer. That was it, so you could call that a performance criteria.

Now when the 3350s will go into volume production, then, of course, clearly there are some performance metrics in stages or milestones that we need to hit, which we are confident that we will hit, and the only problem is that the accounting rules say that when you meet them all, which is also true for the 3300, only they were very limited, then you can recognize revenues.

There is no change. It's just the fact that they are being used for a different purpose and that means that it takes a bit longer.

Wonderful, thank you so much.

Farhan Ahmad, Credit Suisse.

Thanks for taking my question. I had a quick question on nanoimprint technology. There was recently a press release by Hynix indicating that the company is collaborating with Toshiba on nanoimprint technology for MRAM. Also, earlier in 2014, Toshiba and Canon announced that they would be collaborating on nanoimprint technology on production for 1Z nanometer node.

I wanted to check if there is anything that you are picking up from your competitive intelligence in terms of nanoimprint and if there is an increased threat of the technology developing and competing with you.

No, we didn't pick that up, and even with the customers -- those customers that you mentioned, we have extensive discussions on it, and we don't regard it as a competitive threat for EUV, nor for [DUV].

Thank you. The one question on the second-half shipments for the year, it seems like there was a pull-in on their DRAM shipments that caused upside to the fourth quarter, and DRAM spending is pretty high, following the seasonal patterns and with node conversions planned around the first half of 2015. Foundry spending also seems like it's pretty high in first half.

I wanted to hear your thoughts on how the second half of 2015 could shape up. Could we see a similar decline in half and half in revenues as we saw last year, or do you think there are some offsets on the second half of this year that can balance the year?

Yes, let me start. This is Wolfgang. I want to comment on the pull-in. It was not a pull-in. It was incremental demand.

Sometimes when you do guidance, it is not quite clear whether it's in the last two weeks of the quarter or whether it's in the first two weeks of the quarter.

It was not a pull-in; it was just demand strengthening as customers were outfitting the factories. And as you can see from our guidance in Q1, that is up [EUR]100 million or over EUR100 million. It was by no means a pull-in.

As it relates to the second half of the year, I will let Peter chime in, but we didn't give any guidance and we have given some color on the first half. I don't think that we go much beyond what we already said there.

I think I would like to reiterate what I said earlier on the memory market. Every customer is not only talking, but is using their well-earned cash to put capacity in place in different forms. So there is a time component to it.

Now you are -- if I really want to answer your question precisely, I need to know exactly when those time components turn into a tool shipment.

I think for us medium term, the memory business is going to be good because there is a lot of empty fab space that needs to be filled. But now you're asking us to precisely indicate when that is, which is very difficult for us. It's going to be there. Whether it's going to be Q3 or Q4 or Q1 of next year and in what form, it doesn't really matter. It is DRAM and NAND and it looks good, so when we have more visibility, we will inform you further.

Got it. Just one quick follow-up. On the memory demand, incremental demand that you saw, was that incremental new capacity addition or was it all node conversions?

It was node conversions, largely, because it was our newest and the most advanced systems that they needed, but a node conversion is also capacity in terms of memory. It's extra bits.

Got it. Thank you. That's all I have.

Jerome Ramel, Exane BNP Paribas.

Just a question to come back on the installed capacity. You mentioned the 16/14-nanometer node at [one of the] kilo wafer. What will be in your assumption the 28-nanometer node when it will be fully deployed?

That's an interesting one. We said -- we had an initial estimate last year of about 300,000. I think we are now looking at 340,000 and it's still moving. So I don't know where it is going to end, to be honest.

Our customers keep telling us that they see node conversions of existing devices coming from 65 down to 45 down to 28, and where it will end is difficult to say. But it's a big node. Even when you go back to the first introduction of the 28-nanometer node, we are three years further. We're still shipping capacity. That is something we haven't seen very often. Very difficult question, difficult to answer, so I'm not going to speculate.

Okay, and a follow-up. If I understand correctly, maybe only one EUV revenue recognition this year for the 3300. Are you confident with where consensus stands today at EUR6.8 billion of revenue if we don't have this EUV revenues recognition?

Sorry, could you repeat it?

Are we happy with the EUR6.8 billion, given we don't recognize revenue on (multiple speakers)

Again, we're not giving a forecast today, like I said. But if you look at the EUR6.8 billion in the Street and if you look at what portion on EUV is in there, what we have told you today is that we know for one system for sure that it will recognize and for the rest -- the two systems for TSMC will most likely not recognize, and for the rest, it will -- it depends on the terms and conditions, so you've got to make up your mind on it, but it's probably to the lower end of the range because we only know for one system for sure that it will recognize.

Okay, thank you very much.

Amit Harchandani, Citigroup.

Amit Harchandani, Citigroup. Thanks for taking my questions. One question and an unrelated follow-up, if I may. The question that I have is really in terms of -- we talked about service and option sales serving as a strong tailwind for the Company. We have talked about lithography. At the analyst day, there was talk potentially about you looking at other related possibilities in terms of leveraging lithography data.

Could you maybe update us on how your thinking is and what you have -- if you have taken any further thoughts, if you would just share with us how you're thinking about the growth drivers for the Company beyond the sales of system tools and service options? Then I have an unrelated follow-up.

I think that's a good question. I think at the investor day, we did touch on it, and effectively what we said I'm going to repeat.

If you look at the manufacturing process of a semiconductor device, and there are very, very few steps in the process where you can actually use a machine to use for corrective actions. The litho tool is a machine that can do that, and we have thousands of knobs that we can turn in order to change the settings so that the image process -- the imaging process takes account of those corrections.

That means that the more you can feed back into the litho machine knowledge about errors -- errors or deviations from target, whether it's on overlay, whether it's on the critical dimensions, whether it's on depth of focus, whatever, the more information that you can gather, the better you can control it.

Now with the prices of litho machine going up from now over EUR50 million to EUR100 million and over EUR100 million when we talk about EUV, the next generations, then you can imagine that importance only goes up. The quest for looking at data for those elements of the chip production process where the litho tool is suitable to do the corrections, that quest for data will continue.

Now we created one path by creating YieldStar and metrology and integrated metrology. That's a very valuable part of the business, which is growing, and so are we looking for together with our customers and with potential partners to help us, yes, and am I willing to go into detail about what that specifically means, no.

But you have to use your imagination on what that could be, and then perhaps over the years to come we can be a little bit more specific on the work that we are doing in this area, because we are doing work in this area.

But I hope I at least gave you the very clear vision that we have to use the litho tool as the correction mechanism for some of the major challenges that our customers are seeing when they move into the next nodes.

Thanks, Peter, and maybe as an unrelated follow-up, just a quick one, with regards to the customer co-investment program, we get the -- at least in the P&L, we have TSMC and Samsung down there in the other income line and Intel tends to move about a bit. Could you maybe share with us if there was any impact from Intel's contribution at the gross profit level in Q4 and if you're baking in a contribution for Q1 guidance? Thank you.

Let me take this. First of all, there was no extraordinary contribution in Q4 that -- nor in Q1 in the guidance, by the way.

The way how you have to think about the CCIP, first of all, accounting for that is probably even more complicated than the EUV revenue recognition.

The most important thing is that close to EUR1.4 billion will be collected by us in cash. The accounting rules have it that some of it will go through other income, some of it will go through the gross margin, and some of it may go straight to equity. It's very difficult for us. Every quarter, it may also change a bit.

So what we have elected to do is when we give you margin guidance, we will always have the impact reflected in there and we will tell you the other income portion of it so that you can always follow our guidance.

But the most important part, like I said at the beginning, is it's EUR1.38 billion and we will collect the money over the period 2013 to 2017.

Thanks, Wolfgang.

Mehdi Hosseini, SIG.

Thanks for squeezing me in. Two easy clarification questions. Wolfgang, if two EUV system has moved into backlog, how should we think about a commission of the remaining eight systems planned to be shipped next year? This is for booking specifically.

You said an easy question. (laughter). The remaining -- what I said in my remarks and I hope I get your question right is we can ship up to 10 systems this year. (multiple speakers)

Right, and two of those already in the backlog.

Yes, two on the backlog, so we got -- so the 10 systems are -- one has shipped. Then there are six 3350s, of which two are in the backlog and the other four we are negotiating for.

Then there are the three 3300s that our customers have upgraded, have asked us to upgrade to 3350s. Those upgrades are being negotiated. You will probably not see them in the systems backlog because they are upgrades, and as a reminder, these three systems from a cash perspective have also been repaid. We just got to negotiate the price for the upgrade.

Yes, so you will find the 3350s in the backlog when they get booked.

Okay, so of the 10 systems planned to be shipped, I think six are 3350, two already in the backlog and the remaining four will move to the backlog sometime this year?

Why don't you -- I can refer to slide 22 of the slide deck. That actually is simply put out in those three layers that Wolfgang talked about.

The 3300B is already out, so it won't be in the backlog because tools that are shipped are not in the backlog. Now you have the conversion from the 3300B to the 3350 is an upgrade. Generally, we don't put upgrades in the backlog. That's what he said. Of those 10, four you won't find in the backlog for these very specific reasons. The 3350, six should go into backlog, of which we booked two. Four still to book and those will go into the backlog.

Okay. Then the second --

Now we get another easy one. (laughter).

Oh, no, actually I have a big -- the core business, which is much easier than EUV.

(laughter). EUV is our core business, sorry. (laughter).

Your backlog was very impressive, and, yes, I do understand that there is no visibility on shipping on a quarterly basis, but given your backlog and given how you have guided during the January conference call in the past, why not help us with a year-end guide, given your backlog?

You seem very confident in the business. You talk about opportunities in 20, 16, 10, and memory has still got some leg into it. Why not use that to give us a reference for calendar-year 2015 revenue opportunities, and then we could fine-tune it as you report quarterly earnings?

To me, it seems like maybe you are not so confident with the booking trend into next year. Maybe this is why you are hesitant to provide calendar-year 2015 revenue guide, or is it something else?

There's nothing else. There are many examples where we haven't guided the total year, more than -- more examples where we have not guided than we did guide.

Now, and I know when you hand somebody something very nice, they want it next time, also. I can understand that. But we have to be realistic, also.

I said it before, we feel very confident about the -- well, about the backlog, but with all the extensions that the customers are currently planning, the timing in which they want to ship them, putting them in or pushing them out a month or two, is fluid. Like I said, it's 50 million, 60 million tools.

Now if we miss two or three tools, it's 150 million or even more, and I can see the headlines of the reports that you guys write, ASML misses guidance. Well, nothing really happens.

I think we are going to give you what we know that we will ship because we got extra confirmation from our customer, which is this quarter, and then next quarters, we will have a very healthy backlog and we will ship and whether 100 million or 200 million moves into Quarter 2 or 3 or from 3 to 4, it doesn't really matter.

For the total year, it's a bit different. We do not have a backlog that covers the total year yet. That will happen throughout the year. The only thing that we know is that our memory customers are pretty bullish about their future. They have all announced capacity expansion plans in terms of billings, which is the first time since a long time. Some people would argue that is a point of attention, but I would also argue that the memory business currently has a much wider application space than it ever had before and that our memory customers, only a few left, have a much different spending profile as they have shown over the last couple of years.

So I think there is more control in the memory business, that is at least what we hear when we talk to our customers, which is also one of the reasons why things might move, yes, from one quarter to the other in terms of the shipments. These things are fluid, I would say, from a quarter to a quarter point of view.

Now for the total year, it looks good, but I am not willing to give you any specific guidance on the total year. Like I said, there have been many more years where we didn't guide than we did guide. One quarter has to do with the fluidity of timing of some of the shipments.

Got it, thank you.

Johannes Schaller, Deutsche Bank.

Thanks for taking my question. Just two quick ones. Quickly going to -- back to YieldStar, there is a certain upgrade opportunity out there in the field. Could you maybe give us a bit of a sense what you think penetration is already for the tools that are in the field and how much more upgrade potential there is? Some metrics here, maybe even by foundry memory would be very useful.

Then just a quick follow-up. I noticed there has been a bit of a tick up in CapEx in Q4. Is there anything to be said on that? Maybe you could elaborate a bit on the drivers. Thank you.

Okay, the YieldStar upgrade potential, you have to realize that YieldStar comes in two forms. It comes in a standalone form, where basically it's a standalone machine, and it comes in a form of an integrated built into the track. That means that the track should have the space available to put the YieldStar in.

So those upgrade possibilities are limited because you have to have a track that actually has the space there, which is an open gaping hole, which then needs a YieldStar.

Now on the standalone YieldStar, there is always an opportunity, so you could say for every NXT that is out there, that is 28-nanometer node and lower, the YieldStar data can be fed back into a litho tool, which, when you go to an integrated mode where it is in the track, it is almost easier.

But those upgrades are limited. Upgrades for YieldStar are really standalone units that can be shipped, and there are opportunities there, but I would say the biggest focus that we currently have is on the integrated metrology going forward whereby the litho tool and the track are optimized to take the YieldStar, but that is then in there from day one. There is not an upgrade possibility.

Upgrades for YieldStar are there. The biggest upgrade possibility is in the litho systems, where you really take litho systems from a two -- one or two nodes ago, you actually upgrade them to the latest and the greatest. That is where the major upgrade opportunity is.

I'll take the CapEx one, Johannes. Actually, our CapEx for the year came in slightly below what we had estimated the beginning of the year. I think was EUR360 million that we ended up for the full year.

It was a year of higher CapEx for us than usual. It had to do with the EUV factory that we are building out. The reason why it is actually a bit lower than we originally anticipated is because we kept the factory fairly modular and we are building it out as we go. In total, it was 6% of revenue, and also on our investor day, we indicated that our longer-term model is around 5% of revenue, so we will see that coming down. So, nothing extraordinary in Q4.

Ladies and gentlemen, we're going to try to squeeze one more call in, please, so if we could keep the questions short, whoever it is, we would really appreciate it, just to stay on time.

If you weren't able to get through, feel free to give investor relations a call. We will be working through the evening, and with that, Operator, can we have the last call, please?

Pierre Ferragu, Bernstein.

Thank you for taking my question. Just maybe one on your gross margin, so you have talked many times about your aspirational target of getting to 50%, and of course, your Q1 number is very impressive and it feels like huge progress towards that number. But now, given the visibility you have of when you start recognizing revenues for EUV, what sort of timeline do you think we should expect before seeing that number materializing?

Then maybe one very, very, very short clarification. On these four additional orders that you are still discussing with your clients, how many clients on top of the one who already placed orders are involved in the discussions? Thank you.

I'll take the first piece. We said 47% for the first quarter. And without recognition of an EUV tool, I think we said before that the first tools have no gross margin, so in that sense it's accretive to gross margin for the quarter.

Our EUV margin development over time remains unchanged on the 3350s. We are planning to be somewhere in the neighborhood of 20%, and then once we are in higher volume manufacturing, usually like two years after commencing shipments, we should reach the 40% level, and that will, of course, be one of the enablers to get us as a total company into the 50% territory that we talked about in London in November, and Peter, you (multiple speakers)

Yes, on the customers. In the detailed discussions that we have with our customers in executive review meetings, every customer has EUV on their roadmaps, some of it later, some of it earlier. But the customers that we are actively talking about EUV orders number four.

Thank you very much. Very helpful. Thank you. Have a good day.

All right, with that, on behalf of the ASML Board and management, I would like to thank everybody for joining the call today, and now, Eileen, if we could have you formally conclude the call, we would appreciate it. Thanks.

Ladies and gentlemen, this concludes the ASML fourth-quarter 2014 and annual results conference call. Thank you for participating. You may now disconnect.