艾司摩爾 (ASML) 2012 Q1 法說會逐字稿

Ladies and gentlemen, thank you for standing by.

Welcome to the ASML 2012 first-quarter results conference call on April 18, 2012.

Throughout today's introduction, all participants will be in a listen-only mode.

After ASML's introduction, there will be an opportunity to ask questions.

(Operator Instructions).

I would now like to turn the conference over to Mr. Craig DeYoung.

Please go ahead, sir.

Thank you, operator, and good afternoon and good morning, ladies and gentlemen.

This is Craig DeYoung, Vice President of Investor Relations here at ASML.

And I'd like to welcome you to our investor call and webcast.

As the operator mentioned, the subject of today's call is ASML's 2012 first quarter results.

Co-hosting today from our headquarters here in Veldhoven, the Netherlands, is Mr. Eric Meurice, ASML's CEO; and Mr. Peter Wennink, ASML's CFO.

At this time, I'd like to draw your attention to the Safe Harbor statement contained in today's press release and in our first quarter results presentation, both of which you can find on our website at ASML.com.

This Safe Harbor statement will apply to this call and to all associated presentation materials.

As a reminder, the length of the call will be 60 minutes.

And now I'd like to turn the call over to Eric for a brief introduction.

Thank you, Craig.

Good afternoon, good morning.

Thank you for attending the first quarter 2012 results conference call.

Before we begin the Q&A session, as usual, Peter and I would like to provide an overview and some commentary on the first quarter and provide a view forward.

As usual, Peter will start with a review of Q1 financial performance with comments on the short-term outlook.

And I will complete this introduction with some further comments on our current market view update for the long-term.

So, Peter, if you will.

Thank you, Eric, and welcome to everyone.

Summarizing our first quarter, sales came in at EUR1.25 billion, essentially the same as the previous quarter.

This quarter's sales were again significantly skewed toward the foundry IDM sectors with 75% including noncritical KrF capacity additions.

9% of the first quarter sales went to the DRAM sector, representing only six system shipments.

The average selling price of all systems shipped in the first quarter was EUR20.2 million.

And this reflects the high KrF system shipments versus the previous quarter.

Service and field option sales came in at EUR202 million, again, driven by the adoption of performance and housing field options.

We also recognized sales of one EUV system for a total amount of EUR44 million, which is one quarter earlier than planned.

It is due to a faster-than-anticipated installation and customer acceptance process.

This had an impact on our gross margin, by the way, the EUV shipment for the quarter of about 1.5 percentage points.

Updating on our previously-announced share buyback program, as of April 1, 2012 out of the total program, ASML had repurchased 29.8 million shares, of which 4.1 million shares in the first quarter for a total amount of EUR839 million, of which [EUR139 million] in Q1, giving an average buyback price of EUR28.

The repurchased shares have been or will be canceled.

At the end of the first quarter we have close to EUR3 billion in cash and cash equivalents.

The first quarter net bookings came in at 36 systems, valued at EUR865 million.

The booked average selling price shot up from EUR19.2 million in Q4 to EUR24 million in the first quarter.

That is due to a product mix shift toward the most advanced tools.

Our first quarter bookings profile saw foundry increasing slightly to 74% of the total, followed by IDMs at 13%, NAND at 8% and DRAM at only 5%.

Our order backlog at the end of the first quarter was EUR1.6 billion totaling 56 systems, with an impressive average selling price of EUR28.5 million.

The backlog profile at quarter end changed again as combined memory continued on a downward trend to only 23% of the total backlog versus 27% at the end of last year.

As to the outlook, we can confirm our earlier guidance of EUR2.4 billion net sales for the first half 2012, and given a continued solid six months of visibility, we can now say that the third quarter sales are expected to be stable at Q1 and Q2 levels.

Based on this visibility and in combination with our first quarter backlog, we expect orders in the next quarter to adequately cover our sales expectation.

We believe that specific quarterly order guidance has become less relevant over time due to the fact that fewer but larger customers are placing bigger and lumpier orders, making quarterly order patterns more and more imprecise.

On top of that, customers have significantly increased their visibility towards us of their real shipment needs, leaving the order placements as a kind of secondary and almost administrative consideration in the business process.

System demand is currently strongly driven by the continued need of foundry manufacturers to have capacity at the 28-nanometer node in an environment of very strong demand and heightened competition.

We now expect this strong demand trend to continue for the remainder of the year.

The NAND segment continues their aggressive shrink with an apparent moderation in wafer capacity additions in reaction to the current weakening NAND pricing.

The continued difficult DRAM market environment is keeping the DRAM sector's share of our backlog at historically low levels.

As for an update on the expected EUV shipments in 2012, system integration has begun here in Veldhoven, only 11 NXE.3300 systems on order.

Current plans are to ship two of these systems later this year, with the remainder shipping in the first half of 2013.

To compensate for the recent source delivery delays, we have decided to alter the shipment process for the first two systems, which means that the revenue will not be recognized until the systems are assembled and integrated at the customer site.

Therefore, we do not expect the revenue recognition of these tools to occur until early 2013.

And finally, sales for the second quarter are expected to come in again at around EUR1.2 billion with a gross margin that will be about 43%.

R&D and SG&A expenses will remain constant versus the first quarter levels at EUR145 million for R&D and EUR55 million for SG&A in the quarter.

And with that, I would like to turn it back over to you, Eric.

Thank you, Peter.

Well, as Peter mentioned, the short-term, mid-term demand for advance logic lithography capacity is stronger than we expected and guided at the beginning of the year, bridging nicely to 2013 sales when first we expect memory lithography capacity demand to pick up; and two, when we are able to ship and recognize the first EUV pre-production systems.

So, short-term, mid-term looks very good.

The longer-term future looks also positive for us, as articulated often during our quarterly results call.

The lithography intensity should increase further when either of the two lithography technologies enabling the next generations of microelectronic technologies, so whether it is immersion multiple patterning or EUV -- when these two are introduced into production.

Both technologies will cohabit until EUV becomes the only technology capable of imaging the next-generation device due to its lower cost and scalability.

To keep our immersion leadership, which will be used in multiple patterning, we continue to enhance the overlay and throughput performance of our TWINSCAN NXT 1950 systems.

We are certainly the market leader in immersion, as evidenced by one of our NXTs 1950s exceeding another productivity milestone by processing in a productive environment more than 4500 wafers per day with an overlay at about 4 nanometers.

It's a great performance.

We are planning further improvements with the introduction of our next-generation NXT at the end of this year, 2012, with the next level of performance in throughput, overlay, and focus.

We also continue the development and introduction of a strengthened suite of holistic lithography software and hardware products, so as to enable the process window enlargement, process drift management and machine matching, which are all necessary for multi-part patterning requirements.

Regarding EUV, we are certainly confirming steady progress, still with the objective to release a program in the summer for customer production ramps in 2014, so exactly as we said last call.

As highlighted in our press release, we have installed six EUV preproduction tools now, NXE.3100.

They are in use, imaging wafers, at customer site.

They are contributing to the progress of the EUV fab infrastructure as over 9000 wafers have now been imaged on these systems to date.

These systems have also demonstrated an exceptional performance over all with the 16 -- sorry, 6-nanometer on-product overlay and with dedicated shock overlay of less than 2 nanometers.

These are extraordinary performers.

As Peter already mentioned, we are also confirming at this moment the delivery of 11 of our NXE.3300B, which is our production systems, between Q4 2012, this year, and the summer of 2013.

Regarding the source technologies stages, we still do not have a proven source power, the proven source power in situ, which would confirm our target throughput performance.

But again, this was not planned in the first quarter.

This has always been planned in the summer -- at the end of the summer.

But we have made significant progress no matter what.

We have now demonstrated source power at 30 to 50 watts at high duty cycle for prolonged periods.

These achievements are significant, as we only need 100 watts to achieve a 60 wafer per hour machine, which some customers will find acceptable for initial production implementation until we reach our nominal throughput specification of 125 wafers per hour.

In view of this progress and although we have not yet confirmed, as I said, these higher powers integrated into the system, we have still -- we have now started negotiation on the next batch of orders of EUV systems, targeting mainly for customer production ramp expected to begin late 2013 and into 2014.

So in summary, we are definitely encouraged by the market demand development in the last quarter for the short-term and midterm for sales opportunities in the second half.

We are additionally encouraged by our progress in EUV performance and the unwavering support that we are getting from our customers regarding this technology for ramp in 2014 as planned.

With that, Peter and I would be pleased to take your questions.

Thank you, Eric.

Ladies and gentlemen, the operator will instruct you momentarily on the protocol for the Q&A session.

But beforehand, as always, I would like to ask that you kindly limit yourself to one question with one short follow-up, if necessary.

Of course, this will allow us to get as many callers on today as possible.

Now, operator, would you go ahead with your instructions and then the first question, please?

(Operator Instructions).

Mr. Simon Schafer.

We seem to be having some difficulty with Mr. Simon Schafer.

Nic Gaudois.

(technical difficulty)

Nic Gaudois.

Yes, it is Nic Gaudois from UBS.

Just first question is on the comment you just made on the next-generation EUV tool for production, which should start being delivered in H2, 2013 and into 2014.

Could you confirm where you see the initial requirements for volume production between DRAM 1x nanometer and logic as well?

And I guess it is too early to quantify it, but maybe give us some qualitative initial view on how you think this will develop basically in 2014.

And then I have a follow-up question, thank you.

Okay, very good.

So, as we said a numbers of time, there are certain numbers of phases when you introduce such a technology.

The first phase is to have some R&D tools to do some recipes and confirm the technology.

That's what we've been doing.

And although there have been delays in our credibility, by the end of the day, the recipes are being made.

And the windows of opportunity is as you mentioned, and we have not lost any of them yet.

The first window is DRAM 1x or -- so I think you call it 1x or 1y.

They are very near the 20 nanometer.

This seems to be very complicated to execute with multiple patterning.

And the customers are ready to, in fact, wait for EUV to be ready to get there.

And the competitive product that could also be very near a number-one need is the 14-nanometer logic node, which has shown in particular in the last three months, two major pushback by end customers using or attempting to use multi-patterning.

They have seen two problems.

One is a cost issue, which seems to be too high to justify the shrink if they do multi-patterning, and they also have seen a shrink factor limitation.

So in other terms, you go to 14 nanometer but you do not really reduce the die size as much as you should if you were to use EUV.

So we are going -- we are having at this moment heated pressure from the people going into 1x DRAM, the first node below 20 and the 14-nanometer logic.

And at this moment, the race is on to know who is going to be the first one to get into product.

That will create demand in 2014 which we believe will be appropriate in terms of industrial capacity, meaning not too much and not small enough.

So, perfectly, exactly in the numbers of tools that a Company like ours would want to do.

If you do too many tools, it costs a lot of money.

If you do not enough, it costs a lot of money for dilution purposes.

So we seem at this moment to have achieved a run rate target which is being discussed, and which I translated in the press release as, we are negotiating the first orders for 2014.

Okay, understood.

So we call it Goldilocks EUV, I guess.

Just moving -- just a quick follow-up on the more near-term, you added in your presentation of expecting NAND flash to recover later in the year.

Could you maybe precisely, vis-a-vis the timeline, and whether you would expect all major customers to basically come back?

Or some select customers are -- effectively some may have some specific circumstances, in particular in the context of overall memory consolidation?

We are a bit puzzled by the NAND flash business at this moment.

We were expecting, in fact, then to start showing some positive color.

But instead what we have had, and this is in fact a stronger positive news, is the NAND hasn't happened but the logic has, more than we thought.

So this is good news.

We can talk about that later.

The NAND flash, however, will have to be much bigger than the current run rate, if you believe in the tablet numbers, in the smart phone numbers, and in the SSD numbers.

If you just charter what you hear in the market, the statistics expected for those products, you would see a shortage of NAND as early as Q4.

So now we are hopeful that this is not going to even happen in Q4, so that we stabilize nicely our sales between a good 2012 and a sustained 2013.

Okay, that's clear.

Thank you very much.

Simon Schafer.

Yes, thanks so much.

It's Goldman Sachs.

I guess my question is in terms of perhaps the lack of order outlook into the second quarter.

I think I get the point about customer concentration.

It's very evident that the foundries and IDM are obviously a bigger piece in terms of concentration.

But my sense is also that really wasn't all that meaningfully different in the first -- sorry, after Q4.

And you were very comfortable indicating that orders would be at least flat.

So I was just wondering what has changed in that.

Yes, I will be happy to answer that.

I think orders are generally used by guys as a proxy for our sales, which is circling, but that's actually top line and ultimately drives the earnings per share.

Now, I think that is what we are exactly doing.

We are giving you a sales guidance for the next two quarters with some, I would say, very positive consideration for the remainder of the year.

In that sense, I think from an outlook point of view, we are not -- we are very clear, probably clearer than we have been in the past.

Now on the order front, the point is that you have fewer customers, which is indeed true, and those customers are getting bigger.

But also, the order sizes are getting bigger.

So it means you can't have -- between the end of June and the end of July, you can have hundreds of millions of difference, so a very wide spread of when those orders are actually going to hit the books, which is no indication whatsoever of the outlook that we have on our shipment pattern.

So -- and this is the only reason why we are not doing this, because we want to be clear on what the outlook really is.

And what the outlook is, is driven by sales and not by the orders, and that is why we are giving you sales, which we believe is a better indication.

Got it, Peter, thanks so much.

And my follow-up question would be, I guess in terms of orders and deliveries, it seems like actually KrF is perhaps somewhat higher, so clearly there is a good amount of capacity adds happening at the 28-nanometer node.

And I think I would agree absolutely that for the 2x -- I mean 2.0 and 22 nanometer, there's clearly an avenue for intervention and aggressive upgrade paths from all of your customers at the leading edge.

But I guess I'm trying to get a sense as to whether you think there is perhaps a transition risk between the two nodes, just because it looks as if capacity has been very strong, specifically in KrF and some other capacity tools.

No, I would think what we are getting more comfortable with at this moment is that to understand why we've been so positively surprised with logic.

We were worried that there was excess capacity due to the fact that, as you know today, there are multiple players running after the same customers.

So we were a bit worried that there is duplication.

But this doesn't seem to be the case.

And there are two factors.

One is if you talk to foundry customers, you will see a lot of exciting products getting into mobile applications which requires power.

And these things are hugely complex with very large die size.

So this drives a lot of silicon square meters.

So, that, we can test, check, touch.

The second thing that we have now recalculated is the sizes of each of those nodes in compare to what we ship as units.

And I'm happy to report that we are far from having yet, I would say, a stabilized capacity level necessary for normal node at 28.

So it is we have another six to nine months of production and shipment of machines before we have delivered to the market, I would say, sort of mature, not high, level of wafer per month.

After that, there is another engine for 2013 which is to be proven, but we are getting now significant interest, is the customers want to go from 28 to 22 already.

And they do that because there is a drive of product again.

So this is not now a driver due to volume, it is a driver due to product architecture.

And as you know, there is fight in the ARM environment, et cetera, that forces this activity to happen.

So now what we discover is we have a former system that was in the logic arena, a node every two years, every 18 months that gets into a node every year, and therefore a faster ramp to stable numbers of wafers per month, and then immediately not replaced, but prolonged by another RAM.

So that's what we see at this moment.

So therefore, we do not see significant risk of these being wrong excess-type business decisions.

If I were to be a negative pessimist, I would say, however, these are very complicated nodes, technology-wise, and you could always have issues of difficulties to --- of yield, the ramp, et cetera.

But this probably would not impact the litho business.

In fact, it could in fact create upside for us in case difficulties to run those things.

Thanks, Peter; thanks, Eric.

Kai Korschelt.

Yes, it is Deutsche Bank.

Thank you for taking my question.

The first one was just on the -- if I could just cap on this comment, Eric, that you made I think on the second batch of EUV tools you're currently negotiating at the run rate 2014, 2015.

I seem to remember the longer-term target was to ramp up your manufacturing capacity to about nine tools per quarter by 2015.

Are you still thinking in these sort of dimensions?

Or in the process of those negotiations is your thinking on that capacity maybe to increase or lower it, depending on the demand?

That would be our first question.

The second one was just quickly on the potential buyback extension that you talked about last quarter.

Could you just remind us how much net or gross cash you think you need to run the business?

Obviously, your Company is now bigger than it may have been two or three years ago, so if you could just remind us what the minimum level of cash you would need, thank you.

So I will answer the buyback question and Peter is going to answer the EUV question -- no, I think -- (multiple speakers) I wish.

So, the EUV question, yes, I would say that there is three stages off Goldilocks capacity with integration in this way.

I would say the first year we will probably want to be at one unit a month, one unit a month, 1.5 units a month in the first year.

Then we will go to a period of time at three per month, as you said, 36.

And then we probably will -- mature business that we can see now in terms of technology at about 60 units per year, roughly.

So you've got these three stages that is now being discussed, negotiated with most customers, so that we have the proper capacity to achieve this.

So we start with one a month; we then go to three a month; and then we will go to whatever 60 is divided by 12 -- five a month.

Now we'll do buybacks.

Yes, we basically are still in the same position as we were last quarter, no change on that.

I'd just like to reiterate that on the announced buyback program, which also would include the shares we need to buy back to cover for the share plans and the dividends, we still have more than the EUR0.5 billion to go.

So that we will do, of course, first.

Now your -- and that actually means that in the fourth quarter we always review with our Board the distribution policy for basically the next year, which includes dividends and share buybacks and the division between the two.

So that's what we will do by the end of the year.

And let there be no question in your mind, this is a very profitable and cash-generating Company.

So in excess of what we need, we will keep giving that money back.

There is no change whatsoever.

On the -- you could say basic cash balances that we would need a couple of years ago, like you rightfully basically indicated, we were smaller and we said between EUR1 billion and EUR1.5 billion.

It is probably now between the EUR1.5 billion and EUR2 billion.

But above that, you will see us in a cash return mode.

So I think that it is for now, when we would announce the potential share buyback clearly is a function of where we are at the end of the year in terms of our decision-making.

And like I said last time, you do not do a little tiny buyback [EUR400 million].

You want to do it for a big number.

And I would not expect that number to be anything lower than EUR500 million.

So that needs to be a big, big chunk.

So it's not going to be imminent, because we first have to do the first EUR500 million left on the remainder of the buyback program and on the dividend payment.

Okay, thank you very much.

Satya Kumar.

Yes, thanks for taking my question -- from Credit Suisse.

Eric, you talked about 14 nanometer multiple patterning having issues.

Are you referring to 14 nanometer for logic or were you referring to foundry processes?

Well, foundry processes in the following sense, that the foundries are having another challenge that the IDMs would not have.

The challenge is that they have to deliver design rules which are less restrictive.

And they have to deliver a shrink factor which is very aggressive.

So, yes, my comment on 14 nanometer being a competitor to being in the first to go concerns more the foundry environment than it concerns the microprocessor environment.

Okay, can you talk a little bit about the -- it sounds like your visibility on the second half shipment has improved, it seems like primarily from the foundry order visibility.

What are you seeing in terms of 14 nanometer logic capacity requirements?

Is that -- how is that factoring into your thought process for this more confidence on the second half?

Thanks.

We're not very aggressive on that.

We think that this 40-nanometer node is now going to stabilize what we shipped already.

So we are not foreseeing an upside.

We are foreseeing all the positive news that you heard from us in the last 45 minutes is really 28 nanometer.

So if you are a bull, you will now tell me that I should also be positive that the 40 nanometer could see a trend of increase, because the numbers of capacity in store not that high.

But --

He is talking about 14.

Oh, 14, I thought you said 40.

Satyam are you talking about 14 or 40?

14, 1-4, sorry.

Sorry, so--

I also understood 40.

I understood 40.

Because -- no, so 14 nanometer is just R&D at this moment.

There is no impact on production.

Okay, if you talk about a specific customer who calls its next node a 14 nanometer, I cannot comment, of course.

This is proprietary data.

Okay, thank you.

Mahesh Sanganeria.

Yes, thank you; RBC Capital Markets.

Eric, I have a question on your competitiveness on immersion.

You gave us good information on the double patterning layers.

My question is that are you able to meet the requirements of overlay for double and triple patterning?

Or there is more improvement needed in that, that will enable you to keep your competitive advantage in immersion?

We -- first of all, I think it is very, very, very, very difficult at this moment to achieve what the customer wants.

You may have heard me in the speech talking about design window enhancement and process drift management.

So you are now in a point where the customers have to work extremely closely with a litho supplier to define the machine and the software that goes with it in their own design, so that your windows as wide as possible.

So -- but the design now is kind of dedicated to the machine.

And then when that happens, you put this into production.

And you are going to drift, not only the litho, but also my peers in fab, the hatch and et cetera.

And then you correct every drift in production through a new set of parameters that you input into the litho machine.

That is again another set of products that we are shipping.

And we are not satisfying the customers.

That is what I'm trying to tell all of you now that the double double-patterning makes the windows enhancement and the drift a huge concern and a huge challenge.

So yes, we can do it, but it's not as pleasant as EUV.

So the customers will always try to get EUV if EUV works.

That is the first statement.

The second statement is, however, is our competitor capable of doing this better than us?

We've got a feeling that it's not the case, for the usual reason.

One is, we invest a huge amount of money in the packages which are dedicated to windows enhancements and drift management, which are dedicated to the machines.

The machines themselves have 3000 points of controls, which we are well ahead of anybody in the market doing this.

So if you fear something, you are better off with a machine with the maximum numbers of flexible items.

The second bit is that we still have a concept, which is the TWINSCAN, with numbers of measurement points capability which is superior to the paper specification of competitive products.

So we think even on paper we will remain the leader, conceptually, under critical layers.

And I hinted to you the third point, is that the level of risk that is now existing in the fabs do not justify de-focused on R&D.

The R&D to get a machine to work is a year or year-and-a-half activity.

And duplicating it with multi-supplier, with multiple architecture is a non-economic, risky situation, in particular that in production.

Then the machines are not matching, which is you cannot process anymore a wafer on different machines.

That would be pretty bad.

So we think that during the period of time where EUV is not yet fully available and immersion is there, we have a high chance to remain leading.

Thank you, Eric.

That is very helpful.

And just a quick follow-up on your commentary on the NAND.

Is the NAND weakness or lack of orders which you were expecting, is that coming from a specific region?

Or can you provide some more color as to what is driving that?

We heard from SanDisk that they pushed out the Wi-Fi.

Have you seen similar action from others?

Yes, we have -- in fact, I wanted to stop the call here and start asking you questions to figure out what the problem is.

(Laughter) But we are only guessing that these four players or 3.5 players are now more in the business of, call it, embedded NAND.

And embedded NANDs are like logic.

They are not anymore what you want to do, you manufacture.

You just manufacture to the numbers of end products you have been designed in.

And usually in this business for logic, you have always seen this at the beginning, things take a bit longer than expected before you get designed in and the product really works and et cetera, et cetera.

So I would consider at this moment that all the exciting new products on which the new NANDs are being designed in haven't yet pumped in the volume.

And you had all of these people just waiting for the products to come.

So -- waiting for the new PCs, the ultralights, waiting for the SSDs, new generation with a better driver.

I think, if I'm not mistaken, Windows 8 is going to drive those things much more aggressively, so that also you've got to wait to make it valuable.

So I guess that's what it is, because I see this non-aggressiveness to be a bit consistent on all the players of the NAND business.

Okay, thank you very much.

Krish Sankar.

Thanks for taking my question.

It is Krish Sankar, BofA Merrill Lynch.

I had a couple of questions.

Eric, you mentioned that you are having negotiations with customers for future EUV orders.

What type of customers are these?

Are these the early adopter DRAM or are they pretty much across the board?

As a fact, as I said to Nick, as you probably will see numbers of DRAMs and a number of logic people.

And I confirmed NAND not.

And as we only said, NAND will not be interested immediately, because 60 wafer power, 60, 70, which we will -- we are considering as a walk-away type thing, is probably not satisfactory enough for the NAND guys who will want to be a bit above that.

So we will be a year or two later in the NAND business, which corresponds to the Goldilocks strategy, because if we had the three segments together on the first year, we probably will be having some difficulties to ramp all of that.

All right.

And then just two short questions.

You said you had 11 orders for the 3300 right now.

Does that include Elpida?

And then the second one is, in terms of your Q1 bookings, about [74%] orders from foundry.

What percentage of that was for 20-nanometer, if you could help us?

Thank you.

So sorry, so the first question, I think we can't really talk about a situation of bankruptcy and things.

As you know, at this moment, Elpida is in the process of being purchased.

And there are issues about assessment of the assets, which I am advised not to comment.

So I can't really tell you anything of that nature.

Your second question I could not catch, sorry.

In terms of the foundry orders that you saw, strong foundry orders in Q1, I understand a lot of it was for 28 nanometer.

Do you know how much or what percentage was for 20 nanometer?

No, nothing.

Nothing at this moment; one or two for R&D, that is it.

Thanks, Eric.

Gunnar Plagge.

Yes, it is Citi.

I just wanted to come back to the foundry or the predictability of foundry demand.

So on the one hand you have to handle the 28-nanometer designs coming in, but also the year development as it looks.

That foundry seemed to be pretty relaxed because you could, in this scenario, also see a lot of longer backlog of orders coming in.

You have 91% of the backlog shippable in the next month.

So could you a little bit talk about tightness and about predictability here?

Again, we tried different ways to avoid the two interpretations on bookings, but we are genuine in saying that now that we have six to eight customers, all of them expect us to do the business, even if we don't have the orders.

All of these discussions and our body language at this moment is based on the fact that we are having overwhelming numbers of requests to build up the upside, whatever.

These are not orders, but I can tell you I consider these all in orders.

So they are more secured than an order which, in case of recession, would be in any way pushed out.

So the predictability of these things not because you have a PO.

The predictability is because you know why they are useful.

And if I were to tell you they are used for -- because the next President of France is going to be Mr. X, I would be worried about that, because of -- these things are -- may not happen.

But when you say that you have a necessity to ramp the node 28 and then the node 20 to critical mass, which is you reach still a critical mass which would represent about 250 wafer per month (multiple speakers) -- this is the whole industry [per thousand] wafers per month.

I would think that is going to happen, because it has always happened.

It's the minimum numbers of wafers that the industry would bear or consider critical.

So this is why you see us today not really concerned about bookings, and really working our energy to get to the additional upside and logic that have been asked from us.

Okay, great.

And just coming back to EUV, we've talked about this a couple of times, but could you give us your view on EUV gross margins next year?

What would be reasonable to put into a model?

I could give you my view, but I prefer Peter's view.

I think it is a repeat of what we said before.

Our target is to have the first production tools shipped at, let's say, the high twenties.

And we need about two years.

So let's say 2015 area to get to the corporate average that you see today.

So next year, high twenties.

Yes.

Thank you.

Andrew Gardiner.

Thank you, it is Barclays.

Just as a continuation of the last question really related to the customer concentration and order lumpiness, can you also comment on your lead-times?

I take it that these haven't really changed, and it's just a question of the difference in the time from initial indication from your customers with which you can start to plan the tool manufacturer to the actual putting of pen to paper for the official order that has changed.

Yes.

I think the -- we have of course used -- you may remember, because a lot of you -- the people currently on the call were on the call last year also, that we talked about lead-times of our NXT program about 10 months, 10 to 11 months.

That has actually come down significantly in the sense that we are now six to seven.

So that is a lot better.

I expect anything that we currently see not to have issues with lead-times, with delivery in our schedules now.

We'd also tell our customers to please give us the challenge to be under pressure; they want to have more tools, then fine, we'll start pushing the ciphers down in the factory again.

But six to seven months, that's the general lead-time.

But like you said, and Eric also said, we have these discussions with customers about their shipment pattern.

And then we have that agreement with them which, because of their ramp needs, is almost cast in stone, and especially when it is 28-nanometer logic, which is extremely high in demand today.

So that is what we can plan relatively easily.

And when the orders come, it's just very lumpy.

Like you say, only a few customers where you can have per quarter -- depending on when you get the batch of those orders -- you don't get them one by one.

You get a batch.

It could be hundreds of millions difference between one and the other quarter, meaning nothing.

So all in all, about the order level, there is one thing I would say.

We expect very healthy order levels basically supporting our sales guidance for this year.

Okay, that's clear.

Just a quick follow up on some of the comments you made on EUV.

There's clearly work to be done here, but you are sounding slightly better, I think, that you did the last time we heard from you.

Can you give us any information on how you're communicating this progress to your customers, and what they are telling you that they need in terms of what they need to see to make a determination as to which node and to what extent they begin to build EUV into their roadmaps?

Yes, so we have indeed progressed as we expected to progress.

So you can consider this a victory, because most of the time we've told you that it is not as good as we thought.

So today, I think we could say that we have an inflection point that puts us into a situation of a higher level of confidence.

This level of confidence is, however, only based on the fact, as I said three months ago, that we are doing experiments -- more than experiments.

But we're checking the numbers of upgrades separately.

So, each of your grid at this moment are running pretty nicely -- all of them.

So therefore, we have increased our level of confidence.

But this is not sufficient and good enough, because the customers haven't seen all of these upgrades in, as I said in the speech, in situ.

In situ means you try all this into the same machine, and as expected into our business, you discover a bit of issues.

So, on one hand, the message that we give you today is a message of we now confirm our concept and we are absolutely at 60 wafers or more per hour.

There's no question.

The bad news is, in situ we may discover a number of cleanup activities, optimization that may have an impact of months to the situation.

So the customers are not thrilled with this, because their decision has to be frozen, as we said a few months ago, in the summer.

So the summer looks like a time by which, one, we have to bring certainty; and two, they have no choice and go full steam on one or the other.

So this is why the summer is an important date.

Now, some of you will say, is the summer July 15 when we have our call?

Or is the summer October 15 when we have our call?

You do understand that in my view, the summer is going to be mid-June (laughter) because indeed, it is okay for both of us, the customers and us, to struggle our way with the proof until the proof is solid enough that they will release.

So at this moment, the probability of release is high enough that they negotiate the orders, by the way.

It sounds good.

Thank you.

Well, we'll look forward to the summer then.

Me, too.

(Laughter)

For several reasons.

Peter Testa.

Yes, it is Peter Testa from One Investments.

Thank you for taking the question.

I just wanted to understand a bit on how -- what you are describing between the order versus sales forecast, how that impacts how you manage your business.

Thinking about, for example, working capital that might be up and down a bit differently, holding your costs, commitments to suppliers, these sorts of things, given you will be working without, say, deposits and hard purchase orders for the same length as before.

Yes.

Well, it is actually -- we are planning our business based on the shipment commitments that we agree with the customers.

And as I've told you earlier, then the orders will follow later on.

Whether that will -- whether that means that we have a different risk profile with respect to our investment in working capital, that is practically not the case.

We used to have orders in 2008, at the end of 2008 when we went into the financial crisis, those orders were just pushed out.

You know, those were -- they were never canceled.

They are always pushed out to a point where they need them later, because it's all about that -- well, it's all about technology that doesn't go stale.

It will -- NXT.1950s will still be sold in 2013, 2014 and beyond.

So they never canceled.

They just push back.

We only have a few customers, so we accept the push back, which means that working capital is exactly the same as when we would not have the order.

So in that sense, from a working capital point of view, it doesn't change anything.

That's why the order relevance is not that high, either.

It's just a matter of do we accept the commitments of our customers to take a shipment which will support their product roadmap which is clear to us?

And if we say yes, we have the commitment in the chain and have to add to that our commitment in the supply chain builds up over time.

So it means that the closer we are to the shipment date, the higher our commitment is to our suppliers.

So it is an effective, fairly flexible, flexible system that we are using.

Okay.

And the follow-up question, please, is if you look at the reasons why foundry and IDM are lasting longer, can you give us a sense as to how you see these persisting?

I.e., for example, do you think this will hand over to the pickup in memory or overlap the pickup in memory?

Well, we wish that there is no overlap for nicely spreading good revenue between 2012 and 2013.

And there is a high chance that it is going to happen.

On the one hand, DRAM business is only going to pick up after recent decisions are taken with Elpida.

And only by the end of the year driven by the, multiplied by two, DRAM content in the smartphones.

And the tablets has only come at the end of the year, so we should be having the time to get there.

Okay, in NAND, I said I was not sure about the drivers.

But I could guess that drivers are the volume-driven type drivers, which also are a type of Christmas season, new product type thing, which nicely puts us into a situation of getting the orders of these guys probably in the end of Q3, beginning of Q4 for the delivery late in Q4 and deliveries into Q1.

So we've got a high chance here to have a nice flow of those two segments.

And as some of you have already written in your notes, if you add it to these 11 EUV tools that we will in now recognize in 2013, we are going to have a good chance to have now a managed revenue flow which is stable to growing nicely, although it is too early, of course, to guide on 2013.

Right.

But you think the foundry and IDM persists to that point, or persists beyond that point of when the memory picks up?

There is a scenario where they persists, yes.

There is a scenario where they persist.

And at this moment we have huge discussions with all of them about how persistent and real that is.

There was a question before, which was how do we know what is real or not.

And this is why we are paid, to find out if it is real, because we need to prepare all this.

We need to prepare the supply chain.

We need to prepare the level of risk, working capital, and things.

But it's much too early to guide.

Again, if this is true, this will happen.

But I have to give you transparency.

This is not a dead business out there.

There is clearly activities in the logic business beyond 2012.

Thank you very much for your answers.

Sandeep Deshpande.

Sandeep Deshpande from JPMorgan.

Thanks for letting me on.

Eric, can I try to go back to that NAND question?

You said that you could not exactly answer the behavior of your NAND customers, but in terms of what you've historically been telling us is that your customers tell you shipment horizons.

Have you had any discussions with your NAND customers about future shipments, given that some of these lenses need to be built if you were to be able to, for instance, ship tools to them in Q1, Q2 of next year?

Clearly that's not an order, but it means that you have some understanding of their shipments.

So, let's be conservative at this moment; I would say not a good picture in 2012.

This is why the question before you was relevant.

There is a high chance of the NAND guys are starting to talk to us about how big their business could be in 2013, indeed.

That is happening, but not yet to put in until Q4.

Sandeep, all on top of that, if you talk about Q1 and Q2 2013, and you talk about the long lead-time items in the supply chain, we have ample, ample time to react.

So we don't need to have that visibility today.

Okay, thank you.

And secondly, a question again on EUV.

You're talking about that you are starting to book customers for the production version of EUV.

In terms of that, you made some comments today about the next generation node and where this will be inserted for logic, DRAM, et cetera.

So what kind of capacity should we be looking at for EUV in the first full year of ramp?

You are going to 11 tools next year, is what you've talked about in terms of orders.

If we look into 2014, for instance, can you do 20 tools?

Or can you do only 15?

Well, I would -- again, we don't guide, but I know some of you have a spreadsheet right in front of you, so we need to be giving you the proper direction.

I would say at this moment, if we get two customers, one DRAM, one logic to start, and we get some R&D activities, you should do 15.

Let's not be over-enthusiastic about the complexity.

You know, in this transition customers will always ask you for a lot of things in a transition.

And when the transition comes, they find ways to use most effectively the installed base.

So you have to always be usually conservative as to a new technology.

They want a lot, and then they will reduce it by a factor because they can use your installed base.

Now if, in 2014, we have a volume-driven business, meaning that it's good business, like the economy is doing good, then EUV will be even higher, because EUV would be a greenfield.

And then you already have much more than that.

But at this moment, because none of us can call 2014, you would want to just put yourself into two customer plus two or three R&D and make it 15.

Okay, so you're talking about what you can think about you can ship rather than your capacity, correct?

We could ship more, but (multiple speakers)

Yes, okay, I understand.

(multiple speakers) Thank you, Eric, thanks.

Operator, I think we'll try for one more question.

For those of you that didn't get to ask a question and still have a burning one, feel free to call investor relations.

We'd be glad to get back to you either late today or tomorrow.

So, Operator, if we can have the last question, please?

Mehdi Hosseini.

Yes, thanks for taking my burning question.

Eric, going back to this booking issue, would it be fair to say that since that since these eleven 3300 tools have not gone into the backlog, that is the biggest variance looking into the next three to six months, and that is impacting your ability to project bookings?

Is that a fair assessment?

No, the fair assessment is there's upside to what we were saying there.

We were not even discussing these EUV.

They will pop-up in our backlog.

In fact, I don't know how we are going to organize the accounting for that, but no, whatever Peter said before, which is at this moment we're not worried too much about low booking.

We're just worried about giving you a range, which is not satisfactory to you.

So it is not a low level.

And on top of whatever Peter said, EUV will be coming in.

So let me give you my brainstorming view on that, is we need to -- either we book when we ship, which is one way, or we decide at some quarters to say we are going to book all EUV because it's now officially signed off spec-wise, or something.

I don't know.

But we will not use this as a political item to save a bookings quarter.

In fact, at this moment, our body language, as you may have felt it, tells you that we do -- may not need to do that anyway.

And this is a good, nice last question in a sense, which you now have a sort of translation of our positive body language due to structural buys from our logic customers.

So I think that is a good conclusion, if I may.

Got it.

And one of your top foundry customers was making a statement yesterday that given the challenges in migrating to 14 nanometer, they may actually do 18 or do the development for 18 nanometer first until the lithography bottleneck is resolved.

Do you have any comment on that?

Yes, so we are always advising our customers to have a backup, which is multiple patterning backup.

And yes indeed, you have noticed that some of the customers are now making public that getting to 14 with multiple is difficult.

And therefore, they have to try to do something in between.

And the problem with foundry is that something in between has to be economically favorable to their own customer.

And this negotiation and discussion is happening at this moment.

So we are preparing machine for these scenarios, obviously.

But this is not for 2013.

This is for beyond.

Got it.

Thank you.

Excellent.

Well, now on behalf of ASML's Board of Management, I'd like to thank you all for joining us today in the call.

And Operator, if you could formally conclude the call for us, we would appreciate it.

Thank you.

Ladies and gentlemen, this concludes the ASML 2012 first-quarter results conference call.

Thank you for participating.

You may now disconnect.

Have a nice day.