艾司摩爾 (ASML) 2011 Q2 法說會逐字稿

Ladies and gentlemen, thank you for standing by.

Welcome to the ASML 2011 second-quarter results conference call on July 13, 2011.

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.

Go ahead please, sir.

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

This is Craig DeYoung, Vice President of Investor Relations at ASML and I would like to welcome you to our investor call and webcast today.

As the operator mentioned, the subject of today's call is ASML's second-quarter 2011 results.

With me today in sunny San Francisco, California and cohosting the call is Eric Meurice, ASML's CEO, along with Peter Wennink, ASML's CFO, who is joining from our headquarters in the Netherlands.

I mentioned the different locations, should we get separated, one of the other parties we will -- due to any technical difficulties, we will rejoin the call as soon as we can.

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

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

Just as a reminder, the length of the call is 60 minutes and now I would like to turn the call over to Mr.

Meurice for a brief introduction.

Thank you, Craig.

Good afternoon, good morning.

Thank you for attending our second-quarter 2011 results conference call.

As usual before we begin the Q&A session, Peter and I would like to provide an overview and some commentary on our second-quarter and our view forward.

As usual, Peter will start with a review of our financial performance and we will give added comments on our short-term outlook.

We will complete the introduction with some further details on our current position as we look towards the second half of 2011 and into 2012.

So Peter, please.

Thank you, Eric.

Welcome to everyone from rainy Veldhoven.

As Eric mentioned, I would like to take a moment to share observations on the events of the last quarter as well as some details on our second-quarter results.

The second-quarter sales came in at a record of over EUR1.5 billion, second-quarter sales were driven firstly by NAND customers followed by foundry and then DRAM.

As all our customers continue to add wafer capacity at leading edge as indicated by the almost equal number of unit shipments of KrF systems and ArF immersion systems.

I might also highlight that we have now shipped more than 80 TWINSCAN NXT immersion systems.

The average selling price of all the systems in the second quarter was EUR22.7 million.

That was roughly the same as those of the first quarter.

Service and fuel option sales came in at close to EUR200 million, EUR196 million to be (technical difficulty) income set a record of EUR432 million, which is 28% of net sales which also creates a record EPS of over EUR1 for the quarter.

As an update on our previously announced share buyback program, as of June 26 of this year, ASML has repurchased 13.2 million shares.

As we said before, ASML intends to cancel these repurchased shares.

At the end of the second quarter we still had EUR2.7 billion in cash.

The second-quarter net bookings excluding EUV came in at 34 systems.

They were valued at EUR840 million, which is roughly equivalent to what we ordered -- what we received as in orders in the previous quarter.

But this was a couple of systems short of our guided range.

Booked ASPs rose significantly however from EUR21.1 million in Q1 to EUR24.7 million in Q2 and this was due to a product mix shift towards the higher ASP immersion tools and that included more options.

Bookings strength clearly came from the logic sector, 50% from foundries and followed by IDMs at 31%, so memory sector bookings in Q2 were less than 20% of the total with net bookings accounting for more than 75% of those memory bookings.

Our order backlog exiting the second quarter was a healthy EUR2.76 billion, which did not -- which does not include EUV -- totals about 105 systems with an average selling price of EUR26.2 million, which by the way is almost exactly the same backlog we had 12 months ago.

For the third quarter 2011, we expect net sales to be around EUR1.4 billion which includes to NXE;3100 EUV systems and this represents our first revenue recognition on EUV of about EUR80 million.

These systems will be recorded with zero profit margin.

All order sales will be about 44% gross margin such that the combined margin on total net sales in Q3 will be about 42%.

R&D expenses are expected at EUR150 million, which shows the continued support we give to our strategic investments in the third-generation EUV system development and SG&A will be at EUR56 million in the third quarter.

Now, for the near-term outlook, given a growing uncertainty in the near-term device end amount, our outlook for the third-quarter orders is at the level likely not to exceed EUR500 million.

This indicates a significant customer hesitance to commit already today to any significant capacity planning for 2012.

We think this pause by our customers is logical given that we are coming off a two-year period during which we witnessed a gradual increase in litho systems deliveries, where significant feature shrinks have continued and capacity has been added.

The growing uncertainty in end amounts now prompts an ordering pause.

It is not atypical that pauses like these occur under such conditions and this could last several quarters before a reinitiation of wafer capacity buys will take place.

However in the meantime, all buying sectors clearly indicate that they will continue with their aggressive shrink roadmaps for new product development and lowering their manufacturing costs.

With that, I would like to turn over back to Eric for more on our view of the coming year.

Thank you, Peter.

Let's put some business perspective into this current market situation.

As we already guided, we are enjoying in the second year of an upcycle basically in 2011.

This upcycle is strong.

It has brought us to record sales levels.

We grew to EUR4.5 billion of sales in 2010 and we as we have guided, expect to grow to clearly above EUR5 billion in 2011.

Moreover, this upcycle was in fact relatively easy to call, if you remember.

We were able to guide very much ahead of it as the engines of growth behind it was reached, semiconductor content of products like smart phones, PC servers, and tablets.

So the question today is whether these favorable trends of end products will impact 2012 and if 2012 will be a continuation of this trend or a stable euro down year is the key question.

Recent uncertainty, I will call it microeconomic uncertainty about smart phones, PCs, tablet demands are eroding our short-term customer confidence in general.

The gross run rates of these key segments are being readjusted down due to inventory corrections, PC, tablet cannibalization, tablet growth rate maturing, et cetera, et cetera.

It is therefore natural that our customers are taking time at this moment to assess whether these end demand product trends for 2012 before they clarify their overall lithography capacity plan.

A number of positive news regarding renewed interest in PCs like the super slim PCs or the tablet with keyboards, high server growth for clouds due to power consumption and transaction speed, accelerated solid state drive conversions are not materializing yet in clear demand numbers at this point as no customers are issuing solid forecasts for 2012.

We expect this uncertainty to last until Q4 at least.

Thus our view is that bookings in Q3 will not likely exceed EUR500 million.

In the lithography arena however, we remain fairly confident in the future as we will still have to support a significant unfinished production ramp on the new products like the 3x nanometer DRAM, the 2x nanometer NAND, and the 2x look nanometer Logic Technologies.

We in addition to these production lines, we will have to support fairly aggressive and extremely litho intensive R&D effort for those next-generation node sub 20 nanometers which will start into 2012.

In the past we have suggested that this growth was sustained, will be sustained by the numbers of critical layers, mainly immersion layers which are in fact booming.

This is also true today.

We confirm this trend with an average numbers of layers due to double or triple -- due to the double and triple patterning technology to grow by an estimated 20% across all applications in 2012 versus 2011.

In addition these processes, which is very complex processes of the future, will likely see lower yield and lower utilization than the current nodes, which are easier to manufacture.

These will create sustained demand pressure for new lithography system with extreme overlay specification -- that is higher ASP -- whether for immersion tools themselves of course but also for ArF dry and even KrF dry.

The introduction and delivery of the first NXE, the EUV machine NXE;3300 production tools for 2012 are necessary for the process recipe development of the future will further strengthen the lithography demand due in 2012.

Regarding this EUV development itself, we are showing some progress.

As of today we have shipped five EUV preproduction tools called the NXE;3100 system with three of the five already running wafers at the customers.

One final system, six in total, is to be delivered within a short period of time.

The energy source or the light source remains in fact the gating factor for throughput of these units.

We are making progress because we have identified key issues and have now a strong roadmap for continuous improvement to reach the plant power which will correspond to about 60 wafers per hour on these machines.

We are in parallel continuing the industrialization phase with the completion of our new factory planned by the end of year.

So in summary we are, like the whole industry, confirming that there is a customer pause at this moment while they read 2012's end product trends.

However during this pause, we are still fairly confident that the leading-edge lithography intensity -- technical intensity required for the next step productions will serve as a key driver for the next years somewhat adding a potential macroeconomic lull in the semiconductor end markets.

Or if in fact these end markets prove to be sustained with new products in 2012, these technology difficulties will enhance the growth potential of ASML.

So with this perspective, Peter and I would be very pleased to take your questions.

Thanks, Eric.

Ladies and gentlemen, the operator will instruct you momentarily on the protocol for the Q&A session but before hand I would like to ask you to kindly limit yourself if you could to one question with one short follow-up if necessary.

This will allow us to get to as many callers as possible.

Now, operator, if we could have your instructions and then the first question, please.

(Operator Instructions).

Simon Schafer, Goldman Sachs.

Yes, thanks so much.

It's Goldman Sachs.

Just a question about some of the variability we have seen from the foundries.

You've clearly said in the order book that's already showing up, but I was wondering how much of a risk you may have considered or what you are thinking in terms of potential push outs in terms of delivery schedules as well?

Any thoughts on that would be appreciated.

No, at this moment there is no thought of push out.

As I said, there is a pause mainly due to a reading of 2012 -- maybe Peter.

When you do this thing, you have -- you need perspective -- you need a bit more time.

So the normal digestion of the situation will take at least three to four months or so, so you do not expect -- we do not expect a crisis situation to pop up during the summer by which they would comment every orders on books should be pushed out because we know of something new.

Now what we expect is the foundries or even everybody to say we now need to build up the capacity for 5% growth, 12% growth, 10%, 6%, or whatever of the different product, and therefore we will adapt our bookings forward to some numbers.

So at this moment a pause definitively; a crisis not at all.

Understood, thank you.

My follow-up question would be sort of along the same lines, just excluding EUV bookings, what do you think the minimum run rate is for orderly -- sorry, for quarterly orders just for technology migration in the event that customers are not ordering much capacity for some time?

What's the right run rate level just say technology spend?

So this is the order question when we got into the last downcycle what would be the minimum business that we would sustain just to do technology shrinks, R&D plus a minimum numbers of products, not driven so much by the volume but by the fact that you need to have a critical mass fab.

So if you remember at the time we said to you we think a run rate of about EUR500 million of sales which would correspond therefore to bookings would sustain this technology transition.

But as I said in my little script, we expect those technology nodes to be worse than before, which would probably translate into a very bare minimum need for machine that is 750 or so.

So please, I do not guide for 750 and I don't see this at all because we see volume anyway.

But the good news is technology intensity for just R&D is huge.

And I can even add one more comment to this is that you've got in fact two problems.

One is numbers of layers are so complicated.

The process control is so complicated that they need more time to do R&D and they need more tools to get to the yield that they want.

And then secondly, you have EUV coming in, which doubles the R&D to on a certain numbers of layers for a longer time.

In other terms, EUV requires a recipe.

A recipe requires new resists, requires new masks, requires inspection, and this creates two to three years of hard R&D, which usually would have been done in one and a half years.

So now you have two more years to go.

So all this confirms that the technology content of our demand for 2012 will probably be very large.

I would like to add to that, Eric, if I may that in that previous number that we mentioned a couple of years ago, we assumed a 60% to 65% market share, which is currently 75% to 80%, which of course also brings up this particular number.

I got it, so that basically means you don't think that quarterly orders can really fall below the level that you are expecting for Q3?

I don't think we want to guide more than we said in Q3 because bookings are not the same as billings.

If you know, they could be up and down by one quarter, so they do not reflect the real business trend.

They just reflect a decision-making timing process, so sometime they will take decision with six month's lead time.

Sometimes they take decision with nine month's lead times.

Sometimes they take decisions with three-month's lead time.

So if you put this lead time variation into it, it changes your picture of bookings.

Thanks so much.

Gareth Jenkins.

It's UBS.

Just a couple of quick ones or one and a follow-up if I could.

I just wondered considering your order guidance, wondered whether you would characterize Q3 in terms of DRAM orders as being negligible if any indeed?

And then just as a follow-up on EUV, I just wondered if you talked about the number of layers required ex-EUV.

I just wondered if you could talk about if EUV were delayed by say another year, what the quarter patterning requirements would be and what the impact on your business would be from those alternatives?

Thank you.

Okay, so regarding the very short-term DRAM situation, yes, you have seen the backlog today represents 20% DRAM only in units, so that's fairly small and we expect this to continue bad.

We think DRAM is a weaker segment at this moment with a potential going up next year because DRAM although today is impacting first -- because of the PC reduction at this moment, the DRAM players are working harder to get into the tablet and the portables.

There are much more DRAM now activity for these DRAMs with less power consumption.

There is also activities to work out I would say new DRAM JEDEC specs to make them more akin to scratch pad into all these I would say new platforms.

So there is going to be a pickup of DRAM sometime in 2012.

So but not in the next six months at all where the business will be driven by NAND, by IDM and foundries.

Regarding EUV, I would -- what can I say?

I could say to you that I would say the following in the node -- depending on the customer, between the node of called 20 nanometer or the node called 14 nanometer, which is the industry type discussion item, you could see up to 15 layers of EUV, which is enormous.

But if you don't do EUV, if we are late on one of those, the 15 layers of EUV would translate in 30 to 40 layers of immersion.

So there is even a higher factor than two.

So indeed if EUV is delayed, this will require on some node by our customer and inflation of immersion layers.

Right.

Thanks, Eric.

Francois Meunier.

-- position in immersion with actually Nikon potentially doing a bit better at the moment, are you seeing anything around there at this point?

The second question is really for Peter Wennink.

If we are really heading into a configuration like we had three years ago, i.e.

sales going below EUR3.5 billion, maybe EUR3 billion or even below EUR1 billion per annum, how fast can you cut costs at the OpEx level?

How flexible are you at the gross margin level?

Let's say if we peak EUR3 billion topline for next year, what would be your budget for OpEx and what would be your gross margin?

That's obviously just an assumption, not a guidance.

Okay, so I will let Peter talk.

I think the beginning of your question was cut, so I could not hear fully, but if you say where are we in immersion with the competition, we have at this moment in fact still progressing the NXT, our current high-end machine every year with a set of options which get it to the next level.

So at this moment our competition started in the business a bit later than we have.

Have not yet -- if I understand correctly -- qualified in any significant numbers of layers.

But the next generation layers are even more complicated.

We are adding overlay, throughput, and CD performance improvement to them.

The customers are starting to see extreme difficulty to process control, so their appetite to take risk on another architecture which may not even be cheaper than ours just for the sake of diversification does not make business sense.

So you do not diversify when you do not get value.

At best our competition could match us but we are installed base with so much R&D process, R&D invested by our customer that we probably are in a secured situation at this moment.

So you are not seeing any risk of Nikon coming back at key customers like Toshiba or Intel?

Well, Toshiba and Intel are using Nikon's 620 with current difficulty on numbers of less critical layers and we do not expect this share that they have assigned to them in the -- we will call it strategic fashion -- to change.

Thanks a lot, Eric.

Okay, Francois, on your question on how fast can we cut costs, that's -- we have as you know, we have built a flexible shell around the company and we have tested this also basically every quarter we go through the details.

So we could cut on a total company cost base which includes our fixed costs, our depreciation costs close to 20% within six months.

And that is for instance, we have 22% flex labor at this moment and it's all in that same range.

On the gross margin levels, I would just like to -- I'm not going to guide you on gross margins going forward but clearly we have a target of what we have done in the past.

So I will give you as a proxy you could look at what we have done I would say starting Q1 2010 till today, they have gross margin levels that I think fairly reflect what our target would be under those circumstances.

Okay.

Thank you very much, Peter.

Andrew Gardiner.

Good morning, it's Barclays Capital.

I just was interested in a bit more detail around EUV.

You have started to -- or you are indicating you are going to start recognizing revenue on EUV in the third quarter.

Can we expect further recognition on the 3100 in the fourth quarter?

And also in terms of the throughput targets that you've set out, can you give us an update on where we are with the light sources at the moment and the timing of milestones in terms of throughput?

Thank you.

So, Peter, I think you should talk about recognition and I talk about the light source.

Yes, the EUV revenue recognition, like I said, we will recognize revenue when customers sign the papers to finally accept the tool, which they did.

And in the quarter that they do that, we will book the revenue, which actually it happens in Q3 for two tools.

And over the next let's say two quarters after that, we will -- we expect to recognize the other basically four of which I should say one is kind of a lease type, so that's not going to be that visible, and one will be recognized in the R&D line because it's part of research and development JDA with a research institute here in Europe.

So over the next let's say two or three quarters including Q3, that will be recognized.

Okay, regarding the light sources, clearly we recognize that we made a business error with our suppliers, Ushio Gigaphoton and Cymer, in misunderstanding their level of complexity behind those sources.

So most of our suppliers were not investing -- didn't have the size basically to address all the difficulties in parallel.

This was discovered at the end of last year and since the end of last year we have increased our cooperation with all of them, putting our own people into the different subprojects, adding more subprojects, parallelizing the different issues.

And therefore, augmenting I would say the knowledge of each of those subissues and making all those subissues basically run in parallel with multiple trial and errors opportunities.

So now we feel much stronger in understanding exactly how to do things and when to do things.

Thus the more comfort level that we have on the roadmap to getting to our 60 wafers per hour, which corresponds to about 100 watts.

I put -- illustrated by the fact that there are four major development at this moment on the sources.

And the four path have each an improvement from today.

If you add the four, you will even get yourself above 100 watts, so we are fairly comfortable to get there and we will really feel concern -- or know that we made the mistake of underestimated the complexity of the deal and the fact that we needed to have more people invested into it.

Timothy Arcuri.

Citigroup.

Two things.

First of all, Peter, can you give some idea of -- or Eric, can you give us some idea of whether you have revised downward your DRAM and NAND bit supply growth forecast for this year, given how much bookings have come in?

That's the first question.

And then the second question is on EUV and it seems like the throughput of the tool in the past six months or so is still sort of in the five wafer per hour range.

And I am wondering whether or not it makes sense for you to at some point bring that laser capability in-house, so i.e.

go out and buy one of your laser suppliers and would that potentially help bring this tool to market faster?

Thanks.

Regarding the forecast short term, Tim, we think that within our guidance to be well above EUR5 billion in 2011.

There is assumed about 40% to 45% bit growth on DRAM and assumed 80% plus on bit growth for NAND.

The 40% growth, 40%, 45% growth of bit in DRAM really corresponds to about 0% growth in wafers and mainly growth due to the technology transition.

If you remember, customers have converted from 55 type nanometers to 35 type nanometers.

They had two transitions in one, so the creation of bits has been enormous without having to have wafers.

So these are the two numbers we believe are reasonably conservative.

Regarding EUV at the current run rate, so yes indeed we have a single digit type run rate in the field.

We have proven in the lab more than that obviously and the lab is always ahead to -- from the field by about three months.

So we are a bit better than what you suggest in the lab.

We have as I said taken over much more responsibility on management, the introduction of experts, and project management per se of each of our suppliers.

So we see no reason at this moment to go further than that.

We think we have enough control of the situation by having enough of our people in the different place.

Thanks a lot.

Satya Kumar.

Credit Suisse, thanks.

I just wanted to tie up a few questions on the model and how you are running the factory.

I think you were saying that 750 is sort of the minimum order level that you think the industry needs to do from a technology perspective.

I just wanted to clarify if that includes EUV or not?

Secondly, if we get this low order level for a few quarters, like you said, the first half of next year we could be looking at a revenue run rate of call it maybe EUR900 million plus or minus.

Manufacturing capacity I recollect you were talking about ramping it closer to EUR2 billion in Q4.

I was wondering if you are thinking about ratcheting down some other fixed cost or manufacturing capacity as you plan for early part of next year?

I don't think I said this.

I said -- I was asked what would be the technical minimum billings number to make technology transitions and we said 750.

I never said the bookings number.

As I said, bookings can be zero to EUR1.5 billion because one quarter they will buy it and two quarters they will not.

So look, second point I didn't say that 2012 would be EUR750 million per quarter or EUR1 billion or anything.

We said we do not know at all.

There is no hint, no -- there's only positiveness due to the fact that we know we get into technologies which are litho intensive and litho intensive will save us from disasters if there is a macroeconomic disaster.

That's all we said.

But we didn't say we expect a macroeconomic disaster.

In fact we have no idea.

In fact I would push that back on you.

However, Peter said because of this uncertainty, we have to prepare.

We have to be flexible with our OpEx.

As you know, we have been pretty good at having a very highly leveraged OpEx based on flex workers on one hand and contractors on the other.

We use a lot of suppliers, which allows us to swing quarter-to-quarter our OpEx levels so that we can adapt to any numbers.

But I would prefer at this moment not to comment on the 2012.

We don't know.

There could be scenarios, low-end scenarios, fairly high.

Satya, just to add on that, you have to remember that when you talk about EUR1.8 billion to EUR2 billion output levels and you talk about -- you gave a number of EUR750 million, it doesn't really matter.

That in that sales number, 85% to 90% of the cost of goods is outsourced and so we add value 10% to 15% max of which a significant part is in the labor and labor, like Eric said, we have a lot of flexibility there.

So this is always the way we ran the Company and this is the way we will run the Company going forward.

Very helpful, thanks.

Didier Scemama.

RBS.

Thanks for taking my question.

I would just like to dig a little bit in the tablet PC versus, say, the notebook PC debate.

Can you -- maybe, Eric, can you elaborate a little bit on if you can give us a feel for the litho content if I can call it like that of a tablet versus the litho content of a PC?

And if I were to assume that a tablet have obviously flash memory for the storage capacity, 3G baseband, an application processor, and so on and so forth, can you give us a sense at least a relative litho content of one versus the other please?

Yes, I think we could say that at this moment a tablet would have -- if I take the processors out, I would say current tablet would be more like 80%, 70% of a PC in litho content.

But the tablet of the future is probably very near, so in terms of non-processor comparability, you could get DRAM plus some -- plus baseband, plus NAND to be about the same in the tablet of the future, in the tablet of Q1 2012.

The processor as you know is a big deal difference between an ARM processor or an Intel processor.

So you -- at this moment, you could see a small litho processor from standard Intel process architecture versus an ARM architecture.

In terms of DRAM content per tablet, obviously right now it's much lower than a notebook PC, but when you think about quad core ARM CPUs with [GL] low quad core CPUs going into a tablet, what do you think is going to be the trajectory for DRAM content per tablet say over the next 12 to 18 months?

How do you see that playing obviously for ASML?

What I think we see at this moment is both sides, tablets and PCs are going to be richer in silicon because the new PCs will be thinner and will have bigger solid-state drive and the tablet has -- you don't call it a solid state drive but that's basically equivalent, but it's smaller.

So you are going to have a blurried position in the future.

I would think the PC will be [coming] richer than the tablet, the new PC will be richer than the tablet but the tablet itself will grow.

Yes there will be more DRAM into it, more scratchpad, and the more you put in fact graphics in tablets and things, then it would get the DRAM to be a fairly large number.

So both applications will grow.

It's nearly impossible for me at this moment to understand in fact which products are coming, but we are comfortable on both applications.

Thanks.

Lee Simpson.

Lee Simpson here from Jefferies.

Thanks for taking the question.

Just sort of wondering, we have some idea -- was pushing out silicon at the 22 nanometer spot ready and albeit it's sometimes at a different half pitch than others, I was just wondering, when do we think we will actually see orders emerge in volume for the reengineered NXT-1950is?

Will this come primarily from logic or do you think it will be an NAND [fill in]?

Do you mean the new NXTs every year, which layers are more critical?

Or at this moment let me answer it differently.

About six months ago the customer relationship on EUV was driven by the fact that EUV would be cost effective and easier to manufacture and that there would be a transition period where the customer would hesitate between EUV and triple patterning for cost reasons.

That was about six, nine months ago.

Today we have much more pressure by customers, and all of them -- logic, NAND, and DRAM -- to say that they absolutely -- EUV because there are critical layers which are impossible to do in triple patterning or quadruple patterning.

So we are now getting into situations where you have significant criticality in each of the segments which drives people to say whatever the throughput of EUV is, you better make EUV work.

So how does that answer your question?

Well, it basically says that every one of them today have a critical layer which is pretty hard to execute and for which they are forcing the NXT to its most (technical difficulty) performance which is mainly overlay.

Then all of them will push for an NXT with a better focus control when the CD becomes a significant problem and again NAND, DRAM, and logic will ask for that.

And those guys have even asked that the machines that goes with the immersion NXT would also be extremely precise in overlay because the next layers will have also a very aggressive specification.

So in other terms, what could happen to us in 2012 or so would be that a lot of fabs are renewed also in the non-critical layer, KrF and ArF, just because the old machinery cannot adapt to the new NXT or to the new critical layer performance requirement.

So this inflation of complexity plays in our favor.

Great, that's very clear.

Thanks.

I just wanted to ask maybe a quick follow-on as well.

You have mentioned in the past opportunities you would look at, itinerant ones inside nanoscale.

I wonder if there's any update there and the potential size of any opportunity you are looking at?

So you said other segments, other market segments?

So as I mentioned often, we are always interested to look at segments which can grow outside of semiconductor that we choose as our expertise.

We are continuing to invest about EUR5 million a quarter in R&D, so in the numbers that you have, you have about EUR5 million.

And I confirm that we have at this moment one technology that we feel fairly interesting to -- with multiple applications in fact at this moment but we are trying to make work one application which could lead to us developing.

But it is too early to mention.

But it is part of our investment into possibility outside of semiconductors.

Okay, thanks.

Kai Korschelt.

Thank you for taking my question.

I just had a follow-up on the -- particularly on the H1 revenue question.

Clearly the bookings level is pretty low now and you will ship 85%.

So the bulk of your backlog in the second half of this year, which means that if there are no follow-on orders and we have let's say two quarters of less than EUR0.5 billion in bookings, then the backlog let's say at the beginning of 2012 could potentially be very small.

So is it really unthinkable given the technology migration and other plans that let's say Q1 revenues could be below EUR700 million for example?

That would be my first question.

And then the second question was just on the gross margin profile for the NXE;3300, the (inaudible) next-generation EUV tool, should we think about that as being higher than the 3100?

Should it generate closer to group gross margins?

Thank you.

I will leave the gross margin discussion with Peter and on the Q1, we just don't want to give any guidance in 2012.

The only thing I would say is at this moment there is no reason to think that whatever is being discussed in the industry has a lead time of more than six months in terms of the impact at this moment what you dispose that we are seeing is opposed due to the fact that they don't recognize the Christmas season basically, that there is a question in the industry about what is the next six months, how does that fit Christmas season?

What are the adjustment and adaptation tablet PC?

What is the adaptation of these DRAM versus NAND, new product, (inaudible), etc.?

So you've got whole bunch of I would say short-term impact and of course you have the macroeconomic questions about the Greeks and Italians and the Americans going bankrupt at the same time.

But again, there is no -- in the industry obviously there are no customers who say I am preparing for a significant 2012 reduction.

I have not seen this.

There is no facts about it at all at this moment.

Peter, regarding the margin?

Yes, the gross margin on EUV, as I said, the 3100 is a research and development tool sold at just over EUR40 million, which is also the cost which is not going to be the margin of the 3300.

The 3300 is the production tool which we will start shipping in December of 2012.

Our target gross margin as I said before, in the high 20s, low 30s, and within two years after introduction we should be at corporate average.

So it is going to be a gradual struggle upwards, like it has always been also in some of the previous platforms but that is currently the plan.

Okay, great.

Thank you.

Sandeep Deshpande.

JPMorgan Cazenove.

Thanks.

Just a quick question on -- if you look at previous cycles, if you look at the last equipment cycle, it started with strong orders from the foundries, in the latter half of the cycle, it is driven by memory.

Do you see this cycle also playing out in a similar fashion?

Secondly based on your conversations with, say, the NAND Flash companies, how do you see NAND build out in the next few years?

Will that be the main driver of your orders or are we still looking at upgrades on DRAM logic et cetera to be a key driver?

Sandeep, that's a very good question.

I think first of all, we can't answer again 2012 easily, but it is true that 2011 was not the greatest year on memory and that there is a high chance that memory including DRAM coming back and NAND being a fundamental sector would be the big guys of the next cycle.

NAND is definitively driven by these new storage limits or the solid state drive and those server drives which are kind of hybrids.

NAND will also develop sales which have much higher access time which will then allow them to be used in such scratchpad type memory and servers.

There is a case that says that NAND will eat up into the DRAM business, so people in the NAND arena are going to invest a lot.

So yes, we could -- you could bet that this is an engine which has a high chance to be of the level of 50% to even 100% higher than DRAM in the future.

But on the other hand, we have heard that the DRAM guys are saying if you really go 3-D in your graphic processing and your video and if you want to have all of this in a platform which is mobile, you are talking about a doubling here of the bits easily and then now you're back into good business.

If you do this into a -- with the technologies which have fairly low power consumption, you are creating a need.

So in other terms, today those people, NAND and flash, haven't really created -- and DRAM haven't really created a lot of wafers, so you could see them coming back.

But if on the other part, the logic part, you could say they would -- they have invested significantly for the past two years and they may go now into a mode of maintaining, but these people are getting into numbers of layers which are enormous.

I mentioned during the call -- as took an example -- in fact on the 14 nanometer, you can go as far as 17 layers of EUV or 40, 45 layers of immersion.

This is huge technology lithography impact.

So you may not have such volume requirement, but you may have a lot of technology requirement.

So those guys can in fact not so much go down.

So my best bet would be to say 2012 you would see less logic and more memory and then in 2013, 2014 you could see again memory -- logic back in because of the technology.

Thank you very much.

Mehdi Hosseini.

Thanks for taking my question.

Eric, early on you said that in your R&D facility, you have -- your EUV tools have a better throughput.

Can you share with us if you can where's the throughput for those tools?

And remind us what the milestones are?

Where do you see the throughput for your EUV tool as you exit this year?

What are the critical run rates that you need to have before you gain further traction with customers?

Traction with customers, this has two facets.

One is traction for R&D recipe development and one is traction to get decision as to which -- how many tools to put into production for how many layers.

These are two separate questions.

The first question is am I having happy customers with a low throughput?

Well, at this moment, a customer would tell you that if we succeed to do 100 wafer per day, it is sufficient to do recipe.

And at this moment, we are not there yet.

This is annoying them, so they were not exactly giving them the throughput that allows them to have a comfortable recipe building.

And we are trying to get there and we think we can get there before the end of the year.

Regarding now committing throughput which then allows them to take a decision whether they will put EUV into one layer, two, three, four, five, 10, or 17, is sort of committed throughput.

So at this moment we have a roadmap easily defined up to in fact 205 watts which it is 125 wafers per hour.

And we have -- this roadmap is committed between roughly 20 wafer per hour, 40 wafer per hour, 60 wafer per hour, 90, and 120.

And this has been given to customers, but of course we need to have proof now that we are going to make it before decisions are taken.

And the decisions will be taken as a combination of when they have to be taken, that is the customer has no choice.

They have to jump into the swimming pool.

And secondly, we have enough facts.

So these decisions are going to start I would say around the fourth quarter.

In the fourth quarter, the customers are going to say I have now no choice.

You have to commit, ASML, about how big.

So when are you going to ship, and at what speed?

That has to be a firm commitment.

We expect that to be in Q4 and with our response, they will say, okay.

Based on that I am going to buy two tools or I am going to buy 10 tools because I will do one layers or I will do 100 layers.

So that's our decision.

So today they are not comfortable that we don't know, but they're not uncomfortable.

They can still wait another six months before they commit to production I guess.

Great.

And just one clarification.

Should I assume that what you are doing in-house, your [lapse] is well over single digit per hour, does it start with a one or with a two?

The 10s or 20s of wafer per hour?

I don't think I am going to answer that question.

Your question is in fact not that easy to answer because they are -- when you talk about proof, you can have proof even of digit three.

Even, we have seen some possibility of doing more than 30, near 40, but the question is then every other spec around it.

So if you said to yourself, if you are an optimist, you say at least we have proven more than 30 for two seconds and a half in a dark room and the question is can you light on the room?

So there is a parallel path between different things.

What I can tell you however is we feel much more comfortable now regarding the level of engineering difficulties and discoveries.

The discoveries is much less and the amount of work is enormous but the amount of work is not.

So by proofing for one second that something works at 30 watts or some -- (inaudible) wafer power is already a very large positive impact because we say, okay now we know how to do it.

We just need now to do the engineering around it to ensure that it lasts more than a second and it can last a full week.

And then you are in business.

Sure, and then just one rather theoretical question.

Can your customers avoid EUV even for critical layers if worst case the throughput is not improved?

Or would they have to delay the migration to the next node?

I would like to make a subtle answer to that.

Absolutely no.

They have no choice.

In the days of time -- assuring factor will not be -- which cannot be done differently than EUV.

So there you have it.

So their alternative then is not to shrink.

Okay, great.

Thank you.

[David Wu].

Yes, (inaudible) Global Research.

I've got two quick questions.

The first one is I assume that if we have delays in EUV deployment, it's actually from a profit standpoint better for ASML.

And the second one I have is I remember that in the first-quarter conference call you were thinking that all the -- even the 3100 EUVs revenue will be recognized in calendar '12.

I noticed that they are getting recognized in calendar '11.

I was wondering what is the reason for that?

Okay, so, Peter is going to answer the last part of your question.

But regarding EUV delays, yes indeed if we get into a situation where we delay and therefore the customers have to buy more NXTs at -- and for more layers with an NXT with a higher margin obviously, it will be better for the profitability of ASML.

So the EUV project per se is important for us strategically and for the future but during the period of transition, a delay would in fact be positive for the financials, not negative.

Yes, so on the revenue recognition, our expectation was indeed three months ago, it will be 2012 but since customers are running the recipes and of course like you will, they are currently not yet hitting the target 60 wafers per hour.

They are actually using the tool and they have accepted at least two for the third quarter, the full ownership of the tool, which you know, if you have to sign a document, you have to recognize the revenue, and they are using the tool and they are running the wafers and they are running the recipes.

So that is the reason then.

And we do expect with the developments that are currently under way in the program that the remainder of those tools will be recognized over the next two to three quarters.

Thank you.

Ladies and settlement, I think we have time for one more call.

If you were unable to get through and have some questions, feel free please to contact the investor relations group in the Netherlands or here in the US.

Now, operator, if we can have the last caller, please.

Peter Testa.

It's Peter Testa from One Investments.

The question, in the past you had talked about 13 fab projects with EUR7 billion of litho business for 2011, 2012.

I was wondering whether you would give -- hazard a view on how the state of those projects and overall market looks given the uncertainty?

And then the follow-up was the degree to which some of these project steps or even the uncertainty in general is also engendered by the lack of clarity, whether they need EUV for one machine or 10 machines if you set the different layers at this stage rather than just purely being a macroeconomic discussion?

No, I think the EUV discussion as I said, I would not say our customers are happy and comfortable about us I think to push a decision more towards Q4.

But they don't -- they can live with this, so the big decision is about which node and which sort of layers and which technology is not impacting I would say seriously the capacity build plan.

So we are not guilty through EUV of not having the bookings.

So this is not the point.

However, you are correct in saying that at this moment all the foundries like the DRAM, like the NAND people are looking at those fab projects that they have and try to understand what they really need to do and accelerate or decelerate in 2012.

So whether you are talking about Samsung global foundry, TSMC, [UMC] (inaudible), all of those are saying, well, we need to be sure that our next tranche is done based on the 2012, 2013 year, which justifies it.

So this is what the pause is all about.

Okay, so it's the same 13 projects, the same 7 billion litho.

It's just a question of whether they go ahead as rapidly as before rather than a view on the shape of what they do because of whether they need a structure that involves more or less EUV?

Absolutely.

Now there is -- if you wanted to have one more [facets] for your information, there's more questions regarding some of the customers' capability to yield the current process before they get to a next batch of fab.

So in other terms, as I mentioned to you which I think is good for litho is the processes are more and more complicated to build, so it's more and more difficult to yield and when it's difficult to yield, then the customer would say let me delay a bit until I get at least five tools working before I gets to buy the other -- the next five.

This is by the way playing in our favor because remember what we always said, we didn't want people to have too many tools too early.

We would prefer to spread tools for longer period which this issue plays in our favor.

We sell a critical set of tools.

This then is some work to make them work and then after that, we sell another batch and we sell another batch.

So this currently is also being looked at by logic people who says, well, it is indeed difficult to yield new nodes.

Very good.

Thanks very much for the answers.

Now on behalf of ASML's Board of management, I would like to thank you all for joining us on the call today and, operator, if you would formally conclude the call I would appreciate it.

Thank you.

You are welcome.

Ladies and gentlemen, this concludes the ASML 2011 second-quarter results conference call.

Thank you for participating.

You may disconnect now.

Thank you.