艾司摩爾 (ASML) 2010 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by.

Welcome to the ASML 2010 fourth-quarter and annual results conference call on January 19, 2011.

Throughout today's introduction, all participants will be in 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 morning and good afternoon ladies and gentlemen.

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

As the operator mentioned, the subject of today's call is ASML's 2010 fourth-quarter and annual results.

Co-hosting today's call from our headquarters here in Veldhoven, the Netherlands are Mr.

Eric Meurice, ASML's CEO.

He is joined by Mr.

Peter Wennink, ASML's CFO.

At this time I would like to draw your attention to the Safe Harbor statement contained in today's press release and in our fourth 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.

The length of the call will be 60 minutes.

Now I would like to turn the call over to Mr.

Meurice for a brief introduction.

Good afternoon, good morning, thank you for attending our (technical difficulty) full year 2010 results conference call.

Before we begin the Q&A session, Peter and I would like to provide an overview and some commentary on the fourth quarter, and a view on the total year.

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

I will complete the introduction with some further details on our current position as we look back on 2010 and look forward to 2011.

So please, Peter.

Thank you Eric, and welcome to everyone.

Our fourth-quarter and full-year numbers provided us with several financial records, including record sales, profits, cash flow and system bookings.

I would like to take a moment to highlight some of these along with some other details and observations on the fourth quarter and the full year.

Our fourth-quarter sales came in above our guidance and set a record at [EUR1.5 million].

Sales versus the third quarter grew in favor of NAND memory and foundry customers as they began to add further wafer capacity in these sectors.

The number of dry shipments, mostly KrF, represented well over 50% of the total.

As a result of this, the system mix -- new system ASP dropped slightly from around EUR24 million in the third quarter to around EUR22.4 million in the fourth quarter.

Our service and field option sales came in just over EUR200 million, driven upward by further adoption of several of our new holistic lithography products and continued demand for various other system performance and enhancement products and service products.

Q4 net bookings came in at a remarkably strong 117 systems, valued at EUR2.3 billion.

Let me mention here a slight change in the definition of our book systems and backlog.

Over the course of the last several quarters, we have been closing an increasing number of long-term contracts that involve not only the lithography systems but other sales-related items, such as system options and certain services that will create sales after the tool ships.

Through the third quarter of 2010, we did not include these additional items in our systems bookings.

However, due to the growing size of these additional purchases and the growing importance of our holistic litho products, and in order to better represent future sales potential, we have started to do so in Q4.

And we'll continue to do so in the future.

In reviewing the impact of this new bookings definition on the previous five quarters we see a high single-digit percentage as a positive difference.

I'd like to refer to slide 15 of our Q4 results presentations to see where the strength on our Q4 bookings came from.

We see significant increased demand from NAND manufacturers, representing an increase of over 200% in book value versus the third quarter.

Foundry customers showed even greater strength, with a 300% increase in the book value Q4 versus Q3.

And our IDM customers reduced their order values versus Q3 by approximately 50%.

But notably as also that the DRAM sector that reduced their orders versus Q3 by 40%, a likely reaction to a current supply demand imbalance in DRAM devices.

Strength in NAND foundry and IDM should be considered largely in support of new fab projects, as identified on slide 18 of our fourth-quarter presentation.

It's important to realize that customers often commit purchase orders for critical lithography systems for new and major fab projects as early as possible, in order to ensure timely system deliveries against factory expansion plans.

Notably, for the first time in many years, total memory backlog, which is DRAM and NAND, represents less than 50% of our total backlog.

Our near-term demand outlook firmly indicates support for another strong year.

Our existing backlog plus our normal annual service sales, and our expected bookings in Q1 -- while of course less than Q4 -- justifies the expectations of net sales above EUR5 billion in 2011.

The quick final comment on cash.

As reported, we are building a strong gross cash balance and expect to grow our cash balance further in the coming quarters.

Therefore, we announce the implementation of a share buyback program whereby we will purchase up to an equivalent of EUR1 billion of our own shares.

[We're in] maximum period of 24 months.

In addition and thanks to also our strong financial position and operating cash flows, we intend to propose to the AGM a doubling of current dividend of EUR0.20 per share to EUR0.40 per share for 2010.

And this dividend increase reflects our confidence in our prospects for a sustained longer-term profitability and growth.

And with that I'd like to turn it back over to you, Eric -- (technical difficulty)

Thank you Peter.

We look back on a year of sales growth and financial performance for ASML that confirms our earlier view that the semiconductor recovery cycle was, in fact, strong and sustained.

I think it's good to recognize that we had a good understanding of our drivers a year ago, and that we managed to this expectation.

We now have a backlog of orders to be delivered in 2011 which, with the expected bookings to be received in Q1, justifies expectations of further revenue growth in 2011 versus 2010.

The fundamental question is whether these orders will remain if the end demand suddenly collapses, or whether the current demand trends will continue into 2012.

It is fairly common in the history of semiconductors to see two to three years of an up cycle.

This time, we are encouraged by the following factors.

First, the memory lithography capacity buildup in NAND and DRAM at this moment is fairly measured.

And the backlog that we have, that we showed you, does not yet support that analysts' expected and product growth for 2011.

Namely, the analysts are expecting 59% bit growth for DRAM and 84% bit growth for NAND.

The second factor which is positive for the future of 2011 is that the logic capacity built up, which occurs currently at the foundries, is certainly ahead of pure semiconductor markets (technical difficulty).

But, as we noted in the past, it is needed in order to put strategic capacity in place for the new 40 nanometer nodes.

We estimate that the overall capacity built will sustain about 14% logic unit growth versus an expected 8% unit demand growth expected by analysts for 2011.

So 14% capacity built versus 8% demand growth.

So in a sense, if analysts are correct, the foundries are building ahead of the need by some 6 to 9 months, which is real but measurable.

The third positive aspect about the trend is that all semiconductor segments are accelerating the technology transition with all transition being more lithography-intensive.

We mentioned two or three times; but the numbers of lithography layers, in particular immersion, and of course this is continuing.

DRAM vendors are targeting one new node per year now, like the NAND manufacturers from 18 months per year -- 18 months every generation.

Leading respectively to 30 nanometer and 20 nanometer ramps in 2011.

Logic transitions to 40 nanometer in volume and 30 nanometer in prototyping, and processors ramping 20 nanometer in 2011 and 2012.

So this is a sort of positive trend.

The fourth positive trend is that none of these favorable litho forecasts include assumptions of potential lower yield and process control difficulties that will likely accompany these technology transitions.

This is a fourth, a potential upside for us, if the industry takes a bit more time to reach the same levels of yield maturity and stability that they have been accustomed to with less aggressive technologies.

So these are all the positive trends that we see.

And all this capacity being built is also supported by very rich, leading-edge technology content in the end products themselves.

The drive for affordability and power [swift in this device] are certainly not abating.

But additionally, there is a wide array of new Internet connected device, like the tablets, the e-readers, the high-end smart phone, connected TVs, and also the solid-state hard disk drive, which are biasing the demand for high-volume, high-end, high content semiconductor device.

Assuming a continuation of this technology trend into 2012, we will have to support another huge technology transition.

As the new technology nodes by then will require EUV development to start in earnest, with significant investment in lithography capability, only to support -- not manufacturing, but only to support R&D development and qualification.

We will have shipped six NXE 3100, our second-generation extreme EUV system, by the end of 2011, with revenue recognition earliest in 2012.

We have received, as you know, nine orders to date for our third-generation EUV tool, the NXE 3300.

This system will start being delivered in 2012.

The industry still requires [a number of months] of course to confirm the EUV performance, the plans and the roadmaps.

But it is encouraging to think, to envision this as another engine of growth for 2012 and beyond.

In addition and in parallel to all this, we are continuously expanding our product differentiation in immersion in particular with our continuously evolving NXT performance, and the addition of holistic (technical difficulty) suite of products.

We are also improving our KrF product lines with unique throughput and [overlay] performance, all this necessary for sustaining the new node specification and for maintaining our market share.

With all this, Peter and I would be very pleased to take your questions.

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

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

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

With that, operator, may we have your instructions and then the first question please?

(Operator Instructions).

Gunnar Plagge.

It's Nomura.

Thanks for taking my question and congratulations to the results.

Eric, you mentioned the drivers maybe for 2012 that could provide upside to estimates.

On the other hand, if you look at industry tool productivity, that should also substantially go up in 2011 and 2012 with a new product cycle being launched.

Could you maybe talk a little bit about what you expect on the overall installed base?

What are the increases that we can expect in terms of tool productivity over the next two years?

Certainly, thank you.

Everything we [assimilate], the terms we look at, the demand for [casts of] semiconductors, then we calculate back how many machines that are necessary to do this.

Always simulated with the latest and greatest performance of our machines as well as install base upgraded.

So we are clearly -- what we always do not mention to you is that we are selling a lot of options which are served at upgrading install base, and they are usually upside.

But they are clearly used to calculate the numbers of machines necessary to build the given numbers of wafers and of dies.

Could you give us an indication about percentage changes over the next two years?

Yes.

So, you could -- at this moment I would say we would have a family of immersion tools in the field that runs at about 150 wafers per hour, and we are planning to convert this to a family of tools doing 175.

It's not going to happen immediately into 2011.

It will go well into 2012 and even 2013.

But of course when we bring EUV back in, we are going to bring machines which goes towards 120 or 125 wafer per hour.

So we will restart from this type of level, although there'll be less layers processed on EUV.

Thanks.

And as a follow-up, you haven't given us a quantitative number for the Q1 orders.

And clearly you've discussed before whether it actually makes sense to give guidance on orders, but could you give us some broad color would you expect going into Q4 -- Q1, what are ranges maybe that you're seeing at the moment?

Let me answer that question.

Clearly Q4 was an extraordinary quarter in terms of our bookings, that was clear.

And many of those bookings were focused on fulfilling the needs of our customers looking at their fab expansion requirements for 2011.

Now, clearly we were surprised.

Hence, we gave the pre-warning on the orders early December.

But it also creates a situation where we think Q1 will just be a bookings quarter which will enable us to, let's say, fill up 2011 from a shipment point of view.

And the rest of the year we'll really be looking at order momentum, which will be focused on 2012.

And I can hear you thinking of the next question, which you're probably not allowed because it will be question number three, (laughter) but I'll give you the answer anyway.

It's -- on 2012, the way that we look at it, we don't see any negatives for the year 2012.

We do see positives.

We do see measured investments in NAND in 2011.

We definitely see logic.

There's a lot of people think there's a lot of logic budget there.

But the speed with which this logic ramp up will take place in 2011 will be measured also, and will be largely towards the second half of 2011 and will extend into 2012.

That's number one.

Number two is, and Eric talked about that in his introduction, there is a technology trend that is [continuing] continuously which will drive the need for technology upgrades at our customers.

And last but not least, that's also driven by EUV.

So in 2012 we will see the EUV introduction with very high price tools, measurably starting the new generation.

So in short, Q1 will see a finalization of our shipment pattern and shipment expectation of full year 2011, and the rest of the year will be in the eye of 2012.

And like I said, we see a lot of positives.

Thank you, very helpful.

Nic Gaudois.

Nic Gaudois for UBS.

Just wanted to kick off with where you just left things, Peter.

Effectively we're seeing a stand-off between Intel and TSMC, (inaudible) on CapEx.

And we're seeing capital intensity for logic and foundries combined [going out for] first time in ages.

I mean, do you actually think there is a transfer of [cyclical] risk away from memory potentially to that arena?

And how much does this concern you potentially for 2012?

I think, like I said earlier, what -- [well], we look at the shipment pattern for the (inaudible) projects that you just mentioned for those customers.

That's going to be a second half 2011 phenomenon, which will actually lead into 2012.

Also, if you look at the spend pattern for the foundries, don't forget that everything that is leading-edge investments today, build out of any capacity which is sub-65 nanometer, is only by a very few customers, only three.

So those customers are the logic industry, with the exception of the big micro [processing] maker.

So in that sense, to look at that number in the context of what happened 10 years ago is a bit difficult.

And at least we are not looking at it in that way.

But also, we discussed this.

(inaudible) these large CapEx investments in the foundry is in fact to be [relativized].

We only think it will create, if you add 2010, 2011, just 14% capacity growth.

Remember that in this logic business when you go to 40 nanometer, so you go towards the very large numbers of critical layers, so you need a lot of machine to do not so many wafers.

So this capital intensity in dollars doesn't mean that we create a lot of capacity.

So our calculation is indeed that this 14% unit growth is higher than what analysts talk about, 8%.

But, as you know, a difference between 14 and 8 is six months or nine months digestion time, which in the environment where things have significant yield issues and timing capacity, management difficulties, etc.

is kind of not normal.

So in other terms, yes, indeed, there is on paper an overinvestment.

But it is not material such at this moment, and that it will be eaten up into 2012 by the ramp of these new products.

Which will happen -- even if the volume for the total macro economy is not there, the mix would be there to pick them up.

Great, I appreciate your thoughts on that.

And very quick clarification, but you say for NXE 3100, no revenue recognition before 2012.

Did I hear you correctly?

So shipments this year, but all revenue recognition in 2012 basically?

That's what we expect.

[But if sooner], and it's great, then we'll let you know.

Great, thank you so much.

Odon de Laporte.

Odon de Laporte with Cheuvreux.

I had a question relating sales guidance in Q1.

I was wondering why you looked a bit cautious.

I mean, your guidance for Q1 is weaker than actual [states] in Q4, so could you clarify the reason behind this conservative statement?

It's not that conservative, and it's because of the following reason.

We did [EUR1.5 billion] in sales in Q4.

There's about EUR50 million, EUR60 million in there on billable service which was in the fourth quarter.

Also a bit of surprise to us that it was really customers making up their budget for that year, which included system refurbishments at customer site, which we normally did here in Veldhoven.

Now we did it at the customer site.

Customers are trying to monetize some of their assets that they have.

They used to sell those tools to us and then we would refurbish and we would resell.

That's not what we do now.

It's -- so basically we do those refurbishments at their site and we can charge a nice and good price for it.

We had system relocations.

Customers are shifting around systems because they need to make place for their new leading-edge tools, and we had a lot of fast installs.

Customers want those tools to be working fast and they're willing to pay nice service sums to get those tools installed.

So we had, you could say, kind of windfall sales of EUR50 million, EUR60 million in Q4.

On top of that, there are EUR75 million worth of system sales that were actually a few evaluation tools that we shipped to customers in previous quarters, that we agreed with them they could use for a few months or one or two quarters, and then they would buy them.

That happened in Q4.

So you could say that the shipment was in fact, if you peel it all off, it was about EUR130 million worth of business, (inaudible) worth of business.

It was a bit of a windfall business.

If you take that out, then the guidance of [1.4] is not conservative.

It's actually, you could say, equal but even a bit up.

Okay.

Thank you very much.

Francois Meunier.

Yes, it's Francois from Morgan Stanley.

Thank you very much for taking my call.

Just if you could give us a bit more details about what's going on with EUV.

I'm sure you are working very hard to get it production-ready with your partners on this one.

What are the key milestones [that you see] for the next few quarters, and when will you communicate on this, if you can share this with us?

Thank you.

Of course.

This is a very important question for the future.

So we -- positive milestones -- is that we've shipped the next generation tool, the 3100.

Positive milestone is that the customer [himself] have been able to make an exposure that we've published at 27 nanometer.

Positive milestone is this exposure looks good.

The underlying roughness and the process parameters, in fact, looks very encouraging.

That the process issues or lot of process issues are in fact going to be acceptable.

In fact it's pretty good that immediately after the machine is delivered, that they can even show such a good [line of mass] capability and exposure.

So all this is good.

Next item is we do have indeed five more of those machines, and you can see them, touch them, etc.

All of them are being qualified at this very moment.

And they're reaching a level of overlay which is now near the target spec.

This is again difficult because this is in vacuum, so we have passed another test at the level of the machine.

So on the machine side, it now seems to be fairly much under control.

The negative part is we are still late in proving 60 wafer per hour.

At this moment we are nearer to 10 to 15 wafer per hour than we would be to 60.

This is not yet a concerning level because the customers do not need 60 wafer per hour, clearly, to do R&D.

But it would be better if we were to prove as fast as possible that we could reach this.

So the positive news there is that there is now a much more detailed action plan to resolve every problem of the stage of a laser, basically, to get it to the power necessary.

So there are issues on the pre-laser, there are issues on the -- sorry, not issues.

There are solutions on the pre-laser.

There are solutions on the droplet performance and stability.

There are solutions about delivery mitigations.

There are solutions about catching some of [rays] which are, how you say, (inaudible) negative influence [power rays].

So all this is well-defined and there's work to be done.

So I would estimate that we could probably open a bottle of champagne in the Q3 timeframe, by which we would say this is now past the proving time, and going into just executing time.

So it's more a question of actually executing into the next three, four, five months more than the issues which are technically difficult to sort of --

Well, with this statement the engineers are going to crucify me.

(laughter) I have a different -- the point is, I would say, they are two or three -- no, I would say four conceptual things that have to be proven.

But everything else is very difficult because it's not easy to make the machine.

So making one machine is a complex task.

But it is a task that you can -- that are deterministic.

So, all the deterministic tasks will take time and energy and cost.

The one which are less deterministic is what we've done here to improve, for instance, stability of droplet.

Is it going to improve it by 50%, by 75% or by 40% only?

And these are things you have to prove.

So, I would guess there's another six months of work on two or three things that have to be proven to know exactly where you are on the specification.

Thank you very much.

Janardan Menon.

It's Janardan from Liberum Capital.

Two questions.

One is -- can I just have your views on M&A?

What kind of things would ASML be looking for?

And then what kind of timeframe can you give us for when [all] that could be implemented?

Second question is actually -- if you go back seven, eight years, when your order book was quite foundry-intensive, then it became more memory-intensive in the latter part and subsequent years.

But foundries had a propensity to cancel orders once the capacity utilization levels fell, much more so than IDMs or the memory companies did.

And today your foundry orders have gone off quite sharply.

Would you say that there is such a risk?

As we come to the second half of the year when some of these tools are due for delivery, and if something happens in the demand cycle and the utilization levels fall, would you say there's a bigger risk that some of these orders may not materialize?

Or is that more contained because of competitive environment and the requirements for technology transition?

Two good strategic questions.

On the first part of mergers and acquisitions, the first order of business is lithography.

Clearly at this moment you've noticed that we've reached nearly our target of EUR5 billion sales which we set ourselves for five years ago or so.

And it is now time to set ourselves another target for the future, 2015.

And at this very moment we are still calculating exactly what that number could be, because again we have had now success meeting our target.

We'd like to do this again.

So in other terms, the way I say it is to say lithography is still our fundamental market and we still think there is growth to it.

I just will not be able to say it today whether it's EUR6 billion, EUR7 billion, EUR8 billion, EUR10 billion, EUR9 billion or whatever.

But it's still the area privileged for ASML because there is growth opportunity.

But it is -- in parallel, we are always looking at potential expansion to a second leg, which is a leg which will create growth opportunities.

And yes, it has to be in the area that we are good at, which is equipment, precision, high R&D, long leadtime, etc.

And we have been on a regular basis looking at different options.

And we are continuing.

I think that nothing has changed on this.

The acquisitions themselves will not be large in the sense that we don't think we can buy a company that exists today and we can easily create growth out of it.

But we think we can probably buy a smaller entity which has a leg in a technology or a leg into an application that our size and expertise could multiply by effect.

But again, it's not a short-term item at this moment.

We are continuing to investigate.

Regarding the foundry behavior, well, every segment would cancel orders if they didn't have needs for machines.

We all know that.

It's a fact of life.

So I would not say the foundries are worse or better than others.

So if there is business we'll use it.

If there's no business, they will ask us to delay deliveries and everybody will do this.

What has changed from seven years ago is to the time the industry was more -- I need a fab and I will find a customer later, or this is what year 2000 or 2001 or 2002, areas of people got crazy about -- there is no limit to applications, and we just need the fabs.

We don't know where it is but we need the capacity.

I think this you don't see anywhere at all in this industry.

So here, we've got the point about global foundry, GSMC and Samsung logic who are trying to set themselves a space in this business, so they have to have a critical mass size.

But I would say the smallest possible critical mass, so that they can have customers.

So that's what they do today.

None of them have come up and said we think the market is going to be triple what its historical trend is.

In fact, when I was able to articulate 14% growth versus 8%, it's also based on mix, and it's of course based on technology.

And I know exactly each product etc.

(inaudible) technology.

And they are all based on I would say historical ramp of new products.

And you add to this my famous 14% versus 8%, and this is the overspend.

So I didn't say that there is a major catastrophe in the macro economic environment.

All those customers, including DRAM and NAND, will not come back to us and say -- your nice backlog (inaudible) contractually, we have to keep it of course.

But could you please delay deliveries?

This will always happen, could happen.

[There may be] circumstances where the economic would push it.

But at this moment, again, I do not see a significant over-investing compared to the end product needs.

Okay, thank you very much.

Satya Kumar.

Satya Kumar, Credit Suisse.

Eric, I was wondering if you could remind us what your capacity plans are for the year.

In Q4 it looks like the systems shipments were close to EUR1.3 billion.

How do you expect that to [ramp] through 2011?

And I have a follow-up.

I think that the capacity is now reaching towards EUR1.5 billion, EUR1.6 billion of system per year.

So you add EUR150 million, EUR200 million service to it, so it would be EUR1.6 billion, EUR1.7 billion capacity shipment -- (multiple speakers) sales.

Sales capacity between now and June.

And are you looking to ramp that level or is that a sustainable level, do you think (multiple speakers) of capacity?

Yes.

We will ramp through efficiency, so if you say you reached by hard work EUR1.7 billion, we could probably reach EUR1.9 billion to EUR2 billion by December, by quarter, just through efficiency.

And a quick follow-up.

In terms of the Q4 orders, I noticed that the IBM orders were actually down sequentially, which looks a little bit odd relative to some of the large IBM [convex] announcements we heard last week.

I was wondering if you could sort of help reconcile the apparent disconnect that some might think there is between CapEx guidance versus the orders, and specifically as well if you could talk to market share trends at these IDMs.

I don't know how you make conclusions, very difficult with the statistics we give you about trying to understand what the big IDM does versus the other smaller ones.

Remember, in the statistics we give you, in IDMs we put at least three companies.

Each of them could have impact to each other.

And secondly, one of the big IDM you talk about have already booked a lot with a very long leadtime.

So it is absolutely reasonable for them not to have overdone it further.

Thanks.

Tim Arcuri.

Citigroup.

Two things.

First of all, just on that last question, can you talk about leadtimes?

I sort of went back to my notes and we were talking last fall about leadtimes sort of shrinking to that maybe 3 to 6 month range by now.

So I'm sort of wondering where leadtimes stand now.

Thanks.

Anytime you ask that question, I confuse the whole market.

So I'll do another one trial at it.

There is a difference of probably your question, which is to PO leadtime, which is sort of administrative decision about when you put legal requirements on when you want the machine.

And there is your leadtime, which is a manufacturing leadtime, how much time does it really take to make machines.

With the new machines, like NXT, we are getting into a leadtime of assembly and test which is more in the three months timeframe as we go now.

So we are getting our manufacturing systems under control with this type of activity.

On the assumption of course that you have all the spare parts available so that you can assemble the machine.

If you -- so if you're smart and you have been able to accumulate enough spare parts, you can do three months of manufacturing leadtime.

PO leadtime is -- it absolutely depends on the structure of the company.

You have some very large IDMs who know so much about their business and control it very well, that they would put a PO leadtime of nine months, even nine months to a year, and it will continue like clockwork.

And you have PO leadtimes more in the memory business, particularly DRAM, where the PO leadtime could be less than a month.

If you talk about the integral leadtime, which includes the supply chain, that's (inaudible) nine to ten months.

If you put the lens and things (multiple speakers)

Yes.

That's for leading-edge stuff.

So if I wanted to get a tool today, so if I came to you and I was one of your biggest customers, your biggest, most important customers, could you move things around to get me a tool inside of Q2?

Yes.

You could.

Okay.

Second thing for me, of the shipments -- of the more than USD5 billion -- or more than EUR5 billion that you will ship this year, or revenue this year, can you segment how much of that is for technology?

This is interesting because we used to have a clear understanding of that.

At the moment we knew what was a critical size of a technology buy so that you just do a conversion.

But in there you have technology and volume, and it's all mixed.

I'm going to shoot -- I can't answer that question.

We'll have to look at that and Frankie can come back to you on that.

Thanks a lot.

Mehdi Hosseini.

SusquehannaInternational.

Thanks for taking my question.

Eric, if I were to look at your commentary on 2011, (inaudible) just to look at the EUR5 billion, that would give low double-digit year-over-year growth.

What I want to better understand is how much of it is driven by share gain and ASP changes, especially with more immersion shipped.

And also, how does the mix of litho as it overall double your fees changing?

And I have a follow-up question.

The EUR5 billion would be based on share loss, to be honest at this moment.

We suppose that our competitor will be able to place certain numbers of immersion machines to certain numbers of customers, I would say, which would make it a natural place to be.

And so we assume that the EUR5 billion includes this possibility.

So it's a share loss compared to 2010.

In terms of volume and ASP, at this moment the EUR5 billion is based on a higher ASP than 2010.

Because in 2010 there was in foundry certain numbers of 65 nanometer capacity built, which required less expensive machinery.

We shipped this mainly in Q2 and Q3 2010.

And this is not planned in 2011 at this moment.

So, therefore, (inaudible) at this moment the plan for the EUR5 billion is higher immersion mix than in 2010.

In addition, the NXT will represent a higher share of immersion.

As you know, NXT was above 50% in Q4, but used to be 20% of the immersion in Q2 and Q3, and lower.

I think it will be more like 70%-ish in 2011.

Did you say that because your competitor is going to get some business, your overall share is going to be impacted?

Did I understand you correctly?

Yes.

The point we are making, we don't make a surprise of it, we are early with NXT compared to our competition.

And the competition will have certain numbers of machines shipped into the traditional customers, in certain numbers of layers which are not ours, which are less critical layers.

And this is expected, so this is not something new.

But the share in 2010 (inaudible) -- last year in 2010, the machine was (inaudible).

So we were basically at virtually nothing this year.

There will be something -- sorry, virtually nothing in 2010.

There will be a few tools in 2011.

We don't believe there's going to be anything significant (multiple speakers) going from zero to something.

Sure.

And then [we] kind of lift out the litho mix as overall double your fee.

But I'm going to save my follow-up for this question.

Are you worried that foundries are competing for the same customer for the same order, for the same square inches of silicon?

And then to that extent, how does the technology change at the global foundry trying to implement a new architecture?

And the failure -- the probability of any success or failure would impact the business for you.

Are those kind of risk factors or are you not concerned at all?

Remember that when we do a forecast for you, and we said what we think we're going to ship is based on this overall numbers of chips or square meters of silicon of different processors.

And then we say that's what we will ship.

And then we compare to the PO's from the customer.

And this is why we told you, you are correct, in the foundry world at this moment there is an over investment probably to your point, because most of these players are trying to go after the same accounts.

But the amount of orders spent is only -- I make the difference between a 14% capacity growth to an 8% capacity growth.

So in other terms, 14% minus 8% equals 6%.

6% is this capacity which is double-counted or whatever, (technical difficulty) that is for quoting the customers.

So, two things can happen.

One, they all succeed, and the industry may have to live with a 6% too much capacity.

So you start hearing this [on the wall].

That the customers themselves are so keen in having multiple sources for reasons of technology and for reasons of price that they may have to sustain two partners.

And this may make the two partners less efficient than they used to be.

And you could discuss with those people whether that means they will have less profitability or the prices will be higher, because at the end it's a higher service.

This could happen, and it could be that in fact the industry goes towards a less efficient model in using capacity.

And that less efficient model will allow them in fact to manage peaks, which today they don't.

In the newspaper you see TSMC at this moment is at capacity on critical nodes.

And I think it is public in the newspapers that some of the customers are looking for other [AIOs] to have their chips made, and this is a bit of a scramble at this moment, which I am sure the industry or foundry will have to address.

So this famous 6% overspend may in fact be here to stay as a new feature.

To your next question, last question, is -- are we going to be affected if one of the key players has a higher market share than the other or succeed more than the other.

No.

I repeat.

We are going to make machines for the square meters of silicon by the end products -- by the NVIDIAs, by the TIs, by the STs.

So we're not fundamentally impacted by who wins.

Thank you very much.

Kai Korschelt.

Deutsche Bank.

Good afternoon.

I have two questions.

The first one was, you alluded to sort of pretty strong orders already from one of the larger microprocessor makers already early in 2010.

So I was just wondering, were you implying that most of the CapEx raise for 2011 was essentially reflected in your order book in 2010?

That'll be first question.

My second was on EUV, also just a clarification.

I think you also mentioned the quite different throughput rates in terms of wafers per hour for EUV, which will be even at target 60 is probably materially below the current sort of 150 to 200 range of most fab lines.

So I'm just wondering, does that mean that EUV will really only be deployed in sort of brand-new fab extensions or lines?

Or can they actually be used with the existing installed capacity?

Thank you.

Thank you.

So I will not comment on the CapEx question of one customer.

It's their data, so I will not say anything about that.

Regarding the EUV, don't misunderstand me.

The generation 2, which is used for R&D, is targeted for 60 wafer per hour.

Generation number 3, which is the 3300, which is targeted for production, is targeted at 125 wafer per hour, which is still lower than the 150 which is occurring a standard of immersion.

But it is irrelevant.

The question is that in the build up of a new chip, and I'm going to take how we guess an obvious example, in the low 2x DRAM, you will need certain numbers of UV layers.

Probably four, five.

These layers will be done by EUV, because they probably the customer may not cost effectively -- [sorry], the customer may not achieve the criticality of the designs.

So no matter what your cost or the speed is of EUV, they may be forced to make 20 nanometer DRAM using an EUV machine no matter what the speed is.

And then the next layers, layer 6, 7, 8, 9, 10, etc.

will be done by immersion, and next layer by KrF and probably some layers by (inaudible).

So that's how we do [the] work.

However, on some other segments, to take an example of NAND, in the NAND, say sub-20 nanometer.

If EUV is cheap, meaning if EUV as at 125 wafers per hour, then they will all run into it and say this is good stuff, and they will put probably four layers on EUV.

If EUV doesn't run at 125 but runs at lower than 60, they probably will not use EUV in NAND for the [1xI] and they will go for EUV on probably 1x like 15 nanometer or under.

So, in other terms, there are some areas of the world which is cost effectively impacted by EUV.

And EUV gets in when it's cost effective.

And in some part of the world, some layers where you have no choice.

So in order to know how big EUV will be, we have to be sure that we achieve 125.

If we achieve 125 per wafer it's going to be enormous.

If we achieve 120, it's going to be also enormous.

But if we achieve 60 it will be small.

Smaller.

(multiple speakers) But 60 is still okay.

So we're going to have to manage the introduction of this technology in the different [tiers], and catch the different layers in this way.

That's very clear, thank you.

Weston Twigg.

Pacific Crest.

Just wondering, you said the immersion mix should be larger this year, and I'm wondering on a unit basis how many immersion tools can you make per quarter.

And what do you think that might be by the end of the year?

We can do per quarter between 33 and 36 immersion tools per quarter.

And by the end of the year we could go about 40, because by then we would have enough spare parts to do (inaudible) that we need to do.

At this moment we don't have enough spare parts to just do more than one at a time.

Okay.

And on the KrF side, 25 units in Q4, do you expect that to trend roughly flat through 2011 or be higher or lower?

At this moment we expect KrF to be lower, but this is now I would say [an upside] opportunity.

But we haven't called any of that at this moment.

Remember since summer the whole environment basically is pessimistic as to what I would call volume sales -- 65 nano, the Chinese business, etc.

So, we haven't got significant requests by customers to be ready with this.

At this moment we are I would say at the low level of the equation.

Thanks a lot.

Andrew Gardiner.

Barclays Capital.

I just wanted to see if you could answer the capacity and leadtime question maybe a slightly different way.

Clearly, you have seen very strong orders in the fourth quarter.

But outside the official order book, I'm just wondering what level of visibility some of your key customers are giving you and perhaps what that means for leadtimes in particular, given that you have been capacity constraint for the last few quarters.

And we've seen bookings well ahead of shipments.

Are some of these key customers coming to you even earlier than before in terms of providing visibility through 2011, and maybe even beyond that?

The ultimate visibility is the order.

And we announced specifically that -- when we booked in 2011 we expect some more bookings in Q1, but basically 5 billion is clearly identified and [literally] booked.

So you can -- if you're cynical, you say, that means I'm worried about you.

And they put some orders and they plan to cancel or push out in case because, well, these are machines which you need to reserve.

But [unfortunately], well, fortunately for us, that's not the case for multiple reasons.

The first reason is that, I think we hinted this at the beginning and Peter hinted at it also, is the NAND and DRAM business which usually would do this with the reserve.

And to be sure, in fact, we at this moment know that they are under booking.

They are under booking compared to what analysts expect them to deliver in terms of [ramps], DRAM and flash.

So if we expect something, it's not a cancellation or pull out or push out of these machines.

It's more the opposite.

On -- just because they need to machines to do the chips.

On the logic side, we expect -- we told you that it was more if a building up of a fab with critical mass.

So normally a typical line in the logic arena should have eight immersion tools or -- I don't remember; it depends on the technology.

So, imagine eight immersion tools.

They will not go for six because it's not the critical mass.

It's not efficient.

So they will go for a limited critical mass, and all this is well understood by our engineers and being [in stock].

The third piece of data that I think it is important to know is, at this moment, most of -- all of our customers are negotiating specifications and slotting priorities per quarter.

So it's a normal thing we do.

And it means you have an order and you say we are going to ship it in July.

Do you need this machine or that machine?

Do you need this in the first week or the second week or the third week?

When do you have the pedestal prepared?

When it is the trucks coming?

What sort of plane do put it in?

And all this is being discussed with blood and sweat in a very normal way, where you can imagine that this is not the purchasing guy discussing.

It's a production guy building up his logistics to it.

So we were talking about real activities.

We are not talking about paperwork.

So, for this reason, we have the highest visibility I think ever that this Company has ever had.

Thanks very much.

We are going to try to squeeze in one last question -- and I'll ask the operator to assist in here in a moment.

But if you are unable to get through on the call and still have questions, please feel free to contract us, ASML investor guys, with your question and we will help you out.

Operator, can we have the final question please?

Peter Testa.

One Investments.

I was wondering if you could help us on slide 18, the new fab litho opportunity to understand what sort of view and assumptions you put into those spend figures in the different categories.

Peter?

Of course it's -- without going into detail because that would be revealing some of the customer issues, but we are very much aware.

These are publicly announced projects.

We are aware of the size of the project, whether it's a new build, whether it's a refurb, whether it's an existing (inaudible) facility.

We are aware of what they are doing there, the product that they want to make.

So we know the mix of leading-edge stuff, mature KrF, dry ArF, [highline].

And based on that mix we basically say that they need to support that particular output that they want, because all those fabs have the requirement of so many thousand wafer starts per month.

We can calculate the litho need, and we know the price of the tools, so then the end result is relatively easy.

That's the way we do it.

Okay, but are you assuming for example that these fabs are built out?

Are you assuming they'll just build this to minimal critical mass?

No, they are built out.

Okay.

Thank you.

And then we'll take this period 2011/2012.

Yes, okay.

So you're assuming they're built out during 2011, 2012, at this pace.

Okay.

Thank you very much.

Thank you everybody.

On behalf of ASML's Board of Management as well, I'd like to thank you for joining our call today.

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

Thank you very much.

Ladies and gentlemen, this concludes the ASML 2010 fourth-quarter and annual results conference call.

Thank you for participating.

You may now disconnect.