艾司摩爾 (ASML) 2006 Q4 法說會逐字稿

Ladies and gentlemen thank you for standing by and welcome to the ASML 2006 fourth quarter and annual results conference call on January 17, 2007. [OPERATOR INSTRUCTIONS].

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

Go ahead please sir.

Thank you, operator.

Good afternoon and good morning ladies and gentlemen.

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

The subject of today's call is, of course, ASML's 2006 fourth quarter and annual results conference.

Hosting this call today I have Mr. Eric Meurice, ASML's CEO, and he is joined by our CFO, Peter Wennink.

I'd like to draw your attention to the Safe Harbour statement contained in both our press release and presentation, and we will be applying that Safe Harbour to today's call.

You can find the press release and the accompanying presentation on our website at www.asml.com.

The length of the call will be 60 minutes.

And at this point I'd like to turn it over to Eric.

Thank you very much Craig.

Good afternoon and good morning.

Thank you for attending our conference call.

We will be dividing the introduction to this call in two parts.

This time Peter will start with a review of our 2006 performance, with added comment on our outlook for 2007.

I will then try to beat Peter's positive picture with my introduction on long-term growth trajectory.

Following the introduction we will open up for questions.

So, Peter.

Thanks.

Thank you Eric.

Welcome to everyone on the phone.

First of all I would like to review our Q4 results highlights, and in that regard I'm real happy to report that, of course, our 2006 results are the best ever.

In reviewing our results in detail you've seen that we reported records for revenues, operating margins and for the net profit.

Q4 revenues showed a record of over EUR1b, and contributed significantly to the 42% growth for the total year.

Our operating cost only grew 13%, which was completely driven by increased R&D spending.

And this has resulted in an operating margin on the year of 24.2%, which is an increase of 6.4 percentage points over 2005, and that all has resulted in a net income of 17.4%.

Those are very good numbers, and we believe that next to the favorable industry conditions of last year, our 2006 results were significantly influenced by the following three drivers.

Now first, there is our continued market share growth.

This is the clear result of our relentless R&D spending, focusing on time to market of new technology, and on increasing our product value of ownership.

Secondly, we've been focusing on cost reductions and cost containment for several years now.

And I believe the result is clearly showing in the improved gross margin, and in the level of selling general and administrative costs that keep decreasing as a percentage of our sales, and it's currently running lower than 6%.

Now, I would add to that lastly, that return on invested capital has been our major performance yardstick.

And our cash cycle days on inventories have improved from 188 days at the end of 2005, to 111 at the end of last year.

If you combine that with our improved profitability, we believe that the Company is currently comparing very favorably on this metric against many of our industry peers.

2006 was also the year in which we started our first ever share buyback program.

We succeeded in buying back 8.3% of our outstanding share capital.

And since we feel confident in our ability to keep generating a solid profitability and accompanying cash flows, we will use this instrument of the buybacks, also going forward, to return money to our shareholders.

Which brings me to the outlook for 2007.

Industry analysts predict IC unit growth for the semiconductor industry at approximately 10% in 2007, as well as up to 5% growth for lithography equipment sales.

Our own analysis shows also potential for growth in [nautical] segments of our market.

And the implementation of immersion into the production by major Flash makers in 2007 will also fuel our growth, in both market shares and in top line revenues.

All this on top of a record backlog valued of over EUR2.1b that leads us to conclude that 2007 will be a growth year for ASML.

With regards to bookings, our track record over the last three quarters has proven some difficulty in providing accurate bookings guidance.

For Q1 of 2007 there is a certain level of uncertainty about the timing of orders for a few large customers, which is having a big impact on calling the quarterly bookings.

Many customers are making complex technology and product transitions for Q3 and Q4 shipments, which will lead to orders entries in Q1 or in Q2.

Under these circumstances we prefer not to speculate, and that is why we are not specifically guiding for Q1.

I can just confirm, like we said in our press release, that our order book for Q1 will be healthy, and will be supporting our 2007 growth expectations.

That's as far as I will go for the outlook on 2007 and, Eric, if you can take the remainder of the introduction.

Thank you very much Peter.

As you can see, we feel optimistic about -- with our mid-term horizon.

And I would like to focus now on the ASML long-term growth trajectory, which overlaps on the short-term positive picture that Peter described.

As we've said on several occasions, we simulate regularly future lithographic market needs, based on integrated circuit unit market growth rates assumptions.

If we take an assumption of about 9% IC unit growth, and this is a simulation assumption which is an average, and if we put this into our simulation model we project a total lithography market of about EUR7b by 2010, compared to approximately EUR5.5b today.

This overall market growth potential sustains ASML capability to reach its EUR5b revenue target in the same timeframe.

The following three factors contribute to achieving this potential growth.

First, the lithography spent by our customer per IC unit produced increases with the semiconductor technology road map.

Obviously, the expected growth in IC units in itself creates a need for lithography tools.

But another driver is the shrink road maps.

These ever-accelerating road maps create demand for more complex tools, which have a higher ASP and a higher value content, like tighter overlay, [two fold] management for instance, and with relative lower throughputs during the first few years, thus, generating a higher litho value purchase.

Moreover, these additional nodes require more processing layers, resulting in more litho passes, more usage of those machines.

For instance, typical logic designs which requires, today, 35 to 37 layers at the 65 nanometer node will require, in the future, up to 40 layers, that is more than 10% above today's, at the 45 nanometer node.

Similar increases will occur for Flash and DRAM.

These conditions combined are causing lithography system expenditures by our customer to increase, as a percentage of wafer fab equipment spend.

Today that percentage is approaching 18%, so 18% of wafer fab equipment spend is litho tools.

This number was 14% 10 years ago.

The trend will continue as -- to grow.

Secondly, we expect to continue to gain market share.

ASML's revenue market share stands at approximately 61% at the end of 2006, due to significant new customer wins in multiple product segments and multiple geographies.

This represents an approximate increase of about 4 percentage points over 2005.

As mentioned by Peter, we believe that we will continue in growing market share.

Indeed, the semiconductor industry has recognized ASML's compelling value of ownership advantages through the TWINSCAN architecture.

With over 500 TWINSCAN's installed worldwide at the end of 2006, we expect that our accumulative experience, the cost of architecture switching at our customer and our high and growing level of R&D investment will keep us in the lead.

We are particularly confident in the all important immersion segment, where ASML is enabling customer development of future processes beyond 55 nanometer nodes in a variety of applications.

With over 76 ASML immersion systems in the market today, including 23 of the XT:1700i shipped to 17 customers, and with more than 300,000 wafers purchased already on these immersion tools in the market, we are in a clear leadership position worldwide.

Mid-2007 we will start shipping our XT:1900i system, and evolutionary improvement of the XT:1700i of today.

That will -- this new tool will allow [imaging] below 40 nanometer node.

For the technology afterwards, double patterning technology for instance, ASML we recently launch the XT:1450, a high throughput improved overlay dry system for production application at some 60 nanometer node, and double patterning development activities at the 32 nanometer node.

In the field of EUV, which comes next, the expected lithographic technology for volume production beyond 2010, covering probably up to 10 nanometer node, ASML has shipped, in 2006, two pilot EUV systems to IMEC, and to [Albany] Nanotech for early process development work, and industry infrastructure development.

We have already received orders for three -- from three IC manufacturers for pre-production EUV systems, with first shipment in -- starting in 2009.

The third factor which contributes to growth, mid-term growth, is our increased strategic scope.

As you know, ASML announced plans last month to acquire Brion pending regulatory approval.

We see numerous opportunities where combining ASML and Brion will reinforce both companies' growth and value potential in advanced semiconductor lithography.

Brion is the largest and fastest growing pure play company in computational lithography.

This encompasses IC design verification, critical enhancement technology, otherwise called CET, and optical proximity correction, otherwise called OPC.

These markets are small today, but will grow exponentially as critical designs become more and more complex to execute down a more [slow] curve.

Further, customer require increasing advanced litho tools with capability to optimize the scanner settings for a specific application.

The customer benefit is an ability to maximize yield.

It will also help ASML to create competitive differentiation in both businesses; its current business, the litho tools, and the new business from Brion, the RET/OPC tools.

So, in summary, the message is as follows.

We will continue our focus on execution as we've done now for the past years, to secure market leadership, while delivering high return on invested capital.

Two, we are optimistic about 2007 in view of market growth in general, and our own market share growth outlook.

And three, this mid-term, short-term/mid-term outlook, is strengthened by our overall long-term growth trajectory, which is strengthened by our offer of the most advanced technology in line with our customer road maps, and by our expanding of our scope in the market with new opportunities.

Now, I think, Peter and I will take your questions.

Before we do the questions let me just remind you, ladies and gentlemen, that when asking the questions please clearly identify yourself and your organization.

We also kindly ask that you limit yourself to one question with one short follow up, if necessary.

That will allow the maximum number of callers to get a question asked.

So, operator, I think we are ready then for the first question.

Thank you sir. [OPERATOR INSTRUCTIONS].

The first question is from Mr. Antoine Badel.

Go ahead please sir.

Please state your company name followed by your question.

Yes, good afternoon, Credit Suisse.

On the operational expense front you've given us some guidance for Q1 that shows an acceleration of SG&A and R&D spending.

Could you maybe give us a little bit more guidance in terms of structural spending levels for the rest of the year?

Yes, for the R&D spend for 2007 we have guided Q1, indeed, up.

For the total year the R&D spend could be up, as compared to the spending of Q4, between 10 and 15%.

On SG&A we don't expect a major increase as compared to the guides we've given you on Q1.

Thank you.

And as a follow up, you've reported your sixth straight quarter of sequential backlog increase, which is one more than the last cycle in '03/'04.

Should we expect Q1 to see the first downtick in the backlog?

Or is there any chance that it can continue growing?

When we look at the backlog and the order situation of Q1, I think we were specifically clear in what we said in the introduction.

That, we are focusing in 2007 specifically on growth driven by our estimates on what we think the industry will do.

Our market share gains, the introduction of immersion, and specifically the 1900 in the second half of the year, that will lead to growth for 2007.

And it is particularly difficult, like I said in the intro, to exactly find when those few big customers will place their orders, whether it's in Q1 or in Q2 for their Q3 and Q4 ramps.

So, I don't really -- I would not know that, so I think the answer that I gave should basically cover this.

Thank you.

The next question is from Mr. Thomas Brenier.

Please state your company name followed by your question sir.

Yes, good afternoon, it's Societe Generale.

Maybe just a follow up on the question of [Nicolas].

I just wanted to see how different is your environment compared to where it was in October.

Have you seen any changes?

I understand really why you don't want to give guidance for bookings in Q1.

So, I was just wondering if there is any change in the visibility compared to what you are seeing -- what you saw in Q3 and Q4?

And if you could give us the details between what is going on in Memory, IDM's and Foundries it would be helpful.

Thank you.

Yes, Thomas, thank you.

Fundamentally, no big changes.

We guided last quarter that we were confident about the possibility of growth in 2007.

Here we are confirming that growth.

The body language is a bit more positive, more optimistic because of the realization of orders coming from the DRAM segment.

At this moment we see a Flash business which is front-loaded.

We said that three months ago.

There is no change.

We see DRAM business strong and stable in Q1, Q2, Q3.

We see a strong and sustained business of Foundry at the end of Q1 and then Q2, Q3 and potentially Q4.

We see a pretty spread IDM business for -- along the whole year.

And for ASML we obviously see a development in the R&D [built-up] sector, where customers of any segment built their new recipes for the 45 nanometer node and beyond, using the latest technology [of] immersion.

In other terms, second half of the year, they will need the new tool that ASML will introduce, which is called the XT:1900.

And we are obviously bullish that this will create significant sales, significant revenue in the second half of the year.

Okay, I have some factuals to that.

If you look at the backlog at the end of the year, 64% is Memory, 15% -- so 15% is points of the 64% is really only Flash.

So the remainder is DRAM, and what we call the hybrid fabs.

So that actually shows that we have seen strong DRAM bookings.

And if we run those DRAM bookings and the anticipated shipments for next year in our models, we cannot see that we are going to create a big over-capacity in the DRAM space.

So, that is one of the drivers, if you will, say changes.

We will have some more factual information on the backlog.

And, sorry, as a follow up to this, for the DRAM space are you seeing the same kind of pattern you've seen in the past few quarters on [MEM] that corresponds -- the orders correspond mostly to new fabs?

Or are you seeing also some upgrades to existing fabs?

Both.

Probably fairly balanced that's the way we see it.

Thank you very much.

The next question is from Mr. Nicolas Gaudois.

Please state your company name followed by your question.

Yes, hi there, Nicolas Gaudois at Deutsche Bank.

First question is effectively on how you would look at the timing of the introduction of your new tool in dry v the XT:1450?

Whereby, you are expecting to release this tool by the middle of this year, and in the second half of '07 we will see the microprocessor segment starting to use the double patterning dry for critical layers.

So, just wondering if the timing is basically not an accident, and we should see a strong acceptance of this tool for your microprocessor customers?

Because there are only two microprocessor customers in the market, Nicolas, you will understand why I am cautious in my responses.

The double patterning activity at this moment, as we see, is probably more in the R&D lab and it is a bit difficult for us to call production.

As much as we are very, very gutsy in calling the immersion ramps, because we have now certain numbers of official ramps, in fact, that have started already last year and are progressing this year.

The double patterning at this very moment is a bit more difficult for us to call in terms of production ramp.

The 1450, by the way, has two functions.

One is double patterning.

It's the only tool that allows this in view of its overlay capability.

But it's also a fairly good tool in production of single patterning.

So you -- the second reason for us to be a but confused is to the usage of it.

Okay, fair enough.

And just as a follow up, effectively, if I look at your combined logic and foundry backlog in value at the end of December, it seems to be about 10% below where we were in Q4 '04 so at the peak of a prior cycle.

Yet, you gained market share at [TSMC] in Japan, [Intel] amongst others.

Would it be fair to say that although we are close [inaudible], or maybe over spending a bit on Memory, on your side there may be some under-spending on the bookings effectively on the logic and foundry side going into '07?

Well yes, I think you are mentioning not so much value statements.

You are talking about percentage of the backlog.

So, it is true that we are seeing a temporary increase of the Memory percentage in the total business of, I would guess, every equipment manufacturer.

Memory is a bigger segment at this very moment.

It has two factual reasons for that.

One is there is a renewal of DRAM needs due to [Vista].

At least probably 80% of the needs is due to Vista, although, we also see some DRAM needs in the consumer arena.

So, clearly, you've got a nice upside in DRAM which has happened.

And then we talked about the Flash business, which is building up to some level of capacity before it matures, I would say.

So, these are the two short-term reasons why the Memory business represents 60 to 65% of our backlog.

We expect, in the mid-term/long-term, to be back to 50%.

But my question was basically on the value of the backlog.

If you look at the value of logic and foundry again combined.

If you look back at your numbers in Q4, it was actually 10% higher to where we are at the end of Q4 '06, and that seems a bit contra-intuitive, if you look at continuing spending trends in the non-memory sector, but also your own market share gains.

So, trying to [gauge] effectively there is underpinning that some under-spending in lithography in the non-memory segments at this point in time?

Honestly, I would have to be reviewing the statistics of last year to be very, very precise with you.

But, clearly, today the Foundry business is adjusting their capacity and their inventory levels, being weaker in backlog.

But the activity, the bookings activity, allows me to say to you that we expect good business of Foundry end of Q1, Q2, Q3 and potentially Q4.

So, yes, there is probably a timing issue there which explains the difference compared to last year.

We also, clearly in Foundry, have one of those segments that requires UNIX products with the low -- the shortest lead time, which means the Foundry indications is not so much a backlog question, it's really a billings question.

If I remember by memory, in last year, I think 66% of our business of Q1 was [turns] within the quarter, and a lot of this was Foundry.

So, in other terms the backlog is a fairly bad indication of the Foundry real need because they will come on a very short-term basis.

Okay, understood.

Thank you very much.

Your next question is from Mr. Jay Deahna.

Please state your company name followed by your question.

Hi, good morning, good afternoon.

This is Jay Deahna from JPMorgan.

My first question is do you expect to take in orders for more immersion systems in 1Q than you did in 4Q?

I noticed your ASP's of orders were a little lighter than I would have expected.

It would suggest that i-Line and KrF were a higher percentage of the mix.

So, will you book more immersion in 1Q versus 4Q?

And then the second question is gross margins, where do you expect gross margins to trend throughout the year?

But, more importantly, over the next several years, as you get beyond the upfront investment in Japan and start ramping, can we see gross margins in the 43, 44% range two or three years out, something like that?

Thank you.

Jay, two strategic questions.

Immersion, we cannot answer that.

This is one of the reasons why we didn't want to guide [there in precision] Q1.

The lumpiness of those decisions and exact timing, and the choice between the 1700 and the 1900, are reason that says it's a bit difficult to call.

Obviously, our daily life is as follows.

We've got a customer who needs an immersion machine, and they will want to wait until the last moment before they take the decision between the 1700 and the 1900.

So, this is why we cannot say what exactly is the immersion number, bookings number in Q1.

We are certainly optimistic to have a significant run rate of those however.

Regarding gross margin, I confirm with you that the maturity, the maturation of our lithography business, in terms of market share, as well as the expansion of our scope towards value-added, I would call it, software and service, value added around the litho machine, calls for an increased gross margin towards the level that you just talked about.

Okay great, thank you very much.

The next question is from Mr. Titus Menzies, Jefferies and Company.

Go ahead please sir.

Good morning gentlemen, Eric and Peter, this is Titus from Jefferies.

A quick question on the orders which you have for the first quarter.

How much of that is back to the fact that your capacity maxed out [at the moment] and you are trying to work through an expansion efficiency to turn more equipment through the quarter?

And how much -- if we into look the quarter, how much do we maybe try and gauge in [turns] business for the quarter as well?

Good question.

We are getting out of capacity limitations.

Obviously, in our business there is always capacity limitations depending on which product you are talking about in view of the lead time and things.

But at this moment I am happy to confirm that our capacity is above the EUR1b.

You know we have done more than EUR1.06b in Q4 and we have proven that we can go above.

And we are moving our capacity up and up and up every quarter.

So, we're getting out of that.

And that means we are available for turns.

Yes, there is opportunity for turns.

And I did mention that we do expect some pressure by Foundry.

But I don't think it is a wide request at this moment.

So we are not planning for significant upside of billings in Q1.

Thanks, Eric.

Just one more follow-up question. [In terms] -- with lead times stretching, is that a factor which has gauged the product mix to shift down to KrF and i-Line as customers are unable to get hold of the more advanced tools in the timeframe which they are hoping for?

No.

The lead time is stretching due to our little past capacity limitations has resulted in me taking a whole bunch of 'kick in the butt', technically speaking.

But we've been able to manage most of the customers' enthusiasm without losing business.

As you know, this is an area where the switching costs are so important that you don't just change because you don't have it.

So we've been able to manage effectively that lead time stretching.

And now obviously time is, of course, reducing and allowing, in fact, as I just said today, customers in particular in the foundry sector, to be very short in their booking.

Yes, if I could add something to that.

As you might have seen in the presentation, we have indicated that two thirds of the backlog at the end of the year is shippable over the next six months.

That one third that is shippable in Q3 really relates to the new technology with long lead times, so that is a lead time stretched, you could say that.

And it has to do with new fabs which are going to be starting up in Q3 which basically want to secure their first lines.

So I don't think, just like Eric said, that we really see a lead time stretch that the fact that we have now a percentage more bought for Q3, that's outside the six months window, has really go to do with the new technology and the new fab projects which are specifically DRAM customers.

Thanks, Peter.

Just to understand correctly, that means the current environment suggests that we're looking at capacity purchases coming through rather than just technology purchases coming through?

No, we are going to have both.

As we said, or I responded to Jay, there is some discussions at the customer level about which is the optimum mix of machine for the ramp of immersion and 45 nano, etc.

And this usually is something that they even -- they optimize until the last moment.

Do they take some many of this machine, so many of that, so many ERF, so many KrF, etc?

So you've got all that playing with us, saying hey guys you need to take decisions soon and fast because you are going to lose your slots.

And that's what's happening at this moment and that's for H2.

Thank you gentlemen, and congratulations on the fourth quarter.

Titus, I think we need to move on.

Can we move onto the next question please?

Sorry.

The next question is from Mr. Asaolu.

Please state your company name followed by your question.

Good afternoon, Olubunmi Asaolu from Lehman Brothers.

I'm just looking at your implied order intake by customer type again.

Is it fair to assume that when your Foundry orders return meaningfully that they would come in and would be material enough to offset the expected decline in Flash orders in the second half, since your DRAM segment continued to be strong?

Yes.

We expect, in fact, a combination of good news for H2, which is a combination of Foundry, IDM and what we call technology [buys].

Technology buys goes to Foundry, IDM and also Memory.

And that would offset any potential reduction of Flash demand which comes from the normal digestion.

That's very clear.

And then on depreciation, how should we model this on your cash flow statement now going forward with your newly raised CapEx guidance?

Yes.

Well, I think the depreciation will have -- the depreciation resulting from this additional CapEx will have a result largely in 2008.

So, based upon our 2006 depreciation, you could see an increased depreciation charge of around 10%.

But most of the depreciation charge will hit us in '08 because it relates to the factory which -- it's 100 to 120m, is due for completion by the end of this year.

Thanks very much.

The next question is from Mr. Jonathan Dutton.

Please state your company name followed by your question.

Yes.

Thanks.

It's UBS.

Just a couple of questions from me.

You've, in your 2007 outlook, highlighted total spending of 5.5b.

And with market share expanding by 40 basis points a year, potentially 65% in '07, would the right way to think about this be implied equipment revenues for you of about 3.6b?

We do not guide the year [inaudible] we are growing.

And you can imagine that $1 means growth, and you can be more aggressive.

So, no, it's again too early to be quantitative.

But we are clearly on our trajectory [of 5b].

Just a follow up, if I may.

The 30% capacity increase that you're planning, what sort of market share does that imply potentially in the near term if you do expand it to -- from 300 to 400 units?

Again, we use capacity and tools to satisfy customers.

You can't say that capacity becomes immediately revenue.

Capacity allows us to react very quickly to every customer.

And when you have the ambition to be a big partner to every customer in the world and, therefore, reach market shares of the level of 65 to 70%, which is the case, you need to have a capacity level which allows every one of those key customers to come in the same quarter, which means that you have to have a capacity much above the EUR5.5b target that I'm talking about which is an average revenue level on the year.

So that capacity level will be in place, that high capacity level which allows, therefore, to meet customers' needs any time, will be in place by the early 2008.

So as of now up to 2008, we will continuously improve capacity.

And by the beginning of 2008, we'll have a significant capacity which allows us to, well, to take EUR5b of orders if they were to come earlier [in 2007].

And should we assume a linear increase in capacity through the year?

Correct.

Through the year, through the year 2007, as I said earlier.

The factory will be completed around this time next year.

So that means that full access to that capacity will only be next year.

In the current factory we are moving up and up every quarter.

Because at the end, let's be very clear, this is a business for people.

It's not a business for CapEx.

So the difficulties we have, or if there is any difficulty, it is to move enough people with the right skills.

And that's why this continues.

And that's good news.

We can do this in a continuous fashion.

Great, thank you.

The next question is from Mr. Simon Schafer.

Please state your company name followed by your question.

Thank you.

It's Simon Schafer with Goldman Sachs.

Actually, a follow-up question on the capacity issue.

I was wondering whether you could comment on the maximum amount of lenders you're capable of sourcing right now.

I would presume that, obviously, you're working hand in hand with some of your critical component suppliers to get that aligned.

But how are they ramping that?

Will that be just in a linear fashion in line with your own capacity expansion?

Absolutely.

You are absolutely correct.

We are working hand in hand with [the guys].

And clearly we have at this moment, in fact, extremely similar capacity level.

So should we still assume a 70 maximum capability of sourcing lend in the quarter for now, up until your expansion takes effect?

No.

First of all, even if you were trying a guess, it's mix driven.

So you can't even easily summarize a capacity level.

So call it, as I said to you, we have a capacity now which is much above a value of 1.1b per quarter.

And on top of that, Eric said this with the previous answer also, that we are working in the factory to get the cycle time reductions through the factory that will, actually, with the current square footage of the factory, will create up extra capacity.

So that will create extra capacity.

And the same is being done in our -- basically, at our key suppliers.

Understood.

And actually a follow-up question on a previous question from earlier, just with respect to your assumption of mix of non-immersion tools this year.

Are we assuming, or should we be assuming, a handful of deliveries for the 1450?

And, otherwise, what should we assume for KrF versus i-Line on a full-year view.

Does it just depend on how Flash -- how quickly Flash may slow down and what it gets offset with later in the year?

Well, this is again an extremely tough question.

You are talking about the art of our business here.

We have a hit rate of planning the wrong thing, which is pretty good.

So, impossible to answer.

We have got potential fluctuation on every of those segments which are significant.

And even in the same segment, take Flash if you want, you still have some calls about how much KrF, if any.

You have some customers that basically says I want only immersion and i-Line.

You've got some customers who basically says I want only dry.

It's -- and they are not even sure themselves at this very moment.

So what's important, and this is why, in fact, your questions on capacity were very much operational, is we need to create enough bandwidth for us.

And, as Peter said, enough cycle time reduction so that we can easily adapt to any profile of mix.

And the good news, by the way, is most of our margins are consistent to multiple mixes.

So we're not -- it's not material.

The mix we will reach is not material to the profitability of the Company.

Yes.

Thank you, Meurice.

The next question is from Mr. Andrew Griffin.

Please state your company name followed by your question.

Hi guys, Andrew Griffin from Merrill Lynch.

Just another question about the capacity increase and also the increase in R&D.

What does this do to your breakeven levels, if any?

And could you just update us on where R&D can be pulled back to if there were a downturn in the cycle?

Yes, to answer the last question, we have a 20 to 30% -- between 20 to 30% flex in research and development.

Having ramped from Q1 to Q4, 23% in R&D costs for such a complex tool, that you can only do if you have a flexible and an outsourcing kind of model.

And that also means that we can pull that back if we need to.

Now, on the breakeven point, it was -- it's been a couple of years when we said we had a target of 130.

Well, we are lower than that.

And it depends on the mix, clearly.

Now, how does the impact of capacity increase -- what kind of impact does it have on our breakeven point?

To be honest, we have run the simulations and it doesn't give any material impact at all.

And the main reason is because our top line growth.

Basically, when I look at the growth trajectory going forward, the ASP of our tools is giving us a higher absolute gross margin.

And that is covering the higher absolute cost.

So the breakeven point is not that much impacted at all.

Great.

And just one final on the same subject.

What roughly would the headcount increase be by the end of 2007?

Yes, the headcount increase, we have to be careful for double counting because that will be -- if we compare the end of 2005, we have about 5,500.

We could see, we could envisage a 10% growth in 2007 which part of that growth is already included in the increased guidance we gave for research and development.

And also, if we grow the top line, we need more output in the factory.

We probably need a few more people in the factory also.

So you cannot just double count the 10%.

It'll be spread over the factory which will go into cost of goods, which will guide you on the gross margins.

It will be in R&D.

We gave you the guidance there.

And it'll be in customer support because we increased our installed base [with our] customers actually paying for those people.

So that's the way that you should at that 10% level.

Very clear.

Thanks very much.

Thank you.

And on the subject, it's important that you understand that we're not managing the Company on optimism.

So cost of structure for us is a fundamental thing that we have to manage in a flexible fashion.

We're not just saying because we are going to grow we can go ahead and invest.

That's not the case.

We are always cautious that we have the capability to pull back on those different investments.

So, clearly, there is a difference between us being optimistic about the top line and a need for capacity and us managing the costs structure.

We have a costs structure which is highly flexible.

The next question is from Mr. Robert Maire.

Please state your company name followed by your question.

Robert Maire, Needham & Company.

When I look at the two new technologies, immersion and double exposure, or two new techniques I should say, could you give us your sense as to the growth rates of each through '07 and beyond?

And is one accelerating over the other and what do they look like in terms of percentage of sales and differences in profitability?

I'm just looking for some more detail as those two.

And do you see a more, perhaps, finite life for double exposure than immersion?

I'm looking for more detail on that.

Okay.

First of all, a quick note on the technologies because all this is sometimes used with marketing terms by our customer.

Double exposure already exists for long time.

It has been used by people in the Flash memory, in the logic, etc.

It's a way basically to say you pass twice your wafer into the machine.

But your design, your reticule is a normal reticule.

So double exposure, it's a [yield important] technique, or whatever it is, and it may happen with our machines, or it may not, I don't know.

In fact, what the customer needs is machines that are fairly fast because if you do double processing it costs money.

And I do not know [entries] and I don't think that helps you more than just improving yields on given technology.

Double patterning is a situation where you redesign -- you design your reticule, or you cut your reticule in two, and you expose once and then you expose another reticule.

And you hope that the two reticules, basically, will expose an image which matches.

Therefore, you need machines which are very, very precise in what we call overlay.

This technology is the one that says that potentially will allow post-40 nanometer, the 32 nanometer nodes, you will need those type of technology.

The alternative to double patterning at 32 nano is EUV.

So having made that clear, how do we see the things happening?

As of now, immersion is the only technology that will be really ramping in high volume.

This double patterning for, I would say, 40, 45, etc. would be just potentially a prototype-type environment.

But immersion, at this moment at 55 nanometer node or at 45 nanometer node, is the technology of choice of, I would say, most of the market.

That will start -- that has started already at the end of last year.

That is -- 2007 is the year of major starts.

And every other customer has very different speeds.

In 2008 you will have certain numbers of needs for beyond 40 nano, some 32 nano.

And that's where we will have some double patterning happening.

And there is going to be some double patterning dry and some double patterning immersion.

We can't answer which is which.

But we expect, again, to be on a fairly low volume.

2009 could see some production in double patterning while we start introducing EUV.

EUV will be serving to do some R&D recipes and build up of infrastructure.

So double patterning will be a necessity for customers who will already have the design ready at 32 nano, which is not an obvious question, by the way.

So if they do have ready, they have no other choice than going double patterning.

And 2010, 2011 will be the area, the time of overlap.

There will be an overlap between double patterning and EUV.

And we expect EUV to win the battle by 2011 and 2012 in terms of economics.

Just a follow up in terms of double patterning, I would assume you've probably done some modeling or some math as to, obviously, double patterning requires, I would assume, more [sepurs] given that the wafers are spending more time being exposed, even though it's only at the more critical layers.

What does that mean in terms of revenue or number of machines required per wafer or per dye or per whatever in American?

How does that impact your revenue?

Is that more of a revenue driver than immersion or --

In fact, we did, in our famous simulation tool, all kinds of scenarios regarding the different usage of double patterning dry, immersion and/or EUV in different trends.

And it's not very obvious that -- well, let me put it this way, the simulation that we put in place is the one that makes sense.

So what we see is exactly what I said, that is, at some point between 2009 and 2010 you are going to see double patterning as the most economical thing for the customers.

Therefore, we will grow revenue by the fact that there is double usage of our machine.

And then by 2011 or 2012, EUV becomes most cost-effective.

But it's still an expensive technology.

So it is cheaper than double patterning.

But it will still be showing some additional revenue which is, as I said at the beginning in my introduction, where I said the more you go on the [moves] curve, whatever you do, whether you use double patterning or you use EUV, you are going to pay more.

But EUV by then will be cheaper than double patterning.

So what does that mean --

Robert, I'm sorry, we need to move forward.

Okay.

Go ahead.

Thanks.

The next question is from Mr. Mehdi Hosseini.

Please state your company name followed by your question.

Yes.

It's Mehdi Hosseini, Friedman Billings Ramsey.

I have two questions.

First, if you could provide the number of 1900 immersion tools that are in the backlog.

And also, as a sort of a clarification, what do you mean by healthy bookings in Q1?

Does that imply flat, down, up?

I just -- I think healthy is a very generalized term.

So if you could just provide us with more color.

I'm not asking for any guidance, just to better understand you I would like to hear more what you mean by healthy.

I will handle the easy one, which is the 1900 backlog, and I let the specialist, Peter, to handle the healthy bit.

The -- we have in backlog 20 immersion tools.

And at this point I'm not going to tell you the split between 1700 and 1900, also because it's a very, how do you say, moving mix.

As I basically say, customers are hesitating. 1900 is a better machine with a higher throughput, but it's more expensive and it comes --

But would it be fair to say that it's more than one since you booked one in the Q3 timeframe?

Yes.

It's more than one.

Yes, of course.

Yes.

Yes.

Is that less than five?

You are getting hotter, but it's still cold.

But I will stop there, okay?

Healthy, Peter?

Healthy.

I feel healthy.

Right, it's -- I said it this morning during the press conference also, healthy means good, which is the right level of order intake to support our 2007 growth profile.

And that's a general statement.

But that's exactly what it is because if we say it's going to be up, it's going to be down, it's going to be flat, I don't want to go there.

Because of the issues that we have discussed during this call, is that with the 1900 and the 1700 transitions, the choices customers need to make there, it becomes, with those big customers and the choices that they need to make, pretty difficult for us to predict when they will give us the orders.

But if there is one certainty it's that they will give us the orders.

And the orders are basically driven by the fact that sub-40 nanometer production capability becomes available with the 1900i.

We do expect growth from the industry.

We have [warned] several customers in 2006 that want to basically fill up their new fabs in 2007.

So they're all drivers for what we feel growth.

But the timing of when, for instance, that Flash immersion ramp sub-40 nanometer in Q3 or in Q4 will happen, whether that's going to lead to orders in Q1 or in Q2, we don't know.

And that is what we are trying to say.

It's going to be healthy enough to support our growth for next year.

Sure.

When you talk about --

Mehdi, sorry, we have to move on.

We're going to try to squeeze one more question in.

Thank you.

The next question is from Mr. Mark FitzGerald.

Please state your company name followed by your question.

Thanks.

Two questions on the Brion acquisition here.

Logic guys are struggling with their cost of design here.

Do you find it's been able to slow the cost of design going forward significantly?

This is the billion question, the EUR/dollar question.

Yes, we believe that the way to solve customer problems in design is going to try to give them leverage in the market which they can optimize at the end.

We are far from the design part of it, so I can't really answer your question specifically how much we would save in the front-end.

And I also believe that the industry has [to material], therefore, we have to do some looping tests where we see how much flexibility we can give our customers in terms of how much they have to invest in optimizing design upfront or how much they have to invest in optimizing the manufacturing part.

So we're going to do, of course, the best effort to move value into the post-sign off part of the business, if I make myself clear.

When the design is signed off then you optimize it.

And if you can do a great job in optimizing a design, post-sign off, then there will be a reduction of costs on the pre-sign off.

And we will benefit from the value generic, by the famous Jay's question, is your margin going to go up, or by the capability to sell more tools.

So that is the whole logic of the Brion acquisition.

And, guess what, we feel extremely enthusiastic about what we have found so far.

Can you give us any details of what this means for revenues, what they're doing at this point and what the financial model looks like?

Is there any --

No.

As you know, Brion is a private entity.

So whatever is 2006 is under the private data.

And we haven't [closed] ourselves.

So it would be preliminary to give you a business plan-type idea after we close.

It's too early.

Thank you.

Operator, I think we'll close the call now.

That -- my watch says we're out of time.

So I'd like to thank everybody for joining us today and we look forward to seeing you around in the future.

And I'll leave it to you then, operator, to close formally.

Thank you, sir.

Ladies and gentlemen, this concludes the ASML 2006 fourth quarter and annual results conference call.

Thank you for participating.

You may disconnect now.