艾司摩爾 (ASML) 2008 Q1 法說會逐字稿

Welcome to the ASML 2008 first-quarter results conference call on April 16, 2008.

(OPERATOR INSTRUCTIONS).

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

Craig DeYoung.

Thank you, Nicole.

Good afternoon and good morning, ladies and gentlemen.

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

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

The subject of today's call is ASML's 2008 first-quarter results.

Hosting this call today are ASML's CEO, Eric Meurice, and CFO, Peter Wennink.

I'd like to draw your attention to the Safe Harbor statement contained in today's press release and presentation.

The Safe Harbor statement will cover today's call introduction and question-and-answer session as well.

You can find our press release and presentation on our Web site at ASML.com.

The length of this call will be 60 minutes.

At this point I'd like to turn the call over to Eric Meurice for a brief introduction.

Thank you, Craig.

Good afternoon and good morning.

Thank you for attending our call.

As usual, Peter and I would like to say a few things about the quarter before we open up the call to questions.

Peter will start with a review of our Q1 financial performance, with added comments on our short-term outlook.

We'll complete the introduction with our assessments of the current environment and some details on our longer-term view.

Following the introduction, we will open up to questions.

Peter, if you will.

Thank you, Eric.

Welcome to everyone.

First of all I'd like to review, like Eric said, the Q1 result highlights.

The first quarter developed as we expected, leading to the eighth quarter in a row in which ASML has posted revenues of over EUR900 million.

Revenues were EUR919 million, to be exact, from total shipments of 50 systems, 14 of which were immersion systems.

This technology mix has led to a record average selling price for all systems shipped of EUR16.4 million.

40.6% of gross margin, along with operating expenses of EUR186 million, resulted in operating margin in the quarter of 20.5%, which compares to 21.3% for Q4 of '07.

We generated a strong amount of cash from operating activities of EUR263 million in the first quarter, which resulted in a cash reserve at the end of the quarter of 1.4 billion.

We also concluded our latest share buyback program in the first quarter of 2008 of 14 million shares to cover for outstanding employee stock options.

We booked 26 system orders in Q1, 17 of which were new.

The reasons for this relatively low order intake are threefold.

First, the healthy order book at the beginning of the quarter.

Customers simply did not need to order systems for first-half delivery.

Secondly, the bookings delay for two flash memory fabs resulting from recently announced investment pushbacks.

And thirdly, the bookings delay from a certain DRAM customer, who is currently discussing new strategic alliances.

We are told that these delayed projects will happen since they're, clearly, of a strategic nature to the customers.

The delays, however, could be one or two quarters.

Orders booked in the first quarter came in large part from our stable logic customer base as they added incremental capacity to meet growing demand at their most advanced process nodes.

Looking forward, we anticipate a gradual increase in unit order intake, partially driven by clearly identifiable market share gains and the flash 45 nanometer technology ramp.

In euro terms, the increase in the second quarter will be significant, driven by high leading-edge technology content.

Our confidence in this is, by the way, based on the state of our current order negotiations.

Given the impact of the delays on the order intake in Q1, together with only a modest pickup in Q2 unit orders, we expect weaker net sales in the next two quarters as compared to the eight previous quarters.

The result of this is that we currently estimate 2008 sales to be about 10% less than our record revenue in 2007.

The order intake level will certainly depend on healthier pricing environment for the memory chips.

Based on the current low level of lithography capacity additions, this is expected to happen in the course of this year.

The first signs of memory price increases have been reported over the last few weeks, by the way.

We plan to ship 42 units in the second quarter with an average selling price of EUR17 million, which is a 5% increase over the first quarter.

Our gross margin for the coming quarter will be approximately 40%, R&D is planned at EUR130 million, and SG&A planned at 58.

But, given the current market weakness, we are trimming manufacturing and SG&A costs for the second half, while keeping R&D stable at our current levels.

In summary, we do remain cautious with the backdrop of global economic uncertainty and uncertainty in memory demand.

However, we expect that ultimately, lithography demand will particularly be driven by the ramps of the new 45 nanometer flash memory node, which are required in the second half of 2008, the firming up of DRAM prices, and the completion of new strategic alignments between DRAM makers.

That's it for my part of the introduction.

Eric, I'll turn back to you.

Thank you, Peter.

Peter has highlighted the fact that we again posted sales of above 900 million in the eighth quarter in a row.

This is witness to the effect of both system average selling prices and market share growth.

We posted a 15.8% net income in Q1 '08 on the shipment of only 50 systems, compared to a slightly higher 16.1% net income on the shipment of 77 systems in the same quarter one year earlier.

This performance is clearly ASP growth related, and underlines ASML's relative stability in this cyclical industry.

The continued global economic uncertainty has an impact, obviously, on the semiconductor sector, as logic and foundry customers are booking very conservatively, tool by tool, on immediate need for capacity and as memory customers await firming up of their prices, which (technical difficulty) supply chain supply demand driven, before making serious investments.

This near-term effect on the growth of our business, however, does not impact our long-term trajectory.

Flash memory customers, as Peter said, are still developing a new technology node every year, and we continue to see the ramp of 45 nanometer technology in the second half of '08, while anticipating the start of the 38 nanometer node double patterning technology early in 2009.

DRAM customers themselves are currently aligning their capacity to market needs, which is good, developing the necessary industry partnerships, and, in addition, gearing up for a 55 nanometer node technology ramp by the second half of '08 and the beginning of '09.

Logic and foundry customers have not significantly increased capacity for at least two to three years, and the announced transition of major players like Texas Instrument and (inaudible) Freescale to a fab-like model have not yet generated -- generating -- generated the additional demand from the boundaries as the capacity is basically rationalized at this time.

Significant capacity ramp is, therefore, due to happen at some time in the future.

We will certainly adjust our overhead in coming quarters to adjust to this slight top-line correction, but we will remain steadfast with our investments in research and development during the year.

As you know, our R&D effort has yielded significant leadership in the current new immersion technology, with a further 14 immersion systems shipped in Q1 -- which, I think, is a record for us and for the world -- bringing our total immersion installed base to 86 tools, 53 of which are in Asia, 16 of which are in Japan.

Customers are able to run now up to 2750 wafers per day on this immersion machine, which is equivalent to about 1 million wafer per year.

The total number of wafer produced using ASML immersion tool has now surpassed the 10 million mark worldwide.

Steady ramp at our traditional and new customers confirm our product superiority, in particular in terms of throughput, overlay and critical dimension capability.

We confirm, therefore, our previous expectation that immersion revenues could nearly double in 2008 compared to 2007.

To ensure continued leadership in next-generation technology, we're strengthening our development effort on EUV, the next large leap forward in chip production technology.

We report significant progress on this technology, Extreme Ultraviolet.

[It's the world's] first full field SRAM cell (inaudible) chip is made on one of our EUV machines, and announced at the SPIE Advanced Lithographic Conference of February this year.

We have now the final roadmap that will take us down to 11 nanometer.

As I mentioned, customers are now transferring to 45 nanometer.

So we're talking (inaudible) 11 nanometer in the coming decade, and are producing the first five units of EUV technology to be delivered starting early 2010.

In conclusion, the current bookings slowdown that we are seeing is supporting an industry capacity adjust to demand, which is a good thing, which in turn will translate into a sound basis for future investment in the midterm, for which ASML is extremely well-positioned.

So, now Peter and I will be available to take your questions.

Ladies and gentlemen, when asking a question, please identify yourself and your organization.

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

(OPERATOR INSTRUCTIONS).

Simon Schafer, Goldman.

I wanted to ask a question about NAND flash, and specifically as it relates to your immersion bookings.

I look at bookings this quarter -- only two systems booked.

I was wondering why NAND exactly would accelerate.

I think I understand that, clearly, there's some requirement for shrink somewhere down the line.

But, what gives you the confidence in terms of timing that that would really re-accelerate in the second quarter?

There are different reasons for that.

One is you have some new customers who haven't yet booked into the NAND new technologies immersion, which are in the 55 range or different version of shrink.

One of them is, as you can imagine, IMFT.

And these are, we say, different waves of booking.

And you've heard that IMFT have now clarified when they will ship, or when they will start up, basically, their new factory in Singapore.

This also explains why these type of orders were expected in Q1.

We said in Q4 that we didn't -- we're not sure about the timing.

But at this very moment, if you talk to these type of customers, they will ensure that they will get in now during this timeframe.

Because if they don't, they will miss a technology node, which would be extremely difficult for them to catch up in the meantime.

So, first answer is you've got those waves of new people.

Second answer is although you are hinting in your question that the flash prices are weak, they are not as bad as DRAM.

They are on the weak side.

And that has -- that could [question the tail] orders from the current masters of the game, which are Samsung, Hynix, Toshiba, to name three.

We are very confident that these are going to build a bit more capacity in the second half.

We expected, and this has always been the case, that this build will not be as high as the first half for most of them, because they usually build by [book].

But what is being discussed at this very moment with those larger accounts is it seems to be highly solid and very much in tune on understanding the price levels of flash.

My follow-up would be, as a result, would you expect your book-to-bill in immersion to be above one in the second quarter?

The book-to-bill above one in the second --

Just for your immersion systems.

Yes.

I'm being told by Craig I should not give you too much (multiple speakers).

As you said yourself, and I did not underline, in addition to the flash capacity that I just mentioned, you do have the first orders for the next node, which are happening at the same time, and they are starting in Q2.

So in other terms, you've got a bit of 55 nanometer; you've got a lot of 45 nanometer; you've got some 32, and these are ensuring that we restart in flash booking.

Thanks so much.

Nicolas Gaudois, UBS.

Just a follow-up on this line of thought.

Could you give us a bit more granularity on when you expect double patterning flash tools to be required next year, so around which quarter should we start to see that, and when double patterning 32 nanometer microprocessors as well?

And I've got a follow-up.

First of all, I think we are now getting [a little] certainty about which technology we're talking about.

This is new news.

There were a certain number of competing technologies that were being looked at.

And I believe at this very moment in the flash sector, the technologies are being (technical difficulty) and they are looking like a mixture of what we call spacer technology and double patterning.

So this is now clarified, which is good.

That means it clarifies -- it's strengthened the possibility of ordering.

These will be used in what they call the 32 nanometer nodes, and the 32 nanometer nodes we expect to ramp in the Q1 or Q2 timeframe of '09.

We already have a picture of the needs, and we believe that it will be still more -- requiring more capacity in terms of (inaudible) than the current 45 nanometer type nodes.

Regarding logic, it's even better for us and a bit worse for our customer in the sense that they cannot use spacer technology, which requires a bit less immersion than what we call pure double patterning, where you absolutely need two immersion layers on top of each other.

These technologies also now being frozen in the logic sector, and we are going to see certain numbers of customers buying in the first half of '09.

But, we would still expect logic to not represent a very large part of immersion.

It will still be dwarfed by the flash ramp, and by the new immersion ramp of DRAM, which will start definitively, or has already started a bit, but will be in production in Q3 of this year, and more customers will come from Q4, Q1, Q2 of '09.

So DRAM is really the segment that will boost everything for two reasons.

One, DRAM is still larger than Flash, so a transition -- node transition is a big thing in a big sector.

And second reason is DRAM has more immersion layer than flash.

So, for two reasons, we would expect DRAM to be potentially a bigger buyer of immersion than flash altogether in the first half of '09.

Great.

Just a follow-up.

[Obviously, you start with a] question in this environment of whether the tools are needed, but whether your customers can actually buy them at some point.

If you look at your memory backlog, it's EUR650 million at the end of March.

Could you tell us how much of that is non-Samsung, non-Toshiba -- so all of the rest -- which we would qualify would be higher risk balance sheet customers?

It's a bit hard to give that to you on the phone without making mistakes, but I would say this (multiple speakers).

There is no DRAM second tier, as we said in our press release, because we are expecting this to happen sometime much later in 2008.

But it is, I would say, well -- it's well-spread on the top three at this moment.

Thank you very much.

(inaudible)

I was wondering if you could maybe let us know when you think sales will start recovering again.

Would that be in the fourth quarter?

And could you give some more light -- shed some more light on immersion sales this year?

I didn't catch your question.

Can you please repeat?

The first one or the second one?

The first one.

If you could tell us when sales will start growing again, in which quarter.

By saying that we expect two quarters a bit weaker -- in fact, when we say a bit weaker, you've seen our guidance on Q2, these are still big numbers.

We have been at above 900, so this time we declared to be a bit lower than 900, but this is still a good quarter -- weaker, but a good quarter.

Then we've declared that Q3 may be on the low side in view of the bookings -- low bookings of Q1.

This is natural, so this is based on the lead time, and therefore -- and we expect the pickup, as we said, in Q2 of the bookings.

We said the gradual bookings picked up in unit, and we said significant value pickup in bookings in Q2.

So you can imagine that significant value booking pickup in Q2 will translate into significant sales in Q4.

Does that answer your question?

Yes, the first part.

The second part was on the immersion forecast for the year.

As I said, it's still -- we said that we would nearly double.

We've been consistent in this for the past six months, and so I say the same thing.

Double versus 2007.

Sorry; was that nearly double or really double?

Nearly double in Q1 at the end of Q4, and that's a statement that still stands.

I use the same nearly and the same accent.

Thank you.

Satya Kumar, Credit Suisse.

I had a question on immersion.

Can you give us a sense of how many layers of immersion is needed for NAND at 55, 45 and 32?

Is there one answer for that?

I think we can follow-up with you.

But, (inaudible) and more when you go from one technology to the other.

You would add a minimum of 10%, so one layer.

And this is flash, and then DRAM is more than that.

But I think (multiple speakers)

If you want details, I think you should refer to Craig and to Frank, who can give you the details.

I don't think we have them here.

The other question I have is, on immersion, what are the expectations for immersion unit shipments for the market this year?

How much of your doubling of sales is market share versus market?

Interesting question.

We believe that the doubling that is a good number of immersion tools we shipped this year will be -- they will be a number -- I'm calculating fast in my head -- about 15% of them.

10 to 15% of the total number will be market share gain.

And the rest will be market, also driven by the node transitions in DRAM.

So we had flash last year, and we have, of course, flash continuation this year and DRAM transition this year.

It sounds like 10 to 15% sequential market share gains.

When you go from -- looking out into your June quarter, it sounds like you have a pretty big increase in your immersion bookings.

In your non-immersion bookings that you expect in the June quarter, is that likely to be up or down sequentially?

How should we think about that?

As you know, we booked a very small amount in Q1.

So we don't expect to go lower than that in what we call the capacity orders.

Could it be lower, the non-immersion (inaudible)?

I'm not going to be more granular.

We think at this moment the whole industry is not booking a significant or any number of capacity CapEx.

So if Q1 is a bit lower or higher than Q2, it's irrelevant.

The reason why ASML is doing better than everybody is we are doing this immersion business, which at the end is responsible for most of those 900 million of sales that we're doing.

Now, that is good news, what I say is good news.

In other terms, I'm saying that at this moment, we're not calling a pickup of capacity orders.

And this is an important message.

It basically says we are conservative or realistic as to the fact that there is a crisis out there, and that customers are not going to take risk, further risk on capacity.

But by saying this, we're saying that the DRAM business is not building up further capacity.

In fact, you have noticed that in the presentation we have on the Web, we are showing that DRAM capacity built is going to be less than 2% [growth], if our view as I presented is correct.

So, very low capacity built.

I also mentioned in my little discussion in the beginning that foundry is not building capacity, or is building capacity one at a time.

So, very low.

So on this trajectory, which is pretty dismal, we have still, I would say, I think, good body language about our future on 2008, because we have immersion.

On the other hand, you should take this also as a positive statement that if this industry doesn't build enough capacity and then the demand picks up, we're going to have significant potential.

Also as a reference, you could go to slide 25 of the presentation, where you can see that for -- immersion is close to 60% of the backlog, which is basically leading edge.

And the 30% (inaudible) is leading edge for our logic customers.

So it is largely -- 90% of the backlog is leading edge.

And there's just above 10% of capacity orders in the backlog.

And like Eric said, we don't think it's going to change that much over the next one or two quarters.

Thank you.

Harlan Sur, Morgan Stanley.

Nice execution on the March quarter.

As it relates to the market share gains, you mentioned you had 16 immersion tools installed in Japan; that's up from 14 (technical difficulty).

I'm just curious as to what's your expectation on installed systems in Japan exiting 2008.

At the end of 2008 in Japan, it is possible that we will finish with another -- maybe another 10 -- 10 -- 10 to 15.

An incremental 10 to 15 tools?

Excellent.

And then, on that same note, maybe you can touch upon the progress you're making with the one Japan customer, I think, who's currently qualifying your immersion tool.

The question is, how is the call proceeding?

When do you expect them to start taking deliveries if everything works out?

On the technical side, everything is progressing ahead of schedule.

As you may imagine, these machines are complicated, and they have to adapt to a now completely different environment when you go to a new customer.

This is happening very smoothly, and I guess the market rumors may confirm that.

The decision, I think, is still the same.

We have been told that the decision to proceed or not with us, and at which speed and at which size, will be taken starting early -- starting end of June.

So I would say negotiations and planning, etcetera, may take a bit of time within the July timeframe.

So I don't know if by the July call that we'll be able to report.

That doesn't mean that we don't have it, but it's one of those things where you have to negotiate and plan, and you know that we have some leadtime.

That will be happening.

At this moment, we feel comfortable because we haven't seen significant pressure by our competitor.

And we've seen the momentum of our immersion tool at the other customers, who are making, I would say -- or giving a sense of confidence to these new customers about our capabilities.

Thank you.

Great progress.

Timothy Arcuri, Citigroup.

Number one, do you have any way of gauging what percentage -- so, as of the end of Q1, what percentage of the total 45 nanometer system opportunity have you chewed through?

Basically, if the industry is going to book, say, X number of systems, what percentage of those systems have basically already been booked?

Unfortunately -- I would like to have that answer -- this is where the limit of our visibility is, where customers kind of plan things and change every day.

So you do have for each of the customer a mix between the usage of an immersion tool in flash or DRAM.

That's the first thing.

And they would take the decision at the last moment.

And then you have -- in the flash you would have 56 or 55, whatever, 45, or even 32 in R&D.

The decision is really fluid on their side; it really depends on the prices of the chips.

So in other terms, if they want to sell [one 512] or 1 gig (inaudible) DRAM, or whatever, they will change their mix.

So unfortunately -- in fact, I am trying to get this (inaudible) for my own sake, and to manage the model, I would say with certainty our supply chain, but I can't even do this.

And do you have any view on what the incremental cost benefit is as these NAND companies go to 45 and also 32 nanometer?

There's been some chatter that there actually is less incremental cost benefit once you get beyond 45 nanometer.

Have you had any talks with your customers?

What is your sense on that?

What you hear, which is -- which we confirm, is the semiconductor industry in general, whether it is flash, logic, or DRAM, are concerned about the affordability of the technology.

And of course we are one of the guilty people.

Litho costs more money, and as we said, the good news is it has -- it gives us an opportunity for growth.

The bad news is our customers need to invest more in litho, and therefore, they are concerned about the cost of litho, but as I said, the cost of technology.

So people would tell you that for a 1 centimeter square of a given technology, [65], (inaudible) the 1 centimeter square of 45 is much more expensive, and it is dramatic and a major issue, and the industry should look at ways of cost reducing.

So that we hear.

And it is true that there is some kind of acceleration of that phenomenon, so the industry is addressing.

But, if you look at it differently, if you understand that technology does not -- in fact enables you to either shrink the die or to put more feature on the die, then per feature, the technology still is much cheaper.

So -- and it's much, much, by a major factor.

It's even an exponential curve.

So if a customer is capable of selling its chip for the value contained -- that is for the numbers of transistors contained -- then there's no issues at all.

So now, this is theoretical.

What's happening practically when you do a -- when you go from 55 nano to 45 nano flash, the whole game depends on the market price.

And this has happened this year as bad news for the industry, which is fully responsible for some of the problem that we're seeing today.

People were going towards a shrink, so in fact the one that happened was between 65 and 55, because the shrink at 55 supposedly got them more margin.

But it was assuming that the price of the market for the 55 nanometer chip was not going down as fast as the 65.

But, because everybody jumped into this market, there was a supply/demand question, and the chip price of the new technology all went down much faster than historically.

And therefore, the customers came out with a same type of margin at the 65 nanometer, the old technology, or at 55 nano, the new technology.

So these type of things happen.

So in other terms, market price makes more difference than my theoretical view of life, as I explained at the beginning.

Great.

Thank you.

Sorry for the complexity of my answer.

Jonathan Crossfield, Merrill Lynch.

My first question was on the decision to cut SG&A in the second half, but leave R&D unchanged.

What's the thinking behind this, given that you have so much more scale than your competitors.

And what would trigger, in your minds, a decision to scale back R&D?

I think the thinking behind this is that in R&D we have a few major projects that we're executing that we think are vital for the medium-term future of the Company.

You can talk about the advantage we currently have in our current immersion technology, so that is at least the first project that we're working on in R&D.

The second one is the next generation -- that's the next-generation architecture.

The next-generation architecture is necessary for double patterning, and I think I talked about that quite extensively.

That is something that we need to execute if we want to secure our market share position that we currently have.

The third one is EUV.

Eric said very, very clearly that we have shown, clearly, the viability of EUV because it can make four chips, as we have shown at the conference in February, the SPIE conference.

So a lot of things need to be done to make EUV production [worthy] to ship the first preproduction tools in 2010.

You can argue that that is currently not a distinguishing factor in the business that we have vis-a-vis our competitors today, but it will be in two to three years from now.

On top of that, we have litho plus, which is the (inaudible) integration, and it's the in-situ metrology development that we're combining with our scanner solutions, which we believe is together with the new architecture in absolute necessity to do effective double patterning.

So those are four projects -- immersion, the next architecture, EUV, and extended litho -- which we believe are vital to the mid and long-term future of this company for us to be able to sustain the market shares that we currently have.

That is the reason why we aren't cutting (inaudible) R&D.

And to be honest, like Eric also said as an answer on one of the questions, we do believe that we will see in Q4 an uptick.

And that means that we are looking at two quarters that will be weaker in terms of sales as previous ones, but that is not enough for us to start cutting in R&D.

If you say when do you start cutting R&D, we will have to see a very sustained long downturn with severe pressure on the capacity additions and on the technology needs of our customers.

We aren't seeing that.

And by the time that we see it, we'll communicate that, and you'll see our actions in the R&D line.

But not earlier.

Just as a quick follow-up, could you update us on the XT 1000 KrF tool, and has that started to ship yet in place of dry ArF?

No.

It has been announced for the Q3 timeframe for the first shipment, and it is planned to ship as scheduled.

I would not put this yet in my model about our future forecast this significant conversion.

I remind the audience who do not know, this is a product that is supposedly allowing customers to convert back from an expensive ArF environment to a KrF environment, and still do the same type of shrink.

So, in order to do that, you need to do significant R&D.

So we expect the R&D to be minimum six months if not one year before the customer can significantly take advantage of that.

So we will ship some of [that then], but production will probably be due mid 2009.

Thanks very much.

Didier Scemama, ABN AMRO.

First of all, looking at the competitive landscape, it seems like this year your market sharing value would be probably somewhere between 70 and perhaps 75%, and you're taking share in Japan potentially with immersion at the only customer that would potentially deploy immersion that is not using ASML, if that's clear enough.

So, my first question would be, do you think there is enough room left for your two other Japanese competitors in the market outside of the RF immersion?

And my second question is related to the DRAM industry, and whether or not you expect 200 millimeter capacity retirement potentially in the second half of this year triggering maybe more 200 millimeter equipment demand at the same timeframe basically.

Regarding the competition, I guess I would have to have two facets to the answer.

It is true that if lithography is a good business and a growing business, it is possible that a second player at 30% market share could make a good business, and we suppose that will happen.

Can two players share that for a period of time?

Yes and no.

What we've seen with our competitors is they are lucky enough to have another business called LCDs.

So when you look at total portfolio, they may be pretty small on the semiconductor side, but they may be bigger on the LCD side, and therefore, they can potentially remain in the large machinery-type business by mixing the two business models.

So, in other terms, I could imagine that it would be a stable business to have 75% market share, and the rest being shared by two, and those two using LCD to help them having critical mass of some sort.

The 200 millimeter question -- yes; we've seen multiple plans and multiple changes of mind about when each of the 200 millimeter fabs will be closed.

Certain numbers have been announced.

I think I made it clear that there were four that we knew of at the beginning of Q4.

Those ones are being -- are proceeding.

They are being upgraded basically from 200 to 300.

If I'm not mistaken, I think the [walls] are still there.

So it's already a 300 millimeter opportunity for us, so it's happening.

But in the past three or four months, I haven't heard of any new ones.

And this is in line with the fact that the DRAM and flash environment is really at this very moment focusing on what is the price of the market going to look like before they get to another set of decisions.

So today, you've heard the DRAM players are very clear, clearly focusing on the marketing point, which is we will raise the price.

And they are raising the price by forcing it in, and by freezing activities, freezing plans, which is why we didn't book, as I said.

And this is the two things they do.

And they don't think about something more extraordinary, which would be to close further and faster 200 millimeter.

I think what we heard already is significant enough.

Just a quick follow-up.

In terms of share buybacks, you've increased -- or you have announced dividend, which will allow you a pretty significant tax-free share buyback, I think, next year.

I'm just (multiple speakers).

Sorry?

2010.

2010.

Do you have any plans to maybe use -- do you have maybe a reverse stock split, given your excess gross cash on the balance sheet?

And what would be sort of the factors that would drive a new share buyback?

I think it's nothing changed, I would say.

I think we have a clear financial policy in which we say we will do -- we will return excess cash as defined outside the bandwidth of the Company that we said is between 1 and 1.5 billion.

When we have excess cash we will return it to shareholders.

Now, the straight buyback has been closed until 2010 because of the difficult situation with respect to the withholding tax issue.

But there are still options open for us to do things like the capital repayment and the reverse stock split.

That has not changed.

We have said that before, and we keep repeating it.

When and how will, of course, be announced, I would say, when it is there.

So there's nothing to announce today.

We will do that in due time.

But our appetite for buyback is still there.

And as Peters says, it's now a technicality when we execute these (multiple speakers)

Thank you.

Jim Fontanelli, Arete.

Firstly, on inventory, your inventory is up, obviously, sequentially near 30% year-on-year against a backdrop of falling sales.

So I was wondering if you could run through why that is.

I guess it's to do with maybe growing immersion shipments.

But I'd be interested to [follow through] on that.

And secondly, whether you're seeing an increasing use of sale and leaseback by, maybe, some of your more financially stretched customers, to make CapEx budgets stretch a bit further.

I will handle inventory and Peter will handle the sale and leaseback.

For the inventory, obviously, as a former guy from Dell, I do not appreciate inventory, so we do have significant efficiency projects to reduce inventory.

And this is working, I would say, pretty nice.

One part of the project is about cycle time reduction in production so the width is reduced, and another part of the project is to have part of the load being shared with the suppliers, and that has also happened.

So we're doing okay there.

What you see in the rise of inventory, however, is two things.

One is, yes, a mix.

The immersion mix is extremely high, much higher than in a normal environment, because we don't have any capacity orders.

I'll give you an order of magnitude.

A 1900 immersion machine will take about 14 weeks, 15 weeks to be manufactured, that is to be carried on inventory.

And a KrF machine would take about six weeks, four to six weeks.

So, this makes a difference of (inaudible).

The second aspect is voluntary.

We said to you last time that we would boost the inventory voluntarily because of the uncertainty of the market.

And at the type of market share that we are reaching, it is difficult for us to take a risk that we are not able to deliver when the customers start picking up.

And the foundries, very much so at this moment, are still buying a good level -- not a great level, a good level -- and they do this very, very short leadtime.

So we do carry a bit more than in an environment where we have more visibility.

Now to Peter.

I think on top of that, what we also have in inventory is a higher-than-normal percentage of tools that are currently in R&D, which are basically good tools, because we are developing the new immersion tools, so we have more, let's say -- I wouldn't call it prototype, but more prototype tools, where we do all of our testing and we do new development on it, than we normally have.

Plus we're also building our new architecture, which is currently in the assembly phase.

So that is, I would say, typical for where we are today.

If you look at the immersion ramp, which has been the fastest ramp ever, it also means that the requirements the customers have to control their processes vis-a-vis, let's say, the effectiveness of the tool, require our R&D people for this moment in time to keep more tools in R&D.

(inaudible) those are good tools that will be sold when the R&D needs for the immersion ramp-up goes down.

So that is on top.

That's, like Eric said, the buffer of the voluntary decision to take some (inaudible) to react fast, there is -- that is there.

It could both add up to about EUR150 million, so that is quite a significant amount of money.

Sale and leaseback transactions -- yes, we have seen that.

We have clearly seen that.

Actually as a couple of deals have been closed, without going into detail, our -- as a management group who actually helps customers, but also helps some of the leasing companies to get their hands around residual value management, has been extremely active.

And I would say that with the exception probably of a very large memory maker in Korea, almost every memory maker -- DRAM maker is talking to lease companies.

And some of them have already closed some of the deals, and some of are in the last stages of closing.

And some are having more difficulties.

That's also clear because the risk is also evident, and some companies are more prone to getting it closed than others.

But there is a lot of activity in that market.

Mehdi Hosseini, FBR.

Most of my questions have been answered.

Just (inaudible) long-term strategy, as you look into -- beyond 2009, as advanced logic approaches 32 nanometer, and foundries at 45 and so forth, how should we think about your growth prospects?

Will [UAV] be enough to sustain the current growth that is supported by immersion, or should we think about other markets?

First of all, the semiconductor leasehold market we still expect to grow, and we expect it in 2010, 2011 or so, to be fairly spread between different technologies.

We expect some customers really adopting EUV, and we know why; there are some economics for that.

We expect some companies to keep immersion alive and using very complicated double patterning situations.

We expect the low-end i-Line/KrF to be fighting between each other to win this set of easier processes, which is why we introduced the XT 1000.

So, if you would want to ask us what the profile of 2010, 2011, 2012, assuming no -- another -- not another financial bubble in there, it would be spread with multiple technologies.

This is why we're doing so much to develop different architectures.

We believe that there is enough growth in this business that we should not de-focus our attention from there.

But it is true that in the year 2015 or so, we probably would reach a level of maturity which would make sense for us to think about something a bit different than semiconductors.

So we have been working on this since the past two years or so.

And at this moment, there is nothing like [solar] that we want to announce to the group.

But it is a very much strategic [responsibility] for the Company to be sure that we always deliver growth opportunities to our shareholders.

But as I said today, we've got a nice run with lithography semiconductor with enough growth that should be where we should spend most of our time.

I look at your revenue growth; from 2003 to 2007 you were able to do a CAGR of 25%.

Some of it was driven by share gain, some by immersion, and some by capacity purchases.

I guess what I was trying to figure out -- how will -- those growth rates will change over the next three years.

I'm certainly not referring to 2015, but more so in terms of the next two to four years.

And especially given the kind of R&D spending you have.

And I'm trying to better understand how you should think about the dollars of R&D spent, and to what extent is that going to help sustain the past four years' revenue growth.

Let me answer this.

The danger of taking a point in time and then calculating a compound average growth rate is always what's the moment of start.

In 2003, which was -- you could say the very low point in the last 10 years also in terms of revenue debt, so we grew that from 1.5 to about 3.8 last year.

If you would take 2000 or 2001 it would look different.

That for just the record.

If you look going forward, and you don't exclude the cycles, and we anticipate semiconductor unit growth to be around 10%, and you extrapolate that for the different technologies that Eric just talked about, whether it's double patterning extended, or whether it's EUV, and you extrapolate that to 2012.

And you do also -- you include there things like the increased number of litho layers because the complexity goes up.

As you go to 32 nanometer, you need more litho.

You take that all in consideration.

Depending on what kind of technology scenario you take, we have shown at the analyst day in November that the total market litho would be anywhere between 10 and 12 billion.

Now, at our 70% market share target, you could easily calculate what that is.

And then you could say over the next -- over the next four to five years would give you a compound annual growth rate of between 12 and 14%.

It depends on what you take.

So that is what we had presented at that time, and there's no reason to say anything else on the [growth model] of the Company now than we did at that time.

So, I don't have any other information.

But if you say the leverage of R&D to [growth], I think you've got a point that we don't want to commit to you today, but we've said often that we are over-investing at this moment, and that we do expect the R&D load to be lower on us after we've built up all those architectures.

So we are building an architecture called XT 4; we're building an architecture called [NXT]; we're building an architecture which will be using EUV, which has no name; we are building an architecture in metrology; we're building an architecture in simulation tools, etcetera.

So, yes; this is heavy.

Do we need to have so many start-ups in 2010?

No.

So, I do -- I have hinted on a regular basis that we're probably getting at our peak now or the next year or so.

I still believe we can put another 10 million or so per quarter to accelerate those projects, and our duty, I think, is to accelerate them for strategic reasons.

But when they are accelerated and completed, then the level of maintenance R&D is going to be lower.

Thank you.

Ben Pang, Caris & Co.

You mentioned the push-out of the two NAND flash factories.

Those essentially have no impact on your immersion projections for 2008, correct?

That is correct.

That is correct.

And in terms of the total number of units you're looking at for 2008 right now, including all of your immersion units, what's the level that you're looking at for the whole industry?

I guess you have this on page 9 of the presentation which is on the Web.

We have made an estimation which is about 426 tools, lithography, semiconductor lithography tools, for the total year 2008.

On a unit basis, your market share should still be above 50%?

Is that correct?

We have been, I think, in the 45-ish% range historically.

And it's not getting better because of the price tag of the 1900.

So I think you should for your formatting probably keep about 45-ish% of the 426.

But of course our ASP level is growing.

And a quick calculation will tell you that in order to reach our current assessment of the total year, we would need to have 45-ish% of those 426 multiplied by a good average ASP, and we do expect to have.

Thank you very much.

Nicole, maybe we can take one more call.

Janardan Menon, Dresdner Kleinwort.

I just want to dig into the trimming of manufacturing expenses that you're doing, which is -- what impact will that have on your gross margins?

If your volumes stay at this level in the second half of the year, will you be able to exceed the 40% gross margin?

And if your volumes in the third quarter go down to 32 tools, or something in that range, will that enable you to keep your gross margins at about 40% in that case?

Janardan, I will answer your question only if that allows you to change from bear to bull.

I'm trying, but your share price isn't coming down, it's going up.

What can I do?

Because you leave me hope, I will answer.

It is difficult to be scientific at this moment because it depends what the exact range of the revenue is.

When we announce, of course, that we are trimming G&A and manufacturing costs, this is exactly to your point, to try to hit our famous 40% gross margin target, which has been what I said we're managing the Company on.

I may lie a bit, and we may be around 38.5, 39 or so, if we are fairly low on the sales level, and we will be above if we are back above 900.

I don't want to be specific.

I think what I want to clarify, however, is we are now a bit smarter than we ever were in variability.

So we can keep our gross margin within the range, I would say, that you will probably consider acceptable.

We voluntarily will lose a bit of money by continuing the R&D effort, as we said.

So we could have scaled it, but we don't, because we think this -- we think at this moment that we don't need to do so, because the down side is not significant, and at this moment limited.

So I don't know if I answered your question.

But, yes; we are going to try to fight to be around that 40%.

But please accept that if the top-line is significantly under 800-ish, we will probably be a bit lower than 40%.

A quick follow-up if I might.

On your 45 nanometer flash memory assumptions, I would assume that much of your shipments of immersion tools in the last two quarters have been of the 1900 to NAND Flash companies.

I'm just wondering why is it that they can't use those tools, which they have been using at the low 50 nanometer level, at 45 nanometer, and why they would need to buy more tools effectively?

Or is it that they have capacity expansion plans, which is warranting that additional tool purchase?

Yes and yes.

It is in spite of the fact that in particular you are saying that Flash [is not so great a business], but they do have a growth rate of some sort, which is not as great as it used to be.

It used to be 50% unit growth, etcetera.

I think now the industry is all talking about 30, 25, (multiple speakers).

On the 20, 25% range, you still need some capacity.

So they need it.

Secondly, you do have multiple products.

So, you are doing -- and I don't remember the exact integration levels.

But on 55, they are going to continue producing for another two years or something, because that is the die size and the die integration that is current in production.

They are not stopping this, and they are not converting everything to 45.

They are converting -- they're not even converting.

They are putting on 45 new designs.

So the old designs (inaudible) they have been designed in some phones and some cars and some whatever it is, and you can imagine a guy who has done a design with an application (inaudible) is not going to change [this now].

So they're going to have a lifetime of X.

And during that lifetime, those companies will have to continue delivering.

In other terms, these 55 or 65 will remain in production for one, two, three years, or so, and then convert only later.

In the meantime, you need 45 capacity.

On top of that, in addition to what Eric said, if you talk about 45 nanometer nodes, you could say that's your 1700.

So why would they need to buy 1900s if they have 1700s?

[Why should you do that]?

On top of what Eric said is that many of those customers have -- their true node is between 40 and 43, which actually means if you have a 45 nanometer node-capable tool, for instance, a 1.2 or 1.3 NA tool, that would really stretch the process window.

So if you have the 1.35 NA 1900 would bring you largely right within the process window, and that has a direct impact on, for instance, the output at yield.

That is the reason why if you are stretching, which is (inaudible) the flash makers do, they have the most aggressive nodes between 40 and 43, that's why they need the 1900 to be -- to basically have the yield to be profitable.

Thank you very much.

At this point I'd like to thank, on behalf of ASML, everybody for joining our call.

Nicole, if you could formally close this out, I'd appreciate it.

Yes, of course.

Thank you, sir.

Ladies and gentlemen, this concludes the ASML 2008 first-quarter results conference call.

Thank you for participating, and you may disconnect your line now.