艾司摩爾 (ASML) 2007 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by and welcome to the ASML 2007 fourth quarter and annual results conference call on January 16, 2008.

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

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

(OPERATOR INSTRUCTIONS).

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

Craig DeYoung.

Go ahead, please, sir.

Thank you, operator.

Good afternoon and good morning, ladies and gentlemen.

This is Craig DeYoung, Investor Relations at ASML.

Welcome to our investor call and webcast.

The subject of today's call is ASML's 2007 fourth quarter and annual results.

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

I would like to draw your attention to the Safe Harbor statement contained in today's press release and presentation, and you can find our press release and presentation on our website at www.asml.com.

The length of the call today will be 60 minutes.

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

Thank you, Craig.

Good afternoon.

Good morning.

Thank you for attending our results conference call.

As it is now our habit in our quarterly conference calls, we will split the introduction in two parts.

Peter will start with a review of our Q4 financial performance, with added comments on our short-term outlook.

I will complete the introduction with a longer-term view.

And following the introduction, we will open up to -- the call for questions.

Thank you, Eric, and welcome to everyone.

First of all, I would like to review our annual and Q4 results highlights.

I am pleased to report that the quarter developed as we had anticipated.

Q4 revenues were EUR973m, nicely within the band of sales that we achieved in the past six quarters.

The strong fourth quarter results leads us to record high sales of EUR3.1b -- sorry, EUR3.8m for 2007, which is up 6% versus 2006 net sales, being in line with the revenue growth guidance given by us in the earnings call this time last year.

A 40% gross margin and operating expenses of EUR186m resulted in an operating margin on the quarter of 21.6% versus 22.4% in Q3.

And that reflects our increased R&D spend in Q4, which lead to a total amount of EUR129m.

Cash from operations was EUR72m in Q4 and EUR670m for the total year.

We ended 2007 with around EUR1.3b in cash.

Some of the significant events impacting our cash balance during the year were the inflow of EUR600m resulting from our inaugural Eurobond and an outflow of EUR200m for acquisitions, and distributions to our shareholders in total of about EUR1.4b.

Through 2007, the number of outstanding ASML shares decreased from 477m to 436m; that's a reduction of 41m shares.

And consistent with our policy, we will maintain a target for cash reserves of between EUR1b and EUR1.5b.

We will complete the existing share buyback program for 14m shares, 9m of which were already acquired in 2007, and we will continue with our commitment to return to shareholders cash generated above our cash balance target.

ASML's order backlog as of the end of the year remained essentially flat at around EUR1.7b, which included 89 systems with an average selling price of EUR19.1m.

For comparison, the ASML backlog at the end of September 2007 was valued at around EUR1.8b, which included 90 systems with an average selling price of EUR19.7m.

Orders booked in the fourth quarter, they served in large part the intent of our memory customers to move to new [plus] nodes in NAND Flash and in DRAM production.

12 new orders for immersion tools were placed by customers, leaving the year-end immersion backlog at 34 systems.

Interestingly, near equal numbers of ArF and KrF-based systems were ordered in Q4.

That is to support the expansion of leading-edge production.

Next to the memory segment, that I think again led the way in the ordering in Q4, was that we booked a significant number of tools by our IDM customers.

As in the previous quarter, we saw, again, a low level of orders from our foundry customers, while we continued to see the factories run at the highest overall utilization in the last 15 months.

And the 8% foundry level in our order book is the lowest percentage on record at ASML.

We expect to ship 50 units in Q1 2008, with an average selling price of EUR16.3m.

And although the market has recently been nervous, we still expect very healthy sales levels for Q1 of above EUR900m.

That's on the back of the strength of technology transitions by flash memory and DRAM customers, making this, by the way, the first quarter of 2008, the eighth consecutive quarter with sales above this EUR900m level.

Regarding Q1 bookings expectations, we acknowledge the potential weakness due to the global economic sentiment, and that this indeed will cause our customers to take a prudent view on capital expenditures for the time being.

They will likely await more positive market signs before confirming significant numbers of new orders.

However, I think ASML's backlog at the end of 2007 reflects some of the current market trends and supports a strong first half.

Number one is the backlog, at 89 units, of close to EUR1.7b, 80% of which is deliverable in the first half.

Secondly, immersion technology represents 60% of the EUR1.7b backlog, being EUR1b of expected immersion shipments in the first half -- sorry, EUR1b of total immersion in the backlog.

And thirdly, we see solid IDM demand of approximately EUR440m, that's 26% of the backlog, which we think underpins the non-memory portion of the backlog.

And fourth, that we have a fairly limited exposure to the second tier DRAM customers, that being only approximately EUR200m in the backlog, most of which are technology driven.

So, in summary, we still remain positive about our market opportunities this year, in view of the large immersion technology needs.

The high level of capacity utilization, specifically at our foundry customers, and the relatively low level of semiconductor inventory, and based on the independent IC unit growth forecasts.

So, now I would like to turn back to Eric to the longer-term view.

Thank you, Peter.

Peter has highlighted ASML's record revenues and net profit for 2007.

Indeed, we ended 2007 with a 6% increase in revenues and finished the year with a net profit of 18.1% of sales, which is a record profit on record sales.

We also marked a return on invested capital of about 50%, 49.7% for the year, which is also a fairly commendable performance.

This has been no small achievement in the environment of accelerated technology demands and fierce competition.

We've been able to increase our market share further, to our estimate of about 65% for the year 2007 from 63% in 2006.

However proud we might be of our accomplishments this past year, we must now look, obviously, to 2008 and beyond.

For 2008, and as Peter discussed, we are currently operating in an environment of overall global economic uncertainty, which has led to significant adjustments over the past few weeks in different industries.

In these uncertain times, I would like to emphasize, however, ASML's privileged position.

First, ASML operates as a pure player in semiconductor lithography, which is a market which has become much less market cycle dependent, as two mechanisms compensate for each other.

In boom times, demand for lithography is fed by the need for trailing edge technology capacity.

And in slower times, demand for lithography is dominated by leading edge technology needed to achieve the cost levels and the access to new markets required by customers.

We again confirm that this mechanism is currently in action, as 60% of our large backlog is covered by leading edge immersion products, which compensate for the slower trailing edge capacity orders.

That is point number one of our privileged position.

Point number two regards growth capability.

ASML's privileged position, in this case, is related to the overall lithography market's significant long-term growth potential, supported by technology evolution roadmap of the semiconductor manufacturers and the accelerating need for and cost of leading edge technology.

First of all, competition in the DRAM market has caused a dramatic falling of DRAM prices.

In response, leading manufacturers are acting with urgency in choosing to pursue device shrink for lower cost to the 5X nanometer litho node and ramping months before their original plans, while retiring - this is an important point - while retiring older 200 millimeter capacity.

So that's one of the drivers of growth.

The second driver of growth is NAND manufacturers who are continuing on their aggressive pursuit for higher density, to lower the cost of the device but also to address their expansion into markets with increased density.

This involves adding production capacity at the 5X nanometer litho node but also, in parallel, ramping production volume at the 4X litho node.

That is the second engine of growth.

Third engine, the foundry manufacturers, which haven't been pursuing a lot of capital expenditure in the past two years, are starting to see volume increase in the leading edge 65 nanometer nodes and are starting to enjoy very good utilization rates.

The fourth point which drives growth is that we believe that the new technologies required for the future, already in 2008 but significantly in 2009, will require significant use of lithography as double patterning, the technology of choice, will increase numbers -- the number of processing steps.

So what I said, one, privileged situation because of the fact that the market is becoming less cycle dependent, two, that the market is overall sitting on a growth trajectory and, three, ASML operates a winning business model, enabling further growth.

By leveraging its size, in particular its R&D budget potential and its enormous know-how accumulated for now seven years on a differentiated product platform, the TWINSCAN, ASML is able to achieve above-average market share in the leading edge lithography market segment which, as said, currently is one growing very fast.

Indeed, we expect to potentially double our sales of immersion technology in 2008 in comparison with 2007.

Our immersion products are the industry standard now.

Having -- after having shipped 35 immersion systems in 2007, we now carry 34 XT1900i products unit backlog in 2008.

We have now had more than 8m wafers processed on our immersion platform, up from 3.5m last quarter.

These systems are running consistently at production levels about 2,100 wafers per day and per tool.

This is creating a significant experience base for the customers at transitioning into production.

We have 20 customers with active immersion manufacturing or process development programs in place.

So, in conclusion, as we consider a semi long-term trajectory for growth and its cycle robustness, we confirm our optimism for the development of our market position in 2008 and beyond.

Now, Peter and I will be happy to take your questions.

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

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

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

Of course.

Thank you, sir.

(OPERATOR INSTRUCTIONS).

Today's first question is from Simon Schaefer.

Please state your company name, followed by your question.

Hi.

That's Goldman Sachs.

My first question would be you talked about the 200 millimeter capacity retirement in the DRAM industry.

I was wondering whether you could share with us how many facilities you thought were being retired and what type of timeframe you're looking at there.

Thank you.

Yes.

This is a very precise question.

To this precise question, I may be forced to go for a precise answer, unfortunately.

So at this moment these are projects and they are being discussed with different customers.

And we are, at this very moment, working on four major 200 millimeter DRAM retirements.

Four fabs?

Four fabs, not four --

Understood.

Thanks for that detail.

And my second question would be what percentage of your order book is with your top three customers now?

And how does that compare to, let's say, six months ago or so, or nine months ago or so?

Do we have the data?

We're looking for the exact data.

I'm going to try to answer you with a trend at this moment.

As usual, in the first half of the year the flash activity will be higher than normal, because this is when flash over invest in the full year.

They - how do you call this - they front load their investment and therefore you would have a significant portion of the flash vendors of the world.

Now, because there are more flash vendors in which we participate, if you mind my drift, that probably reduces the percentage of business that we used to have on our top customer.

Is it still -- is there any way we can quantify the percentage number, or should I call in later?

I think, Simon, if you call in later we can (multiple speakers).

Thanks.

The next question is from Nicolas Gaudois.

Please state your company name, followed by your question.

UBS.

And apologies for a pretty bad voice today.

Just a first question on your degree of uncertainty on orders in Q1, and you don't confirm anything.

Should we attribute this pretty much solely to the second tier Taiwan DRAM makers and maybe a couple of others being still very cautious in their ordering pattern, or if anything having stopped ordering activity for capacity?

Or is there anything else as well contributing to your degree of uncertainty in the near term?

Obviously a good question.

The first point is I want to repeat, and I will repeat and I'll repeat again, we've said 10 times that we are trying to avoid to guide on bookings because of the complication of the timings of bookings.

And sometimes we give you a hint about up/down or etc., depending on whether it is obvious that we have enough data to document these things.

Today, as we've done two or three times, we don't have enough data; we clearly have an uncertainty.

The uncertainty is based on the fact that we've had the first batch of, I would say, flash orders, which are the obvious ones, as I said at the beginning, which forced us to prepare deliveries for Q1 and Q2, which is one of the reasons why we said H1 is going to be robust.

And therefore, in the Q1 timeframe, that's when we receive the orders of the next batch of customers, which will include the DRAM large ones, the foundries, IDMs and DRAM of, I would say, second tier.

I would say at this moment we are fairly comfortable about the logic segment, where we think we know, as usual, what is happening, and so there is a run rate of some sort.

But on the other side of the equation, the other segments -- three segments that I mentioned, you do have a question mark about which technology, how fast and what is the impact of the overall economy on my decision for CapEx.

You've seen this, obviously, yesterday with the announcement of Samsung, which some time ago in, I think, December timeframe, they mentioned that they may go down in CapEx.

You've heard them yesterday say that they will be stable, a bit higher.

This is obviously a decision made, and I think they underlined this.

DRAM, which requires a bit more thinking about which DRAM you are making in terms of designs, which DRAM you are making in terms of integration, and therefore which DRAM you make in terms of technology.

So, there is a significant amount of technology question and significant amount of how much CapEx they can really afford and when.

Obviously, as I repeat, the reason why we say we are uncertain is because the world is uncertain about exactly how those numbers will pan out.

But what we've correlated, as we usually do, is notwithstanding when they take decisions is what are their needs.

And this is where our optimism comes back.

We are confirming the needs.

That is we know that there is an acceleration of DRAM towards the 5X or 6X technologies that requires a lot of machines.

We know that foundries are reaching the level of utilization that is not sustainable any longer.

And we know the second tier DRAM vendors, who obviously are worried about price and things, still have to go and make some technology transition, of course lagged, but we will call them capacity but for them it is technology transition.

So we know these things are happening.

And if they are sustained by a simple growth of 10% or so unit of semiconductor, then there is growth in 2008 which will force those booking decisions to be taken.

Great.

And if I look at your indication on the immersion revenues, which could nearly double, I get to about EUR1.8b, EUR1.9b of immersion revenues for '08, potentially.

What will this do to ASPs overall in '08 trend wise versus 2007?

Yes, we do -- as you know, where we have been good and the reason why we have been, I would not say recession proof, this is a too strong word, but pretty stable at managing different situations is we do manage different scenarios, Nicolas.

So we have a scenario where a lot of people will start double patterning a bit earlier than first quarter of 2009; that will require major immersion.

We have other scenarios where spatial comes in and then we are going to go for much more, what we call, trailing edge.

So, the answer to your question is significantly difficult at this moment.

We are comfortable that there is square meters of silicon.

We are comfortable that these square meters of silicon will be handled by ASML machines, for good reasons.

But it is difficult to know if it will be one 1900 or one ArF and one KrF at this very moment.

I'm talking, by the way, about second half of the year.

In the first half of the year it is very clear, because we know the technology that is being required and which is being ramped at this moment.

What's going to be up, so ASPs are going to be up in that sector.

On average it's going up but how much is difficult.

Yes, yes.

Okay.

Understood.

Thank you.

The next question is from Jonathan Crossfield.

Please state your company name, followed by your question.

Yes, hi.

It's Merrill Lynch.

Thanks for taking the question.

My first question would be on leasing arrangements for tools.

Given the financial restraints we seem to be seeing from the second tier DRAM players, are you seeing any acceleration in a move towards leasing tools to those sorts of vendors?

Yes, I can be very short on that.

That's the case and that's because leasing companies are trying to engage with us to see how residual values in the end can be effectively matched.

So we're aware of many opportunities for leasing companies to get engaged with semiconductor companies and I think we've seen quite a significant increase in the activity, clearly.

Okay, that's great.

And then just as a follow-up, you've talked in the past about the flexibility of your OpEx.

And given the sort of uncertainty we're now facing, at least through the first half order levels, what sort of criteria do you look at to determine whether you scale back OpEx, because obviously you're guiding to maintaining fairly high levels of R&D into the early part of '08?

Well, the early part of '08 is, we say, also covered by what we call a robust level of sales and that's covered -- that's basically covered by the order backlog situation.

So, in that sense, that's perfectly in line with what we have said in earlier quarters.

We have about 30% flexibility in R&D and about 10% to 12% flexibility in SG&A which we can call basically within a quarter.

And we would be calling that, and we've said also that in the past, when we cannot sustain our sales levels.

Now, as Eric pointed out earlier, when we look at 2008, and even at modest IC unit growth, looking at the utilization rates of especially our foundry customers, relatively low inventory levels, the need for advanced technology transitions, we do not believe that 2008 is currently going to be a year where we will significantly call upon that flexibility.

Okay.

That's great.

Thanks very much.

The next question is from Mr.

Jay Deahna.

Please state your company name, followed by your question.

Hi.

JP Morgan.

Good afternoon.

Could you just expand a little bit on your thoughts on the foundries?

You said something about 65 nanometer utilization finally coming up.

And with seemingly every IDM in the U.S.

and Europe going [fab lighter], at some point you would expect that the foundries would have to do something for logic capacity expansion.

And then the second question is did your market share in i-Line and KrF go up or down in calendar '07?

And when do you expect your high throughput KrF tool, and please reiterate the model name of that if you could, start to ramp significantly?

Thank you.

Jay, regarding the foundries, yes, obviously they are becoming more important; SGI, Freescale, AST, NXP are now going to use them more and more often.

At this moment, however, as you noticed, we still haven't received significant orders in foundry and they only represent about 10% of our backlog or something and have been fairly slow for, I would probably guess, two years.

So, obviously, something has got to happen.

And you know that significant numbers of analysts now and colleagues in the industry are saying they are reaching a level of utilization and need for capacity which -- as I have mentioned as being an upside.

But they are managing their business very well and they will move only at the moment where they clearly see that the top line is moving.

And as long as they can still squeeze efficiency out of their fab, they will be doing and this is what they do.

Now, obviously, we've started to see some requests for short-term orders and that is being processed at the -- at current -- in the quarter, so we start seeing the first signs.

Regarding our market share in i-Line and KrF, I think in i-Line the numbers are big enough that I could say yes, we have increased the market share significantly and I guess we will be able to give you the numbers in, whatever, three weeks from now they will be public, I guess, so three or four weeks.

KrF, I'm not 100% sure because I don't have the data, but I think it has been improved.

And the machine called the XT1000, which is a high throughput and high [NA] machine, is, as planned, due for July, I guess it is, and that has not changed.

Does that answer your question, sir?

I think he dropped off.

Okay.

We go to -- the next question is from Timothy Arcuri.

Please state your company name, followed by your question.

Hi.

This is Srini calling for Timothy Arcuri from Citigroup.

My question is basically on the overcapacity that you have talked about last year.

You were saying that the overcapacity would end in the fourth quarter, but if you listen to the Samsung conference call they were saying that the overcapacity in DRAM might last till the first half of this year.

So have you done any adjustments to your model?

And when do you think overcapacity in DRAM and NAND will be alleviated?

Yes.

This is a very good question and which leads me to explain a bit more the terms that is being used in the industry.

I think we are very compatible with what Samsung says; it's just the terms.

In business of memory, DRAM or flash, in fact there is no overcapacity per se because every players having a fab of memory would run it 100%, and then would be able to find places to sell.

So, quote/unquote, it is not an overcapacity per se because they don't stay in inventories anywhere.

They are being dumped and the machines are being working on and people are developing those chips.

So there is never an overcapacity.

Now, what we all call overcapacity is whether you have too much capacity or not enough capacity at a given price level.

So, whenever somebody talks to you about this model, you would have to ask which price level are you talking about.

And when we put our simulation tool here in ASML, we put the price level of the beginning of the year.

So we say - imagine that the price level stay as it is; there is overcapacity or under capacity to it, which in other terms means if you have overcapacity that means you are going to have a price reduction.

If you have an under capacity, you are going to have a price hike.

So Samsung's message is very similar.

They basically think that the price will be remaining under pressure for another six months, which is their point and I think we agree with this.

And we also agree that from what we've seen a significant, in fact, under capacity or a possibility for price to go up in the second half, in particular in view of the technologies being moved.

I can't be much more detailed but there is a mix activity at this moment in DRAM which will favor value to volume.

Thank you.

I have just one other question.

When you said you have some percentage of uncertainty, could you quantify like the percentage of uncertainty in your bookings in the first half and perhaps the second half?

Oh, the percentage of uncertainty, it has two facets - one is mix and one is volume.

As I said to, I think, Nicolas at the beginning of the call, we are a bit unsure about which technology some customers are going to go for high volume on and some others, and this is still being discussed.

So there is a significant risk of the ASP average and the numbers of units.

We may come up with much more, let's say, trailing edge and less immersion or vice versa.

It doesn't change too much about the whole picture because the margin is not that significantly different.

So there is a part of the uncertainty is here.

The other part of the uncertainty is the unit but it is really based on the timing issue.

What my worry is all about is if the type of activities, as we see today in Wall Street and everybody discusses about whether there is a recession or not, etc., you can always -- in any business you can always decide to delay a decision by one, two or three months.

It's not going to impact anything.

So, [a PO] is very much delayed.

Now, in our industry we usually know these things - and I'm going to be very clear - even if we don't have a PO, there is definite working together with customers about preparing things because the lead times are so long.

So, yes, we are concerned about situations where a purchasing agent would say "Well, I haven't received the go ahead by the big guys.

We will receive in a very short lead time, and we know you guys have nine months lead time so we'll have to find a way to deal." So these are the types of activities that can happen and has regularly happened, by the way, in the foundry business.

And it's happening any time it is necessary, as it is today, where everybody says "Oh, the American consumers are going to buy more or less, etc." And obviously every foundry, who by the way have lost more and more the visibility of the customers because they are now doing business for TIs, etc., etc., who have also a set of customers.

You've got so much uncertainty in the line.

But as I say, at the end of the day it is only the need that counts.

And if you plug into your program that you would expect about 10% PCs growth, you are going to discover that you need a lot of lithography tools.

No matter when the booking comes in, we have to be ready.

And we don't intend to lose market share, if you know what I mean.

What Eric basically said is the -- there is a weak correlation between the time that we get an order and the time that we ship the tool.

And that has got to do with the fact that we anticipate the very narrow agreements we have with customers on that technology roadmap when they need certain tools.

And you could say "Well, is ASML then going to take a particular risk by ordering less?" Answer is yes.

So, because we know at a certain moment in time the customer actually needs that tool.

Now, is that risk going to translate into extra inventory?

No, because we have agreements with our supply base.

So, that particular correlation between the time we get an order times the standard lead time into a shipment that is not there.

Thank you very much.

The next question is from Harlan Sur.

Please state your company name, followed by your question.

Hi.

Morgan Stanley.

Thank you for taking my call.

Is it a fair comparison to think back to Q2 of last year, where you did see a pause in orders after a pretty strong memory spend in the first half of last year?

Is that a good way to think about the pocket of uncertainty you're seeing right now, here in the first half of this year?

Potentially, yes.

I think I would also like to add to that -- to your question because it's very relevant.

Think about Q2 of last year, which was a year of just over 30 orders -- sorry, it was a quarter of just over 30 orders, which if you add to the average six months lead time would say, well, then your fourth quarter's going to be 30.

Well, the fourth quarter was 55.

Yes?

So, I think that is to my point of what I just said; it is a weak correlation between when we get an order and when we know customers need the tools.

Yes.

Again, then, it is difficult for us and I don't know how much you trust the depth of the analysis we're making here.

The fundamental point for us is to understand the needs, the economics.

And today our message is bullish.

From whatever we hear, with or without this recession discussions and things, we haven't seen any customers or any statistics that tells us that we will not need those machines.

And we can correlate them with numbers.

It's not like we trust people.

We just go and count and see them.

And we had a discussion at the beginning of the call, where we discussed 200 millimeter fab retirement.

Well, these are big deals.

That means -- you know what that means.

That means you've got X fabs going out of business because they are not cost effective, not because you don't need the chips.

So, those chips have to be replaced by high technology machines, which come from us.

So, these types of activities are happening now.

So, if you add up those opportunities and you put this into a normal economic environment of about this 10% unit growth of semiconductor, which is not high, clearly not, then we show this growth potential.

But it is true the bookings -- it's a disaster for us to -- because it is, as Peter says, not correlated.

But if you try to make a trend on the bookings, you will be wrong.

There is no -- it doesn't represent the business trend.

Okay.

And then a follow-up question.

The team did a great job of bringing down cycle times for products in 2007.

Maybe you can just give us an update on the progress of bringing down the manufacturing cycle times for the immersion products?

What would they be now and where does the team expect them to be six to nine months from now?

At this moment, we improved cycle time of immersion machine by about 25% last year.

This year we probably will become -- we'll get towards under a quarter on the three months lead time.

I should not say this to a customer because they will [expect] orders now within the quarter.

I made a mistake; sorry.

The manufacturing time.

Yes, the manufacturing time, I'm sorry.

They -- of course, lens take always more time, but these lens we have them into fridges.

So -- and will we be able to go beyond a three months lead time?

Yes, but that will be -- manufacturing time.

Yes, that will be some time in 2009.

We have a new process of manufacturing which is being worked on at this moment, which gives opportunities for another, say, four weeks of lead time reduction.

Great.

Thank you for taking my question.

Yes, thanks.

The next question is from Jerome Ramel.

Please state your company name, followed by your question.

Yes, good afternoon.

I just would like to know what can be -- what is the split between the flash and the DRAM in your backlog right now.

If you have one -- if you have another question and I'll just --

Okay.

And the follow-up, actually, is I understand why Samsung is increasing CapEx but I have a hard time to understand why the rest of the industry should accelerate the migration to the 50 or whatever, because simply if you look at that cash situation now it's becoming very tough.

So, I can understand that the top three could continue the race, but for the others aren't you afraid that at one point it simply cannot cope with the price?

Well, first of all, the very good news, virtuous circle of this is set when you go and shrink, like this, you go to the next technology, in fact you build less chips at the end.

Either you put more integration to it or the machines are not as effective as the trailing edge in throughput.

So, you reduce, in fact, the numbers of units and you create more value.

So, at the end, just by the fact that there is not so many units you raise the price.

And the moment you raise the price, it happens that the trailing edge guys, the second tier, may get a bit of life by staying in the lower end of the business for a little longer, and that's probably what we expect, before they also jump to a leading edge.

So, if your point was do we believe that the second tier are going to run as fast as the other guys to shrink, I think the answer is no.

But they will benefit from the other guys ramping a year to two years ahead of them on the next generation, and that's very nice for us.

And then also, with regard to the question that was about leasing, I think our estimate is, in talk to the leasing companies, that the addressable market for leasing companies in the semiconductor space will double this year.

I think they will find other ways of financing.

I think we've seen clear evidence of that.

And there's a lot of activity in that space, especially when it comes to, like you say, very expensive tools like 1900.

So, you see a doubling of the leasing activities this year?

Yes, yes.

And one question -- I'm sorry, the answer to your question on the split of memory, it's about half DRAM, half flash.

Okay.

Thank you very much.

The next question is from Andrew Gardiner.

Please state your company name, followed by your question.

Thank you.

It's Andrew Gardiner from Lehman Brothers.

I have a question on the competitive environment.

You've highlighted in your presentation and the press release some of the gains you've been making recently in Japan.

I was just wondering if you could provide a bit more detail, given some of the noise that was coming out of that market late last year.

And it's quite clear that earlier on the 1900i had a clear advantage over your competition, but that was being drawn into question in certain camps around the end of last year.

I was just wondering whether, from your perspective, you think that's true and whether -- or whether feedback from your customers in that market are counter to that, and perhaps any further light you can shed on incremental gains in the Japanese market.

Thank you.

So, first of all, we have got two competitors, Canon and Nikon.

Canon has not yet introduced immersion machine and it is being prototyped and the prototype hasn't yet delivered a wafer.

So, clearly this is not the one you're talking about.

You're probably talking about Nikon, who has a valid machine.

That machine is based on a single stage platform, so it has between 30% and 40% throughput disadvantage compared to ours.

This is not new.

This is the definition of the two products.

And they are still on the lens which has an NA of 1.3.

Ours is 1.35.

And we have an overlay capability which is better than theirs.

So, our machine is still the only machine in the world capable of exposing below 45, say, and Nikon's machine is capable of doing above 45.

The -- that leads us to still being able to grow our market share.

First, because again we have capability they don't have.

And secondly, the cost of ownership is not bad, because our throughput is much better.

That translate -- and this is now difficult for me to be specific customer-wise, but what can I say?

We are growing, obviously, significantly our immersion market share on the customers we've already won.

I don't know how we -- but I'm sure you will have the official data soon from the different analysts.

So, you will see that this is clear.

And there is now one more customer who has been -- and I think it's now public, who has been interested to evaluate one of our 1900 and this is being -- this is in process.

So, one more customer is being at this very moment worked with, [very high weight].

On top of that, just to talk a bit about the 1900, the 1900 introduction has been the smoothest and the fastest introduction of new technology the Company has ever seen.

The 1900s are all built within the targeted manufacturing cycle time.

They're being installed in the targeted time.

They are producing wafers very fast after the sign up by the customer above 2,000 wafers, which is basically the result of all the cumulative learning we have had on the 1700.

So, as you see, the 1900 is a very successful introduction of an ASML product.

And also, if we look at the first six, I would even say nine months of the year, we cannot make enough 1900s.

We are shipping lower than the customer demand and that's simply because of the ramp up of the lenses.

So, I think, all in all, the 1900 has been a very successful introduction and it still is.

And I think it will reflect in our customer market share.

Thanks very much.

The next question is from Jim Fontanelli.

Please state your company name, followed by your question.

It's Jim Fontanelli at Arete Research.

Just circling back to the question of less correlation between orders and revenue, your guidance for near doubling immersion sales in '08 obviously implies that you'll be taking orders -- immersion orders over the next three to four months of maybe circa EUR700m to EUR800m, given an eight/nine month lead time in immersion currently.

Given all the provisos and the lack of visibility that you've been talking about so far on this call, what's the degree of confidence you have that that EUR700m to EUR800m is real and it's not going to fade away in part over the next quarter or so?

Yes, well, the biggest driver is - and I think Eric talked about that - is that our memory customers, largely memory customers of flash and DRAM but also, I think, to a lesser extent foundry and our IDM customers, they have a roadmap that in time differs quite significantly when they need the first new leading edge tool.

So, we have seen the most aggressive customers in that sense coming into the backlog over the last two quarters, but we haven't seen everything yet.

And that means that either those customers decide that they're going to give up on their technology roadmap and their shrink, or they're going to execute.

And they will execute on roadmaps, like I said, that in time will differ.

So, going back to the discussions that we have with those customers, we just don't walk in and we get an order.

Those are discussions that takes quarters, months.

Yes?

Just to make clear, when they need a specific tool to serve their roadmaps.

So, that gives us the confidence, those kind of discussions.

The inclusion of a 1900 in their roadmap at that particular point in time gives us the confidence to make that statement that we think we will close to double the immersion sales in 2008.

Yes, so in other terms, if I put a slight amount of details more, in the second half of the year there is going to be, as usual, it has always happened, the tail end of the first tier, because they buy a lot in the first half and they always forget one, two, three, four, five or so.

Then you're going to have the second tier.

But I should -- the second tier and then a third tier.

And so, there is a complete new ramp of, call it, second tier customers, but they will be annoyed if I made that so clear.

So, it's not the first, the top three.

The most aggressive.

It's the next ones.

That will come up in Q3.

And then, in addition to that, you're going to have the immersion starting from -- for the Logic guys, because again Logic hasn't started, by the way, in production.

So -- and that, we have some orders.

So, yes, there is new activities coming up in this environment and I didn't want to confuse anybody.

The immersion for the third tier DRAM is not even a 2008 story.

So, I'm sorry to confuse people with second tier, third tier, etc., but I don't want to put names to it.

There is clearly a necessity for what I call the second tier DRAM makers to be there, because if not they would absolutely leave it to the top two or three.

And it's also in that space where we see a lot of the leasing proposals.

Okay.

Okay, that's pretty clear.

And just to follow up on the DRAM front, I guess it's all the sub-60 nanometer capacity that's coming on this year that's going to drive your specific DRAM immersion demand.

Do you have an idea of what proportion of global DRAM capacity you think is going to be sub-60 this year?

In fact, I think we can send you that, but unfortunately I don't have it in front of me.

That's again something we can forward to you, Franki, Craig, so.

Yes, it is.

We have the data on the node percentage.

Great.

Thank you.

We'll come back to you.

Thank you.

We'll send it on, but we don't have it here.

Yes.

But you can call Frankie and Craig.

Will do.

The next question is from Mark FitzGerald.

Please state your company name, followed by your question.

Great.

Good morning.

Thank you.

If you look at the DRAM market as being probably the highest risk at this point, can you give us an idea of what you're thinking about the DRAM business is going to adopt immersion in 2008 in terms of units?

And then how many of those units go into the top three customers?

First of all, I think, yes, like you we thought that it was a risk, because it's so -- it's a very large business and they have such a price problem that at some point everybody was worried that they would disappear or stop, or try to reassess.

But today, clearly, they have come up with this solution, which is ramping much faster their 5X or 5X equivalent.

So, this is where I would say the enthusiasm is.

They are building up this plan where immersion is getting them out of a price problem.

And they have discovered also by doing this they -- only the big guys can be first.

So, there is significant hope for them.

So, you see this as a risk.

In fact, we saw this as a fairly big, I would say, dynamic that saved everything.

Now, in terms of units that that would create for next year, would we be wrong by saying about half?

So, if we said the business of immersion is about in the EUR1.9b-ish, as Nicolas said, and EUR1.9b to EUR2b, EUR1b of it may be DRAM.

Would that be the equivalent of 25 systems or so?

Divided by 30-ish, but there may be some 1700.

No, it's 30.

30 systems.

And the top three customers dominant in that mix?

That I don't think I can give names, can I, or -- no.

I'm not looking for names.

Of the top three customers.

Oh, the top three customers.

Don't want to give their names.

They're heavily present in that mix, yes.

Yes, of course, yes.

So, does than mean something like 70%, 80% of the mix?

DRAM customers, you mean, on the 30 units?

Yes.

Of the top -- of the 30 units, how much of that would be the top three?

Yes, I understand what he said.

Yes.

About three-quarters.

It's big.

We are hesitating because we are ramping somebody new, but to be honest it's difficult.

But let's not try to answer your question, because we don't understand.

So, you may want to have a little technical discussion with Craig or Franki about exactly what you're asking.

We may give you a wrong impression.

But clearly, our top three customers are significant at this moment in time in the order backlog, our top three memory customers.

So, it must also be for the immersion tools.

I don't have the exact data here, but they are significant.

If you want an exact percentage, I can give you that, but they are very significant.

Thank you.

The next question is from Weston Twigg.

Please state your company name, followed by your question.

Hi.

Wes Twigg, Pacific Crest.

Just wondering if maybe you could help explain the jump in service revenue and maybe tell us whether or not that number should carry forward.

Yes, the jump was largely in the field options.

Like I said, that's a lumpy business.

Sometimes we have big upgrades of fabs, where we do productivity enhancement packages for the whole fab and then you suddenly have a big jump.

So, on average, the service revenue is about EUR100m, but it can vary between EUR85m and EUR110m, EUR120m.

So, it's really when we have these bigger projects which are focused on system enhancements and system upgrades, and have happened in the fourth quarter.

It's difficult to predict.

Those are really orders that come in and that we need to deliver within the quarter.

Okay.

You don't necessarily expect that number to --?

No, [it doesn't] be at that level of EUR120m.

No, this on average is about EUR100m.

Okay, great.

Thanks.

The next question is from Dan Ping.

Please state your company name, followed by your question.

Dan disappeared.

The next question is from Sandeep Deshpande.

Please state your company name, followed by your question.

I'm from JP Morgan.

Just a quick question on -- there has been quite a lot of discussion on this, but in terms of your [SIC], you've talked about tier one, tier two, tier three DRAM customers, but there are tier two NAND customers as well.

Have all the tier two NAND customers placed orders already, or are we all completely relying on DRAM to be the next wave in the immersion orders?

No, the smaller NAND haven't.

So you do expect that some of the smaller NAND customers will be placing 1900i orders at some point this year?

Absolutely.

Correct.

And then, finally, regarding your options program, we've talked about this in the past.

We talked a lot about the possibilities.

Do these options still exist and could they be exercised at short notice, the second quarter, for instance?

Can I give you the truth, Sandeep?

Yes.

In fact, we're doing this with a - how do you say - a commercial system.

But in addition, we have to do it also with, I would say, aggressive lead time programs, which sometimes are not options.

So, we have two sets of programs in place, i.e.

a commercial program that says you accept to 100% guarantee that we will be ready to ship within two months.

And then there is another program which is you accept to be getting those units but your probability is not 100%.

And those numbers are limited, the number of --

One is cheaper than the other and to be honest the one which is popular at this moment, extremely popular, is the cheap one.

But -- so it puts us into a difficult logistics situations, but we shall see.

But as an answer to your question, very clearly, yes.

That is still out there, so those options have been called.

And, like Eric says, we have the second program for a limited number, where customers try to get ultra short lead time but we can do it only for a very, very few number of tools.

And both are there and as a matter of fact we have seen, like Eric said earlier, especially in the foundry area, in the leading 65 nanometer capable tools, we have seen those spur demands coming over the last week or so.

Thank you.

Ladies and gentlemen, I think we have time for one last call.

If you're unable to get through, the IR department will be here for a while, so feel free to give us a call in Veldhoven.

You can find that number on our website.

Now, operator, if we can have the last question.

Your last question is from Janardan Menon.

Please state your company name, followed by your question.

Hi.

It's Janardan from Dresdner Kleinwort.

I just want to go back to the immersion question asked earlier.

When you say you're getting about EUR1.9b or so of revenue in the year, you're talking about 62s or thereabouts a little bit higher maybe, at about EUR30m a piece.

You have 34 in your backlog.

So, does not that guidance imply that you have some visibility on taking in orders of about 10, 11 a quarter over the next two quarters, or is that visibility not really there?

It will be natural to do so and we are working on those.

That's why we are declaring that we can do these tools.

But, as I said -- in fact, I'm going to give you a story.

We booked, I think, three 1900s on the last week of December for delivery in April.

So, that gives you a feeling about things.

So, I -- what you say is, in fact, the logical answer with a bit of a stretched lead time, but that's -- I think that will be okay.

But it is not obvious.

The three deals on the 1900 were, in fact, a significant activity for us to discuss, I would say, terms for good reason, so there was no giveaway or whatever, but it took a bit of time.

But obviously the machines were already started.

Thanks.

And just a last follow-up, if I can, sir.

On your Q1 orders, can I assume that if things were good or things are bad, let us say, as different order levels will go, that in your mind the biggest delta between what is good or bad is whether the foundries will order or not?

Or is it -- can that delta come from somebody else as well?

To be honest, at this moment, we do not need the foundries to have a robust year.

We may need the foundry to have a growth year with -- but we haven't put into our scenario, in order to give you an optimistic view of the year, a surprise booking by the foundries.

In Q1.

In Q1 or in Q2 even, but not -- just a very small amount, as they have been doing.

So, we're not dreaming.

In fact, I think it is a good conclusion.

The message we try to give to everybody is we're not dreaming.

We know that there is a potential recession or etc.

out there.

We are working our CapEx and OpEx accordingly, obviously.

But we also understand that there are economics that requires our customers to ramp their production a certain way, in particular in order for them to cost save and etc.

or introduce new products.

So, we have to do them.

We know about the level of units that they want and we are optimistic about it.

But we're not saying, oh, great, we are going to have all of them come in and the Taiwanese are going to have a lot of money, and etc.

And, therefore, thanks to all this, we are going to have a year 2008.

In fact, we do not at all simulate any of that.

We simulate on a very normal behavior by the very large players, who are doing two things - technology transitions and retirement of some 200 millimeter fabs.

And on this notion, we finish the call with a fairly optimistic message of 2008.

Okay.

Thank you very much.

Well, I'd like, on behalf of ASML, to thank everybody for calling or listening in and we look forward to talking to you next quarter.

And, operator, you can go ahead and close out the call for us.

Thank you, sir.

Ladies and gentlemen, this concludes the ASML 2007 fourth quarter and annual results conference call.

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

Thank you.