艾司摩爾 (ASML) 2009 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by and welcome to the ASML 2009 fourth quarter and annual results conference call on January 20, 2010.

Throughout today's introduction, all participants will be in a listen-only mode and after ASML's introduction there will be an opportunity to ask questions.

(Operator instructions).

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

Craig DeYoung.

Please go ahead, sir.

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

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

I would like to welcome you to our investor call and webcast.

As the operator mentioned, the subject of today's call is ASML's 2009 fourth quarter and annual results.

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

Eric Meurice, ASML's CEO; and Mr.

Peter Wennink, ASML's CFO.

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

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

Let me remind you or advise you that the length of this call will be 60 minutes.

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

Meurice for a brief introduction.

Eric?

Thank you, Craig.

Good afternoon and good morning.

Thank you for attending our Q4 2009 results conference call.

Peter and I would like to provide an overview and some commentary in the quarter.

And as usual, Peter will start with a review of the Q4 financial performance with added comment on the short-term outlook.

I will complete the introduction with some further details on our current situation as we close out 2009 and head into 2010.

Following the introduction, we will open up for questions.

So Peter, please?

Thank you, Eric.

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

To start with that, our fourth-quarter sales were EUR581 million.

This was slightly higher than our sales guidance for the quarter.

We shipped a total of 25 systems, 19 of which were new.

New system ASP was EUR19.7 million versus EUR23.4 million in the third quarter, which indicates a substantial blend of immersion and non-immersion tools.

We shipped 11 immersion tools and 14 non-immersion tools, reflecting an increased demand from our foundry customers for capacity-related tools.

The average selling price of new and used systems combined was EUR17.3 million.

That was, therefore, also lower than in the third quarter.

Our service and field option sales were high at close to EUR150 million.

This was primarily driven by productivity upgrades.

Such upgrades are not unusual as customers look for incremental capacity when a rise in their product demand suddenly becomes evident.

And, the company fourth-quarter gross margin was 38% versus 34.4% in the third quarter.

This reflects an improved coverage of our fixed manufacturing cost, cost reductions and stronger service and field option sales, latter generally carrying higher gross margins.

We have continued our focus on R&D with a spend of EUR115 million in the fourth quarter, which was, by the way, the same as our spend in Q3, and SG&A came in at EUR37 million, which is about the same as what we had in Q3.

As a result of the aforementioned, operating income improved from 7% in the third quarter to around 12% in Q4.

And last quarter of the year, we generated EUR19 million in net cash.

With that, we were able to close out 2009 with a cash balance comfortably over EUR1 billion.

Q4 net bookings were strong at 40 systems, value at EUR956 million.

Net immersion bookings were 25 systems; 23 were new and two were used.

This demonstrates a continued focus on technology investments, driven again by the DRAM and foundry market segments with Flash investments just beginning to emerge.

In addition, we sold 12 new KrF bookings, indicating a continued need for more non-critical systems in support of the leading edge production ramp of our foundry customers.

Fourth-quarter bookings pushed our backlog to EUR1.85 billion, containing 69 systems at year end, that being the highest level of the backlog in 10 quarters.

1950.

As for the outlook, we continue to increase factory production capacity in order to meet the growing demand as indicated by our fourth-quarter and the anticipated first-quarter bookings.

However, due to the fact that the normal production lead time for our more complex and new leading-edge products is still fairly long, we will not be able to completely satisfy all demand in the desired delivery time.

As a result, some first-quarter delivery requests will be shipped early Q2 when our ramp-up capability will have accelerated.

And as a result, we expect the first-quarter sales to be about EUR700 million with the second quarter coming in around EUR950 million.

Gross margin in the first quarter is expected to be about 40%.

R&D expenses are expected to be EUR120 million and SG&A EUR40 million.

Furthermore, we expect orders in the first quarter to be at approximately the same level as the fourth quarter of 2009, mainly driven by memory customers and foundry customers, again.

And, finally, in respect of 2009, ASML will submit a proposal to the 2010 general meeting of shareholders to declare an unchanged dividend of EUR0.20 per share.

And now I'd like to turn it back over to Eric, who will give us some more views on the upcoming quarters.

Thank you, Peter.

So we closed the year, as planned, with recovering sales and strong Q4 systems booking of EUR956 million, confirming the semiconductor segment upturn.

This upturn is currently mainly driven by technology buys from the DRAM sector, by a fairly modest technology set of buys from Flash sector and capacity buys by the first-tier foundry at 40- and 45-nanometer node.

We believe that we will see a continued high level of bookings for the beginning of the year, so the numbers of engines of growth are contributing to this.

First, the buildup of 40-nanometer DRAM capacity will continue well into H2 2010 as the second batch of DRAM players will start ramping then.

Second, the long-awaited 30-nanometer Flash ramp and 20-nanometer development Flash will also start contributing more heavily in Q3 2010.

Third, we expect structural foundry built in H2, including the first-tier, but also the second-tier customers.

With our current 2010 shipment run rate, we have calculated that the litho installed base will support semiconductor unit capacity increase of only 12% in 2010 versus 2009.

Upside to this semiconductor unit growth number, as discussed by most industry analysts at this very time, could justify improved lithography shipment run rates during the year.

In other terms, the structural capacity catch-up that we are seeing today, the technology buys and the potential capacity gap to demand growth support, at minimum, the current litho shipment run rate while certainly not contributing any surplus capacity.

With this demand ramp as a background, we are continuing to focus our product development on product development and new product introductions.

We have now converted all our mid-range TWINSCANs to an upgraded platform with a higher throughput, better overlay and better CD uniformity.

We are also ramping our new NXT platform with a state-of-the-art performance.

We are deploying our holistic lithography product suite, and we continue development of the EUV technology.

Regarding the NXT, our new platform, we now have installed base of five systems and a backlog of 17 systems, 11 in the memory sector, five in the IDM sector and one in the foundry sector.

Customers will be ramping up these -- these machines, the NXTs, into volume production in Q2 2010.

Our holistic lithography product suite, including the customer accepted software solutions producing higher process yields, is being further enhanced by FlexRay, our flexible illumination system, and our YieldStar standalone and in-line metrology systems.

This hardware, when married with software of the suite, provides even better scanner performance and greater factory process control.

Regarding EUV technology, we are making good system integration progress.

That, along with a continued source performance development, allows us to confirm our ability to ship the first preproduction systems in the second half of 2010, systems for which we plan to close the [sixth] order very soon.

We believe that the renewed industry interest for product and technology transitions that we are seeing today in conjunction with ASML's stronger position mainly due to improved cost structure but also differentiated and diverse products has positioned ASML very well for its next significant growth opportunity.

So on this summary, I would like to leave it to you to ask Peter and I questions.

Thanks, Eric.

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

But beforehand, I'd like to ask you to kindly limit your question to one, one question with one short follow-up, if necessary and if possible.

This will allow us to get as many callers in as we can.

Operator, could we have your instructions, and then the first question, please?

(Operator instructions) Didier Scemama.

Good afternoon, it's RBS, thanks for taking my question.

I'd just like to ask you a question on the Flash memory market and, more specifically, on SSD.

I'd just like to have your opinion on basically the impact of that technology on your revenues.

If I'm correct, the adoption of SSDs is quite modest at this point for netbooks and notebooks, probably in the low single-digit levels.

So what I'm wondering is, if SSD's adoption in notebooks and netbooks were to reach something like 10%, what would be the impact, in your view, on the demand for lithography equipment?

And related to that, what sort of technology would be required, when you talk to your customers, to get to the sort of inflection point on pricing to really drive adoption of SSDs?

A very good question, because SSD is a very large driver.

Clearly, there has been a bit of a disappointment in some ways that SSD has remained in the areas of, I would say, 3%, 4%, 5% mount rate in the PC world.

To get to about 10% you would have to have about EUR1 billion to EUR1.5 billion of lithography equipment.

The customers, I think, have started to indicate to us that the mount rate is probably going to increase, due to two things.

One, they are improving the drivers, the software drivers, which I think is more or less related to Windows 7.

I understand that it is not yet perfect, but I understand it's improving and therefore will increase the speed of adoption.

In addition, it is true that there is a sort of new agreement that SSD cost target will be so much more supported by 20 nanometer, which is the current development being put in place by the first year.

So in other terms, SSD didn't really contribute anything at this moment into billings and into bookings, and we expect this to be one of the upside opportunities of the second half of 2010 and into -- significantly into 2011 in 2012.

And in your view, when Google announced that their Chrome operating system would be only supported by SSD, do you think that that could kick-start the market?

Or do you think that it's a bit further out?

No; it is certainly another obvious position that says the world will probably get beyond 10%.

I think that's the message that the whole industry is expecting.

The key question is an economic question, really, is exactly when do you have an inflection point and an acceleration point.

And to that, I'm not a specialist of the industry, so I cannot really answer to you.

But, clearly, we expect to go beyond 10%, and we look forward to, say, 2% to 3% mount rate on that, correspond to about EUR1 billion of lithography.

So we feel this is a significant driver of the business, and we feel good that it will last next five years.

Gareth Jenkins.

Gareth Jenkins from UBS.

Just a quick one on service revenues, very strong in Q4.

And I just wondered what the sustainability of this is; and obviously, related to that, whether the outlook for gross margins beyond Q1, given the service sales.

Yes; the service sales in Q4 were particularly strong, which is not uncommon, like I said also in the opening comments.

Customers have underestimated the capacity ramp; they are looking for upgrades which they have not done in downturns, and especially productivity upgrades are then very popular, which we can do relatively quickly.

So in the first quarter, just to take away some of the output pressure, this is a very popular product which actually brought the revenues to about EUR150 million.

Those revenues throughout the year, the first three quarters, were less than EUR100 million.

Normally, the normal run rate would be anywhere between those, so between EUR90 million and EUR115 million, around EUR115 million, EUR120 million, EUR125 million; that would be a normal run rate.

Thanks, just follow-up on that then.

The gross margin outlook post Q1, you had very solid gross margins in Q1.

I just wondered if you could give us a sense on how we trend through the year.

Well, we talked about it also on previous calls.

The gross margin is a function of, of course, our ability to cover our fixed costs.

So the order intake that we saw in the fourth quarter and we guide for the first quarter is a good indicator that fixed cost coverage will not be our major issue going forward; as a matter of fact, it will contribute to the gross margin.

Also, the cost reductions that are coming through, which were the result of the cost programs that we had last year, will come through.

So the guidance that we gave you, 40% on EUR700 million, clearly gives you upside when we move to EUR950 million.

How much that will be, will be more specific at the end of Q1.

But clearly, the movement is in the right direction, and the main drivers I just mentioned.

Gunnar Plagge.

Yes, hello, it's Nomura.

Could you talk a little bit about the market in Japan?

It looks as if, the last three quarters, you haven't had a lot of orders, or almost no orders.

Revenues last four quarters were also basically zero.

Then you still have EUR230 million of backlog, which I think is mainly Toshiba.

So maybe, first, with regard to Toshiba, are we expecting a shipment here soon?

And could you maybe also qualify whether this is likely XT or NXT?

And then, with regard to Elpida, an update would also be very helpful.

Thank you.

Well, as you know, it is impossible for us to be specific on customers, though I will try to [rush] an answer that makes sense for you and your modeling.

Yes, Japan is clearly, at this moment, a DRAM vendor, a Flash vendor and a very, very small market.

For the rest, as you know, the business of logic is not booming at all, at this moment, in Japan.

So in the DRAM sector there is definitively a split business between Japan and Taiwan in terms of Elpida, and we are happy to have won their confidence.

I think that is public.

And as you know, they have announced themselves that they are ramping their production as of Q1.

So this production is going to be split between Taiwan and Japan.

The Flash environment in general, as we said, I would (inaudible) disappointed, but slow.

And this is because it is true that the whole industry have built up significant overcapacity, which took 2009 to be digested.

And it is highly possible that Toshiba, in fact, it is clearly the case of Toshiba.

So we have also said that the Flash business is going to come back with a revenge, driven by the applications we talked about plus additional needs in mobile phones and additional needs in service.

And this is starting to build up momentum as of Q2, and that will increase our business in Japan.

Generally I would add to that, that looking at these regional sales, with the importance of, let's say, global, I would say, alliances, especially in the memory world, where technology choices are actually made but could be executed in different countries than the what I would call real originating technology, could come from country A and could be, then, made in country B.

So you have to be careful with just looking at the regional split because those designs could actually be made in these, let's say, joint corporation structures in other places of the world.

And that's definitely the case for the region that you just mentioned.

Okay, thank you.

And then on the supply constraints, it seems that you are now at about 1100 flexible labor.

I think you were at 1700 [end] 2008.

So my question really is the supply bottleneck.

Is that more within ASML, or is it more within the supply chain?

No; it is definitively the supply chain, although we thank our suppliers for their effort because, in reality, the constraint is because we have converted all our products to the new products.

The mid-range used to be called XT 3 versions; this is internal code.

It is now an XT 4 version, the new machine called NXT.

So you are talking about having to rebuild from scratch, from zero, a complete supply chain, which usually would take between one year or nine months, at least, to be set up including -- inclusive their own CapEx, etc., etc.

So this is the main constraint.

Our business is one of integration where testing is more important.

And yes, it is easier for us to build up temporary workers or using some of our engineers to do more.

So we definitively are far from being the limits.

The limit is because we are moving up all those new products, and it takes time to build up the pipeline.

Satya Kumar.

Satya Kumar from Credit Suisse.

A quick question on your orders from memory.

I presume, obviously, most of these orders are coming from DRAM.

How much of these orders are for technology shrinks versus capacity?

And, at this point, do you have a sense of what portion of the installed base in DRAM has already placed orders to shrink that's in your backlog in Q4?

Well, the first question is easier to answer.

In the DRAM bookings that we got, 100% is technology transition.

In fact, this is one of the things we've discovered.

It seems that now the customers have this capability of using all the installed base in what we call the non-critical layers of the new technology, and therefore the only thing they need is the critical layers, and that's why they only buy the latest machine, which, obviously -- immersion.

So in DRAM, clearly, 100% is in immersion, the critical layers and the noncritical are obviously the installed base.

Regarding the percentages of what remains to be booked by the guys who have already started, I would say we are at 50% of the first tier, and we are low on the second tier.

Okay.

If I look at your -- I mean, continuation on this DRAM theme again.

If I look at your 2009 shipments to memory, in aggregate, that was less than a fourth of your shipments to the sector in 2007 peak year.

Obviously, DRAM supply growth was constrained, and clearly, you are guiding below consensus for Q1.

One of your customers, Nanya, this morning actually lowered their supply growth projections for 2010 because of the lack of equipment availability.

Supposing we get another 40% or 50% demand growth for DRAM in 2011, would we need to start to see capacity orders from DRAM to meet that demand because you are sort of using up all the shrink spend this year?

Yes.

In fact, first of all, it is very difficult to explain why would a customer be reducing or increasing its output.

It may be because it's litho-driven, but it may be because there's other reasons for that.

But clearly, at this moment, everybody is trying to ramp as fast as possible.

In my opinion, I think the biggest issue that they all have is that they also have to ramp new technologies.

And process is not only CapEx, it's -- we do not know about the yield, we do not know about the complexity of making it work.

So you are going to have, for a period of six, nine months, tons of difficulties in this segment of ramping to demand.

So that, in turn, will create potential need for additional bookings because some of the guys who will have more difficulty to hit the customer needs will, the customers will go for their competitors.

So you are going to potentially have, in fact, an acceleration of bookings on some areas to try to win the places where you cannot deliver.

Our own constraint is, we are -- yes, we are indeed putting constraints on our customer at this moment because you start from scratch.

And at the moment in this industry where the customers come and say, I need to have delivery in three months, and we all know the theoretical lead times are nine.

So there is obviously some haggling about who is right, wrong, etc., and the terms were creating some constraint.

But at this moment I would not consider these constraints are life-threatening to anybody, nor stopping you to buy any nice end products.

I guess my question was, in 2011 -- I understand the 2010 technology shrink part of it, but in 2011, if we again get demand growth in DRAM -- (multiple speakers).

Oh, yes, yes, yes.

Sorry, yes, I missed it (multiple speakers).

The node today in DRAM is 40 nano, and in 2011 there is going to be another node.

In fact, we are at this moment reviewing the requirements for the new nodes.

And it seems that the DRAM industry is putting itself on a one-year type lead time for new nodes from an original 18 months.

So there is something, at this moment, which is happening for 2011, a new set of technology.

It is not clear yet whether that shrink will better be satisfied the bit growth demand in 2011.

If that's not the case, we need new fabs.

That's what you were looking for.

That's unclear right now.

That really depends on the bit growth numbers that we should try to find out how much that could be for 2011.

Now, those numbers have been going up.

The industry analysts have, over the last two or three months, have actually upped their forecast for bit growth.

And we are currently trying to figure out what that would mean in terms of real additional capacity in terms of new fabs.

That's not clear yet.

Sandeep Deshpande.

I just had a question on the competitive position of the NXT versus Nikon, mainly because of a lot of noise in the market associated with some customers trying out the new Nikon tool, etc.; how you see your positioning in terms of what you've shipped, as well as what you have in terms of orders from the major customers.

And I have one follow-up.

So definitively, our competitor is a very valid competitor working on value generation.

So our competition is clearly on the capabilities, on the performance of the machine.

At this moment both of us are running our machines, our self NXT, and Nikon the 620, to try to meet the customers' requirements.

We think that with five installed base and a backlog of 17 with certain numbers of committed production, we have proven to the customer that we've reached the level of performance that fits their needs.

We understand that our competitor will continue improving and improving and improving their machine.

This looks very, very similar, to us, to the competition that started two years ago under their 610 and our XT.

At the time, we were fighting or targeting the same specification.

Today, it is about the same situation; the specification is basically the customers' specification, so we have to try to make it.

It will have to be proven who can make it, which are the products which are, at the end of the day, have more extendibility, more capability and better cost of ownership.

We feel comfortable about that race because we are still using a TWINSCAN concept which allows a machine to be measuring its wafers while it's exposing, which allows, in fact, you to decide how much you want to measure, which allows you to decide to, in fact, obtain an extremely good performance of chip resolution.

Our competitor is using, still, a serial method by which you measure and then you expose.

By doing this, it is logical to understand that if you measure a lot to achieve very good exposure performance, then it takes time, and then you wait, the lens waits until you can expose.

So it has a lower throughput and potentially higher cost of ownership.

So we are comfortable that our concept is still different than our competitor's and allow us to think that what we achieved when we introduced TWINSCAN in 2000 will be projected in the next decade because we are still using a differentiated architecture.

Thanks, Eric.

Just one question for Peter, on the margin.

In a previous call you've talked about hitting, going towards the 30% margin at the peak.

How would we be looking at gross margin then?

Given EUR950 million in order intake, approximately, in the fourth quarter, by the fourth quarter of this year, possibly you could have revenues north of EUR1 billion, EUR1.05 billion or so.

So if you did your numbers sitting today, what would we be looking at gross margins at that sort of revenue level?

Well, I think clearly you understand that I'm not going to give you any guidance on the fourth-quarter gross margins.

But to give you an indication, we said drivers for the gross margin development were adjustments in the cost structure, as a better product mix, and especially focused on those products that will enhance the value of our tools, i.e., the holistic litho products, which we are making progress with also in terms of shipments and ordering.

That will find its way partially through, let's say, field service and sales, but also in the value that we can contribute to the tool.

Those are elements next to the higher volume which will give us a better load of the fixed cost, of which we said also last year that our target is to go -- if we start moving towards our focus on EUR5 billion, that we won't have a target of 45% gross margin.

Now, clearly, we are not going to be there from one quarter to the other.

There's going to be a trajectory that's going to take a bit of time.

But when you talk about sales levels, you just indicated clearly the gross margin will be, definitely, significantly over 40%.

And it would be heading towards our target.

Whether we will be there in the fourth quarter, it is too early to say.

But that's definitely -- you will know what our target is, and we should be heading in that right direction.

What that would mean for the operating margin you were referring to -- clearly, we said also in previous calls that that could be heading towards the 30% mark.

And that is easy if you take, let's say, the EUR5 billion at 45% and you subtract our R&D costs and our SG&A costs, you can make that calculation easily.

So, yes; clearly, the gross margin and the operating margin trajectory is quite clear to us, and higher levels of sales to the level that you just mentioned will definitely help.

Tim Arcuri.

Citigroup; two things.

First of all, If I take your June quarter shipments that you're actually guiding to and I assume that litho is going to be roughly 25% of wafer fab equipment, which is up fairly sizably versus what it's been the last couple of years, so I'm sort of building in some share gain there for litho, it implies that the industry is going to spend about EUR50 billion.

So that's sort of at a EUR50 billion CapEx run rate, which is odd because all the companies -- if you add up what all the companies are guiding CapEx to, it's more like EUR35 billion or EUR37 billion.

So I'm wondering, is it that people are forward-buying lithography, or is it that the companies just aren't coming clean with how much money that they are going to spend this year?

And then I had a follow-up; thanks.

Will definitively, I think, historically litho had been forward booked because lead time is enormous on our side.

And secondly, it takes also very significant time to qualify your litho tools.

So the wafer's out, it takes a long time.

So yes, there is a timing issue.

Secondly, we cannot really comment on your 25% share of litho to the total CapEx, in particular on the quarter.

You are now talking about an environment where the customers have been able to only buy equipments which help them on critical layers and using their installed base for noncritical.

Even for litho, it's kind of new; we used to have a more balanced situation by which a customer would come and say, I've got a new fab.

To make the fab work, I need 3 -- I think the normal line would be 3 or 4 immersions, 5 to 10 KrF and 5 to 10 i-Line.

And we would take an order for that.

That was the original way of doing business, about two to three years ago.

We haven't done that for two to three years; we are only shipping the leading edge.

So it may be that your ratios have to be refilled completely in view of this new model.

I can't really help you on this one because you are now having to talk to Applied Materials, KLA, etc.

But for us, we are certain that the guidance that we've given correspond to spots in those fabs.

And if we were not to ship, the customers will definitively not be able to build even the same amount of wafers of the new product than they even had in 2009, which would be impossible.

So in other terms, you can question whether the run rate will last more than four quarters, but you cannot question that the run rate will not last for two quarters.

Secondly, I think you should look at another ratio if you wanted to do macroeconomic ratio.

You should take the diffusion rate, litho CapEx divided by sales of semiconductors.

And you will see an interesting curve.

You will see that the curve was hovering around 3% for a long time, and then it went up in 2007, when the DRAM business got crazy, and it went from 3% to, say, 3.2% or 3.3%.

And then, since 2007, that ratio is down.

Last year it was 1.5%, and this year it may be less than 2%, or at 2% if some of the analysts are correct.

So you are talking about a significant number, yes, EUR950 million in the quarter.

But if you divide this by the total expectation of silicon revenue, you are going to be surprised that we are much lower than historical [trends].

I guess the gap is just so significant that it just sort of indicates that maybe there's going to be a fall-off in bookings later on this year.

But as a follow-up, it seems like the weaker shipments in March came out of nowhere a bit.

And I'm a little surprised because, if it's related to the technology platform, it seems like you would have known about that before.

So I'm wondering whether there was something that occurred between your meetings with folks during the quarter and now that sort of came out of nowhere that led you to be more shipment constrained in March.

Well, yes and no.

The discussions for the fourth quarter with customers have been, do you want XTs, or do you want NXTs?

When you usually come up with this question, the customers always want NXTs.

And then, of course, then you get yourself into a mode that says, well, an NXT has a nine-month lead time, NXT has a six-month lead time.

And then you have a compromise discussion because, of course, a customer wants an NXT because it has, supposedly, a longer lead time -- sorry, a longer lifetime, with more extendability.

So the customer is ready to pay more money, etc.

So you've got that debate.

And to be honest, it lasts until the last moment, usually, in particular because they are excited about the NXT, so they want to wait until the last moment to switch between XTs and NXT.

So what has happened, if you will, is that we were expecting or we could have expected some customers to be more, I would say, realistic and ask for XT shipment in the Q1 time frame.

But in reality, we could not deliver the NXT.

And they asked us to wait -- or, not -- well, they didn't ask us to wait.

But they basically agreed to that they had to accept that certain of NXTs would not ship in March, but would ship in April.

So now, remember, an NXT cost EUR40 million.

So you can imagine how many NXTs it makes to have this slight change of expectations, Q1 to Q2.

So that's basically what happened.

I think it is nothing special regarding the trends of business.

It is just the timing, plus or minus one month, on four or five machines.

Simon Schafer.

Yes, thanks so much, it's with Goldman Sachs.

I wanted to revisit your theme about being able to support 12% unit growth in the broader industry, and maybe ask a question that was sort of asked before.

What sort of DRAM bit growth do you think your installed base can support today?

In other words, how much of an incremental upgrade path do you think people really need to get to some sort of bit growth target that is in line with the historical averages, if you look at DRAM bit growth coming out of a downturn, historically?

So we think that our 12% DRAM, if I'm not mistaken, corresponds to about 50% bit.

48%.

48% bit growth in DRAM.

And that's about it.

And in there you have also NAND, which would correspond to a 69% bit growth.

So when you talk about, now, the excitement about Windows 7, you will hear some analysts talk about more than that now.

And I'm trying to find out the latest analysts -- in fact, I don't have it with me.

So you may have to ask them.

But there is now more interest or more excitement about those numbers.

Thank you.

And my follow-up question would be, there seems to be some sort of discussion about when or if your own order book perhaps may be peaking at around EUR1 billion or so.

But I was just wondering, in your view, given that capacity adds, as you just discussed, in DRAM at this point are still relatively low, the same as seems to be the case in foundry and then NAND, if anything is still relatively dormant.

So I was just wondering what that means from your point of view in terms of the potential order peak then the cycle, because historically speaking, of course, your quarterly order peak was not too far from current levels.

Well, this is, in fact, why today, for the first time in, in fact for me, for five years, we have guided six months.

And we are also guiding the possibility of upside.

So that's a way for us to say, we believe that there is enough engines of opportunities of growth.

And what you mentioned about Flash not being there already but should be there, DRAM bit growth is not that big at this moment, etc., etc., are as much opportunities to sustain this type of run rate that we are talking about.

So let me try to summarize it differently.

If you -- and it's not us to answer; it's an economist to answer.

If you believe that we are back into a re-build up of no more semiconductor cycle, so you look at your curve with semiconductor and you were down now, and you are putting yourself back into an historical curve.

If you do that, clearly, we are talking about a sustained growth profile for 2010, 2011, 2012.

And I do not know if we are going to hover plus or minus EUR1 billion or whatever, but you are talking about a very healthy opportunity, which is why we have continued talking about our corporate target to reach EUR5 billion when -- in fact, during this next cycle.

Of course, if an analyst would tell you, no, this cycle will, in fact, be impeded by sort of a W-curve of some sort; that is, the world is going to have a hiccup due to unemployment or etc., and all the good news, all these engines we talked about, the fact that there is tons of applications coming about, the fact that there has been underinvestment for a long time in some segments like some foundry segments; the fact that, as Tim Arcuri was -- like we were discussing, surprisingly, people are buying only the leading edge, they are not buying anything else, that has to change; the fact that there hasn't been any fab lately, but you are starting to see global foundry, etc., etc.

So if all those good news are going to be impacted by an economic W-curve, then you would have a slightly less aggressive -- well, you won't have the sustainable EUR1 billion thing, but you probably won't be very far.

Also, I would like to add to that, if you make a comparison to previous cycles as to what Eric just said in this particular cycle, there's a lot of emphasis on the leading edge products, in which our market share currently is significantly higher than it was in the previous two or three cycles that, at least, I have seen.

So I think there are a couple of things that are definitely different here.

Philip Scholte.

Yes, good afternoon; Philip Scholte at Rabo Securities.

A question on your potential resumption of share buybacks, because I see some things on the news wires passing by.

Could you maybe update us on this call as well as to when you will decide on restarting share buybacks?

Yes.

What was said is quite clear, that we will have a good year, we'll be profitable, we'll be generating cash.

But it could be that in a year like 2010, we see, like we've indicated, from Q1 to Q2, we see a growing sales trend which will, the way we run our business, meaning that the investments in working capital, like in receivables and in inventory, will clearly precede the generation of the cash.

So it means that you first have to collect it in the end before you can distribute.

So that was the only comment that was made.

If you think about cash generation, that will be good.

But you have to keep in mind that there's a cash conversion cycle which first has to go through working capital and through receivables in a, let's say, year where you could see growing sales.

So that would limit, let's say, the cash generation this year would definitely show in 2011.

That was only comment made.

Right, and the comment on the EUR1.5 billion cash position; is that correct?

We always said, we have a policy which is we target our cash balance anywhere between EUR1 billion and EUR1.5 billion.

So we don't want to go lower than EUR1 billion, and we don't necessarily need to be higher than EUR1.5 billion.

So there is no change.

And that means that if we go higher, we will return the money to the shareholders, and which will be a combination of dividends, which we announced also today, and of share buybacks.

Right.

A short follow-up, maybe, on operating cost.

I was a bit surprised by the ramp-up in R&D and SG&A.

Can you already comment a bit on your expectations for the second quarter, in those costs?

Can we expect them to continue to go up?

No.

They will -- basically what we are currently seeing, they will be at approximately the same level as Q1.

All right, all right.

We certainly decided to put more money into (multiple speakers) R&D, as you've noticed.

This was a definite voluntary move.

We can afford it, and we have some ideas.

So we are going to use it.

And the SG&A was mainly driven by selling-related costs, but those selling costs are going up, and also a bit of inflation of the existing workforce.

So there were some modest salary increases, but that's also part of that.

Jonathan Crossfield.

It's B of A/Merrill Lynch.

I just had a question on pricing.

I wondered if there was any pressure on pricing resulting from the constraints you are seeing with the customers, or whether the pricing you are seeing on the NXT machine was affected in Q4 by the sale of the two evaluation systems.

Well, in this business pricing is mainly a very long-term negotiation.

Often the customer who will choose you chooses you for a long time, so you price based on the technical performance, and that has happened.

There is no technical pricing where we would try to increase the price for one quarter or reduce it, depending on availability or pressure.

So, no, we are on the same trend of pricing that we talked about, which basically is targeted at offering the customer the standard cost of ownership improvement per year.

We're doing this at this moment, and that should allow us to move our gross margin, as we said, inclusive of the volume effect and every other volume north of 40% as we go.

Okay, thank you.

And then just as a follow-up, with the capacity spending coming through from the tier-one foundries, what is your sense of how much further they can go before they need new [shelves], or is there a risk that we see a pause in that capacity spending when they have to build more capacity?

Well, you have now, I would say, officially a position by a global foundry that they are going to have at least one, if not two shelves.

It is my understanding that TSMC has also published their new shelves' decisions.

I think it's a bit complicated to understand because they use the same numbering, if I'm not mistaken, fab 12 or fab 14 or whatever is it.

But these are new shelves, discussion or decision.

I am not aware of any activities at UMC, obviously, but these are two very large members of the foundry community which are going to put shelves in business.

Mehdi Hosseini.

It's Mehdi Hosseini from FBR Capital Markets; two questions.

First, if you are not constrained in manufacturing, what would have been your Q1 guidance?

I don't know if I will be put in jail if I want to answer your question, but I would think demand -- the customer would have been happier with EUR1.2 billion or something.

But be cautious with my number.

Life is always a comparison between dreams and reality and we are always constrained in the business because our lead times are so enormous.

And so, if you ask me more the question what is the natural need that does not constrain the real output, I think we are not far.

Of course, we are pissed off about our [one months] than -- because we take comfort out of the equation.

But at this point, I repeat, your sons and grandsons will have their [products] intact.

And I look at the past few quarters, like Q4.

The actual reported booking came in about 23% above your guidance.

I want to know what happened that caused this upside compared to what you provided.

And could we see the same trend in Q1, primarily because of the increasing lead time?

Why wouldn't memory, or especially DRAM guys, wanting to get into the queue to secure shipment later this year?

Yes, I think on the guidance, we gave you our best estimate at that moment.

Three months is almost like -- it feels like a century from time to time in this business.

So it's very difficult to -- let me go back to Q3.

In July and early August we saw no trends of any booking activity, really.

And then it just -- we came back from the holiday period, and it was all on fire.

So these things happen fast.

It's difficult to predict.

We base our estimates on our current contacts with our customers.

But, as we speak, there are days when we have an estimate, and e-mails from top executives of our customers comes in and they are changing the demand.

These things change.

And it could be a few tools.

It could be, like I said, Eric said, 42 million, 40 million tools.

If they want four extra, it's 150 million.

So it's that level of granularity that has an impact.

And that's basically what it is.

And what could happen going forward, I would like to refer what Eric said earlier.

There are clearly -- when you look at the three sectors in our business -- DRAM, Flash and foundries or logic -- it was quite clear about where we are and that we feel there is an upside potential, and that upside potential, somewhere down the line, will translate into orders.

And when it will happen is difficult to say.

And at this moment there is no overheating in the order book, overheating in the sense that you mentioned it, which is I'm a customer, I'm getting worried about not having enough capacity, so I'm overbooking or I'm booking ahead.

That hasn't happened, and we are working very hard that it doesn't happen.

And the fact that most of our capacity constraint at this moment is due to an introduction of new product but not capacity constraint, per se.

(Multiple speakers) it's a learning curve, so you get to know it because of a delay, but you don't get to know it because you are just not getting it.

In fact, all our customers are going to get their product, and we are not intending to be limiters at all.

Sure, I'm just trying to understand the [longevity] of this cycle.

On the one hand, the memory guys, DRAM, Flash, have all placed orders for only shrink or technology.

They have yet to come back for capacity.

On the other hand, as one of the callers has suggested, maybe orders have peaked.

But it seems like two conflicting thought processes.

So if the memory guys have yet to come back with the real capacity orders, and given the increasing lead time, how should we think about the timing of those?

Is that just the Samsungs of the world are waiting to have more confidence on the world economy and PC unit shipment, or something else?

Well, we have, unfortunately, by good performance, I suppose, convinced our key customers that they can come up with very short-term reaction.

So I do expect, in fact, when they will have a need for capacity, which could happen, with the example of a foundry guy who is today focusing on 45 nanometer but can get a capacity order on 65, that will be on very, very short-term lead time.

So you will not have a preparation of that.

It will happen, and we are going to have to be ready for these things.

Just also one comment.

Just over 70% of the backlog is for shipments over the next six months.

So if you would see a real overheating, there would be a difference in the numbers.

So customers still have [what I accept] confidence that we will be able to ramp up, and we have ample capacity in Q3 and in Q4.

And we believe the same.

So they don't need to rush order and double-book, which, by the way, for those very expensive tools we have pretty good intelligence also at the customer side of where those tools are going.

So we have also some reasons, and we have some, I say, some good sources to check whether this is true or whether it's just fantasy.

Thanks for the detailed comment, thank you.

Ladies and gentlemen, I think we have about time for one more caller, so if you're unable to get through and you still have a question, feel free to call the investor relations department and we'll try to help you.

And now, operator, if we could have the last question or caller, please?

Janardan Menon.

Yes, hi, it's Liberum Capital.

Just a question, just following up on the previous set of questions, which is, by when would you expect to get the lead time on the 1950 down to around six, seven months?

Or rather, by when would a customer feel that delivery of this tool is not a sort of bottleneck for his capacity ramp?

And an associated question is, when you first came out with the 1950, the assumption was that your early customers would be from the NAND Flash market.

But now it looks like the DRAM industry has gone ahead of NAND Flash and placed a lot of orders for this tool.

So why do you think the NAND Flash industry is actually holding back on placing orders for this tool?

And do you think that the IDM industry, especially microprocessor, will also adopt this tool in a fairly wide way?

Well, yes, yes and yes.

So in terms of the lead time for the NXT, I would say it is now normal, that normal means six months.

And we expect Q3 to be the time where we are taking as many orders as the customers are ready to give us.

So the difficulties, I don't say, are fully passed because [we] -- but remember, the machines take a long time to be tested.

So most of the problems we have are there already, the machines are there.

The buildup for Q3 is at this moment reasonably under control.

So Q3 shipment, July 1 would be, I would say, on this time that -- a situation where we don't control -- constrain the market.

Your questions about NAND before DRAM or vice versa, I don't think it is product related.

It is, when do they need the capacity?

And I think, again, the reason why there is no urgency or there was no urgency on NAND in the past six, nine months or so has been because the big players have over-invested and have been digesting a fairly good set of ASML machine, by the way.

And the next batch is at this very moment being planned.

And I think it is, again, applications-driven and/or volume-driven.

And I think it is because they may have been disappointed about SSD's and are getting the SSD's to re-ramp back into the Q3 time frame.

So it's application-driven.

It's not us.

And is the NXT machine a good machine for logic?

In fact, we have been told by certain numbers of leaders of the logic arena that the challenge that the next generation nodes give them an overlay is such that they are very highly worried about process control, and therefore the more tolerance that you have or the less budget that you have for the machine drift is fundamental decision-making point.

So we feel extremely happy with this difficulty.

I should not say this in a cynical way, but we think the NXT is offering, in fact, a significant solution to this environment, logic environment, which is even worse in requirements than DRAM, which is a good user of NXT.

So the only negative part of logic is it takes longer for them to build up the recipe and to ramp into production.

So we would expect R&D to be done now.

In fact, you've seen in the backlog it's mentioned that we have a lot of IDMs who are in the NXT now in R&D, but those guys will ramp, probably, production in 2011.

Okay, to finish on behalf of ASML's board of management, I would like to thank you for joining us on the call today.

And Nicole, if you could formally conclude the call for us, I would appreciate it.

Thank you.

Of course, sir.

Ladies and gentlemen, this concludes the ASML 2009 fourth quarter and annual results conference call.

Thank you for participating.

You may disconnect your line now.