艾司摩爾 (ASML) 2006 Q3 法說會逐字稿

Ladies and gentlemen, welcome to the ASML 2006 third quarter results conference call on October 18, 2006. [OPERATOR INSTRUCTIONS]. I would now like to turn the conference over to Mr. Craig DeYoung. Please go ahead sir.

Thank you, Operator. Good afternoon and good morning. This is Craig DeYoung, Head of Investor Relations at ASML. Welcome to our Investor and Media Call and Webcast. The subject of the call is ASML's 2006 third quarter results according to U.S. GAAP. Here with me today are Chief Executive Officer, Eric Meurice, and Chief Financial Office, Peter Wennink.

ASML draws your attention to the message on today's press release and presentation regarding forward looking statements. The call will be 60 minutes in duration. On our website at asml.com you will find our press release, a brief presentation, and financial reporting with figures in U.S. GAAP and IFRS. And now I'd like to turn the call over to Eric for a brief call introduction.

Thank you Craig. Good afternoon and good morning. Thank you for attending our conference call, but I will correct Craig first. I will be a bit more lengthy on the introduction today, as there are obvious questions which we felt it would be more efficient if I were to try to address them in the introduction discussion.

So I've got, in fact, three messages in this introduction. First of all, our execution performance. You can see from our Q3 results that we have again delivered on key performance areas, namely, growth, profitability and liquidity. In Q3 we have been able to increase our sales by 80% year-to-year, leading to a projected 40% growth for the full year 2006 versus 2005.

According to analyst sources, ASML's market share has now reached above 60% in value in the first half of 2006, the latest period reported. This market share development looks sustained, as Q3 net bookings have reached, as you know, 95 systems, bringing our backlog to a record value of EUR2.1b, or 151 systems.

In terms of profitability, we see results from our scaling costs structure, in view of the high sales, as well as from our program for reduction of the cost of goods. We delivered operating profit of 25% of sales, a record net profit at nearly 18% of sales, on a growth margin above our operational target which is, as you know, 40%.

In terms of liquidity, we have significantly improved our manufacturing cycle time which has been our focus for the past two years, allowing a reduction of inventory, even in view of next quarter's forecasted growth. We used cash to buy back nearly EUR150m worth of shares during the quarter, as our cash generation outlook for the next quarters is strong.

In other words, another solid and consistent set of results, underlining now our maturity.

My second point relates to ASML's technology accomplishments. At the lower end of the market, our i-Line business is expected to double, year-to-year, 2005 versus 2006, but also 2006 versus 2007. This growth is enabled by our market leading overlap.

On the Immersion front, we have shipped 28 Immersion systems to date, including shipments of 15 1700i systems in the past two quarters, the 1700i being the industry's only system capable of producing 45 nanometer technology. We are well ahead of competition in the next node, DRAM, which went into production by our customers. The Immersion data now confirms appropriate defect density for two to three layer production which is basically necessary for the Memory business. On data, the appropriate defect density is about under 30 defect [per] wafer. Also, our current Immersion backlog, at 15 units, and the strong current bookings activities confirms the capacity build-up for production for our customers. At this point, we expect Flash manufacturers to ramp first, followed by manufacturers of DRAM and then of Logic.

Although it is too early to call the numbers of Immersion units to be shipped in 2007, we stick by our previous estimate that the new node of this nature involves a typical ramp of 70 tools, roughly, with 18 months. By mid-2007, ASML will offer the current 1700i for 55 nanometer and for 45 nanometer nodes, and a new TWINSCAN, the 1900i, which will enable volume production at 40 nanometer node, with the industry's highest wafer throughput.

To stay ahead further in the technology race, ASML also delivered in Q3 the industry's first true full-field EUV development system to [Albany] Nanotech and to IMEC in preparation for potential ramp of pre-production EUV units for mid 2009.

In view of the current favorable market conditions and to reinforce our position as a technology leader in lithography and to leverage our size, ASML will continue to invest further in R&D. We now intend to raise our R&D expenditure [EUR205m] in the fourth quarter 2006.

Finally, my third point will address the state of the semiconductor industry, as of course we see it from the lithography point of view. As we said before, the industry is maturing, as five key trends are contributing to its relative stability. The first trend is the semiconductor market is comprised, as we said, multiple number of times, of five different segments, market segments, namely DRAM, Flash, IDM, Foundry and R&D. These have different dynamics.

The second factor is that no single product application, TVs or mobiles, PCs, Playstations, etc, is now large enough to influence the whole industry as it was previously the case when a winning product cycle was overwhelming the total market.

The third element of stability is that the management of the supply chain in general, that is the management of the IC inventory, but also of the manufacturing capacity by our different customers, is more cautious than ever, as ROIC, return on invested capital, is a [positive] measure of results for the whole industry and obviously draws attention to these fundamentals.

The fourth element of stability is about Semiconductor FAB CapEx, which are now managed by tranche, as opposed to by very, very large new FAB as it used be. Whereby the overall building investment can be made, of course, on one hit, but the equipment itself is only brought in by small increments on a need basis.

And the fifth element of stability of our market is very much lithography driven. Yes, lithographic complexity raises its share value of the total FAB equipment CapEx as new nodes call for more and more and more layers to be processed with systems with higher ASP.

These five factors result, on average in shorter and shallower overall industry cycles which we have been living through for now the past three years or so.

And, and this is not a surprise, we confirm that this seems to be the case for the years to come, as we see it from the different data in front of us. On the negative side of what we see in front of us, there is, arguably, a bit too much inventory in some selected channels, a bit too much IC inventory in some selected channels, and there is definitively over-capacity built-up in Flash.

The IC inventories are being kept in check as the capacity utilization that we monitor shows some slight reduction in the related segments, but at levels which still enables lithography purchases. The Flash FAB capacity build-up, however, does indeed call for a correction at some point.

But if you insert all those negatives into the different models, and if you add the single digit IC unit growth assumption in 2007, the industry analysts are calling for a roughly stable lithography market overall in 2007 in comparison to 2006.

On the positive side of this equation, and positive for ASML, the 65 nanometer production ramp-ups which are now galloping, I would say, or increasing at a higher speed at our customers, their start-up of 45 nanometer with Immersion, the R&D cycle which will start mid-2007 for the development of the 40 nanometer processes, our increasing penetration in the mid-critical layers at our customers, we talked about i-Line, but we are also growing and our KRF business in customers that we freshly won in the past year or two, all this will translate into share gains for ASML and are significant engines of growth.

So in summary of my three points, in view of our execution track record, our strengthened technology leadership, and the lack, and this is an important message, lack of significant market peaks or drops foreseen. And analyzed by the different analysts of the industry, ASML is well positioned to execute on its secular growth target in the short-term but also in the mid-term, and the long-term. Now, Peter and I will take your questions.

And before I turn it back over to the Operator, I'd like to ask that you identify yourselves and your organization when asking the question, and in order to allow as many individual questions in the time allotted, we kindly request that you ask just one question, and you will find that we will be stricter than normal on this particular point, again, in the interests of allowing as many questions as possible. Now, operator, we're ready for the first question please.

The first question is from Mr. [Matt Shearer], Lehman Brothers. Go ahead please sir.

Thank you very much and good afternoon gentlemen. Eric, I just wanted to ask, actually, beyond the 32 nanometer -- oh, sorry, beyond the 45 nanometer node, am I reading it correctly that at this stage you can take the 1900i down to 40? With Immersion you're unlikely to be able to get to 32 unless you have some additions coming from EUV type technology. Is that the right way of looking at it? Thank you.

No, it is true that the 1900 or any machine with the maximum NA possible which is going to be about [technical difficulty].

Mr. Shearer, could you maybe mute your line, while you're listening to the answer?

Yes, sure.

Thank you.

Yes, so on every machine at 1.35 NA, which is the maximum possible at this moment, we could reach only 40 nanometer. However, you can do lower than 40, up to 32, and even beyond if you use double patterning, or double illumination. Therefore we expect customers as early as 2008 to start introducing these possibilities to go beyond the 40 nanometer, as a bridge towards EUV. And then EUV and double patterning will compete against the cheaper players, will win.

Thank you very much, and at this stage, you're the only guys in town with double patterning that works.

Thank you. Yes.

Thank you very much.

The next question is from Mr. Nicolas Gaudois, Deutsche Bank. Go ahead please.

Yes, hi there Eric. You seem to be very confident in particular, probably incrementally confident, on low-end so non-critical [part of your tools]. You just said, I think, i-Line to double again in 2007 versus 2006. Is this coming primarily from market share gains or to existing customers? And with that, do you actually see more customers following what GSMC, IM Flash or to have [inaudible] which is effectively to move to a single source for [inaudible]?

Nicolas, good question. This growth doubling twice, in fact, in two years, is about, I would say, nearly 50-50 the new customers and traditional customers who are expanding our market share with them, and yes, we go nearer and nearer to a, let's say, stable 80% to 100% market share at those customers.

Okay, and beyond the ones I've just quoted, basically, as we think you're nearly there, do you see similar opportunities at other existing customers?

Oh, certainly. The reason of our success in the low-end i-Line is because of this overlay. This overlay is beyond our competition capability as we understand it at this moment. And therefore if you continue, I think as a 65 nanometer, etc, growth, in some applications you definitively need this overlay on the non-critical layer, and the solution that we offer seems to be the unique solution with, in addition, high throughput. And therefore we expect the customers who have never bought from us to be also interested at the low-end.

Great, thank you, and as a quick follow-up, could you --

We're going to have to move forward, sorry.

Okay, fair enough.

Thank you.

Mr. Simon Schafer, Goldman Sachs, go ahead please.

Good afternoon. Eric, you mentioned that maybe at this point it seems as if the Memory segment is facing somewhat of an over-capacity situation, and therefore obviously your order guidance for the fourth quarter is a little bit downwards. That segment for you is about 75% of orders right now. As we look into the first half of 2007, when would you expect that to see a recovery, or is that a pause that is likely to be taking longer than a couple of quarters?

Okay, this is obviously a good question. So first of all I want to say one more time, and I think you've heard it from a lot of different companies, and I think research in particular, be cautious with trying to find trends into bookings. Bookings, of course, are correlated to billings at some point. But the factor that influences booking is also lead-time, and you could have sometimes very high quarter bookings and very low quarter bookings. That does not really translate into a capacity need or a billings trend, and you should be highly cautious. And I also noted since we introduced results today that some people were talking about bookings peaks and etc. This is a superficial measure of a cycle. So I will try to answer you now into billings activity.

In terms of billing, that is leads from the customers, we didn't say there is over-heating in memory. We said there is over-heating in Flash, which represents to that a bit more now than half of the total memory segment, and it's a bit more because there is over-heating. Normally it should be about the same.

So in DRAM itself, you see good strong bookings, but at this moment, billings, or a forecast to bookings, at this moment, the DRAM capacity build-up is very much a parallel with projected needs by the different analysts. Clearly in 2006, there was a gap, not enough capacity, a bit of a build-up to try to recover 2006 already, and a projection of the needs for 2007 which, as we say, is nothing exciting but this growth requires capacity. So DRAM, we feel pretty comfortable that there is not significant over-heating.

On the Flash, we all know that the different players have recognized that Flash is a segment of growth, or very high growth, which has a very positive characteristic which is if you can squeeze, if you can shrink, if you can integrate more transistors, you create more applications, which was not the case of DRAM. DRAM was basically limited, or is limited, by the numbers of PCs or boxes, etc, in which they are put in. Flash has the opportunity for a good long period in front of it, to be limited by only the capability of integrated chips, which is why they are all building capacity, and all running extremely fast to the next node.

But that said, yes, it is true, there is an over-building of that capacity, in view of what the analysts expect the Flash total business to grow in 2007. If you hear some analysts, they will talk about 30% growth of units of Flash compared to this year. Some will talk about 50%, but we estimate that the capacity which is being built up is above those numbers, and therefore that's why we call it an over-heating.

On the other hand, that over-heating has been assessed by us as being reasonably limited in view of our total portfolio. We have mentioned often that this over-heating has been estimated to be about 10% to 15% of our total sales. So in that case, we are considering that if there is a major correction in 2007 of this over-heating, that means a stop of the bookings for a period of time, well that period of time may not be very long. And that confirms our position that we don't see major downturn or drops, or froth, or whatever in this business because of this over-heating which is well understood, well-known. It is only on one segment.

Understood, thank you, and maybe a follow-up on the 1900i. Initial orders have gone out over the last couple of quarters. Could you confirm how many they are and whether all of those will turn into shipments in 2007? Thank you.

Oh, definitively, we're not booking for 2008, so yes, everything is going to happen in 2007 at this moment, but I remind you what we said last time. We said the customers will have to go first to a proved and qualification of [recipe] before they go and buy the further production units in volume. This is what happened in Q3. What I'm very happy to report is that, and in fact, you will read this in multiple newspapers and public statements, etc, the big guys are now saying that they are at the defect density and throughput and [overlay] etc that allows production.

So that is now creating the wave of orders negotiation. Today, the numbers of orders we have for next year are limited. It's about five-ish, as I mentioned by saying that we have a backlog of about 15. So it's five for delivery in 2007. But the numbers of activities, that is, of, call it, open orders for 1700i and 1900i, and I mentioned in my speech that we had the two, because we've noticed now that those correspond to a marketing. This is being discussed, and we are encouraged, and this is why I mentioned this 70 units in a period of 18 months as an outlook that we confirm.

Thank you.

The next question is from [Herman Betten], [Betten Bears] Media. Go ahead please sir. Mr. Betten, are you there?

Can you hear me now?

Yes we can.

Yes, I was muting myself. Sorry about that. I'm afraid I'm going to have to continue a little bit longer about the Memory and the whole Flash thing. If my numbers are correct, you received orders in the last quarter in the Memory segment of more than EUR900m and I was wondering, given in the light of over-capacity in the market, first of all whether it's sustainable, and second of all, how you think the over-heating is limited to only 10% to 15% of revenue?

Again, to answer your last part of the question, we have about estimated that on the period of nine months or so and the projected Q4 bookings, we will have a total Memory backlog, I would say, at about 55% to 60% of our sales. In there, we estimate that the Flash-only business is about half of it, and so that's, say, 30% of sales, and on which a very conservative person will see that half of it is over-booking, which is 15%.

And that's a very rough statement. We went into much more details to come up with those numbers, but this is the order of magnitude so I should tell you, very conservative way that this over-booking of 10% to 15% is not that exciting at this moment.

Remember also one key word of our press release, which may not have been seen very well. We said 65 bookings expected in Q4 with upside. At this moment, we are seeing still a level of booking activity in the different sectors, Foundry, IDM, Flash and DRAM, that would sustain higher than 65. We have, however, considered that 65 would be a conservative number, which corresponds to a conservative projection of the numbers of ICs that to be built in 2007.

Okay, would it also be fair to assume the 50-50 spread between Flash and DRAM also for Q3?

Yes we were trying to, in fact, help you with this, but we came up with a complication that a big, now nearly a third -- no, more than a third, it is a third and a half of our Flash holders could be DRAM also. They are switchable, so at the last moment, it's very, very complicated to know for us where it goes.

Okay, thanks a lot.

The next question is from Titus Menzies, Jeffery & Company.

Good morning gentlemen, this is Titus Menzies here. A quick question on lead times, how have they evolved in Q3 since Q2? Have they stretched, and are you seeing any double ordering coming through at all?

On the lead time, on the one hand, we made very good internal progress, and that has been translated, as you can see, from an inventory perspective. The bad news is, and I'm going to be very clear with you, because we are booking a lot, we are very near our capacities on certain numbers of critical manufacturing areas. So we have not fully reflected to the customers the possibility to do what we can technically do, which is less than six months lead time. So in other terms, on some orders, we are asking customers to put orders with nine months return or so, which normally we would be able to do better than that.

Do we have double booking? From our position at this moment, with all the parameters we look at, like utilization rates and the identification of where those machines go, we would be highly surprised to discover double booking.

Remember also what we said to you last time, in fact two times in a row. We are now identifying new lines, new FAB lines. Normally identifying FAB lines is easier for us to size the typical needs, and all our orders correspond to, at this moment, a real project. So to your point, if you guys expect that something bad is going to happen, and know everybody will cancel the orders, it would be really, if the IC business is collapsing by a major factor.

Okay, so in terms of the i-Line and your [inaudible] --

Titus, I'm afraid we're going to have to move on. Sorry, you've asked your question there. Thank you.

The next question is from Mehdi Hosseini, FBR. Go ahead please sir.

Yes, thank you. If your unit expectation for next year is flattish, how should we think about revenue, and to that extent, to what extent blended ASP should improve? Thank you.

Well, we usually, and you heard me last year, and the year before, trying to avoid giving guidance on the next year, because again, I would like to give guidance only on facts, and it's very difficult sometimes to project a full year. But it is true this time on the press release you have seen us talk about 2007, which you should translate as a fairly comfortable body language which I would not do on a regular basis, and I would try to tell you about uncertainties of life, etc, etc, but we have given you a guidance.

Let me read to you the guidance so that we understand it very, very carefully. We look first at a nominal forecast done by analysts about the number of units of ICs being projected and you heard single digit IC growth in 2007. So on these numbers, say between 7%-ish and 9%-ish, those same analysts have assimilated a lithography total business which is between 0% and -3% or something, so very stable. That is the first message we do.

If we believe these guys, you should immediately think that with our position in immersion, our position in those new customers, our position in i-Line, our position altogether, we're translating such a market share increase that you see 2007 in revenue as a fairly significant growth.

But again, this is only assumption of the analysts that I talked about. If you are the pessimists, as I heard two or three of you trying to tell me that there could be a major overheating that I don't see. So if there is an overheating with a correction -- significant correction which is negative to what the industry analysts say, then we believe that anyway, with our compensation due to market share, we do not expect to have a bad year in 2007.

So in other terms, these are the two sets of guidance we gave you.

Next question.

Next question is from Jay Deahna, JP Morgan. Go ahead please sir.

Good evening your time. Eric, given your prepared commentary before the Q&A, especially your comments towards the end, the more stable cycles, the Company's position with technology, etc. Would it be fair to want to expect ASML's earnings '08 to be greater than '07, especially as immersion kicks in? I think it's pretty clear that '07 is up over '06. So what are your thoughts on that?

And I'm coming up with seven immersion units shipped in Q3, and the likelihood of 10 in Q4. Is that about right?

[Literally] we have always said 25 -- 20 to 25 immersion tools in 2006. We are not -- we are never lying, Jay. But honestly, maybe 24 or 23. It's a question of execution. But we do have 25 planned and it [is impossible].

Regarding the 2008 point, obviously the type of scaling we are talking about with the new [method], very positive forecast for the total IC industry for 2008. So if we continue growing our top line, our scaling will significantly increase our gross margin.

You've seen that already on this quarter. I do apologize, we have been much higher than our 40% target, at 40.8%. So yes, there is significant upside on profitability in 2008 if we continue our [secondary] growth model and our immersion/leadership in technology in general.

Thank you very much, Eric.

The next question is from Robert Maire, Needham.

Yes. If you look at the opportunity of double exposure, obviously the number of scanners, steppers needed and light sources and edge tools increase, is that factored into your equation? And how much more upside do you think is there for the smaller geometries that are created by double patterning?

And as related to that, does that perhaps push out requirements for immersion, in other words, doing double exposure for dry. You seem to indicate that as a bridge to EUV. Does that suggest that there might be more double exposure and less EUV ramp up?

First of all, I don't think I'm wrong with this one, I don't think there will be any double patterning in production next year. And double patterning is sometime in 2008. Next year there is no forecast from our side that anybody would go in significant volume with double patterning. You know the big names who have announced some double-patterning work. But this is not production.

Two is you've got multiple ways of killing the cat, which is getting to your next generation node. You can do double patterning dry. You can do double patterning. When I say double patterning, double patterning or double exposure, and people, there are different ways of doing it. But it's okay using twice as a definition. So double patterning dry, double patterning immersion. Then you have EUV, as options.

Then you have some people playing with different ways of designing, more layers for vertical structures, you are talking about the special technology which allows you probably to have less critical lithography tools, but much more lithography tools, etc., etc.

So every year those companies are playing with this. And you would have potentially one saying my solution is better than another, etc., etc. What we're sure is that in the year 2010 or '11, if you are conservative, EUV is the cheapest way.

Okay. So your assumptions going forward are assuming the industry goes to EUV and is not as expensive a user of double patterning?

No. I said that between now, that is between 2008 and 2010, you are going to see people in double patterning. Good news, by the way, for all of us. And during 2010, 2009, end of 2009, 2010, you probably will see Flash people going full steam with EUV, because EUV is much cheaper for those guys who have limited number of masks, longer -- very high volume -- high demand on throughput, versus the Logic people who may prefer double patterning.

So, in other terms, 2009/10 is a transition where you would have X% doing one thing and Y% doing the other. This is great news for us because there is no way we would be able to have the capacity to take 100% of the business in EUV in the first year. So this staged-in approach is good.

And that plays into the secular growth model that we have. If you put in your model in 2008 some double patterning, in 2009, some double patterning and the start of EUV, and 2010, a lot of that [with the price] you see, it looks good.

Thanks, Robert.

Thank you.

The next question is from Thomas Brenier, Societe Generale. Go ahead please.

Yes. Good afternoon. I'd like to know if the EUR105m R&D for Q4 is going to be the run rate for 2007. And also if we should consider higher guidance than the usual EUR75 to EUR80m for services and [fleet] options as this has been over EUR100m [technical difficulty].

Okay. So Peter will handle the options questions. Regarding the R&D, the decision we've taken is to put as much R&D as our P&L can afford so that we can put in our developments and introduce products earlier and create, therefore, a major differentiator in this business. So we will put them on it if we can afford. At this moment we think that there is, say, a good vision on 2007. So we therefore believe that the number will be higher than EUR105m before.

For me, a specialty question. The service and options sales, yes, we've already done about EUR75m over the last couple of quarters. You've seen that that has trended above that number. The fixed part of that sales, which is basically revenue from service contracts, which has always been about EUR50m, is trending up. It's about 10% up. And the same is, I would say, true of the option sales.

So if you want adjust your models, then I would suggest that [inaudible] 75 between 80 and 85 per quarter is probably, going forward, a bit safer now.

Thank you.

Next question is from Mr. Amin, Citigroup. Go ahead please.

Yes. Hi guys. A quick question. As I believe it, you are going to be ramping capacity by around 10% a quarter next year as your current potential maximum new tools a quarter is around 75-ish. What gives you confidence in increased shipments next year and beyond that this capacity will be needed, given that currently you're shipping around the 60 new tools per quarter? I mean how soon will you need this new capacity, do you think?

I am hesitating because I know some of you will put in the newspaper immediately my key long-term forecast. No, I think we have published often our target to be EUR5b of revenue in the foreseeable future. This is not a forecast. This is simulated in the numbers that says if people go and do the double patterning, a bit of EUV and grow by, say, 7% or so of units per year, we should get there soon. So in other words, by 2010.

So yes, we absolutely need the capacity. And yes, we are a bit ahead of the roadmap.

Okay. Thank you.

The next question is from Marcel van de Hoef, Bloomberg.

Yes. I'm double checking something actually. Let's say the market is concerned about a downturn, would you say that risk is mostly in Flash and that the maximum impact on revenue in 2007 would be 10 to 15%, or is that completely incorrect?

The total here is 10 to 15%. So if you have a correction on one quarter, you could have much more impact than 10 to 15%. But remember we are now talking about overall quarters, about EUR1b or so, EUR1.1b. So these corrections could still get us into areas of magnitude of EUR500m a quarter, which is why we are not getting too worried.

Okay.

Thank you.

The next question is from Didier Scemama, ABN Amro. Go ahead please.

Yes. Thanks for taking my question. My question is as follows. Basically, we are starting to see companies getting out of the NAND Flash market, such as [Kimonda]. Admittedly small players. A lot of companies have gone in NAND Flash, attracted by the volume growth over the next few years, and they have really been beaten down by some of your customers.

So my question would be in the event that the market in NAND Flash would actually reconsolidate around maybe three players, two of them being major DRAM vendors, what would be the impact on your business going forward, do you think?

In fact, today, I guess we are in every Flash manufacturers. I think it is now not a secret that we are working with Toshiba. Obviously, we are working with Samsung. We are working with IM Flash. We are working with Hynix, etc., etc. So it will be completely irrelevant for us if there is consolidation.

Okay. Thanks.

The next question is from Janardan Menon, Dresdner. Go ahead please sir.

Yes. Hi. Thanks. I was just wondering about the -- going back to the comment that if the industry grows at single digit in 2007 it should translate into flattish lithography demand. You said that that's the view from analysts. I'm just wondering what your own view on that number is because just to put that -- the reason I'm asking is in 2005 the industry grew at about mid-single digit, but your shipments fell by about 30% year on year in 2005. So I was just wondering how to reconcile that fall with a potential single-digit rise next year.

Be cautious, we are always talking about sales value, because units is not the right number to look at for us. So we look at units of semiconductors, and that translates into value.

Yes, we did simulate to check what the analysts are saying. And we have a simulation tool ourselves. And we put the single-digit forecast for 2007. And the result is not very far from the analysts'. I will not be very, very specific. But it's not very far. We believe we can tweak a bit the new model if it goes up or down depending on how fast some nodes happen.

And remember, this is why in my speech at the beginning I said we are very hopeful the 65 nanometer is ramping in even more volume. In fact, you should try to translate that sentence in ASML terms. That means hooray, we are selling a lot of 1400. And it is now a 1400 business compared to what is an ARF business at EUR20m, compared to a KRF business at EUR10m.

So these things make a big difference in your simulation tools. But, I repeat, we are very comfortable with the mechanics of those. Peter?

Yes, I'd like to just add one thing because you are comparing -- you're bringing 2005 into play. In 2005, you are right. The numbers were, indeed, down. But our sales were up. And sales were driven by market share gains and by ASP growth. That sounds a bit more doesn't it.

So you're talking revenue staying flat rather than unit?

In 2005 revenue was up and you talked about units were down.

Sure. Absolutely. Thank you very much.

Thanks.

The next question is from Mark FitzGerald, Bank of America. Go ahead please.

It's a two-part question. Can you give us some idea of ASML's position in the microprocessor industry for 45 nanometers?

And the second part are there any -- was there [turns] business in the current quarter just reported [more than] in the December quarter?

Could you repeat your last question, what was it on the turns business?

Turns business in the September quarter, was there any [turn]?

The issue of microprocessor, I am cautious because, of course, we are talking about two customers. What can I tell you which is okay?

Do you expect to gain any layers in terms of your penetration at 45 nanometers?

This discussion is in progress. And it -- let me say it this way, if our product is superior to the competition which, in this case, is Nikon, there should be no reason why our current market progression -- market share progression doesn't continue, if I may say this way.

I think it's also public to say these decisions are not yet taken by the big guys. And a sure thing that I can tell you is you should ask them which machine they used to show double patterning 45 nanometer sample. That could tell you why we have some hope.

Regarding your issue of turns, no. Minimum turns in Q3 and Q4, but maximum pressure to put orders. But we could not basically do it.

Are you ready to take the next question, sir?

Yes please.

Thank you. [Karen Holshouse], Credit Suisse. Go ahead please.

Yes, hi. It's actually [inaudible]. I've got a question on NAND Flash. We think there is a lot of capacity on memory that's still running on 200 millimeters, especially on NAND. How long do you think chip companies can push NAND on 200 millimeters? And specifically on what level of activity are you seeing with immersion lithography on 200 millimeter? And I have a quick follow up.

No, immersion is going to be only on 300. Maybe a bit of R&D is at this moment made on 200, and we help. So clearly we helped these customers that had done that. But production at this moment, I don't know if any would go in volume production with 200. It would be 300.

And the rest of the 200 millimeter will be continuing on the old technology Flash. And, as you know, some of those things are being moved into China.

Okay. And a follow up. Your comment about Flash, that Flash is weak. And given that that will be the first segment to ramp immersion and volume production next year on 300, am I correct in thinking that over half the immersion systems that you're shipping this year are perhaps R&D for chip companies? And is there a risk that if Flash capacity [admissions] come down, that the number of immersion systems that you ship next year may not grow as rapidly as the 70 unit number that you were talking about?

I think, again, I would say no chance to be -- to have a bad performance here. If ever there is a correction, as people would expect if the overheating continues, but that correction happened after you go in battle. So the battle hasn't started. Capacity hasn't been set up. And most of the tools we have shipped, and we shipped 25 tools, will be used and will be half in production. There is another batch of [ex] immersion tools which will have to be put into production before the battle starts.

Then the battle starts and then somebody says either you go out of business or I stop buying, and this is when I am calling for the potential famous correction which we expect not to be too large anyway.

But there is no doubt in our minds, let's be real, that the 40 nanometer tool we are going to introduce next year, that everything we can make we will sell. There is so much demand for that additional shrink. There is no doubt in our mind that whatever the overcapacity would be that that would put a lid on our capability to actually sell those. There's no doubt in our minds. We will sell those.

By the way, to be clear, some of our tools -- immersion tools are today in production at multiple customers, like 65 nanometer in immersion. So you can imagine that if immersion -- the effect of density gets to the right level, immersion will increase the performance of the chip at 65. So people will use those [inaudible] to differentiate themselves, bring more performance into each chip. So we're not very, very worried about immersion to be only limited to 45 nano or 55 nano.

Thank you.

Next question is from [Alan Sogne], [General Capital]. Go ahead please.

Thank you for taking my question. I wanted to get an update on your thoughts on share repurchase. It looks as if you shrunk the shares down about 3% in the quarter, and the cash performance is very good. So I would hope while in order for investors to be focused on the short term, you'd take advantage of the stock price here and continue to repurchase stock.

Today?

I'm sorry?

I had a joke. Today I think we should jump.

I think on that question I think we can just basically reiterate what we have said before. We have this EUR1b of our net cash target which we started the year off with, and cash flow generated we have put into share buyback programs. Now absent any major M&A deals, which are currently not on our radar screen, we will continue what we are doing.

And this is not the moment in time to announce the next share buyback because there are, of course, disclosure requirements. So we will do that at the right time and based the right place. But clearly the, I would say, buyback strategy of the Company has not changed.

And with the, like we said also in the introductory talk, are looking forward to 2007 quarters being strong in terms of generation of cash. It's not difficult to imagine what the execution of that strategy will be for our Company.

Great. Thank you very much.

The next question is from Mark Bachman, Pacific Crest Securities. Go ahead please sir.

Thank you and good evening. This is Mark Bachman at Pacific Crest. Two questions for me here. First, Eric, you stated your 20 to 25 immersion tools as per for 2006 still stands. But you also mentioned that you are building additional [end] tools here for early 2007. Is there possible upside potential of the 20 to 25 estimate with additional shipments in Q4?

And then just a follow on to that, can you discuss the technology mix of all sales in Q3 and how that mix is going to be changing into Q4, and maybe specifically call out the effects of i-Line? Thank you.

Excuse me, the last part of the question is? [Inaudible]. So first of all, no upside in Q4 for immersion. We just don't have the capacity to do more at this moment of exceptional capacity demand, etc. So we could have shipped more. There is the demand for that. But we will not have upside.

Regarding the overall mix of the Company, you have seen that i-Line represents only 10%, 9%-ish of our backlog. We expect, as I said, to double that business. But we -- as we said some time during the call that we also expect to grow the total [inaudible]. So I wonder where we will -- where i-Line will reach. So, at this moment, you should probably think that i-Line will not go to 20% of our sales, but probably will go around 15%-ish.

Thank you so much.

The next question is from Stuart Adrian, Morgan Stanley. Go ahead please.

Yes. Hi there. Hopefully a relatively straightforward clarification. Just looking at the wording for the 2007 guidance, you basically say litho demand [price] now falls in '07 expected to be about the '06 level. Now, does that mean that the 265 units that you think you're going to ship in '06, if I just take your fourth-quarter guidance, you think you're going to ship 265 units in '07, which basically implies, one assumes, ASPs go up and therefore you're going to see some pretty decent revenue growth in '07? Or basically are you saying that '07 revenues are going to be flat with '06?

No. I'm so sorry about the confusing sentence. And I think I tried to say at the beginning of the call, you have to put yourself into assumptions. We don't call a forecast, we call assumptions. If you take the assumptions of the whole litho business as stable, as the analysts say, then there will be significant growth.

If you are a naysayer and you feel that the analysts do not understand that Flash is really going down the tube, etc., and there is a correction, you should put us about flat in revenue.

Okay. So if you think litho market is flat in '07 versus '06, you can still take market share in that environment. So therefore your revenues will grow in '07 relative to the '06 level?

Absolutely.

Okay. Thanks.

And the fact that you foresee to say these things is, again, a vote of confidence. Usually, I remember, I have often said we don't know, we don't know, we don't know. But the current things we know, the long-term backlog we have has enabled us the discussion we have with the customers for the project, allows me to -- forces me to tell you that we are on a range of success, or good success for 2007.

Okay. Thanks.

The next question is from Antoine Badel, Credit Suisse. Go ahead please.

Yes. Good afternoon. I want to come back to your order guidance please. You've said that you're already seeing upside to 65 orders for Q4. And you cautioned against extrapolating that decline into a trend. So my question is how big is the swing factor for Q4? And what are the possible scenarios beyond that? Are you saying you expect orders to stabilize in the early part of next year or to rebound?

This is a problem of those bookings. You never know. We can again surprise all of you and say 65 plus is a significant amount. We can just be at 65. But that does not change the fact that, to answer [inaudible] question, that there is significant engine of growth for 2007 in billings, and that at this moment obviously yes, we see them for Q1 and Q2 because this is within our lead time. And there is a bit more uncertainty in Q3 and Q4 which is why then you have to simulate stuff.

But the bookings trend is going to be a dangerous thing for you to try to make value of. Now we all know this number is going to psychologically impact our evaluation, etc., but it doesn't really change the picture at this moment. Whether we have 65 units, which is still very strong bookings at a quarter, whether we have 85 or more, this will not change again significantly the picture of the first two quarters and will just solidify Q3 and Q4. But then most people would have put orders ahead of that.

And when you say 85, is that a realistic upside scenario for Q4 or is it just theoretical?

No. I just meant mathematical numbers if you -- we -- our guidance is 65 and upside. And we have no idea of the [upside non booked].

Thank you.

Operator, I think we have one more question.

Thank you. Jan Willem Berghuis, Kempen & Co. Go ahead please.

Yes. I'll make it a short one. Maybe you can comment a little bit on where you will spend the additional R&D or, in general, where you're spending R&D, whether it's the lithography market, are you thinking of some additional areas which you are investing in at the moment? I hope it's not too confidential.

To be sure, everything we're talking about in this EUR105m is IC.

Okay. And is there a big portion going into EUV?

Not yet.

Not yet, okay.

There is a portion. But it's not -- at that moment we're spending much more time on other things.

Okay. Thank you.

Thanks, operator. I think that that concludes our third quarter results conference call and I'd like to thank everybody for attending.

Thank you very much.

Ladies and gentlemen, this concludes the ASML 2006 third quarter conference call. Thank you for participating. You may disconnect now. Good bye.