ASML Holding N.V. (NASDAQ:ASML)
Barclays Capital Global Technology, Media and Telecommunications Conference Call
May 22, 2012 02:45 pm ET
Craig DeYoung - VP, IR
Andrew Gardiner - Barclays
Andrew Gardiner - Barclays
Good afternoon everybody. My name is Andrew Gardiner from the Barclays European Equity Research team covering the technology hardware space and it's my pleasure on behalf of Barclays, on behalf of the European team as well as our US team here covering the semiconductor capital equipment space headed by CJ Muse to welcome ASML. There we have Craig DeYoung and Peter Convertito from the investor relations team at ASML and we are just going to do a fireside chat, we should have some time for questions at the end, but I will just run through a few hopefully salient topics for Craig, but first of all he is going to give us perhaps a quick update on the most recent quarter’s latest trend.
Sure. Thanks Andrew and thanks everybody for your interest in attending today. Yeah I talked -- maybe I would take about five minutes to kind of give you an update. There are have been some things that have changed over the course of the last three months, not so much since we've reported our results, but prior to that, so I wanted to bring everybody up-to-date. So we entered the year with a different view than we have now, probably almost in all sectors except maybe one. So it was our thought that as we entered the year that we would see a digestion period, probably in the second half of the year based on the number of tools that we shipped to the foundry space.
So we started to ship in real volumes to the foundries in total about middle of last year, so we shipped a significant amount of equipment. So again it was our opinion that we would probably see a bit of digestion and that hasn’t happened and we will get a chance to talk about more of the details in a minute. So today, we view the second half of the foundry activity of shipments at least from our standpoint to be stronger than the first half of the year where the opposite was true just a few months ago.
And we've been a bit surprised also by the NAND space. The NAND space, kind of the opposite has happened. We anticipated based on industry analysts' forecast about 70% bit demand growth in 2012, not sure that’s going to happen at this point in time. 70% bit supply would’ve required both shrink and capacity addition. So, we calculate that about 50% bits can be NAND bits, can be grown by the shrinks alone, the planned shrinks of our customers and just to remind you, 20% of our backlog at the end of the quarter is NAND.
So we continue to supply to that, to the shrink roadmaps of our customers, but on top of that, the other 20% or so that was expected from the supply side would had to be added by wafer capacity additions and this is where we see the moderation in behavior of our customers. So, within the 20% backlog, there is an amount of capacity that is going to be added, but not to the extent that we expected likely because the demand expectations for the NAND space are not quite through the year what was originally intended. So, there is rational and reasonable reaction from the NAND players in postponements of capacity additions in almost every location or transition from NAND to some other process in certain cases. So that was a bit of surprise.
So in the second half of 2012 coming in to the year, we expect that the NAND would be slightly higher in terms of us supplying to them than it appears they are going to be at this point in time over the first half. DRAM, really no surprises. Unbelievably, only 2% of our backlog at the end of the quarter is DRAM. DRAM can go bit by, somewhere approaching 40% by just the installed base that they have now or minimal additions to it in order to support the technology, no transitions that are planned by the aggregate DRAM players this year.
So in both the NAND and the DRAM spaces then we’ll have to look to expected demand growth in 2013 for either of the parties to come back to us for a greater supply. So again we'll have to let things develop a bit and let our customers get a view of 2013, we think at this point in order to get any real serious activity from them.
And then just to finish off on the foundry space there has been an increase interest in continued supply of both area of immersion, the critical layers tools as well as KrF tools to build out capacity in the 28 nanometer node. So you’ll notice in our reporting that we’re selling a significant number of units of immersion as well as the non-critical tools which fill out a brand new fab capacity.
So we know that behind that is real demand and you all can feel that and read about it et cetera. On top of that there is a renewed competitive slant to the sector such that that's stimulating probably an amount of additional capacity as we look forward into the at least this year and we’re starting very recently to get the view of the first half of next year. So let me just finish off by saying we can measure the utilization of the tools as an indication of supply demand balance if you will. So if we have a very high utilization it suggests we are a worse case in balance and best case if you will or worse case for the customer but good case for us where in an under supply situation. So we watch that, we also look at the number of tools we’re shipping into that sector and based on a number of challengeable variable inputs to the model, we can calculate the number of wafers starts that are going to be added.
And so our calculation show that by the middle of this year the aggregate foundry players will have installed about 70,000 wafer starts of 28 nanometer capacity. So then we can look at that and we can compare that against a mature installed capacity for the previous nodes, let's call it 65 or 45 nanometers and if we look at the two mature installed capacities of those combined on average it is about 250,000 to 300,000 wafer starts.
So if we use that to measure of what might be installed eventually in the 28 nanometer node for lack of anything better to look at, we'd say we were only about 25% through the ramp to a mature capacity of an average node, so it does appear that we are in any risk in the short term of real over capacity which is something I know you would be concerned about and us as well. So again that kind of gives you a feel over the things that have changed, what our current view of the different sectors are and with that I will throw it open to question from your end.
Andrew Gardiner - Barclays
You know perhaps if we continue on the foundry side of things, we have heard various players within the fab list were complaining about not having enough capacity at 28 nanometers beyond allocation, not getting what they need and that having an impact on their immediate second quarter ability to sell, what are you hearing from those customers, what is it that you are seeing in terms of ability in the foundries to get the technology process up to speed quickly enough to solve that.
Yes, so there I say that we haven't really seen a real sense of urgency from our current customers with regard to but they already have plans. So if you look at our backlog $1.6 billion roughly at the end of last quarter, 60% of that so that's about almost $900 million in backlog, that is a planned deliveries over the course of the next two quarters to the aggregate foundry suppliers.
So I think when they are confronted with questions about supplies they are already prepared with a known set of equipment deliveries against that demand. Now that's not to say we haven't had new inquiries about our capability to supply but any time a customer or two or three are frustrated by the amount of supply they can get, they are going to look at alternatives and I think part of the discussion we've had over the course of the last probably six to eight weeks is some of our customers inquired about our ability to respond to them and they in turn potentially responding to the customer that maybe frustrated by supply and so but I think over the course of the last, so we did see a number of new inquiries but I think that's settling out a bit at a level that's probably less than the initial pays as guys decide whether they are going to be able to supply and their customers decide whether they are going to award business or not.
So but one thing I would say is that the larger foundry players are starting to give us a longer view of the requirement. So as I think I mentioned our expectation now is that our business in foundry will be relatively strong from a delivery standpoint probably through the middle of the next year at this point.
So we see a ramping from a 70,000 wafer start to middle of this year to this 250ish level which is a mature note through the course of first half of next year and it’s a pretty linear ramp that is indicated by our customers. So they are starting and this happens also for two reasons when there's real demand, or real loan demand as longer router visibility grows within our customer base then their visibility to us obviously increases.
Also there can be a bit of competitive element there also so you know if one guys knows that his competitors are getting in line, his interest then will be to get in line sooner, quicker, before, longer, after just to secure slots in the delivery planning and production planning of [ourselves]. So there is again not, I wouldn't say a huge sense of urgency regarding what's on order but there is a growing visibility given to us by our foundry aggregate combined foundry customers.
Andrew Gardiner - Barclays
And what you just described indicates that your visibility is clearly a lot better and improving relative to how it had been in the past as compared to your official backlog that you are reporting. So just in terms of the market’s reliance on that quarterly order figure and how are you guys going to continue to communicate that how you think we should be thinking about it. It is time to try and think about it differently?
Yes, I think what we, what ASML always tried to do is just tell you exactly what we know but not try to guess based on what we think we know, what the future is going to hold, because the customer base that were especially in the logic area specifically that's the buying group right now is very compact and the number and the intensity of litho within the foundry and logics base in general is very intense and its probably
And the number and the intensity of litho within the foundry in logic space in general is very intensive and is probably anywhere between three and four times the intensity of the memory sectors for the same number of wafer starts and creates what we see as a real lumpiness and either further potential lumpiness in our business.
That isn’t necessarily indicative of how our revenues will flow. So we decided that instead of trying to, we would have to provide such wide range of bookings, guidance that would have been meaningless to you guys. So we decided instead to trade that off for our revenue guidance.
So we guided two quarters of revenue of 1.2. Peter Wennink, our CFO was with me in Boston at a conference last week where he suggested in a similar format that there is no reason to believe that the fourth quarter won't be along those same lines as well. But we’re getting the sense now, qualitatively, what (inaudible) opportunity is further out than they guided two quarters and we just have to wait for the real data to come in for us to be able to quantify that in the future. But we’re going to give you and continue to give you, tell you what we know as far as we can. We’re just going to translate, this lumpy booking stuff into a revenue guidance instead.
Andrew Gardiner - Barclays
And in your introductory comments, you mentioned foundry, you mentioned both sides of the memory space. You did touch on the IDM, one player in particular, with coming up the time for then, almost planning for the next node. Can we expect further orders to come through there for that later this year?
Yeah, again we hesitate to talk about specific customers but I think it’s well known that the big idea in here in the U.S. starts in to a new cycle if you will. So, there you’re going to transition to their next node.
Our history shows that at the beginning of those nodes is a little bit heavier weighted and in another words the first year than the second year. So we would expect generally that with litho spend would be you know above and I think they have indicated this is well captured in their CapEx guidance that they are going to spend pretty heavily next year on non-building or equipment related spends.
So yeah, but given but people understand, I think about our lead times then if they are going to want equipment from us have to contemplate orders here and certainly within the next quarter or so.
Andrew Gardiner - Barclays
So that does it in terms of walking through different customer groups. Can we talk a bit about the technology side of things and in particular the next generation using development we have some load blocks hurdles to clear. So far this year in terms of development particular on the light source side, so what's the latest that you can tell us there into this power and throughput?
Yeah, I would say we’re not feeling uncomfortable about the progress we’ve made all in the last quarter or quarter and an half actually we’ve you know spend a significant amount of time, energy and money ourselves in support of multiple source suppliers but with that particular focus on one over the course of the last six months or so.
And there has been significant progress made in the source power which we translate into wafers per hour. So our customer buys from us wafers per hour. We’ve been talking about source power because the focus is of course then on the source as it's been the troublesome area in terms of getting power up on the system. We're fully involved in the development activity on that firsthand, and we're actually guiding a lot of that activity. Again, not against the will of the suppliers but with the full support of the suppliers.
So it's brought (inaudible), it's brought return, and I think actually in this timeframe the early adopting customers, which are known to be now the DRAM players, will get into the point where they're satisfied or near satisfied with the wattage performance that we translate into throughput for them on wafers per hour.
So the level of input to our systems are projected in input based on what the source suppliers is enough, but their focus now is moving away from the minimum number of wafers per hour that we'll probably commit to at the end of the year, to the reliability of their systems.
So you recognize that when a customer ramps their process, they need a minimal amount of throughput in the beginning that's matched to their early production ramp. So in time, we need to grow the ramp and grow the throughput of the tool. So we have kind of reset a new throughput timeline with at least the early adopting customers, whereby a 40 wafer per hour target at this end of the year with a 70%, 80% availability or reliability is now the focus.
So that minimum 40 wafer per hour is -- appears to be, again, kind of I'll call it a magic number for lack of a better word, whereby if we can get reliability up, they will find their way to adoption of this in the early stages of their production ramp, which by the way will be late '13, early '14. So we actually have the 12 to 15 additional months to work on, getting the power and the reliability up.
But I'd say clearly the focus on our customers' part is moving from the throughput to a reliable throughput, whereby they can make a commitment to this in the process. So the timing for a commitment to us is kind of in the third, early fourth quarter at the very latest. That's based on two things, two important things. One is the lead time that we need in order to prepare for volumes that might be required beyond the 11 tools that we're going to ship in any case.
And I'll come back to those in a moment. But as importantly or more importantly, they need to make a decision on which route they're going to take, because currently they're doing tandem development with multi-layer patterning immersion and EUV, and they will prefer EUV for a number of reasons but they need to get security, again, in the reliability and the throughput.
But they need to choose one of the other to continue their development work over the course of the next 12 months or so. So that's why there's a critical timeframe milestone for both us and them in the third-ish quarter of this year or so.
So just to remind you, we are going to deliver our 11 tools. So we have a batch of orders of 11. We've talked about that over the course of the last 12 months at least. It's grown a bit from seven or eight to 11, but we owe those to customers, starting very late part of this year through the third quarter of next year. So the second batch of orders that we're talking about today with our customers are not deliverable until the fourth quarter of next year into the first quarter of 2014.
So again, that's consistent with their need in ramping the 20-nanometer node in DRAM. So the timing is not completely mismatched, even though we've seen some delays in the source power. So we expect that adoption by the other sectors will be staged. Likely that the second adopters will be the foundry players as they will be challenged to do cost-effective imaging at the 14-nanometer node.
And again, the 14 nanometer is out in 2014 and '15, but they need 24 months development time. So they need something in hand in early 2013, and they will get one or more of our 3300 early versions. And then, the NAND and the big logic, IDM are not looking to implement EUV in to the 2015/16 timeframe. So you'll see a nice, rolling out, snowballing if you will, of demand between end of 2013 and 2016 timeframe.
What are your expectations in terms of production capacity for EUV at the moment?
Our production plans today, our capability today is about one tool a month. So through the course of 2013, we'll maintain that one system a month capability. And again, that will take us again through the third quarter of next year in order to supply, let's call it, if we ship through this year nine additional tools in 2013 at that rate, so we'll have a couple, two or three available at the end of 2013 to ship.
And then, we'll move rate up to two systems a month in 2014. Most of that will come, if not all of it, through cycle uptime reduction on the tools. So we need (inaudible) square meters of factory space in order to double our capacity to the 24 per year in 2014. And then we'll have to take a look and see what the demand might look like in '15, given the fact that we'll see pretty much everybody come on board in the 2015 timeframe.
We're looking at a plan to hit double again the throughput to, let's say, nominally 50 systems in 2015. And that should be not usually mismatched with what our model show demand might be at this point in time.
We're drawing near to close here, but are there any questions from the audience?
You mentioned 50 systems, EUV systems in 2015. Would you expect your ASPs to be similar to what they are today? In other words, that's a big chunk of your or what your revenue would be this year. So -- and I expect you'd still filling some immersion, so --
Yeah, I would hope you wouldn't hold me to the 50 in 2015 exactly, but in the '15-'16 timeframe we can imagine that those -- that we could approach those numbers in terms of demand. The 11 tools that we have on order today, the 3300 in its current configuration, have an ASP of about €75 million. I would expect in 2016 that it would be, I'll use the word significantly, but it will be above that as we enhance the performance of these tools.
So there is a roadmap that will enhance the imaging capability to overlay in the throughput of these systems in time, which is our habit on all of our platforms. So -- and with each of those performance improvements comes a level of price increase. So I'd rather not throw a number out there but it will probably be well above the 75 million ASP.
Any other questions?
One more? I think we have a breakout. We do, yes. We have a breakout. I'm not sure which room it is. It should be on your schedules. So we do have a breakout after this.
Any other question we'd be glad to handle them. Thank you very much.