ASML Holding Management Presents at Goldman Sachs Technology and Internet Conference (Transcript)

Mar.12.12 | About: ASML Holding (ASML)


Goldman Sachs Group Inc. Technology and Internet Conference

Feb 14, 2012 17:40 PM ET


Simon Schafer - Goldman Sachs

Simon Schafer - Goldman Sachs

Well thanks everyone for joining us here with ASML Holding, Goldman Sachs Conference just for the sake of the webcast as well. I am Simon Schafer, I cover ASML and Liotech for New York Goldman and other.

Greg we have done this for a number of years, so really great for you doing this again. Thank you.

Unidentified Company Representative

It's my pleasure.

Simon Schafer - Goldman Sachs

We have not had from a great deal of companies in Semicap today, I think Ram presented and in Tarzian we had from two and they sort of shared this outlook of look (inaudible) adequately and broadly flat and I think ASML has also provided an outlook for the past but you haven’t quite ventured into the territory of providing full year outlook. Maybe just remind the audience as to what your latest expectations because you did talk about just a couple of weeks back on your quarterly.

Unidentified Company Representative

Yes we were in a fortunate position a year ago to have a huge backlog coming into 2011 whereby we could actually guide with some level of confidence for the full year and in the end things worked quite well in that regard. We have posted of course a record level revenue of 5.6 billion. There was some uncertainty as we traveled through the third and into the fourth quarter of last year regard what the first half of this year may look like. We had a fairly clear view given to us by our customer regards what their requirements might be in the first quarter at least but through the course of the third quarter and again into early fourth quarter they were hesitant to place orders against those their forecast.

So, just to give you an idea ASML bills to a forecast, so we do a bottoms up and we do a top down and then we try to bring those together as far as we can and based on how well those match we go ahead and do production planning.

So, we are not so interested to use a word in kind of the CapEx announcements of our customers because we are forced because of our lead time to prepare well in advance of when our customers might decide to broadcast what their CapEx looks like. But in the end in the November time frame we saw our customers attitude change a bit, they grew in their confidence level, certainly in the foundry space and they came through with orders or certainly gave us reason to believe that they would through the course of this quarter and next you know live up to their word relative to their forecast.

So at the end of the quarter with our 1.7 at the end of year, with our 1.7 billion in backlog in addition to the discussions we continued to have with our customers, we were able to guide a 2.4 billion first half but that is the extent of visibility and we are just not in a position now to really know what the second half might look like to the course of this discussion we can talk about our view on the different sectors if you would like but maybe I will leave at that for the time being and we will proceed with your next question.

Simon Schafer - Goldman Sachs

And as you said, I have somehow actually I am like many other Semicap businesses given your lead time, you actually have given some of this is, it's made to manufacture you know, people know what their strengthens, forecast are roughly, I mean the subject change but broadly speaking people have a good idea. So when you look at your own forecast and your customers project trends maybe just walk us through what you are seeing, one maybe in memory in times of supply and demand, what is that people are projecting?

Unidentified Company Representative

Yes so currently we have models that allow us to given the certain bit growth expectations, so demand bit growth we can calculate the number of tools that are required to support that given the nodes and the way for start at each node that our customers discuss with us and make public otherwise.

So we are able to kind of forecast a system demand based on that, so currently the industry analyst forecast for DRAM bit growth this year are in the low 40s and we know that given the shrinks that our customers have in place or that they expect to do through the course of this year that they can grow bits to about that amount. So there is no need for the DRAM industry to add wafer capacity to grow bits to the current demand forecast.

But it does require them to invest in continued shrink at the rates that they have already indicated to us and to again in some cases publically. In the NAND sector and by the way I am sorry let me finish off on DRAM. If you look at ASMLs backlog today only about 10% I believe is for DRAM. So off our 1.7 billion that’s roughly a 170 million euros worth of equipment and if you think about emerging tools that 40 million euros each it's only four tools.

So, the plan today collective demand plan from the DRAM guys is only about four tools over the course of the next six months that’s not to say that they couldn’t come through with, some orders on a shorter lead time but it gives you an indication that they are set at least for the first half of the year to grow bits by the amount that’s expected.

In NAND our plugin number is a Gartner indicated number, as I think about the 73%, 74% mid-demand growth expectation for this year and as we saw through the course of second half of last year our customers in NAND continue to order and continue to take delivery to that level.

NAND bit growth through shrink alone as planned by our customers is about per our calculation about 50%, so that means to get to a 70% to 75% bit supply growth this year they will have to add N amount of wafer capacity, not as much as they added last year and so our calculation that they added about 300,000 wafer starts last year to grow bits to the mid-70 level.

This year they don’t need add quite down much publically actually about half of that so it is our expectation that they will execute on buy through the course of this year pretty evenly to meet that level of demand.

Simon Schafer - Goldman Sachs

And what about the foundry in a sense that a lot of the spending growth is in my mind anyways happening in essence because of the clash of the Apple Ecosystem and sort of the legacy if you will and you powered to be mostly Intel who is driving a lot of incremental catch up spend and you can either take the way that well that’s just adding excess capacity or you know like Lan (ph) was saying this morning well actually there is enough secular reasons to believe that that amount of spending growth can continue at that pace. So your view on that?

Unidentified Company Representative

Yes from a list of spend perspective I think there is a couple of things to understand, I believe there is real demand and in essentially all of the litho buys are for the 28 nanometer node. TSMC has publically said they are going to ramp this node at twice the rate of the 40 nanometer node or the previous node that appears to be happening we have to assume that is against real demand.

They also announce 30 to 40 design wins, so there is real demand out there certainly Apple’s demand is growing and so they are requiring more investment in capacity from Samsung for example. So there is real demand behind this, you will also have to remain that the foundry model certainly in this atmosphere of heightened competition there with Samsung and TSMC and global foundries competing in parts for the same business.

The foundry model says I have to have process capability but as importantly I have to have an amount of capacity because that’s what I sell, process capability and capacity and if I have the process capability and I am engaged with a customer that wants capacity and I don’t have then I am going to lose it to the next guy who has the process capability and the capacity. So I use the term, certainly in this heightened level of environment and with real demand behind it and investment in strategic excess capacity.

And I think it's really, really difficult to say how much of each customers purchases are in that strategic excess versus the real demand because there are some complicating things out there regards trying to measure that.

Number one there is real increase demand in dollars then due to the dye size increases. So if you take the current product the Apple A4, not even a 28 nanometers, that the 40 nanometers node versus an A5, you have a 15 millimeters square dye in the A4 and a 120 millimeter square dye in the A5 at the same feature size and I understand that what the A6 is going to go to the 28 nanometers but the dye will be equivalent or even larger to the A5.

So double the size and the dye means twice the number of wafer which means twice the number of litho tools. On top of that you have the complexity of the 28 nanometers devices which shows itself in two areas, one in the increased number of layers which requires more litho tools, the nice thing about litho is every layer is imaged not necessarily at your depository foundry implant so that’s the nice thing about litho, the increase in layers directly translates in a need for more litho tools.

On top of that with the accelerated development pace and process introduction pace of 28 nanometers, our customers are confronted with yield issues, it happen every new node, it's to be expected but it could be by some people’s measure and some reports exaggerated at this point in time.

So if you just imagine for the sake of this discussion, if we have a mature process running say conservatively 80% yield probably more but let’s say 80% makes the math easy and you have a new process that’s running 40% yield in the beginning.

You know you need twice as much equipment if there is real demand behind it, to meet that demand. So if you look again the increased number of layers, the increased complexity which causes yield issues, the doubling of the dye sizes in the particular case that I mentioned and there are more examples of NVIDIA and TI products of a doubling of the dye size, so it's not a singular example.

That all argues for a significant spend, so if you try to look at historic spend versus current spend you have just be aware that there are a lot of drivers for a significant increase in spend and let me suggest that it's even going to be worse or better for ASML when you get to the 20 nanometer node where you really start to implement these multi-patterning techniques another increase in the number of litho layers that need to be processed and in the microprocessor guide is a similar example where with the train comes more complex imaging, more layers et cetera.

So again I'd be hard-pressed, and ASML is hard-pressed suggest that there is a is a gross excessive capacity in this time frame. It's very difficult for a company like the ASML to model that by the way because of the large deviations in dye size for example because the dye size is a plugin number to the models.

But let me finish off on foundry by saying it's probably to be expected that the foundry guys take a pause here in the next couple of quarters regards to the delivery of tools and may be orders as well as they digest the capacity they have put in over the last few quarters including the first quarters of this year. It just makes sense as they compare the installed capacity against the real demand against the yield, et cetera.

So we would be surprised if Eric Meurice, our CEO said on our conference call in subsequent meetings at the second half, it's a little bit weaker. But question is it might be you know if there is a pause how long is the pause, a quarter or – and we are not in a position to know that at this point in time.

Simon Schafer - Goldman Sachs

So I will take it that’s the primary reason as to why, unlike Lan (ph), who historically actually it's bad worst visibility than you. You have decided not to get full year guidance just because the foundry spend perhaps is an area of…

Unidentified Company Representative

Actually all of the spends, probably the most well understood we think spend could be in the NAND space, and so uncertainty about the DRAM in the second half and largely that the memory spend this year will be predicated in part on their view of 2013 demand.

So as we get towards the middle of the year it's likely that the memory makers all of them will start to get a view of 2013 whereby they will have to invest in anticipation of growing demand if that’s what they see or conversely not invest if they don’t see the demand.

As we saw last year specifically in the NAND area there was considerable amount of capacity added in the second half in anticipation of need for this year. So, that’s kind of the way the memory works you know they have to build capacity at the very end of the year to be prepared for the at least for the first half of the subsequent year.

So again our inability to guide the whole year is strictly related to the lack of visibility given to us by our customers and so they just don't know at this point in time.

Simon Schafer - Goldman Sachs

Makes sense, and I wanted to go back to this discussion of perhaps growing capital in terms to, I think majority is out there as to whether the overall industry as in Semicap is seeing growing capital intensity but I think it's certainly true to say that about litho, I mean the number speaks for themselves you know over a 10 year period maybe litho was in the 19% - 20% of equipment spend and that’s probably because of 25%.

But according to what you just outlined you I guess I think that number is still going up. So do you have any sense as to whether that is 30 or is it 35, how much pricing power do you have, you almost have a duopoly and not…

Unidentified Company Representative

Yes of course that’s a function of the revenues from our customers, but in terms of capital intensity versus semi-conductor revenues. No, I mean it's very difficult for us to try to suggest to what level it would go but it is clear that again that the intensity in litho especially in the logic area is going to be intense. The other thing to recognize, we sell a very basic critical layer tool basic, I mean it's a tool that doesn't know what it's imaging, so the same tools is used essentially for all the different application. There is no real customization in the basic tool, so the 40 million euro tool is usable by NAND, DRAM and logic, but process complexity of the logic, I will use the word demands that the customer load these tools up with very significant value adding software and hardware features but the NAND guy does use.

So I have seen purchase orders for a NXT:1950i immersion tools that have exceeded 50 million euros and again these are loaded with the with the Brion software options which help in process definition, process window creation and process control as well as some hardware features like our advanced illumination system that again allow tuning of the process in time.

So the point of being the logic demand grows because of the more features, the logic ASPs for our tools grows because of the same type and difficult process requirements.

Simon Schafer - Goldman Sachs

Let’s talk about that sort of complexity a little bit more because one of the big steps and actually in my mind one of the biggest steps the overall manufacturing industry is facing is this at 70 nanometer or somewhere in that ballpark it actually becomes pretty difficult. It is not difficult now it certainly is becoming very difficult sub-17 to manufacture with traditional litho.

So one of the things you as a company in your ecosystems and work knowledge in EMV but the problem has also been that perhaps the throughput but laser is achieving has been much harder than people would have expected, it's actually is caused due to take a lot of the risk in the R&D resource back into your own hands and perhaps change some of the R&D resources from your suppliers back then.

So maybe talk to us to some of these challenges and how you can overcome them and so on.

Unidentified Company Representative

So certainly there is a recognition by the industry of the limitations of 193 immersion tools at roughly the level you suggested, there are some other drivers in NAND that require that architectural changes take place because floating gate it doesn’t below the level you suggest so there is two things going on in NAND, one is – there is an extendibility of the current processes and litho down to about the 15-16-17 nanometer layer but they are significantly expensive but extendable.

So the driver for them in the near-term let’s say in the next couple 2, 3, 4 years is more of an economic driver that it is I am talking about just from the mature litho perspective not talking about the need for an architectural change beyond that point in time. But it is an economic driver, the logic guys on the other hand are driven and the DRAM guys are driven by technical drivers, so in DRAM today one of our DRAM customers is telling us for their next node which is the 20 nanometers node that they require EUV for a couple of layers and it's not an economic issue it's a technical issue. They haven’t apparently found a trustworthy if you will way to do the couple of layers otherwise.

So that technical challenge can be met by EUV potentially EUV alone so that’s the driver and that’s the driver and that’s the nearest-term driver we have for it appears it today for the adoption of EUV and the general logic and when I am talking about logic I am talking the broader logic like the foundry guys are doing 28 nanometers today with the some techniques in single layer and double layer patterning but they are having to constrain the designs in many cases of their customer.

So I can print this 28 nanometer feature Mr. Customer but you got to give me X amount of space around it, and more space than you'd like to add as a very gross example of the design constraints that are placed on these guys, well okay I have no choice so I'll design that way but I would much prefer that I will be able to design the way I want and for you to be able to print this the way I want and the solution for them, then after 20 nanometers and certainly in the regime you're talking about in the mid-teens is per-EUV, you know probably the logic guys will have to do the next node, the 20 nanometer node with this complex double patterning techniques with the design constraints. But they are looking for a relief from that in subsequent time frames.

So today you are right, we have delivered six immersion tools all of them being used in development activity at five major customers and one major research facility in Europe today which means the hardware that we delivered is doing what it promised in terms of imaging and overlay requirements such that the customers can do process development.

The fact that they're delivering only five or six or seven wafers per hour is not so problematic in the near-term from the development standpoint but it suggests something about a capability in the future. So really the next I will call it milestone if you will separate from the development milestone of power from the source suppliers is this year when we deliver our capability whatever it might be to the customers in order to determine what the next batch of orders is going to look like. So we’ve shipped six used in development activity with one customer for over a year with another for just a matter of a month or two at this point in time.

Processes are being developed and decisions need to about implementation of EUV in the near term memory processes that I mentioned before in the middle of the year, why? Because we have to prepare and we need a minimum of year of preparation time for the supply chain, so we have delivered six and let me now remind you that we have eleven orders for EUV tools to be delivered starting in the third quarter of this year.

Those 11 tools will be shipped, I will say for the sake of this discussion regardless of the throughput capability of the tools and so those are committed, so the first available tools if you will are not until the late and the third quarter of 2013. So what we have to do mid-year this year again is determine what the demand will be from third quarter 2013 on let’s say for the first half of 2014.

And whatever we are able to demonstrate in the middle of this year will determine the level of orders that we get from the industry. So again we are still targeting for the 60 wafer per hour in the middle of the year, it's challenging, it's extremely difficult stuff and the unfortunate part is nondeterminent, so when we are asked what kind of progress we are making, well we can be great progress through a week or two weeks or a month but then we could encounter something that will take us another month to solve.

But there is nothing fundamental today that says we can't get to the 60 wafers per hour and then eventually to the 125 wafers per hour that we‘ve promised the industry but a critical time will be in the middle of the year when we sit down again with our customers with the data that we have and present that to them in order for them to feedback to us what their requirements will be for the second half of 2013. It will be determined based on the throughout capability we demonstrated at that time.

Simon Schafer - Goldman Sachs

And sort of a broader question actually on some of these challenges in the sense that unlike Japanese competition you are actually not vertically integrated, you have done only online supply, you certainly don’t own stakes or participations in the laser guys and you work very closely but with 2 billion in net cash and I guess a broader appetite in the industry for consolidation what’s ASMLs appetite either participating or broadly yours to perhaps consider becoming more vertically integrated just because some of these technology challenges are really becoming that much more apparent.

Unidentified Company Representative

Yes I think to the specific question about vertical integration of a source supplier, I think we are content at the moment with the relationship we have with the source suppliers, We are heavily involved with both Zimmer and (inaudible) in helping them to understand the issues and solve whatever problems that we can. So maybe we were a little late in understanding the complexity of the tasks that each of the source suppliers were confronted with their size and these both certainly Zimmer is a relatively smaller company and the focus today is clearly on Zimmer for a number of reasons mostly related to configurations that match our 3300 but yes so we were probably a little bit late in recognizing that we could contribute in a meaningful, so today we are fully engaged with both suppliers both but specifically Zimmer.

And I think we are content for the time being to have this kind of virtual joint venture where we have people in their facility and they are accepting of our help and our support. ASML our system integrators in these big sources now are involve a lot of system integration. So just by a way of an example we have taken over the relationship with their suppliers which happen to be suppliers to that we have a history with. They happen to be European suppliers so they are in the same time zone and in some cases speak the same language or near the same language so that makes it easier so we have taken over some of those things just as an example so there are three European suppliers the key suppliers that we are in front (inaudible) regarding the relationships but so outside of that kind of acquisition activity potential we continue to, you have heard this many times before and we haven’t found the magic thing to do but we continue to search for things that use our core competencies and that can create a growth opportunity for us in the future.

There's nothing again near-term, we have an idea about one but we want some cooperation from customers and these customers that we've identified are not in a position today to make any significant commitment towards these projects, so we will put them on the back burner but otherwise we have a lot of growth opportunity ahead of us. This year is going to be a year of the product for ASML no doubt about it.

The market will deliver to us what it delivers and we are not overly concerned about that we are prepared you know with a production capacity of somewhere between 1.4 billion euros, 1.5 billion euros per quarter.

We have demonstrated that last year, so we are able to certainly go to that upside and we can manage cost to certainly to a 1.2 level where we are today, easily we have a lot of flexibility in the organization to go with that flow, so the focus again for us is on execution on the continued improvement, performance enhancements to the immersion tools which aren’t going to go away by the way.

I think some people were concerned couple of quarters ago when we said we are going to employee this dual product strategy and you know as I think a lot of people took that as we were concerned about the progress we are making in EUV and whether EUV would make it or not that’s not truth, that’s not why we did it, we did it because we perceive the concern relative to us and our competition in the immersion and we wanted to let you all know that we have a continued roadmap of performance, improvement certainly an overlay and in throughput that make ASML a moving target in this area.

So whatever I like what Tim Cook said today in a session he said regarding the tablets, people are developing the tablet against our iTablet 1 and we are developing Tablet 2 I mean it's the same situation that we have with our competitors in this space so now we are moving target and we intend to be and we always have been, I mean we are always with all the products that I've ever seen us launch and I have been with the company 20 years, we've always had a roadmap for continued performance improvement in fact we highlighted that a couple of quarters ago creating the amount of concern about EUV but don't take it that way and we are obviously fully committed and believed strongly as the rest of the industry does in EUV but we recognize also that immersion products will be used for the next 5, 10 years at least in significant way so it's important that we meet the needs of our customers in that area as well.

Simon Schafer - Goldman Sachs

And my final question actually would be, well if you are right on EUV will work and even some of the what litho is going to be running the right product a long time really and go scale manufacturing, you got 2 billion in net cash now you don't really want to redeploy on acquisition at some place or vertical integration. So what happens to that.

Unidentified Company Representative

Actually the policy regards excess cash above our nominal 1.5 gross cash target is return it to shareholders through a combination of dividends and share buyback. So we propose an increase in the dividend of 15% to 46 euro cents I believe in it so that will involve a couple 100 million in payout in dividend this year, last year we bought we announced a 1 billion euro buyback program, we bought back roughly 700 million plus again and little bit lesson than 200 million so we paid out about 800 almost 900 million last year and it's our intention to do the same.

We announced with the last quarter, results there was a limitation on what we can do from a tax-free share buyback that's related to the dividend payments. It will be I think about 300 and some odd million this year so we increased the buyback amount to 1.5ish billion and so will buy back the amount that tax free and we also reminded that you available to us this synthetic buyback, is which is kind of quirky kind of thing but we have done it before in 2007. It's an efficient way to return cash to shareholders, it's essentially a capital of repayment in a reverse split and we can do that again so we wanted to let the markets know that we had vehicles whereby we could buy back in amounts of the tax free amount allowed by Dutch law so no change in that again anything that we have in excess of that amount will be returned to those two vehicles.

Simon Schafer - Goldman Sachs

Great. Well actually it's a great talk, thanks so much. I think we are out of time, so thanks for coming and thanks for doing this again.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to All other use is prohibited.


If you have any additional questions about our online transcripts, please contact us at: Thank you!