ASML Holding N.V. (NASDAQ:ASML)
Credit Suisse 2012 Technology Conference Call
November 27, 2012 5:30 pm ET
Craig DeYoung – Vice President, Investor Relations
Satya Kumar – Credit Suisse
Satya Kumar – Credit Suisse
Our next presentation is from ASML. From ASML, we have Craig DeYoung, VP of Investor Relations, and also Pete Convertito from Investor Relations. Craig’s going to make a presentation. We’re going to save the bulk of the questions to a breakout. There will be a breakout. The breakout will be in Room G, but there will be a few minutes at the end that Craig’s promised for folks in this room who are not going to attend the breakout.
With that, let me turn it over to Craig.
Yeah, thank you, Satya. Thanks, everybody, for your interest and coming this afternoon. I’m going to try to kind of go through this quickly, but this does anticipate hopefully a lot of your questions. So I thought we’d get through this, and then, as Satya said, leave a few minutes at the end for any other questions.
Got a big, long Safe Harbor now, given the fact that we’re involved in an acquisition, so I won’t read it all to you, but there, it is what it is. Just a quick business summary, talk a little bit about the merger agreement with Cymer, a business update, and then our environment that we’re in, little bit of technology status, and then an outlook. So this is, again, you’re used to this, this is our normal template for the quarterly results, but we’ve indicated here total revenue at about €4.7 billion for the year given our €1 billion, roughly €1 billion guidance in the fourth quarter sorry.
So I think it’s interesting to note for the moment that this will be our second highest revenue year in our history, so it’s not a bad year for us at all. Just by way of where our business is going to come from at least over the next six months 75% or 76% of our backlog is deliverable in the next six months. You see that down in the lower left-hand corner. We always give you that as a measure of the level of certainty that we have in the backlog.
An interesting breakdown. We still have, I think it’s kind of curious 20% of our backlog are almost €0.25 in backlog in both NAND and DRAM, so we are, even in this environment, delivering small quantities across at least three customers in this particular space. We expect that, a maintenance level of business in memory, going forward. Foundry is at foundry in idea, meaning logic total obviously is dominating the end used pie charts here in terms of the backlog, and we expect that to continue.
Just to remind you that we’re involved here in a reverse bid capital payout as it relates to our co-investment program, but we’re also in the background doing €160 million worth of share buyback. This shows the value of the buyback to date. So we announced a year ago in January a €1.13 billion buyback, and we’re still executing on the final portion of that. And again, this quarter it will be €160 million, and that buyback program will be complete at that point in time.
Our cash return policy remains the same. Everything over gross cash of €2 billion will return to investors through a combination of dividends and share buybacks. This is a wordy document, but I wanted this here for the record. We get a lot of questions about the status of the co-investment program. So just to remind you, we have three new investors, Intel, Samsung, and TSMC, and we’ve received about €3.9 billion in share purchases from them. Along with that, we’ve gotten a commitment for €1.83 billion in R&D investment over a five year period.
Actually, as we speak, we’re determining what the incremental spend will be based on their contribution for next year. We’ll be announcing the details of that probably we will with our fourth quarter results, again what the incremental spend will be and what particular projects it will be against in order to meet the needs and requirements, desires of these three co-investors.
We’re proceeding now, as I mentioned, with the capital repayment reverse split, so we will be paying for every €9.18 per share. Translated today, the rate was set – I don’t remember, Pete, I don’t know if you remember what it is, but the DR Group from JPMorgan put out an announcement today. So those of you that hold New York shares will be paid in dollars at the conversion rate that was announced today. So if you have trouble finding that, contact myself or Pete, and we’d be happy to send that to you.
And again all the record dates and ex-dates have been announced, but we’re in the midst of it. It is a little bit complicated in that the ex-date for the ordinary shares was Monday, and the ex-date for the New York shares, based on the way that the two exchanges operate and operate differently is a different date, and it’s actually midnight the 28 of this month, just for your information.
Again all of that is available also through JPMorgan, who is our depository receipt manager. So we’re on our way doing that. I think by December 3, it will be done. The share count going into the co-investment program and coming out will be the same at about 410 million shares.
In the meantime, the EPS share count will be about 450 this quarter due to an averaging, because we shared before the split. So again it’s the intent was to make this thing non-dilutive, and in the end it will be. And then also the shares will be reduced by the 160 million in share buybacks, we’re doing such that Q1 is expected to be at about 405 million shares.
So the Cymer merger agreement, we announced with the Q3 results our intent to acquire Cymer. The idea here is that we’re able to maintain and hopefully accelerate the development of the EUV source. We are delivering and will deliver 11 of the systems next year, and those systems are performing according to specifications on imaging and overlay.
But as most of you are aware, we’ve been troubled a bit by the development of the source power required to provide the economic incentive for the broad base of customers to use EUV. So this acquisition is focused mainly on bringing EUV forward as quickly as we can. Some of the co-investment R&D dollars are also focused on the near-term and long-term source power development, so whole lot of work to do there. But in the acquisition, there also will be around EUV and around the main business, a lot of efficiencies to be gained, not many of which we can talk in detail yet.
The merger, the acquisition, should be complete probably not before the end of Q1. It could extend a bit based on regulatory approval processes, which are not that predictable in terms of timelines. They are political processes in part, so it’s a bit hard to determine the exact timeline on these. But we don’t anticipate significant issues with any of the regulators, wherever they might be in the world. But it is a process that we need to go through, and it’s essentially just started with most of the regulatory agencies.
The transaction details, if you’re not familiar with them, it is a cash and share offer worth about $2 billion. So every Cymer shareholder will receive $20 in cash and ASML shares at a ratio of 1.1502 per share of Cymer. It’s for all their operations. Again, we also need shareholder approval, of course. I think that’ll take place, I’m not sure if it’s January or February, but that’s on the docket. And then, of course, we have the regulatory approvals.
And as it says in the last bullet, we do expect that it will be in the second year after the closing that it will be accretive on a non-cash purchase price accounting adjustment basis. Business environment, so currently, our fourth quarter revenues, as we discussed before, will be about €1 billion, that’s in service and system revenue. That’s supported by stable demand from the foundries. And the expectation, going forward, is for a nominal, or let’s call it a maintenance level of spend from the memory sectors, both NAND and DRAM.
For 2013, we expect, again, large part of our business be driven by the Logic Foundry Group, actually foundry and microprocessor. We have a good feel today for what both of those sectors will be spending next year based on their forecast to us, which is kind of a normal process that we go through at this time of the year.
There’s some levels of certainty within some Logic customers who are awaiting clear forecasts from another one of the foundry players. But in the end, we expect that our business will be largely driven by these two sectors, or the combined sector, Logic. The expectation for memory, as I said, is not high. Our estimations, our calculations are that the DRAM players collectively with the installed capacity at the different nodes by somewhere between 35% and 40% next year. So bit supply based on the current capacities can be at that level.
I don’t think there is a forecast today that exceeds that level. So the requirements for next year from our view are minimal for the DRAM guys. Again, it’s just a maintenance level of support, filling out certain completing transition, node transitions, and filling out some spots in capacity, but a very probably in a minimal spend for the sake of this discussion.
And then slightly different in that we expect at least for the next couple of quarters, that the NAND customers collectively don’t need to spend beyond a maintenance level. Our estimates are that they can grow capacity with the installed wafer capacity at the different nodes by somewhere between 30% and 50%, and that large variation is dependent upon the assumptions you make about the bit per cell assumptions.
So with the higher three bit per cell assumptions, you can grow about 50% with the lower ones about 30%. I don’t think there is a demand forecast today that is really above the 50% level. So again for the next couple of quarters, we don’t anticipate any level of significant demand from the NAND players. So in other words, they can go a couple of quarters at least for that in an investment.
If bit demand is going to be above this, let’s say 50% level then we would expect it would trigger capacity additions within the collective NAND space. So we’ll have to wait and see. Nobody again really knows what the NAND demand will be. There are some people that have relatively high expectorations for NAND at the latter part of next year maybe into 2014. But again we’re not in the business of predicting bit demand. We’re just able to calculate, based on demand expectations, what kind of litho requirements there are.
In terms of the technologies, as I mentioned, we have the NXE-3100s. Those are the first volume tools, the systems we shipped six up to the field. They’re in use today, printing wafers in the development mode, in a learning mode, for five of the top IC manufacturers in the world, and one industry supported R&D facility in Belgium.
As I mentioned, we’re on target to deliver 11 of the 3300, which are the production tools. Now, the total of the 11 are still going to be used in development efforts. None of those will go into production next year. There are eight customers that will get the 11 tools, so you can imagine that there’s a couple of guys or three guys that are going to get two of these tools for their development work.
The image in the overlay, as I mentioned, also are exceeding our expectations. The performance is excellent on the lens and the system in total. We’re currently demonstrating 50-watt capability at the supplier. We have 30-watt capability in our facility. As I mentioned, the target is for 70-wafer per hour in Q1-ish timeframe in 2014 to meet an early adopter, DRAM adopters’ requirements for tools for a 20-nanometer process.
So that’s the next real milestone for us. We’ll do our very best to update you on progress, but week-by-week or even a month-by-month update is impractical in a lot of cases. So bear with us if you ask a question about where we are and we’re non-specific about it. It’s not that we’re not working. It’s not that we don’t know. But again, to translate the performance and the improvements that we’re making into a wafer per hour improvement is sometimes quite difficult.
By the way, I must say that our customers are focused on us getting to the 70 per hour for production requirements in 2014. It’s not that they don’t care about what the throughput might look like when we deliver the 11 tools, but they recognize that these are R&D tools. And for them, the key performance metrics are in the imaging and the overlay, which allows them to develop their processes. Again, throughput is secondary. They need to know for sure, though, that we’re making progress in the background of what we deliver to them.
So they’re showing a level of patience as long as we’re communicating performance improvements in time to them. Just so you know, or this is exciting to our customers. It may not be exciting to you, but what this is, is this is a 14-nanometer node device printed in one image at one of our customers. So today, for you to do this without EUV would probably take about three different masks, or triple-patterning, so a litho etch-litho etch-litho etch, which is quite expensive, and it introduces a lot of yield challenges, as well.
So again results for you may not be that exciting, but for our customers, they’re quite exciting. From an imaging standpoint in the bottom left-hand corner, that’s a distortion map, or an overlay map. So it’s just, it’s showing performance that is quite acceptable to our customers for their most advanced nodes, let’s say, and Logic to 14-nanometer node.
Seven machines are in the buildup right now against the 11 that we owe, so just for information, this is real stuff. All of these, if you were to visit our factory today, you can see these seven tools in different stages of build. The EUV scores, I already mentioned that we are going to ship 11 of these tools, We expect to recognize revenue against these 11 shipments of about €700 million next year.
The important thing I think for you to watch in terms of our performance, against the throughput requirements of our customers or the source power developments is really in the orders we get from our customers, because our customers are communicated on an ongoing basis about the status of the sources in detail, and it’s based on our communication around our progress that they are willing at a certain point in time to place orders.
So as I mentioned, one big DRAM guy has made a commitment to us for four systems for delivery in 2014. We’ve made a commitment in return for 70 wafers per hour. We have another four to eight systems that are expected to be ordered through the course, probably over the next couple of quarters, and our progress against the 70 wafer per hour will be very important in triggering these eight additional orders. I’ll mention some of those are for DRAM and some of those eight are for early adopters in the foundry space.
So make no mistake, these guys are at our doorstep almost every morning asking for an update on where we are with this. They desperately need EUV to maintain Moore’s Law, maintain reduction in cost per function. So they’re confronted, especially the Logic guys at the 20 nanometer of the foundry guys with an ever increasing cost total, but it’s specifically with litho cost increases.
So we estimate to go from a 28 nanometer node to a 20 nanometer node, you need about a 1.7 times increase in the cost of the litho systems to support, so it goes from about €650 million for 45,000 wafers starts to almost €1 billion for 45,000 wafer starts from the 28 nanometer node to the 20 nanometer node, and that’s getting to a point where some of our customers’ customers are reconsidering their product rollout plans, and that’s not something that we want to have happen and they certainly don’t either.
So with that Satya, I think we’ll go ahead and have seven or eight minutes for some questions.
Satya Kumar – Credit Suisse
Are there any questions from the audience?
So the question was if we see a delay in EUV and the industries confronted with this escalating cost of lithography, what will they do? Will they move forward or will they pause or wait and what might happen? There are alternatives, but they’re costly alternatives. So it appears that the foundry customers are prepared for the 1.7 times cost of the 28-nanometers, so they are moving forward, some of them with some aggression. So, apparently, they have customers that can create enough value in the device that justifies the incremental costs or the additional costs.
So I think it’s on a, but then you see some of our customers’ customers that are saying we can’t, we’re not seeing the cost per function reduction that we’re used to, so they’re not embracing the 20-nanometer node in the first instance. 20-nanometer, which I think the industry has accepted that the additional cost and the solution being this multi-pass patterning, double-triple patterning, will really get into production in 2014 going into 2015.
So the real requirement, if you will, not that they wouldn’t use it if they had it today, but the apparent acceptance of the increased cost of the 20-nanometer will get them and us through the 2015 timeframe. So and we fully anticipate that we’ll have high volume, high throughput tools, but in that timeframe. So it will meet their sub-20-nanometer timing requirement the best we know right now.
So then you can talk about what happens if it doesn’t, and I’m not sure all the answers to that. It will vary by customer and by customer of our customers, but I think we’re okay for now through the 2015 timeframe, or it seems to be an acceptance of the cost, or an ability to justify the cost based on the value of the devices.
Craig, as we are waiting for another question, I guess, there is a lot of interest in knowing how the model’s going to look like, I guess. I know you mentioned that you’re going to talk about the R&D part in January. But just from a framework standpoint, how will the monies that you get for R&D from your customers flow through your financial statements?
So we expect that the incremental spend will probably approach, but not exceed, about €200 million, so €50 million a quarter, so our base spend if you will, ASML’s, is about €150 million a quarter today or €600 million a year. The incremental spend in the first year probably won’t exceed €200 million. And again, we’ll define that a little bit more precisely towards the end of the year as we again, set the milestones, set the deliverables et cetera, et cetera. From an accounting standpoint, the Samsung and the TSMC investments are fairly straightforward. So that cash will be delivered to us, go on our balance sheet, and be there to offset incremental spend for those two in the R&D line.
The Intel accounting is a little bit more complicated based on U.S. GAAP accounting rules. And so what it will look like is, because Intel with their ownership of ASML being above 10%, they’re considered in accounting terms, a significant related party. And when you’re involved with a significant related party in a multi-element transaction, another accounting term, I believe. So remember, we’re involved in the R&D investment, the equity deal, and our ongoing contract to deliver systems that we’re delivering today. So those are the multiple elements.
Then, from an accounting standpoint, you have to consider that all is one. So what will happen is the cash we get on the co-investment program from Intel will go on the balance sheet. It won’t enter the P&L, or it will enter the P&L as an offset to discounts that we give to Intel on a normal ongoing basis. And it’s based on incremental tool sales. So I won’t even build a scenario for you.
So whatever discount we give them, the money will come off the balance sheet as an offset to the discount, which will raise the price of the tools to them, which will create a higher gross margin for us on those tools and overall, which then is intended to offset the incremental R&D for them.
Satya Kumar – Credit Suisse
So ballpark, I guess Intel represents half the R&D contribution or a little over?
Over the five year basis, I don’t know it will be next year. But it will probably be at least that next year.
Satya Kumar – Credit Suisse
So the €200 million that you talked about, so I guess that Intel part of it for next year will be incremental to revenue line and obviously that doesn’t R&D. The €200 that you talked about is in addition to that setup?
No, the €200 million will be in addition to our €600 million. So we’re at roughly €150 million per quarter basis right now that the base R&D. The incremental R&D will probably not exceed in the first year €200 million and that will be the collective contribution from the three co-investors.
Satya Kumar – Credit Suisse
And a part of that, not all the €200 million is going to show up as a net increase in R&D, right?
Show up as a what?
Satya Kumar – Credit Suisse
That €200 million is not…
Will show up as an increase in R&D, and will be offset by the cash contribution from the Samsung and TSMC. The other will work its way through the P&L as I described.
Satya Kumar – Credit Suisse
It’s a little bit contorted and complicated. The hope is that next year, our CFO Peter Wennink, said he hopes to make this P&L neutral next year, but just so you know the mechanics of it. So it may not that way on a quarterly basis, because it will be dependent upon incremental shipments to that one large customer, and we don’t determine the shipment timing to them.
How should we think about gross margins, then, as we progress through 2013 and EUV revenue kind of mixes into the P&L?
Yeah, so we’ve stated before that the EUV gross margins coming out next year will be in the high 20s, so probably easiest to look at the base business and the EUV business. So the €700 million will carry roughly a 27%, 28% gross margin. The base business at the €1 billion level where we are today is, what have we got, 41%.
So let’s assume for the moment that we continue on through the course of next year at the €4 billion rate, I haven’t done the math, Stewart, but you can, if you wouldn’t mind doing that, it’ll be the combination of that. I don’t know what it is, yeah.
And by the way, while we’re waiting for this question, it’ll probably take us about a 24-month period as stated by Peter Wennink a couple of quarters ago to get that 27%, 28% gross margin up to the corporate average. And that will be through a combination of value enhancements to the system, which would drive the ASP up, but also through a, which is our normal redesign for cost. So every time we put a new system out, the most important thing is time to market, and then, through time, we redesign for cost.
Two questions. One is on ASML’s market share, how’s the change versus Nikon and Canon over the past year. And then, the second one is for 28-nanometer and 20-nanometer, so I guess many people say there’s only one, for the foundry side, there’s only effective 28-nanometer provider today. And I’m curious, how many other foundry customers of yours can make it to 28-nanometer, and then what will the number be in 20-nanometer in terms of the ability of these firms to continue down the path?
I’m sorry, I forgot the first question already.
Market share, yes, between Canon and Nikon.
Oh, market share, sure. Yeah, our market shares have probably been stable over the course of this year and last year at, I’m going to call it 80%. I don’t know what the VLSI or the other guys have measured at, but we estimate it to be at roughly 80% total litho share. Recognizing that it’s higher in the more advanced tools and a little bit less in the non-critical tools, but overall it goes about 80%.
I would say that on the 28 nanometer node, there are four foundries that have some level of 28 nanometer capacity installed. I’m not in a position at all to talk about the competitiveness of those processes, but there’s four involved. I think, publicly across at least three customers have talked about some level of 20 nanometer capacity additions sometime next year in 2014.
Satya Kumar – Credit Suisse
I think we need to wrap there. There is a breakout session in Room G.
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 www.SeekingAlpha.com. All other use is prohibited.
THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.
If you have any additional questions about our online transcripts, please contact us at: email@example.com. Thank you!