Some excerpts from Lam's Q1 CC of a few weeks ago. A fair bit on 3D, as Lam is providing a lot of important tools for it. My bolding.
seekingalpha.com
Martin Anstice opening remarks
Still, today, the NAND segment represents the most difficult to predict and largest swing factor of spending. But this is largely a debate about the final quarter of this year and the first quarter of next, we believe. Similar to DRAM, we're seeing a couple of NAND manufacturers pull in the timing of their next-generation investments. We continue to believe that the demand for flash memory is an attractive growth opportunity in the industry, for our customers and for Lam. We have essentially maintained our full year calendar '13 view of NAND's WFE spending at this time as we still expect manufacturers will be receiving a similar level of capacity, as anticipated in our last call.
The dominant capacity additions are playing up, but relative to 3D NANDs, we continue to project initial capacity shipments occurring towards the end of this year....
--------------------------------------------------------------------------------------------------------------------
We believe that we are well positioned for these opportunities once production buys begin later this year and anticipate that our market share in etch multi-patterning will be at least as high as our average market share position overall. Similarly, we've established strong positions in critical dielectric etch with NAND manufacturers to support developments of their 3D NAND's devices. These vertical NAND structures are built up from 24 or more layers of alternating films, which creates significant challenges for etch suppliers to deliver uniform processed results at high productivity. Lam has demonstrated the ability to address these challenges, which supports our confidence in expanding our dielectric etch share.
==========================================================================================
Q&A excerpts
Timothy M. Arcuri - Cowen and Company, LLC, Research Division
Two things. First of all, obviously, there was a pretty big memory pull-in into June. If I sort of back into what the overall WFE pull-in was using your share, it's probably like $1 billion worth of WFE. It looks like it might have pulled into June. But yet you didn't increase your full year number. Is that just due to the Intel cut? How should I think about that? And then I had a follow-up.
Martin B. Anstice - Chief Executive Officer, President and Director
Yes, I think in the context of memory, we, for sure, don't have all of the answers to the question, Tim, but there seems to be a remaining consensus around 25% to 30% bit growth for DRAM and the kind of mid-40s to high-40s level for NAND flash. And that's kind of the governing condition for capital. And so lacking any data points around a demand curve, we felt like it was appropriate for us to kind of keep the WFE number for memory essentially where it was previously. And things change and that gets stronger, then clearly, we'll revise that later. But we didn't have a basis to necessarily conclude life was going to be stronger than we previously anticipated.
Timothy M. Arcuri - Cowen and Company, LLC, Research Division
Got it, Martin. And then just a quick question on the timing on 3D. Some of the big producers of NAND have recently said that there's not going to be meaningful 3D production for another 2 to 3 years, yet you're seeing some orders begin to happen and some production that might start to ramp the end of this year. So can you help bridge that gap? Is there just differences in timing among all the producers? Because there seems like a big difference, and it really sort of underpins some of the thoughts around just the NAND industry.
Martin B. Anstice - Chief Executive Officer, President and Director
Yes, I -- we definitely expect and I think the industry expects to receive orders and ship product into, at least, 1 3D NAND fab this year. But I think you described the situation quite accurately. There is a reasonable divergence of timing, as well as ultimately, process flow decisions for our customers. And that doesn't help any of us in terms of kind of answering some of these questions concisely. But uncertainty is there. It does appear there's a reasonable timeline difference between first adopter and last adopter.
------------------------------------------------------------------------------------------------------------------------------
Martin B. Anstice - Chief Executive Officer, President and Director
Yes, I think the answer to the last of that question is not really. I mean there's decent amount of SaaS space, whether it's in the right location at the right time, is obviously, an important question. But I don't expect there to be a clean room limitation, per se. And we're assuming in the range of 100,000 wafer starts of capacity being added this year and some part of that is kind of a very original investment, new capacity and some part of it is a transition from the DRAM space to NAND flash. If you look at the spending, so we're assuming a little more than $6 billion of spending for WFE. The additions we would estimate around the $3 billion level and the conversions we would estimate around the $3 billion level. So that's kind of a math of the data points we shared previously. I think the demand statement for NAND flash is a very positive one and certainly I would expect it to be more positive next year over this year when you look at the tablets, when you look at content, when you look at the smartphone, when you look at SSDs and the intercept points and generally assumes penetration for those products. We're going to end the year ready tight again. I mean, if these plans play out, there's not going to be whole bunch of anything sitting there waiting to be utilized and I think inventory levels generally have kind of gotten to a normalized level. So it would seem rational to conclude that spending in the NAND flash space is more likely than not and time will tell again.
Vishal Shah - Deutsche Bank AG, Research Division
Martin, I just wanted to clarify the comments you made on NAND flash 3D. I think you said previously that you expect 20,000 wafer starts at 3D capacity to be added in the back half of the year. Do you still expect that sort of magnitude of capacity added in the back half? And then, when you make your comments about OpEx for the year, do you assume Peter Wolters in that number? Or does it exclude and how should we think about that business -- I think you were looking to split that off some point this year.
Martin B. Anstice - Chief Executive Officer, President and Director
Yes, until it's not part of the company, it's part of the guidance of the company. It's not the most material components of the financial statements so you might want to remember. So relative to -- what was the first part of the question, now it's gone?
Unknown Executive
3D NAND.
Martin B. Anstice - Chief Executive Officer, President and Director
Sorry, 3D NAND. So I would say the 20,000 wafer starts, plus or minus, 5 is a reasonable assumption today and when we make those statements, we're talking about capacity we're shipping, right? So what form it eventually takes in terms of being available to service demand, there's a bunch of questions still remaining on that. And that's about qualification, processes, and it's about the very specific decisions of the customer that we don't clearly influence.
-------------------------------------------------------------------------------------------------------------------------------------
Our next question comes from the line of Jim Covello with Goldman Sachs.
James Covello - Goldman Sachs Group Inc., Research Division
Martin, sort of -- we've been around this issue, but maybe not addressed it directly. So you guys and ASML have pretty much articulated the 100K plus NAND wafer starts in 2013. People like Samsung and SanDisk and HInext have all said they're not going to add any wafer starts. Is that just you think those guys not going to show their hand, which I know happens sometimes? Or is there any reason for the delta there?
Martin B. Anstice - Chief Executive Officer, President and Director
Yes. I'm experienced enough to know that answering that question directly gets me into trouble. But so you got to make it what you can. I mean, we are doing the best we can to articulate the world as we see it. Which doesn't mean it's perfect, but it's the best information that we have at hand. I've been very clear and I think consistent again today that the NAND space is clearly the most difficult to predict for the reason that you've just described. And we're doing the same thing, which is to give you the best view that we can and unfortunately, the reason you get the big box is you get to triangulate around our comments.
James Covello - Goldman Sachs Group Inc., Research Division
I wish. And maybe one sort of logistical thing around that. And part of this is, I know it's going to be tough to answer another part. Your industry perspective could be helpful. Somebody like an ASML where the lead times are 9 months, didn't really see a lot of NAND orders, saw very good DRAM orders and maybe part of the answer is DRAM is making up the difference. But if we didn't get a lot of lithography orders for NAND in the first half of the year and the lead times are 9 months, it's hard to understand how you could order the tools in July, get them installed by sometime in the second half and then bring that stuff up online such that you then need to order etch equipment to support that litho equipment. Is there something different about that dynamic or again, maybe part of the answer is that we should be thinking about memory as a whole and DRAM is a lot better?
Martin B. Anstice - Chief Executive Officer, President and Director
No, I think having a conversation around NAND is just fine. I mean, I -- there's a decent component for the industry of conversions, right? I mean there's about 0.5 million wafer starts to conversion likely in NAND flash this year and that's a lot of spending, that's approximately $3 billion of spending. What is very clear to everybody is, as I just stated, the 3D components of NAND spending is going to be probably quite limited at the beginning for lithography generally because the entire scheme is more relaxed. And so the [indiscernible] opportunity I would imagine is quite significant. We're not the experts speaking to that, we are the experts speaking to etch, depth and clean. And certainly, I think there can be some situations where you will not see lithography spending in NAND when you do see other segments. And to be fair, I'll spin it to the other direction in logic, for example, you should expect to see some spending in the lithography space when you might not always see to the same extent spending in other segments. So the segment stories are really important this year. The 3D NAND and NAND generally is beneficial to the etch and depth space nicely and we got some really nice messages, like patterning, like TSV in the logic space as well. But in terms relative to lithography, it's tough to keep up with that success.
===========================================================================================
Mark J. Heller - Credit Agricole Securities (USA) Inc., Research Division
Okay. And then with 3D especially with the first question, there's a lot of confusion. Does it make sense for your customers to have a pilot line, 20K and then it's going to take them some months to figure out the recipe because there are different technologies, there's an IP issue and at the end of the day, that 20 pilot line may not be producing too many NAND chips for quite some time. Isn't that a fair assessment?
Martin B. Anstice - Chief Executive Officer, President and Director
Well, I mean, there's a limit to how specific I can get because there are only -- very limited kind of customer presence in this space. But I mean, it is a very complicated technology, pilot line is an important place to validate that. But I would say, although there is clearly a lot of choices to be made in the industry on process, you can reasonably assume when we're this close to capacity hitting a fab, any customer where that's true is pretty clear what path they're on.
Mehdi Hosseini - Susquehanna Financial Group, LLLP, Research Division
I see. Is it just the customers haven't shown their hands because there are few of them.
Martin B. Anstice - Chief Executive Officer, President and Director
Well, they've shown their hands to us in terms of the engagements we have, but I'm not at liberty to describe them. |
