02.26.2010 SanDisk Investor Day- Eli Harari
[This transcription is not complete. I transcribed the portions that seemed most interesting.]
Eli Harari, Chairman of the Board, and CEO
Eli: … Last year we didn’t have this analyst’s conference. Last year it was very very difficult, to see anyone one of us stand here- Man on fire – talking about how negative things looked.
But we are very very forthright. We faced sets of conditions and circumstances that afflicted us and our industry. We took very decisive action and you will see some of the results of those actions today.
But really what I want to talk about today is what we are seeing for the next decade. I’m not going to focus on this quarter. Even though Judy will talk about 2010. I really want to talk about where we were in the last decade. The last decade was a decade of phenomenal growth for flash. Flash is now everywhere.
Flash is the most pervasive technology. Far more pervasive than any other memory technology and able to create new markets, enable new markets, enhance existing markets, and has done so just across the board. I’ll talk a little bit about that.
But what really excites us is this coming decade. What’s happening now is basically, we think, that the growth that we are facing today with new markets, and I’ll talk about those markets, may very well dwarf what we saw in the last decade.
This is really what we are preparing for. This is what SanDisk is, I believe, as well positioned as any to play a major role. So the theme of today is in the coming decade, flash will be bigger than you think.
And I must admit that I kind of borrowed from Mary Meeker at Morgan Stanley, who said just very recently that in the next decade mobile internet will be bigger than you think. We think that the pace of data and content and applications and cloud- just about anything that is user generated data or content is just taking off exponentially and that is representing a very substantial opportunity for our industry and for SanDisk.
This is really what I am going to be covering today.
[talks about problems with projector]
This is I must admit, an optimistic outlook, but we have been optimistic in the past and correct in the past. It is not like we are just a bystander in this industry. So the agenda today is I am going to really cover first how we did in 2009, but really just in headlines because you are going to have that covered ??? . I will talk about the past decade, get into again, in a general sense,
[talks about problems with projector]
Anyway, talk about technology, competition and our plans for capacity additions. Then briefly talk about intellectual property. Then move onto the growth opportunities that we see in the years ahead.
[talks about problems with projector]
2009 started as you know pretty abysmal.
[Slide 16- NAND Cost Reduction Trends]
… In the coming five years, we expect annual cost reductions to be moderated, to be quite substantially less. So the industry is already 90% MLC- going to 3 bits per cell where we think that the utilization of 3 bits per cell may be closer to 50% because you cannot use 3 bits per cell across the board and the fact that 3 bits per cell over 2 bits per cell has less of an impact than 2 bits over 1 bit per cell.
NAND technology in a very short time will get to 1x. 1x is 20 nm or a little bit under. 1y is closer to 15nm. There is a debate in the NAND flash community, if we can take NAND down to around 15nm.
I personally feel we can, but of course, what do I know? I mean nobody knows to be honest.
EUV lithography will become a critical for going much below 20nm, and that will be a difficult transition for the industry- very expensive and it will take several years. So I feel that the cost reduction ability is going to decrease over time and that will have its consequences in terms of driving new demand and adding new supply to the market.
[Slide 17- Memory Technologies’ Cost Curves]
I want to show you some of the [pause], if you take a look through 2016, where the technology is going on the cost, you can see the red curve, which is the NAND curve, beginning with the dotted part of the curve, beginning to slow down in terms of cost per GB, but still continues. I mean its not that it just hits a wall and stops. It continues, but the slope is much less than in the better years.
I’m showing on this slide, other technologies- how they stack up in terms of cost per bit as far as their ability to disrupt NAND. You have Ferroelectric RAM, MRAM, Phase Change Memory (PCM). A conservative and [said “or”] optimistic two curves and then we have 3D- which is 3D read/write, which is what we are developing- optimistic and conservative- is basically when does it come in.
And then we have BICS. BICS is a different technology. Its a vertical NAND, that has been prepared by Toshiba and also a variant of that by Samsung- that is also under development.
All of these dashed lines, are not for sure. All of these are in development and all of these will take several years to be perfected and good enough to start production.
And then the real test is can they can do better than NAND?
So NAND, as much as it is slowing down, its still a pretty darn fast bullet train. And derailing a fast going bullet train, even if it is not accelerating, even if it is decelerating, its not going at 200 mph, its just going at 120 mph, you try to stop it. Its not that easy.
So, we want to give you an idea of what we think is going to happen in this decade, in that 2011 to 2016/17 timeframe, because I’m not going to cover what BICS can do and what PCM can do. I do believe that PCM cannot meet the challenge to derail NAND. I think that BICS and 3D are better candidates, but both of them have a lot of development still ahead.
Both have their advantages and disadvantages.
[Slide 18- 3D (Read/Write) Memory- Scalable Cross-point Diode Array]
The technology that we are of course banking on, backing on, is the 3D diode arrays, scalable cross point diode arrays, and those, we acquired a company- Matrix Semiconductor- seven years back and [missed words check stream] continuously ??developing this technology.
We now have joint technology development with Toshiba. It is now proceeding in Yokkaichi which is good because we are now beginning to develop structures that are at the same kind of node as the most advanced node for NAND.
We are making good progress with R/W layers which is the key to switching, we are not yet ready for production.
We believe that 8 layers which is 8 layers of stacked cross-point diode arrays, which is equivalent to 8 bits per cell or 8x is going to be feasible and given our tremendous knowledge and IP in 3D diode arrays, we believe that this is the most likely successor to NAND in the coming the decade.
Not only do we think so, we think that once we can deliver the first technology node into production, we will get back into the scaling. Scaling of 3D, we believe should be less constrained than further scaling of NAND. You are not talking about declining numbers of electrons, this is really a resistive memory that does not work on how many electrons you still have left.
Now, if we are successful with 3D R/W, and by all means I say if, because until we do it, we haven’t done it. If we can succeed in that then 3D R/W would with the projected cost structure, will definitely give a second wave on the SSD.
Basically it has the economic [pause] you can kind of see it, if you take the either the 3D optimistic or conservative, you can see that the cost, this is an exponential cost curve, is substantially lower than what can be done with NAND, and we also believe that issues like endurance will be far less, frankly, will not be an issue.
[Slide 19- NAND and Post-NAND Fabs in the Coming Decade]
So what about this post-NAND? What’s going to happen to all those fabs? The existing NAND meg-fab is a very large fab- 200,000 wafer’s/month, 300mm. Will go to 1x nm.
By 2012/13, they will be highly depreciated. They will be highly cost effective. They will generate, I believe, very attractive margins for a long period of time. They will be throwing off a long tail, a long fat tail of cash, I believe, and very likely to the end of the decade.
And the reason for that is, because at 1x, you have such a good cost structure, particularly when you are depreciated. You can do 128Gbit chip and with that- 16GB on a chip- and with that you can cover a very very large train [??] of legacy products that are growing- the markets are growing.
So I think that these investments are going to have a second life, in fact benefit, from the fact that there is a slowdown in NAND and therefore there is far less need to continue to invest- to upgrade them. So basically you get to a certain point, you cannot go any further with immersion lithography, you are down at the edge. No one can do any better. And basically you have the market for that.
New NAND mega-fabs on the other hand, will be very pricey, will require, people are talking about $8- $9 billion for 200,000 wafers per month – fab will still be [??] 300mm. They will certainly need to support EUV and they will also need to support production of post-NAND technology.
So fabs will be at NAND and over time be converted to post-NAND or maybe the two co-exist for quite some time.
Timing-wise, I want calibrate you, the industry transition from SLC to MLC took 3 years. Some of you who have tracked this industry in the early days of MLC, Samsung was pooh-poohing MLC- no good- poor quality- by 2005 Samsung was mostly MLC.
By the same kind of wording about x3 today from people that don’t have it. I think that will take a while. But the point is that moving from NAND to another technology that is not NAND is far more complex than moving from 2 to 3 bits per cell; from 1 to 2 bits per cell which is basically a design concept on basically the same technology.
So the industry transition to post-NAND definitely will be more complex. We expect that it will be to a certain extent over a period of time it will be quite prolonged. There will be overlapping of NAND designs and post-NAND designs in production for most of the decade.
We of course are driving both. We are driving NAND at a ferocious pace, mostly on 3 bits per cell but also [??] we are driving 3D. As soon as we can bring it in, we would like to do that.
[Slide 20 How We Stack Up Against NAND Competition]
I would like to talk a little bit about how we stack up against the competition…
[Slide 21 Intellectual Property: Status]
There is nothing here that is really new…
Flash memory portfolio keeps growing.
On the card side, …
We do think that we have untapped potential for licensing and that is with our controllers, our SSD, x4 technology, mobile security, and of course fundamental patents in the 3D diode arrays, which we believe apply, extend well beyond the way we develop 3D. I mean other people will have to do that.
[Slide 21 Our Flash Capacity Plans]
[Slide 29 Flash SSD: Massive Market in Coming Decade]
Let me just say a few words about SSD. We think SSD is going to be a massive market in the coming decade. SSD adoption as we know is still in the very early stages, but we do expect the market to accelerate 2011 with the 20x, 24, 25 nm NAND even in 2 bits per cell. The tipping point could well be reached with the 1x nm NAND cost and pricing structure.
SSD will come [pause]. Today most SSDs are basically coming in a 1.8 or 2.5 inch package, but that will change. There are coming all kinds of form factors. At the end ?? you will see some. Later Yoram is going to talk about that.
But basically the next generation of SSDs will be optimized for specific usage models, fostering new architectures and innovation. And I am very happy that last week we did introduce our first 3G, third generation SSDs. You can actually buy them today in retail.
And you will hear some more about that. But one very interesting point about SSDs is that once that market becomes very large for the foreseeable future, in my opinion, highly price elastic, which means that if there is any period of excess supply I think that the SSD will be readily able to absorb that, basically become a shock absorber for a period of excess supply.
We see SSD as playing to our strengths in both systems, architecture, IP and the fact that we have captive supply. Captive supply of SSD, of flash is critically important to OEMs. Third party resellers or architects ?? can buy flash and architect their systems, but OEMs need to know that you control your own destiny in terms of your supply. So we expect to be a long term participant in this market…
[Slide 32 Summary: Flash Is Everywhere, Will Be Bigger Than You Think!]
I will conclude here now and say that we see that flash is everywhere, but we think that flash will be even bigger in the coming decade. Flash is already ubiquitous, and growing. Smartphones will continue to grow at a very fast pace and particularly there are major opportunities in emerging markets. The mobile internet is in its very early innings and will be bigger than you think. This is from Mary Meeker, Morgan Stanley.
Cloud computing: We feel flash ?? as the last secure node on the network and it is going to be in your pocket. We think we have what it takes to provide that secure storage.
On the other hand, highly scaled flash is getting very hard to do well. You’re required system solutions and this [missed words] is absolutely playing to our strengths.
So this is why we feel so optimistic. On the one hand, unprecedented growth opportunities, tremendous proven track record, very well poised to take advantage of this next decade.
So with that, Sanjay my cofounder…