Micron Technology Inc. (MU)
August 24, 2011 9:00 am ET
Brian Shirley - Vice President of DRAM Solutions
Glen Hawk - Vice President of Embedded Solutions
Steven Appleton - Chairman and Chief Executive Officer
D. Durcan - President and Chief Operating Officer
Mario Licciardello - Vice President of Wireless Business Group
Unknown Executive -
Ivan Donaldson -
Thomas Eby -
Unknown Analyst -
During the course of this meeting, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company and the industry. We wish to caution you that such statements are predictions and that actual events or results may differ materially. We refer you to the documents the company files on a consolidated basis from time to time with the Securities and Exchange Commission, specifically the company's most recent Form 10-K and Form 10-Q. These documents contain and identify important factors that could cause the actual results for the company on a consolidated basis to differ materially from those contained in our projections or forward-looking statements. These certain factors can be found in the Investor Relations section of Micron's website. Although we believe that the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee future results, levels of activity, performance or achievements. We are under no duty to update any of the forward-looking statements after the date of the presentation to conform these statements to actual results.
Okay. We're going to get started. Let me welcome everybody. I apologize for starting at 6, but we couldn't get the room at 5. So we do appreciate everybody being here and allowing us to share some thoughts about the company with you. I also welcome everybody on the webcast. We're going to start out with some commentary from myself about where the company is today, and then you'll hear from the business unit guys and some discussion from Mark Durcan around the option technology. And then I'll come back up for a little more discussion around the industry, and of course, we'll open it up for Q&A. Consistent with what we did the last time, all the Q&A will be at the end. So we'll leave a lot of time for that, and everybody will be available for the Q&A as well.
A couple things that you'll note, Mark Adams is not actually doing a presentation. Ron Foster is not doing a presentation just by virtue of where we're at and in the cycle in terms of the quarter for us and they will be here, though, available for questions and we'll all be up here at the end and you can ask on whatever topic that you want to ask on.
So with that, let's get started. I just quickly wanted to summarize, again -- the company's done a lot of things to get where we're at. We've pointed out that we've been pretty efficient in our capital deployment. We have been through a number of M&A activities over the years, and our portfolio is also very broad now. And I'm not going to spend a lot of time on this because we've talked about it before. But I just think that it's important to point out that we are in a pretty good place.
Now set that aside from what we're experiencing in the economy in the industry, Micron in and of itself over the last decade has done a lot of things to make sure that we would be able to either take advantage of opportunities as they come or that we did take advantage of a lot of opportunities that have occurred in the past.
With respect to how that's being reflected in the company, our end market diversification is really very good. This is the most recent data around how it breaks down for our end markets. Obviously, we have a greater presence in some of the more specialized markets. And in correlation with that, we still have a presence in some of the more difficult markets in particular today. Clearly, we still participate in the PC business. We still participate in some of the consumer products. So the point here being that as compared to what we used to be, our diversification is actually is very strong.
Now I've made this comment before around how the DRAM and the NAND cycles look compared to each other. And if you remember back in the '80s, the DRAM really was very price elastic and had more significant movements in terms of its pricing. In other words, we would have an upcycle that would last somewhere between 2 and 3 years and we'd have a downcycle that would last somewhere between 1.5 and 2 years, and it was pretty consistent as we went through the cycles.
And I used to say that I had -- when we went into the cycle last time that it was my seventh cycle. If you look at DRAM right now, and I'll show you some data later on, they clearly were in a cycle. I guess, for me, this is my eighth cycle now watching it, and there's a lot of similarities to what used to happen in the DRAM as to what is transpiring in the NAND. And this graph is just there to illustrate that if you overlay really what happened in the early decade of the DRAM and you overlaid what happened essentially the last decade in the NAND, it's amazing how similar the charts look. And I'll come back to the significance of that in a little bit.
So when you look at what's driving the NAND and the consumption of NAND, now this particular estimate has the demand a little bit lower than others. I've seen it as high as maybe 80% or 90% for the year. But the point here is that if you look at the diversification of the applications in NAND, this is already at a point in time where it took the DRAM essentially probably 20 years to get to. So if you think about the DRAM and the applications that the DRAM went into, it was almost entirely the computing world up until probably around the year 2000. In fact, if you look at Micron's revenues in the '90s, we were 95% going into the PC computing space. And it's obviously a lot different today. And it took us a long time to get there. I mean, it was 25 years. And if you look at where the NAND is today and the diversification of the applications, it's actually accelerated to the point where it's happened much, much quicker than it ever happened in the DRAM environment.
And from Micron in particular, if you look at the diversification of our own product line, this is the first quarter where we've crossed over. NAND is actually going to account for more of our revenues than DRAM did in the quarter. And that's the first time that we've had that transition. Now we've been talking about that coming, but it's actually happening for us right now.
And by the way, I don't point this out because we don't like the DRAM business. We're obviously, very heavily involved in the DRAM business. And I'll speak later about why I think that, that business will have opportunity at downstream. It's just to point out that we've also been able to take advantage of the evolving memory markets that have consumed NAND. And you might expect this chart to crossover anyways a little bit because the DRAM pricing has been down more recently as compared to the NAND pricing and obviously, that affects revenues.
But the real story here is that as you look at Micron's diversified product portfolio as compared to where -- and we only go back a few years and show you what -- where we were before. If you actually went back into the early 2000s and the late '90s, of course, this chart would be much, much different around DRAM and somewhere in the neighborhood of 95% dependency. So we've really transitioned the company to have a broad memory portfolio, and that's -- it is being reflected in the results of the company. Some of which, I'm going to show you a little bit more data when I return back up to make some additional comments.
So to set the stage for the individuals that are going to come up next, we restructured the company. We talked about creating these business groups, and we really have somewhat of a hybrid model. Just to remind everybody, if you think about the way the company used to be organized, we were really primarily focused around technology. We really have a hybrid model today. We have 2 groups that are very focused on markets, and we have 2 groups that are also very focused on technologies and they interact in a way that's very advantageous to both of them. And we've now had the structure for about a year, 1.5 year, and we like it. We actually think it's working very well. We're not looking to do something different than what we're doing right now on our structure. We think it's working well. And you'll hear from all the leaders of these groups, and we think it's been very effective in terms of what we're trying to accomplish in terms of market penetration. And so those business units, you'll hear from. And then after them, you're going to hear from Mark talk about the left-hand side of that chart, which is around all the operations and technology that feed into that.
So with that, I'm going to turn it over to the next group. And, oops, did I kick it out, Ross? You're just joking, great, to Glen Hawk.
Thanks, Steve. This mic on? Okay. Steve showed a graph that demonstrated the nice exponential growth that we're seeing in the NAND industry. And the cloud is definitely one of the things that's driving that growth. But I'll go a step farther than that and say that flash is really driving the cloud. The cloud that we know and live with today is enabled by flash. And in fact, the cloud of tomorrow that we'll all use and benefit from, flash is required for that. And we're right in the middle of that.
First, what we saw was flash was put behind a conventional SaaS or set [ph] of hard disk drive interface is very easy to put that into some of the existing data centers and immediately reap the benefits that flash technology had to offer. More recently, though, there's been a lot of action on a more disruptive side of flash in the data center. And that's in a memory on the server side with flash behind a PCIe interface. And I'll talk a little bit about that later.
From the numbers' perspective, what I think is interesting is earlier in the year, some large cloud computing companies made some pretty interesting announcements that caught my attention. One company came out and said, "Well, we've shipped 10 petabytes of flash bits into the enterprise." Ah, it's good. It's always good to see one of our customers of flash make such an announcement. Another company came out and said, "Well, we've shipped 15." So for the fun of it, I got with our team and said, hey, how much NAND flash have we shipped into this industry over the same period? We came up with a number of about 65 petabytes. So a large amount of our flash is already going into this segment, and we're participating in it in some very complicated ways, which I'll explain in a second and, hopefully, add some clarity around.
But it's not enough to just have flash in the cloud. People need to be able to use that. And while we have desktops and laptops and notebooks that we could all use to access data from the cloud, clearly, over the last year or so, one of the big stories has been the emergence of tablets. And this gives people another way to interact with the cloud in a very interesting way. We've taken a look at our projections over the next 2 years, and we believe about 55 billion gigabytes of NAND flash are going to be shipped into tablet and tablet-like devices over the next 5 years. Tremendous opportunity for us. If you refer back to the chart that Steve showed, this accounts for about 15%, maybe even 20% of the flash bits at the end of this 5-year period. A tremendous amount of flash. So from a Micron perspective, that's good. We benefit by providing flash for these edged devices as well as the cloud.
I would like to speak a little bit about how we've been participating in that. And this -- to illustrate that, what I put up here is just a snapshot of the sampling of the companies that are participating in the ecosystems in and around flash-based SSDs. And over on the right, of course, you have the NAND manufacturers. You have some third parties who don't make NAND, but have become very proficient at making solid-state drives based on flash. And then on the far left, you have the consumers of these devices, some more traditional ones, HP, EMC, of course there's others, IBM, Dell. Just didn't have enough room on this slide for everybody. And of course, we have some new interesting ones, Google search capability and also social media companies like Fusion. So there's a tremendous amount of activity in this area, and we're participating in all of this through selling our NAND flash directly to some of these companies, as well as our fully integrated solutions, such as solid-state drives. And I'll talk more about that later.
But I'd like to underscore this a little bit by also rolling a video here that you can hear from our customers directly what they're doing with our products.
Okay. So that was just a sampling. We wanted to provide that additional color to give you a feel for the diversity that's in this space. And to kind of further illustrate that, I brought some "show and tells" with me for those of you that are here in the audience. And what I'm holding up right now is one of our what I would call raw NAND flash devices. You can barely see it. That's kind of the point. It's very thin, less than a couple millimeters in terms of thickness. Inside of this package, there are 16 8-gigabyte NAND flash devices. There's 128 gigabytes of flash storage within this package. An amazing amount within here.
Now some customers can take this just as it is, use it in their systems quite effectively. What I like about those 3 videos is you saw 3 examples of customers that are doing interesting things with their flash, some of which are capable of just taking our very simple raw NAND or more highly integrated solution. There's another product that we make that's called ClearNAND, where we insert a controller in the same package. It physically looks the same. You wouldn't be able to tell the difference.
And what that ClearNAND controller does is it manage -- it takes care of all the issues associated with error correction on our flash. So it's one notch up in terms of system level integration that our customers can benefit from. Now other customers need a more fully integrated solution, of course, like a client solid-state drive. So here's one of our client solid-state drives. As you can see, it's the exact same form factor as a magnetic hard disk drive. Of course, it's a lot bigger than what I've just shown. This one happens to be our C400 device. It's 256 gigabytes. So you can see that there's a lot of other value that's packaged around that flash that we're adding.
This is -- the shipments for this have been going great. On the next slide here show what has happened over the last 5 quarters. Not a lot of numbers on this slide, but you get the basic idea. Over the last 5 quarters, we've gone from about 1% of the client SSD market to about 10%. We've been working for several years on solid-state drives. We're up and we're rolling now. And now we're penetrating the market. Yet to come are some other interesting versions of that.
For example, what I'm holding up right now is an mSATA version of an SSD. And you can see right away a much smaller form factor. It's going to find a nice place in laptops and some of the portable computing devices that are coming out. And other companies use other form factors for solid-state drives. This is frequently referred to as a Gumstix, and you can see that this is yet another form factor for a client SSD. There's a tremendous amount of variability, a tremendous amount of variation, a big opportunity to differentiate in this space.
The other thing I'm holding up here, and I'll talk more about this in detail. This is one of the exciting products that we just announced this year. This is our PCIe card, a NAND flash-based PCIe card.
As you can tell, just looking at it visually, it looks a lot like something you would see inside a server, a computer, and that's exactly what it does. Even more value wrapped around our flash on this device. I'm particularly proud of this device because I think this -- it best illustrates what capability Micron has created and we're delivering to the market today. Not only do we make the NAND flash devices, we also -- it's our controller. A lot of people out there are not convinced that we've got the controller expertise required. I'll show you some of the specs on this device later that should put that question to rest. The components -- we do all the manufacturing on this, so from end-to-end, this is truly our product. We're very proud of it. And I think this is a glimpse of what you'll see from us in the future.
Again, we deliver solutions at multiple levels of integration, and that's what's going to be required to win in this space of interesting products in and around the cloud. Of course, it's not just about client solid-state drives. As a lot of you know, this is an updated version of the slide that we had shown at the conference in Arizona 6 months or so ago. As you can see, there hasn't been any change in the upper left-hand graph here from a qualitative perspective. No surprise, client SSDs are still expected to outship enterprise SSDs in terms of units. But if you move to the lower left and you look at that same story from a revenue picture, you see a different view of things. Obviously, there's more value being added to the enterprise SSD solutions and so as a result, the revenue opportunity is comparable. And so, obviously, that's an area that we've set our sights on. I'll share more about that later.
On the far right, you also see from a price per gigabyte perspective what we're seeing in terms of the uplift for these enterprise-based solutions. And if you go over here to 2011, what we continue to see is a nice uplift for enterprise SSDs over client, about a 4x improvement. And we expect that to continue out in time.
I thought I would put together one slide that would really give you a good visual feel for where we've been with SSDs at Micron and where we're going. What are shown here in the 3 regions are: Where we've been, what SSDs we have had in production and have delivered, that 10x growth that I'd shown earlier. And as you move to the right, it shows more what you can expect from us.
So far, to date, the P300, the C300 are both SATA-based drives for enterprise and for client. And our recently introduced C400 SATA drive, which is based on our 25-nanometer flash shown there in blue, those have really been the workhorses to date that have allowed us to get up and running in manufacturing, get our core competencies dialed in. And over the last couple of years, we've had a lot of teams working in parallel to rapidly proliferate these product lines around our SATA-based client products at the bottom. I have shown you a sample of our mSATA offering, which was a nice, fast, efficient derivative from our C400 drive. We also plan to announce sometime later this year a self-encrypting drive for clients, which is a nice feature that a lot of our customers have been asking for.
On the enterprise vector above, our P300 SATA drive, we will be offering a 25 nanometer version of that shown in blue. And visually, you get a nice feel with what we're doing with the NAND lithographies and our fully integrated solid-state drive solutions. You see, as we progress from the purple 34 nanometer to the blue 25 nanometer and ultimately, the yellow shown there in 20 nanometer.
We have mentioned this publicly before, but we'd like to be much more clear about it that in enterprise, we know it's not just about the SATA interface. We know that SaaS is very important. We have some very exciting things that we'll be announcing in the near future on SaaS. We'll be talking about a 6-gig per second interface, as well as 12. As we've done with other introductions on the SATA drives, you can expect to see leading-edge performance from us on the SaaS versions that we plan to announce. And I just talked about our PCIe solution. And as you can see as you move to the right, in addition to the single-level cell version that we're offering today with our Gen 2, we'll also be offering a multilevel cell version.
So it's a very nice product portfolio that is now emerging, that you'll be seeing more from us in the future. We're very proud of the work that we've done there to date.
I'd like to speak a little bit more to the features and the specifics of the -- of our P320 drive. Again, this is our PCIe interface for a server-side enterprise storage. The thing to think about on server-side SSDs is IOPs as well as latency. And I'll speak to the latency PCR first. It's very fast to access the NAND in a caching application, which is what this is designed to do. Our specs that are out there, our spec sheet says less than 50 microseconds. But in some cases, we can achieve much faster, depending on the exact usage conditions. Sometimes it's faster down to 17 microseconds. There is a range of latencies that you'll see. They're all fast. They're all significantly faster than SaaS or SATA hard disk drives, which are orders of magnitude faster than some of the other options that customers have available to them today.
The other story is also on the IOPs. So I'll skip this slide. It's going to talk a little bit about the importance of benchmarking. At Micron, we are very proud of the way that we're setting our specs. We think it's based on real-world usage conditions. And so when we compare our specs that are out there versus the other PCIe providers that are out there, some of which, by the way, are our customers, you can see that we have leading-edge performance from both the random read and the random write perspective.
The point here is not for us to show that we are significantly better than some of our -- some of the other folks that are providing solutions in this space. These are all great solutions. Again, some of these are our customers, and we're very excited about the disruptive nature of PCIe interface solid-state drive for cloud computing between what we're delivering and the customers that we're working with that also deliver these drives. It's absolutely changing what's happening in the cloud today.
In conclusion, I also wanted to provide a little glimpse. I'm talking to a lot of people here that also are interested in where Micron's revenue and margins are going to be coming from in the future. And I'm very happy to share you with you where we see things going. What we're showing here over the 2011 to 2015 period is where the revenue has been going and where we see it going in the future. As you look to the far left of this graph, you see that the vast majority of the revenue for the NAND Solutions Group within Micron was going to large OEMs, it was going to the channel and it was also going to USB and cards, especially through our own retail channel or Lexar brand.
Those are still good businesses for us. But as we move to the right here, we're going to see our participation in the cloud and in and around the cloud with the client devices associated with it in a much, much larger way. We'll get to the point where well over half of the revenue for this business unit as Micron is directly going toward those types of customers. And we're very excited about that because, as I shared with you, there's a lot of value that's wrapped around the NAND flash there. And we're uniquely positioned to deliver some of that value because we understand better than anybody else how our NAND technology works. And as NAND scaling progresses, that expertise is going to be one of the key things that differentiates us from some of the other providers that are out there.
So with that, that's basically the NAND Solutions Group. At that point, I'll go ahead and turn it over to Tom.
Thanks, Glen. Is this working? Okay, great. Well, good morning. I'm going to spend the next few minutes giving an overview of what's happening in the embedded market and what we're doing within the Embedded Solutions Group. Give an update on the market, the growth opportunities from both a segment and a product perspective. The fact is the embedded market is a relatively more stable, relatively higher margin opportunity compared to many of the other segments in memory. And I'll talk a little bit about some of the peculiar specific needs of certain segments that contribute to that greater stability and higher margin opportunity and some of the things that Micron is doing to address those.
Talk about a couple specific growth opportunities. One, building on some of the comments that Steve and Glen have made with the growth in NAND and in solid-state storage, the other more product specific example. And then I'll close by talking about how in addressing one of the particular needs of many of the embedded segments, that being longevity; in addition to doing a better job delighting our OEM and chipset partners, we're also going to extend the useful life and optimize the utilization of Micron's manufacturing assets.
Okay. So just starting at a very high level, overall market opportunity in 2012, about $7.8 billion. We see a growth rate between '11 and '14 of about 8%. I'm going to -- I've talked about what contributes to the stability and the margin opportunity for embedded. One of them is diversity. Diversity across market segments, across technology requirements, across a very broad customer base, as well as some other particular requirements I'll talk about later.
I'm going to say more about the segment opportunity in a coming slide. I'll only point out IMM -- Steve talked about industrial -- IMM stands for industrial and multimarket, where multi-market is the convenient way to say everything else.
Taking a look at the technology split. Today, the market opportunity is relatively evenly balanced between NAND, NOR and DRAM. However, the growth rates vary quite a bit. The strongest growth is in NAND. Again, some of the growth that Steve and Glen talked about, and I'll detail that in a little bit, low-teens growth there. DRAM, very slightly exceeding the 8% growth we see in the broader market. And the NOR market is essentially flat at about $2.5 billion.
However, within that market, it's a very different story between something called Serial or SPI NOR, which is going in the mid-teens and parallel NOR, which is declining in the high single digits. And I'll talk a little bit about some of the things that we're doing to take advantage of that growth in the SPI NOR marketplace.
So now let's shift gears from the technology perspective and talk a little bit more from a segment perspective. Again, overall growth in the market is 8%. If we look at the markets, including automotive, IMM, and personal systems and servers, those are growing roughly at the same rates as the overall market. Consumer, again, driven by some NAND in particular, eMMC opportunities, is growing a little bit faster. Networking storage coming in a little bit lower.
From a margin opportunity perspective, if we look at this, these are actually ranked bottom to top, from highest margin opportunity to the lowest opportunity, from a margin perspective. Automotive, IMM and networking and storage are all above the ESG average, which has been contributing low 20s in terms of operating profit margins over the last couple of quarters. And while consumer and personal systems are below that average, there's certainly still attractive margin opportunities for the company.
From a share perspective, we're on a run rate of about $1 billion over the last couple of quarters. If we look at that versus a little bit less than $7 billion of opportunity in 2011, that's a mid-teens overall share. If we look at where we sit, we're well above that mid-teens average in automotive. We're a bit above in networking and storage, a bit below in IMM and consumer, and substantially below in personal systems and server. Although, as we ramp our 65-nanometer SPI parts which service the PC bios, which is a major opportunity here, we do see a good growth opportunity.
So beyond some of the more specific share growth opportunities that I'm going to talk about in some further slides, there is a general trend that we've seen in the best margin opportunity segments, that being automotive, IMM and networking. And that is -- those are the customers that place the greatest value on the stability, the predictability and the breadth of relationship, and where we're seeing the greatest uptake in terms of the value proposition of our broad offering. And so we're seeing customers where we may have been very deeply engaged in DRAM, and we're seeing significant design activity on the NOR or the NAND side or vice versa. We may have had a very deep NOR relationship. And we're extending that into DRAM.
The final comment on share, from the first slide, you see that IMM is a many customer. It's responsible for the vast majority of the customers in that 50 through 5,000. And it really is a channel play. And we've continued to invest working with Mark Adams and his team on the rep and the distributor network on those programs. And I think we're building a significant advantage versus some of our bigger competitors, who just don't focus on the needs of those smaller customers as well. And we think that will also contribute to our share opportunity in IMM.
So now I'm going to switch gears and talk a little bit about some of those unique requirements that exist in the embedded world and some of the things that Micron is doing to address them and take advantage of market growth opportunities. And I'm going to focus today on some of those requirements in automotive. And probably the first thing that many people think of in terms of unique requirement in automotive is longevity, and that's certainly important. I'm going to talk about that. But it's actually a lot more than that. There are things like very specific product specification requirements, specification that I won't bore you with called AEC-Q100. Very stringent quality requirements. They expect single part per million quality. They expect that quality at the ramp of the vehicle through something called safe launch. And they expect very quick 48-hour initial response when those failures do occur. They are very specific documentation requirements, very stringent logistics. One day early, no days late is about as much flexible as we get. And they even look into the business processes that their suppliers are using to drive their business through something called an ISO/TS specification, which we now have underway and we have a 2-year roadmap to get implemented, which the automotive customers are taking as a very strong indication of our continued commitment to this marketplace.
But now back to longevity. Typically, we will start engaging at least 2 years in advance on design, 2 years before the ramp of a vehicle. And from the time that we do ramp, it can be up to 10 years that we need to be providing a fit, form and function equivalent product to do that. And from a change perspective, the requirements are also quite extreme. It will typically take anywhere from 2 to 4 quarters for 1 of our automotive electronics customers to requalify our component. And an automotive manufacturer will typically give a subsystem supplier a 1- to 2-week window each year for any given range of models in which to provide a change. And so it's a very lengthy and very expensive process, which again favors the ability to provide longevity and stability.
And so what is Micron doing about meeting these needs? I've talked about some of the nonlongevity needs. But one of the things that we've launched since our last gathering in Arizona in February is something we refer to as the Product Longevity Program. And so what is it? It is a 10-year commitment to provide fit, form and function equivalent products. And if we -- this is not the same technology, not necessarily. But if we do shrink apart, we commit that it is backwards compatible and will give a 2-year window in which the customer can requalify.
It is targeted, of course, at automotive, but also at industrial and medical applications and really designed to provide a safe set of choices that those chipset partners and customers can make, knowing that they won't get disrupted by an unexpected, end-of-life.
And so why is this important? Well, I've talked about some of the challenges of change in automotive. We hear from some of our networking and telecom customers that a requalification at a telecom provider of a piece of network gear can cost $250,000 to $0.5 million. And there can be extremely high regulatory barriers to any change in something like a medical device.
And so because of that, this has been extremely well received, again, both by our chipset providers that are often making choices as to which memory components to put on their reference designs, which are very often adopted by OEM customers, as well as by the OEM customers themselves.
Okay. So talked a little bit about one area of unique requirements. Let's talk about a couple examples of growth. And I'm going to build on some of the general NAND bit growth discussions that Steve led off with and certainly Glen built on by talking about, again, an example from the automotive world. And in this case, it's the automotive navigation and infotainment system.
Now if we look back to the very initial navigation systems, a bit over 10 years ago, they were a single purpose, a navigation system with relatively modest memory needs, under half a gigabyte of memory, and they represented about 5%. The memory was about 5% of the bomb.
If we look out at high end, these are probably going to be model year '13 infotainment systems, they have really morphed to a general purpose, applications-enabled platform, where we're seeing up to 64 gigabytes of NAND and 4 gigabytes of RAM, equivalent with what you're seeing in a high-end tablet, all right?
So why is that? Well, it really is a battle by the automotive manufacturers to capture value that would otherwise go to makers of devices, like TomToms and iPads. And they want to do everything that they can to capture that value in the cost of the automobile and not have it go to third-party manufacturers.
And this value capture is actually happening in other embedded devices as well. If you look at digital TVs, again, if you look back a little over 10 years ago, those were maybe 32 megabits of NOR and 32 megabytes of DRAM, and that represented single-digit percentage of the bill of material. The smart TVs of today have upwards of 5 gigabytes of NAND and DRAM, and it represents over 1/5 of the bill of material. And again, it's about value capture. The DTV guys are fighting with the cable operators and the set-top box manufacturers, as well as the builders of devices like iPads and trying to capture value. And one of the ways they do that is to convert to again, an applications-enabled general-purpose platform that has very good -- it's very good opportunity for Micron.
So beyond the general platform opportunities, there are also a few more specific ones. One that we're focused on, nearer term, in the gaming world, game cartridges that have traditionally been ROM-based are migrating to managed NAND. Again, another good opportunity for growth. You can see what we see as the overall -- there's millions of gigabytes growth in the embedded world over time. Our initial focus is again on gaming and automotive. Going forward, we'll be expanding that into IMM consumer and networking.
Final comment on this opportunity. Because these managed NAND opportunities as Glen described place a controller in between the chipset and the raw NAND, we're able to be a little bit more aggressive about the technology that we use and migrate the NAND technology a little bit faster because we can hide the complexities of the evolution of the technology node behind this controller. And we'll get the best of both worlds. Meet the needs of embedded and do a better job leveraging leading-edge technology from Micron.
So second growth opportunity that I want to talk about is a little bit more product oriented. I said at the beginning that we saw the NOR market overall being flat. However, within that, a very strong shift towards Serial. Now why is that? A traditional parallel NOR has 48 pins to talk to the chipset. A so-called Serial NOR has 4. That means the die is smaller, the package is cheaper, the PC board is less expensive, and the chipset is less expensive. And so there's a slow inexorable march towards SPI as a percentage of the NOR market. And in particular, as advanced technology is beginning to enable higher densities of SPI, the growth within that is in the higher densities. And in fact, between 2011 and 2014, virtually all of the growth in SPI is coming at densities at or above 128.
And so to address that, we're announcing this week the world's highest density SPI products, up to one gigabit in density with the highest performance interface from a clock perspective and read performance. And very importantly, consistency from a hardware and a software footprint perspective across densities. And while this is at 65 across technology nodes, as we migrate to 45 nanometer in the future, we will maintain that compatibility. And we see very significant opportunities across the full spectrum of embedded markets to leverage growth with this new product offering.
Finally [ph] , I want to close on was how in meeting one of the unique needs in embedded, that of longevity, we also do a good job of leveraging and extending the manufacturing assets that Micron puts in place. This shows -- one example, it's our 34-nanometer NAND and looks at how the revenue ramp is split across various business units. Glen's group is, of course, the first to ramp in storage removable PMP and ramps very quickly, but over a 3- or 4-year period, it's moving onto the next technology node. In the wireless space, it's a little bit slower to ramp and to qualify. But again, over a slightly longer period, they also are ramping down. We're a good 5 years in until we're really getting to peak revenue. By that time, depreciation has substantially rolled off. And while we certainly don't utilize the full footprint that got put in place originally, there's a very healthy footprint of low depreciation capacity requiring minimal incremental capital investment yet still providing very attractive returns. And so from an asset utilization and an asset return perspective, it's a very good play for Micron.
And so going forward, we're going to keep focusing on longevity and the other unique needs to take advantage of the growth opportunities, contribute a growing and we believe better-than-average, both margin and stability stream to the Micron financial network and do it in a way that takes good advantage of the assets in place, extends their use life in a very profitable way.
And so with that, I'd like to turn things over to Mario, who's going to talk about the Wireless business.
Thank you, Tom. This is -- okay, good morning. I will start the presentation just summarizing some of the major trends shaping the wireless market. We basically in a market, which has been growing for several years in the row and, even if at a slower pace, it is still continuing to grow this year. The growth is basically driven by smartphones and tablets, which are the more successful application of the wireless technology appearing recently in the market. The convergence is the factor, which is shaping the growth and driving the growth. We certainly want to be able to be connected while we're moving around and want to be able to while we are changing our place to take care of our personal things, as well as taking care of our business activities. This is the convergence of multiple technology, which is enabling us to do so.
So the traditional market, which was started from telecom players, which was just communication, is translating to a more and more complex market environment in which multiple technology are playing the role to determine the evolution of the application. And of course, this is driving a lot of changes in the ecosystem, as I will show after, attracting many other players that be not present there, many of them, by the way, coming from the PC market and the PC experience attracted by the evolution, which is making the application more and more alike, if you see, rather than just a telephone.
Within that rapidly changing environment, vertical integration with the alliance is taking place, acquisition taking place is a common experience. Many things have been happening in the last several months. Just to mention a few cases, we know about the Nokia case merging from the technical perspective with Microsoft to adopt for the next generation smartphone, the Window 7, Window 8 mobile. And even more recently, the Google acquisition of Motorola mobile, which is basically changing the perspective for the Android operating system.
Going more continually on the market, as I mentioned, we see on this left side, the segmentation of the market. There is a poor range of application coming from the -- going from the very low-end entry phones, which are typically characterized by bow [ph] which is maybe in the range of $25 to $30. Moving up in the scale up to the level of the smartphone, the upper handle, which are characterized a bow [ph] , which is in the range of $120 to $150 in a bow [ph] . And the tablet, which is even positioned beyond that.
Beyond the growth in terms of volume, which is very significant in the coming years, because we see the market moving from around 2 billion terminals up to the level of 2.5 billion, 2.7 billion terminals, all in all included, we're just not talking about foreign only, we're talking about all the wireless application, for instance, application like the machine-to-machine communication, which is a wireless communication using the GSM technology. So all in all, this is a global market dealing with wireless communication. And in combination to the growth, we have the average memory content growth, which is a kind of doubling in the [indiscernible] every 18 months. And I will comment a little bit more in the coming charts what is driving that growth. But the conclusive point is that thanks to the combination of the volume growth and the memory content growth, we see a dollar market which is growing very rapidly. We are today in the $12 billion, $13 billion market size, and the expectation is if the trend continue to be the one we see to have a doubling of that market size in the next 3 to 4 years, basically.
Switching to another subject, what I want to show here is a confrontation between what is the situation in a mature segment, like the PC segment, versus what is the fragmentation and the complexity of the wireless markets of today. The PC market has gone through many years of changes, adjustment, acquisition, merger, et cetera, reaching this point of today, in which basically there is few players in the operating system, just a couple operating system. Actually, there are only 2 providers supporting the intelligence of the motherboard. And we have a few manufacturer left into the global world market.
Of course, the price for that is that there is very little differentiation. You can choose one PC or another PC, you're getting plus or minus same performance. The only differentiation part might be in the service area, how the PC is supported, how the service overall structure in the world is available. But in terms of performance, you're basically getting the same performance.
Completely different is the situation in the wireless market. We're still in a very young market, which has been undergone many threshold mentioned in the past years, but a lot is going to happen also in future. We have many operating system contributing to the market. Some of them are proprietary operating system. Several others are open systems. But overall, it's a crowded market in which each operator in that area is either focusing on specific segment or maybe a multiple segment as well.
In association to the operating system, there is a multiple supply of chipset. Again, we have several -- I would say the majority of the company playing in the asset area, trying to be engaged in the chipset area. Some are, of course, is strong positioned as a leader, just to mention one, Qualcomm, which is a global player as a chipset manufacturer in many segment of the market. Some are more focused on specific part of the market. For instance, Infineon is focused at -- with a large priority in the entry level, low-end portion of the market. The same applies to MTK. ST- is a mixed -- ST-Ericsson is in a mixed mode is serving the entry-level market, but also supporting the media processor part for the smartphone area. And so on and so forth. So there is a multiplicity offering in the chipset area for the various application. But the main players in the manufacturing sales organization are still a large number. Of course, the one dominating the majority of the market are probably less than 10. You can recognize the names there, the Nokia, the Apple, the Motorola, the Sony Ericsson, the Samsung, the LG, all those players together are basically accounting for probably 80%, 85% of the total market. But in combination with that, there is a larger variety of supplier, small or medium, which are particular in the Asia part of the world, in China in particular, which are participating to that market sometime creating also disruption in the market through at a label offering or a price disruption, et cetera. So overall, a kind of evolutionary market, very fragmented, which is a distant from stability, which you will eventually achieve over time.
Another point to mention is that the carrier in that market are playing a very important role in the sense that they are really determining the quality of the demand, what kind of services they want to provide to the network and therefore, what kind of performance and features they need to have from the smartphone or from the cell phone in general. So this is a unique situation whereby the carrier are a very key part of the ecosystem determining the requirement, determining the architecture of the solution that we are providing to the market.
Just a few more details to indicate how those things are happening in the various part of the ecosystem. If we talk about the operating system, clearly, the operating system is dictating the overall requirement in terms of memory size and performance. The bandwidth, for instance, is very much linked to the operating system. The amount of memory that you require is linked to the amount of memory that the operating system requires for itself to be working properly, but also to the kind of functionality the system is generating. Therefore, how much memory is required to support all the functionality.
If we go into chipset, there are similar requirement on top of that, there are additional contributions that can be from them. For instance, the interface between the memory and the chipset. You can have a freedom interface. You can have a parallel interface. You can have an eMMC interface and so on and so forth. So really, this kind of way of activating the memory through the chipset is coming from the chipset makers directly.
Also, the packaging requirements. For instance, is the chipset requiring to have the DRAM mounted as a Package-on-package on top of the chipset to maximize the performance? If this is the requirement, you need to develop a package, PoP package, which is responding to the chipset maker specification.
And so also, based on that, the signal integrity, the thermal requirements, all of that is really coming from the original specification agreed upon with the chipset vendor.
The customer also are contributing upfront in defining the performance of the basic terminal they want to develop. Also they are contributing to finishing [ph] of the package requirement and they adding the architecture based on standalone solution. They're having the architecture based on stacked multi-package, multi-chip package solution. What is the bore out requirements, what is the footprint requirements, all that is it really something which is coming from the customer.
And as I said before, also the carrier are contributing upfront in the definition of the global architecture and the global functionality of the phone. All that is just to mention that in order to be successful in that market, there is a real need to be able to manage this complexity and to be able to interface with all those players in the ecosystem environment. And this is one of the major strengths that Micron has, thanks to our many years experience in that area, we have developed a 2x strong links with OEMs. We have developed 2x strong partnership with the people developing chipset. We are starting working together with some of the major carriers, particularly in area like Japan, for instance, where the carrier are in reality determining the system architecture by themselves and so on and so forth. So this is a complexity on which we can capitalize using our competence and our long-lasting relationship with many of the player actively working in this market.
Just to show in a graphic way what is the difference? So this is just 2 commercial product. It's not specifically referring to Micron itself. This is a Nokia [ph] product which is a cellphone, typical cellphone architecture. While this is one IE and smartphone architecture. It's not unique, but one of the most frequently adopted architecture.
As we can see, in this case, we are talking about a very simple board with a very few components. And one processor, which is basically managing all the functionality of the phone itself from the voice to the few features that are embedded in the phone. For instance, the camera features or the messaging features, all these kind of basic features, which are part of the phone. And you see that we have a very simple memory content there. We have just a stacked package composed by a medium-range NOR density 500 to 1 megabit and NOR density is kind of this average density adopted today in this kind of application, combined it with a pseudo-static low density RAM. So we're talking about few dollars kind of a combination in a phone where the bond, as I mentioned before, maybe in the $25, $30 maybe $35 range.
Very different is this picture, and of course, between this and that, there is a full variety of solution which are taking advantage over the multiplicity of the offering that we are making to the market. This is a much more complex board. This is more like a motherboard of a PC, actually. It's not the same complexity, but is approaching that level of complexity. If we look into the architecture, we see that instead of having one processor, we have 2 processor. One is merging the basic feature of the phone, again, similar to that, the modem part and the communication part et cetera. But on top of that, we have a second and more important processor, which is the media processor, which is basically managing all the applications that a smartphone is providing to the users.
And of course, to be able to offer that capabilities, linked to it, there is a very significant amount of memory. This is -- indicated here is a managed NAND, e-MMC. I think you're hearing many of us talking about the managed NAND. Certainly, managed NAND is the biggest prospective of the storage capabilities nowaday in the several application. So we're using here in the e-MMC, which is a managed NAND, basically is a basic NAND with a controller, which is managing the feature of the NAND itself and the performance of the NAND itself with a multimedia, embedded multimedia interface to the processor.
And of course, to explore that capability, there is also the need for a significant amount of mobile RAM. In this case, we are talking about 6 gigabit memory, DDR2 memory. And actually, in that type of high-end application, the range of memory is moving from the 4 gigabit up to the 8 gigabit density. So that's the typical range we are seeing there.
While the e-MMC density, we're talking about in this case is an 8-gigabytes, but the typical range is moving up today to the level of 32 gigabytes and the next generation more in the 64-gigabyte density. So this is the reason why we see on top of the growth in volume, a big growth in terms of memory consumption, therefore, a dollar value of the market.
Why Micron is strongly positioned in that market to be very successfully in the market itself. First of all, because we have the competence in general, but beyond that, we have the widest product range in the market. We are, as you heard also in the previous meetings, we have the full range of product based on NOR technology, including the parallel interface and the serial interface product. We have a full range of product SLC and MLC in the NAND technology. We have a full range of mobile RAM ranging from 512-megabit up to the 4-gigabit density. And we are combining all that to generate the multiplicity of a solution which are supporting the different needs of the different applications. So this is the major strength that Micron has with compared to any other competitor in the market.
I should say that as part of the product offering, even if it's a little bit premature, we are almost to the point to start introducing the market phase change technology-based products which will be complementing in the medium-low density range to NOR offering of today.
We have -- one of our major strength in the global partnership, I mentioned before that thanks to our long-time experience, we have a complete network of relationship on which we can capitalize and build our success story. And of course again, we are leading the technology with a combination of technology and you will hear from Mark Durcan more comments related to that.
Overall, we are capitalizing on all those factors, and we are gaining positioning. We are the #1 by -- since many years in the NOR environment and the NOR offering. We are working hard to become the #1 on the rest of the market, which is requiring different offering as I was mentioning before. We are starting to be very successful.
Our product are designed in a large variety of smartphone. The 190 number should not be very impressive in the sense that this is not a different architecture. This is different models. As you may know, that the OEM are introducing on a basic platform tens of different models with a little differentiation one versus the other. So overall, we are designing in many different architectural platform, which are generating tens and tens of different model introducing to the market.
And we are starting being successful in the upper end of the market, with tablets and e-books version in many part of the counts, particularly in China, which is one of the area, which is much active in that area as well.
To conclude, I think we have proven to be a very successful player in the market through our many years of history. We are today in presence of more than 550 million phone using our memory products. We are expanding our basis on smartphone as well. And as I already described, we are leveraging in all our key strengths, which are product technology and ecosystem relationship to build a successful story for Micron and to contribute in a strong way to the success of Micron to the profitability of Micron, which is a very important part of the global game. Thank you. And now I will pass to Brian Shirley, which will go through the DRAM part of the presentation.
Yes, thanks, Mario. Thank you. Okay. Good morning to everyone. For those of you that I haven't met, my name is Brian Shirley. I head up what we call the DRAM Solutions Group here at Micron.
And we are focused really with our computing DRAM portfolio on a variety of applications here. The way we think about this is really centric starting with the core infrastructure, servers, storage. We'll talk a little bit about cloud computing as well.
Moving out through the edge, talking specifically here about the networking gear, everything from IP routers, switches, backlog capability as well as infrastructure, all the way to the premise, DSL gear, set-top boxes that help make this entire computing infrastructure work. So I want to talk to you a little bit this morning about some of the trends we're seeing out there. I'll update, this is a brief update from some of the trends that we spoke about in Scottsdale back in February. And I'll give you a little bit of market commentary as well and I know Mark Adams can speak as well during the Q&A.
Server. You know I think it's fair to say we're not seeing a big slowdown here. What we are seeing is a different partitioning happen in this space where there is a move, a little bit of what I would call softening of the server resellers, the traditional developers of server systems to more of a direct model. And in general, that's something that's pretty favorable to a company like Micron. We're selling a lot of memory direct applications these days, companies that are using that memory specifically for their own server usage. That's a level of innovation, as well as frankly, I'll call it a removal of the middleman that's beneficial to a company like Micron.
Cloud computing. Well, we'll speak more about this. It is real. It's happening. Its significant growth in this overall server and storage space. And I would tell you that what we're seeing from the virtualization side, cloud and virtualization generally means more cores per system and more cores per system means a lot more memory. So our shipped standard module density into the server space now is north of 8 gigabytes. That's been a pretty significant move up. We're shipping an awful lot of 16 gigabyte and even 32 gigabyte modules into this space.
Networking and storage. Fair to say, there hasn't been much of a slowdown here either. This is a space that's been very, very good to Micron. What I would say interestingly here, well, traditionally, we have focused really on what I'd call the high-performance aspects of this space. The infrastructure networking with our increased 300-millimeter capacity, a whole bunch of shrinks that we executed on 50-nanometer, products that we moved off of 200 millimeter, we are now participating in what I'd call the lower-end segment of this space, generally, the core customer premise gear. This is DSL boxes for instance moving into the home.
What's notable in this space as we're doing an awful lot of product development in these upper end systems is that our competition is not necessarily so much the other DRAM developers as it has been in the past. What we're finding is because of the logic requirements in this space, the very, very high speeds, we're increasingly finding that our primary competition are logic vendors, companies that happen to be good at high-speed SerDes links for instance.
And frankly, Micron competes very well there. When you have the memory technology, obviously, you're not having to use some kind of an embedded DRAM process. And with our in-house logic expertise, we're finding more and more of these design wins moving to Micron.
Graphics and consumer. Fair to say, a little bit of a slowdown here, specifically in the DTV segment. I'll speak more about what's happening in game consoles as well. A pretty good push for more memory coming up in the Game Console segment as a level of redesigns. We'll start to hit it over the next couple of years.
DTV, the memory requirements continue to go up just a little bit soft in that market right now. Personal computing, certainly soft out there. I don't think that's news. What that hides is a pretty good space for innovation here that's giving companies like Micron with a focus on low power and opportunity for tablet, thin and light participation, frankly using some opportunities to reduce power with new architectures. In this space, I'll also say that with some of the current market weakness, we have moved relatively recently quickly up to a shipped module density of 4 gigabytes in this space. So elasticity is still working. We're seeing some favorable trends in that sense.
So all in all, we like how the portfolio is positioned across these 4 key segments that we spread our computing DRAM portfolio across.
Now specifically talking about the cloud. What we've done here is on the left-hand side, giving you an overview of really what things look like directionally for the shipped application of memory. So this is really how memory is making it out to the systems out there divided up between the public cloud and traditional enterprise computing.
We've further broken out the public cloud between really what we call, platform as a service, software as a service and traditional infrastructure as a service, various forms of cloud computing and how those applications make it out there.
And in the enterprise computing space, really our traditional segment there focused on upper end servers for high-performance computing and some level of IT infrastructure that is considered the cloud, although really more private cloud usage.
The key thing here is that in the space with this growth, it's actually, I'd say, nontraditional memory usage. Certainly, very focused as we point out on this being a cost center. So quality and reliability matter. It's certainly fair to say that making sure that the CPU is operational all the time, meaning memory latencies, peak bandwidths, get the memory to the CPU, is very, very critical to reduce that total cost of ownership.
And across the bottom, I'm showing you 4 technologies that are key to make sure that, that happens. Starting with load reduced DIMM on the lower left, either simply modules that allow us to get a lot more density into these servers, generally 32- and 64-gigabyte modules.
We then go to 3D stacking. This is actually stacking DRAM memory chips using through silicon via technology and buffers that we've placed in one of these DRAMs to take that density up a step further. So all of a sudden, you can now go to 64-gigabyte and 128-gigabyte type modules. This is real demand out there driven by the cloud. X DIMMs, really consider this flexible module form factors. There's a number of companies out there that because of different blade and server form factors are asking us to develop modules for them that have different profiles. This used to be really the province of companies, module companies, I won't go through them right now, increasingly, the direct OEMs are coming to us saying, "Look it's better if you just take your memory and get this module done for us directly." And that's been good for us.
And finally, something called Hybrid Memory Cube, which we spoke about in February. I'll conclude with a picture of where that's at but really, a upper end computing platform for high-end systems.
Now briefly, back in February, we announced a product that we call, reduced latency DRAM3, RL3 for short. Micron is the world's primary supplier of something called RL2 to the networking space.
Reduced latency in this context, meaning an upper end computing DRAM memory optimized for networking. And with our RL3, I just wanted to give you a brief update. Since February, we reached initial silicon on this. We got it sampled out to the market. And critically, we're working with a few companies here, of which Alcatel-Lucent is one of these. About 1.5 months ago, Alcatel-Lucent released something called a 400G chipset. This is a 400-gigabit ethernet switch, a networking processor and chipset that's optimized for 400G, the world's first. And you can't pull this off without RL3, and we've got it. So this has been a huge win for us, a great chance to do some joint press with Alcatel-Lucent and show the value of an innovative product like RL3 in the networking space.
Thin and light really, a category here in between tablet and the traditional notebook, although I think some of these same trends apply to tablet as well. There's a lot happening here both with the CPU vendors, the operating system vendors. And the nice thing is that across memory, there is a level of innovation here that we're able to capitalize on. In particular, these are applications that require more DRAM than traditional cellphones.
As of today, less memory than a typical notebook, although we're seeing those trends push pretty quickly notably in this ultra book space to a 4-gigabyte type loaded systems. That means that generally, you have to go out with a different architecture than what LPDDR, DRAM, in other words, traditional cellphone DRAM has been optimized for. We're taking that trend and capitalizing on it by making sure that in our standard DRAM designs, we have the features, the low-power capability and a way to test these parts for lower power that allow us to get out and really capitalize in some of the trends there. So some nice uptake, specifically in the Thin and Light segment.
And talking about consumer again here. I thought it'd be beneficial to show you across a couple of key applications how this looks in terms of megabyte per system. On the left, what we have are game consoles. This is a space that's been pretty flat for a number of years in terms of the average shipped density per system. That's going to be changing here pretty quickly. I think everyone realizes that these systems are somewhat clumpy in their development. The next generation of system is under development now and that because of 3D and some of the bandwidth requirements, drives the megabyte per console up fairly quickly. So we're anticipating some good growth here.
We've worked with a number of these vendors specifically on both custom and semi-custom solutions in that space.
And then in DTV, again, a little bit softer out there right now from a market segment, but some great growth overall that we see continuing because of 3D and the needs of these systems as they pull in Internet connectivity that drives up that megabyte per box.
Now something that nearly, every one of these segments has in common is a focus on power. Power in our space is very critical. What we've done here is try to put side-by-side measured silicon data across 30-nanometer, 2-gigabit solutions, Micron on the left compared to our 2 largest competitors in the middle and right column. Suffice to say, lower numbers are better here. These are measured. Current readings of the device in milliamps across a variety of operating conditions. I won't go through the minutia. Suffice to say, these different IDD numbers correspond to different ways to access the memory that JEDEC, meaning our standard setting organization has set out.
And across the board, I think you can see that Micron, we've bolded the lowest number in green and the highest number in red. We're pretty pleased with how our 30-nanometer, 2-gigabit solution stands out here. Power matters in these systems from a total cost of ownership, not just in servers, but also in consumer gear as well as desktop notebook. Lower power is a key competitive advantage. And we're pleased with how we're situated here against the competition.
Lastly, we spoke in February. We rolled out something called Hybrid Memory Cube. This is not 2011 technology, but it's something in development that we're working on now to get enabled in the industry for upper end platforms. We're pretty pleased with this.
What it is again, as a reminder, a stack of Micron DRAM memory chips sitting on top of a logic device that we have designed. That's all connected with something called Through Silicon Vias. And when you do that, you open up a level of bandwidth that is just simply unobtainable any other way.
So the next generation of high-performance computing as well as frankly, the upper end networking gear, when they saw this solution, their eyes opened wide up. We, in the intervening months, we've been working on a large number of design wins, again, not 2011 technology. It takes some time to get this enabled because of, frankly, just the CPU changes that are necessary to take advantage of that bandwidth. But when you do that, you get a level of performance that equals somewhere between 20 to 25x the amount of peak bandwidth that you can get off of one high-end, industry-standard DDR3 module today. I'm not just talking about an individual chip. I'm talking about the entire module, 20 to 25x the bandwidth out of one single placement of these Hybrid Memory Cubes.
And to that point, we've got a roadmap here. You'll be hearing more about HMC with some public forums coming up in the next months and some articles that will be hitting the press soon. I'm very, very pleased with how this is going across the landscape.
This is our roadmap here, which gives a picture of the DRAM core component that gets stacked. Again, we take 4 to 8 of these chips, this 1, 4 or 8-gigabit DRAM arrays, stack that on top of the memory for a total solution of anywhere between 4 gigabytes and 8 gigabytes. We have the SerDes links underneath and the packaging type. That generation 3 up there, as we talk about 320 gigabytes per second out of one cube, that's again 25x the total bandwidth that you would get from an industry-standard, high-end DDR3 module today. So we're very happy with how that's gone.
And in conclusion, I'd say that both from a segment perspective as well as today's product portfolio and finally, looking at future innovation, we feel like we've got the products necessary to win in these key segments.
So with that, I think we're going to go ahead and conclude the business unit section and take a 15-minute break, at which point Mark Durcan will come back, and we'll go ahead and tear into operations. Thank you.
Fairly brief. We're a little bit behind schedule, but 15 minutes or so, and we'll get you back in here, and finish with the presentation.Thanks.
Good morning, everyone. My name is Mark Durcan. It's great to see so many familiar faces here again this year. So thank you all for coming. I've just really got a few brief slides on technology development and some of the new and interesting things Micron's working on beyond our current product portfolio for the future as we move into an era where Moore's Law becomes more and more tenuous and an update on what we're doing from a capacity perspective in some of the more significant manufacturing operations. So without further ado, I'll jump right in.
You've heard from all the business units that are driving down the current product roadmaps to support all the various market segments we address. I'm going to really address the things on the left here that function across those to prepare us for the future and to deliver the products into the marketplace themselves.
So those of you that are familiar with Micron know that we've always used a technology that's a significant lever on our business to try and position ourselves for the future vis-à-vis our competitors. Not necessarily always being the earliest adopter of new memory products in the past, but always being the early adopter of new technologies that can drive cost efficiency in our business. And that hasn't changed today even as we drive down to the point in semiconductor processing where some of the wheels are starting to fall off at some level relative to lithography and advanced materials development.
So for Micron today, we continue and invest in new and innovative technologies that will keep us ahead of the competition, and we actually are embracing the world where things get more and more difficult as opposed to more and more the same because we think that plays to our strength in terms of technology development.
One of the areas where we feel like we were early adopters here 4 or 5 years ago was in the area of pitch multiplication, double-layer patterning, et cetera. We continue to drive innovative techniques to drive our lithography roadmap down below 15 nanometers in terms of our ability to pattern tight pitch lines and spaces as well as isolated contacts. And as we look to the future, we believe that EUV plays a place in the semiconductor industry, probably first adopted for isolated contacts. But for Micron, we are not letting that get in the way of our roadmap and our ability to continue to drive to smaller geometries.
And this photo up here at the top, it's not an EUV photo. You can tell actually those of you that are familiar with lithography, if you look at the line edge roughness we achieve with some of the techniques we're using, we can drive very, very tight pitches with very, very good edge definition and really remove this as an obstacle to Micron's ability to continue to drive smaller geometries.
As we move to the future and think about what core competencies does a leading-edge semiconductor company have to have in order to be successful deploying new memory technologies for the future and really reconfiguring memory for advanced computing architectures, our consumer devices et cetera, novel material development has never been more critical.
Those of you that have known Micron for a long time know that we've had a long history of chemical synthesis, organic, organometallic precursor development, et cetera that have driven capacity development over the years.
Looking to the future, there are other things that matter as well, things like combinatorial material deposition techniques that allow you to screen very large numbers of very complicated ternary material solution sets for some of the new types of memories we want to build in the future, and I'll come back to that here in just a minute.
And finally, as you work with all these new materials, you also need, all sorts of new chemistry methodologies as well as all sorts of new metrology techniques to really understand what's going on with the materials you are depositing in a much more profound way at the atomic level. And so Micron continues to invest in those areas as well.
The future memory technologies I'm talking about are very wide and dispersed in terms of the different types of memory solutions that the industry is working on. I've selected 4 here today that Micron is actively engaged in for different segments of our business. And you can see in each case, I've listed the type of memory technology, but also where we think that first fits or is most applicable in terms of memory hierarchy.
So starting at the top, phase change memory. You guys, I think, understand that with the acquisition of Numonyx, Micron had a significant infusion of experience relative to working with phase change memory. Although we have a long history ourselves in the phase change area, I actually worked on it myself I think in the early 90s, but this is truly coming to the memory industry now.
We think first as a NOR replacement, but we see lots of applicability beyond that and have made really significant progress here over the last year since the merger with Numonyx in terms of improving cycling reliability as well as yields and really feel like this is now poised to become a NOR class replacement technology in the next year, as well as with increased development in scaling, potentially a successor in the storage arena.
There's been a lot of talk in the press recently about Spin-Torque MRAM technology. And most recently, with the acquisition of Grandis by Samsung, but also you've heard about other startups in the STT arena. Micron has a long history in MRAM technology as well. We worked a number of years ago on a magnetic tunnel junction technology. We currently have a significant program in the Spin-Torque arena. And this, we really look at as a technology that can fit
really into the system main memory arena. It's complicated technology. It's very, very complicated. If you look at the picture here, you can probably make out 6 or 7 laminates in that metal stack there. But in reality, these things can have on the order of 20 different complicated material junctions associated with them. So very, very complicated technology from a materials control and process development perspective, bringing a lot of challenges in the manufacturing arena, but also a very interesting technology in terms of the space it can cover.
Capacitor-less DRAM. Some of you may understand that Micron has acquired intellectual property and capabilities from a number of different companies over the last couple of years. In future DRAM replacement technologies that utilize noncapacitor storage in order to form a Dynamic Random Access Memory.
One of the nice things about the again, the Numonyx merger is we have some significant assets and technical capability in Agrate in Italy that can be directed towards these types of not materials challenges that require really truly a state-of-the-art materials development, but towards projects that are more of an integration challenge and a device engineering type of challenge. And we've actually made some very, very significant progress in the capacitor-less DRAM area here over the last 6 months. And we're starting to feel comfortable, but there's actually maybe a roadmap here that makes some sense for the future of DRAM with the development of fully functional capacitor-less DRAM arrays. And it's just something to watch for the future not coming to disrupt our business in the next year, but certainly, a reliable path for the future.
And finally, in the resistive RAM area, Micron has a long history in this area as well. Many of you may recall that we licensed technology in the ionic memory area roughly a decade ago, did a lot of work with some material sets after being in the sort of 2002 to 2005 timeframe. We've been more quiet recently about the types of our RAM materials we're working on. But we have significant programs on multiple material sets in the resistive RAM arena that we think are very promising. And in this, for this particularly type of memory, we think it's most applicable probably to a storage type of application.
So a lot of very significant innovation going on at Micron today in order to position the company for some of these future memory technologies as we continue to crunch down more as well. And we actually think that this is a good thing for Micron. Even though there's an inflection point and potentially a disruption, this actually plays to what Micron does best, which is a real true core competency in complicated process and materials development.
In order to make sure we have the resources to deal with all these things, we're well down the way in construction of a new addition to our 300-millimeter R&D facility in Boise, Idaho. This is roughly a 60% floor space increment, so we're adding roughly 50,000 square feet of R&D space, should be completed in the January timeframe and available to start adding incremental tools to support this additional work we want to do in our future memory technology areas.
Okay. Moving on to operations. The Flash Singapore operation continues to execute very well. We told you that at the last meeting that we intended to be roughly at the 60,000 wafer starts, sorry wafer ounce per month level near the end of the year. We're still on track to hit that plateau with the capability to ramp the facility on up beyond that as we see market conditions support that kind of incremental investment. And that's still our plan as we move forward into the new year. Now one of the questions that always comes up is so what's your capital budget now look like for 2012? And this will be variable. We'll be flexible, as we always are, relative to how we see the marketing, how the market develops. But our current plan is to target a roughly $2 billion CapEx as we move through the fiscal 2012 for Micron. That's down from the roughly $2.9 billion we told you we would spend and that we are generally on track to hit in the fiscal 2011. This capital will, of course, fund ongoing tool installations in the time fabs complete the ramp to the 60k level as well as incremental investment in the existing 300-millimeter fabs to drive technology migrations.
A brief update on the Inotera facility. I actually did not get this slide updated to reflect where we currently are, so I apologize for that. But it is generally indicative of where we are today. We're well into the 42-nanometer ramp. In reality, the 42-nanometer ramp is now steeper than projected here. On a starts basis, we believe we'll be fully ramped in the next month to 6 weeks. In the 30-nanometer silicon, although you don't see it as a significant piece here on this chart, we actually have 30-nanometer silicon running in the Inotera fab already. So we're down that path and that ramp will progress as we see early results on the first engineering lots and as we evaluate market conditions moving through 2012.
Relative to the maturity of 30-nanometer technology, we're making great progress. This wafer out of our Virginia fab where we're now ramping 30-nanometer technology significant rate to support not only the PC business in this current quarter, but also the server business as we move into the first calendar quarter of 2012.
And finally, an overview of how we look at our capacity moving forward on a global basis, looking by product category. And as you look at this chart, you can see that we have really significant growth in the overall capacity. But also a migration of technology to advance nodes at a pace that is not necessarily driven entirely by driving to a lowest cost per bit going into the most commoditized space, but driven by the demand to the end applications or the BUs we've been talking about and driving to the highest margin products that we can that we want to support.
With that, I'm going to turn it over to Steve and he's got a few things to wrap up on industry status. Thanks.
Thanks, Mark. So I thought in this final section, I would share some data around competitive benchmarks and then spend a little time around some industry dynamics. And then wrap up with a few more comments around that and then have the rest of the group come up for Q&A.
So to start off, if you look at the competitive metrics, we had some discussions in previous times around the P&L statements, and what we tend to believe now is that given the various ways that companies treat their accounting, you really have to look at a couple of basic metrics around the balance sheet, of course, cash is one of them. And the debt structure is another one. And when you look at what's happened with us relative to others, you can see that even in cash covering short-term debt that we are far better off than everybody we compete with, and it's obviously because we manage our cash pretty well, but also because we don't have the kind of debt structure that most of those people that we compete with do. Now I would think, the caveat by the way, that in all of these data, we don't have Samsung because we can't get the broken-out data. But it's worth highlighting when you look at a number of these metrics that I pointed this out earlier.
When you look at the P&L statements of the various companies, it doesn't really always give you a true picture. An example would be that Elpida adopted 9-year depreciation. They decided to go to a 9-year depreciation instead of 5-year depreciation, which we do. And if we were to adopt that policy, our net income would go up by $1 billion for the year, for every year. Our net income would be increased by $1 billion if we were to adopt that approach. And so you have to be very careful when you look at the P&L statements based on the data of various ways that companies have to do their accounting. And the fact of the matter is, at least with respect to DRAM, Micron is the only company left in the Western world that produces it. And so you have to keep that as a backdrop. But when you look at all these metrics, you can see that when it comes to the performance that Micron really has signed, so to speak, compared to most. And I'm not trying to suggest that at all these metrics, we are better than everyone else. But clearly, we're either at the top or very close to the top when you look at a lot of things that matter in our business around our ratios and debt and cash.
And then when you look at the gross margin story, again, comparing Micron's gross margin over time, if you look at this chart here, it gives you a sense that our model change, and we went from being a little bit lower than the industry average to a little bit higher. I actually want to point out that when you look at the historical perspective when our average was lower, remember that we were transitioning the company at the time to a different model. And so we had a penalty to pay as we invested in other things that didn't generate much for the company like NAND, as we were going through this period of time. But ultimately, our portfolio and our model caught up with us and our margins are better on average than the industry now.
And in particular, if you want to look at a shorter period of time and you want to get a little bit more granularity around that, clearly, I think what this demonstrates is that when the market's worse, our gross margins are much better than the industry average. And when the market is doing well, then obviously, the more commoditized stuff tends to be equivalent in terms of margin to the more specialized stuff just the way the industry works. And fortunately, we participate in both, but we are able to mitigate the negative impact on the margin with a much better product portfolio.
And then cash flow. I was pointing out earlier, cash flow mix has big difference. Our operating cash flow is pretty good, and we've continued to be very focused on that as we've gone through these cycles.
Now this chart, I thought was fairly informative and I think pretty instructive in terms of what's going on with where we're at in the industry. This only goes back a few years but most of you remember, the financial collapse, so to speak, or the economic meltdown in the 2008 time period, a lot of people forgot that for the DRAM industry at least, the downturn started way back in late '06, or early '07 when we started having the collapse in the ASP. And so by the time when we got to the world meltdown, we are going to be experiencing our own downturn for quite some time. And this shows the ASP erosion through time. This is an average. This is an average cost curve for the industry. And what you see on these blue bar charts are, this is the range of the cost of the companies that compete within the industry. And clearly, we're taking a variety of data and coming up with this chart.
But a couple things to take from this. One is, we think that the industry on average, at least in the DRAM space, is pretty much more cash cost effective, on average. Now there's some companies that continue to be lower than cash cost and there are a number of companies that have higher cash cost in the selling price today. And you're seeing that reflected in a variety of reactions that are going on around the world with respect to what they're doing in DRAM. That's the first point I think that's worthwhile noting.
The second, is that I would point out, that we have not -- that the selling price has not only been in decline in the last quarter. Certainly, we've had some ASP erosion that's been pretty significant in the last quarter or 2. But if you look at where the ASP ended up flattening and then peaking here, it's quite some time ago. So we've been in decline on the average selling price in the DRAM business for quite some time.
And the reason I point that out is because I can't tell you whether because of the industry dynamics, and I'm going to show you a little more data around wafer supply in the industry, I can't tell you whether or not we're going to bottom right now or whether for a quarter it goes down again or 2 as we've seen in a couple of prior example. You see here it kind of lasted for a quarter or 2. Here, it was maybe more like 2.5, 3 quarters. I don't -- it's impossible to predict exactly where we're at in this curve.
But I think another way of characterizing it is the pressure that exists in the DRAM business today is really untenable for a number of players and that something will have to happen here in the next quarter or 2. And the perspective around what's causing this, it really has been I think from our perspective, a demand-driven scenario, not a supply-driven scenario. And that's evidenced by even what's been happening in CapEx in this area is not compared to what happened in CapEx a few years back.
And in fact, this 2012 number, given what's going on in the DRAM industry today, is probably a little aggressive in terms of CapEx. But clearly, 2011, is going to end up quite a bit down from 2010, which by the way, still only reached somewhere around 60% or 70% of what happened in the peak in 2007. This is for DRAM.
Now NAND is pretty steady, as you would expect, and that's because of the growth rates, which I'll get back to in a second. Now on the DRAM front, this is the industry wafer capacity. And obviously, there are some forecasts here, but I mean think of it this way. We pretty much know the capacity that is going to exist in the DRAM world over the next 1.5 years, 2 years because all that stuff is in play. It takes a while to construct the facility, not really new facilities under construction. There's optimization, which I'll show you in a second. But the bottom line is that the wafer capacity that existed in the DRAM space clearly, it fell off. There's a bunch of 200-millimeter capacity that came off. Some of you will probably remember that. And then there's been re-utilization of some capacity that existed. In other words, this stuff that was sitting out here and it's just kind of slowly come back online and trickled. But there's nothing dramatic going on here with respect to the supply in the DRAM industry.
It's primarily been a lack of demand which has caused the most recent ASP erosion. And I think that, that actually will speak well to something settling out here in the near future because we don't have any of these large chunks of capacity that come along on this space.
Now on the NAND, it's a little bit different story as you might expect because the growth rates are so much higher. And remember that we're sitting in here and it's expected by the way when we get to really the kind of the first quarter of 2013 at the wafer capacity in NAND in the world will exceed that of DRAM, and that will be the first time that, that's happened. And here, we clearly show growth rates and capacity, both in silicon and in technology advancement. And I think, by the way, that's why the NAND will behave more like DRAM did in the earlier years where it was very price elastic. In other words, we know right now that we're experiencing much more price elasticity in NAND than we are in DRAM for all the reasons that the prior speakers touched upon.
If you look at what's happening with the capacity that exists in this space, first going to DRAM. There's a couple of things on here. One is the expected growth rate of the bit supply. And the other is what is the source of that growth rate in the bar charts. And historically, I just want to point this out again, that if you go back into the '06, '07, '08 timeframe, when we were basically peaking in CapEx and when the market turned down, we had large amounts of the new capacity that was coming online in the form of a new wafer. So not only did you advanced the technology and get more bits because you're using a more advanced technology, we built new wafers into the industry. And that caused these kinds of bit growth rates.
If you look at what's happened out here and particularly since 2009 time period,
clearly, there was some capacity that came off-line. Everybody knows that. And that capacity came online in 2010. But if you look at what drove this 45%, 50% growth rate, it was either advancing the technology or it was bringing underutilized capacity back online. And that's a far different picture than what you see here, which is why I'm actually a little surprised today that we have as much pressure in the DRAM business that we do. But I always make the caveat that we can't predict the world economic climate. And clearly, given the weakness in Europe and given the weakness in the U.S., we're just -- and by the way, the success of a couple of products like the iPad, which are drawing -- it's good on the NAND business, it makes it more negative on the DRAM because it doesn't have quite the DRAM content. And all those factors are playing into what's going on in the market, I'm a little surprised that it's as weak as it is in DRAM because when I look at the underlying data around capacity and supply, it just doesn't demonstrate the kind of new bits into the marketplace that we experienced in the past to drive this kind of imbalance. So I think that that's also a reason why it will settle out fairly quickly and then we will have some kind of extended period of time where it's under the kind of pressure we're experiencing right now.
Now on the NAND side, what's also interesting, and I would not have really expected this, is that the capacity, the new wafer capacity, there is some. But it's also nothing like what was occurring in the prior periods of time. So we are getting more bits into the marketplace based on new wafers coming in. Micron, as you heard, being one of them with the IMF facility that we have in Singapore. But as a percentage of total capacity growth in terms of the bits, it's still dominated by advancing the technology as opposed to bringing new silicon. So we're not in this environment where we have crazy amounts of CapEx being spent in the memory world. We're in an environment where we just being much more subjected to the world demand being somewhat weak, in obviously, a couple of the major markets.
Now my final slide, if you consider all the players in the Memory business, and I like to show this slide. To any of you who have heard me speak on this topic before as really my final slide around what's going on in the industry, there's a couple of things that are worth pointing out. One is, we always talk about where we're at in relationship to Hynix and Samsung, and remember this is a total memory perspective and that's what we started doing a couple of years back because of the product portfolio. That it's really pretty amazing at the differences that are starting to occur as you go down this curve in terms of market share of the world memory market. And we've said for sometime that the model will continue to drive this to only having a few players that survive. And the pressure that's in the marketplace today, and particularly in DRAM, will probably be a catalyst, the longer it exists like it is now, the more pressure, the more catalyst that will be in play in order to try to consolidate this industry a little bit further than it is right now. And I don't need to speak. I think I'll answer questions on it. I don't need to spend too much time speaking about what's going on in a couple of the countries in Asia, but there's a lot of weakness out there and many of these companies, in particular some of the companies towards the bottom of the chart are in really bad shape. And I think that something ultimately will have to happen there. And it's a little bit uncertain as to how that plays out. But I think the current situation is pretty untenable in the DRAM business and something must happen.
At the end of the day, when you look at the product portfolios, there's really only 2 companies: Micron and Samsung. They have the portfolio that we have, and I think that companies will have to continue to evolve in order to survive in the kind of environment we're in right now, where there's additional consolidation.
Okay. With that, I'm going to wrap up. I'm going to open it up for questions. And I'd like all the rest of the team to come back up.
Okay, questions. And by the way, it would be very helpful if you have a question that you can use the mic so that the people on the webcast can hear it too. Right here.
Unknown Analyst -
Probably a couple of questions. Steve, can you start off by updating us on what you see as happening today in the PC markets by way of demand including the back-to-school? I know at the last conference call you talked about 5 weeks of inventure [ph] where are we now with that, perhaps that's my question.
Yes, Mark, do you want -- Mark's probably -- do we need a hand a mic? I'll comment too, but, Mark, why don't you go and comment on where you see PC market and what's happening there?
Generally in the PC market, not a real shock in terms of the current state of lower demand than expected. We started to see that in July in the forecast into us from the major customers that play in that segment. And we talked about the macroeconomic conditions. That's obviously a big driver of it, and certainly, alternative computing solutions around tablets and even smartphones, when you look at what the teenagers are doing with computing on the go and whathaveyou. So we started to see significant shortfall of demand in the PC segment in July, and from the bigger branded players. And so we continue to watch that dynamic as we look at the holiday season. Obviously, still big numbers, but off of the growth number that we thought going into the period.
Yes, I think that's been confirmed now by a number of people that participated in the supply and in the PC market, whether it be Intel or others that the -- when we started out the year, I think a lot of people were talking about 15%, and others talking about 2% to 5% or somewhere in that neighborhood. And that's what we see playing out right now.
Unknown Analyst -
[indiscernible] 4 gigabytes modules for PCs? How widespread is that? Is just is it -- and by regions if you can just talk about what you see and obviously, emerging markets have been very strong. With China now being probably larger than the U.S. in terms of unit demand. Can you talk about regional mixes in terms of bit growth just for us to get a sense of how much run rate there is for [indiscernible]?
Well, in the, clearly, in the desktop notebook space, we have seen average shift densities from our calculations pushing north of 4 gigabytes already. I think the latest figures are roughly 4.3 gigabytes per shipped system. And that's prior to really some of the latest market weakness, and again elasticity is starting to occur there. So our shift module density has just pushed over really about the last 3 months up to a 4-gigabyte average module. So we're pleased with the content per box on the desktop notebook side in tablets and in ultrabook. Obviously, these have started out with lower content, but I think you'll see with some of the developing platforms coming out over the next year or 2, a pretty strong push up to 1.5 gigabyte type numbers in the tablet space on average and in the ultrabook space, somewhere between 2 to 4 gigabytes. So we are pleased with the elasticity there, partially a function of the soft market, but by all accounts, there's still room to grow that.
Unknown Analyst -
Maybe a question for either Ron or for Steve. Modeling the company historically has been extremely difficult, and now you've got this mix element that's coming into the fray both intra-memory technology and inter-memory technologies. How should we think about mix in any sort of given quarter? Are you always going to try to maximize profitability within the quarter? How much control do you have over that and what kind of guidance will you be giving us around mix going forward? And then I have a follow-up.
Yes, well, mix is always difficult to predict too far into the future because the market dynamics are changing. But I think companies like us, and I think probably Samsung is in a similar situation, will always get to just optimize for profitability a particular segment. As an example, if you think of a company like Hewlett-Packard, they buy product that goes into the PC business, and they buy a product that goes into the service and some of their higher-end business. And most companies don't allow you to just participate in one and not the other. And so we're consistently under pressure. If they have any broad business like that to supply the entire line of product that they need. And so we can't just say we're only going to serve product in the server business and otherwise. That's kind of 1 thing that affects mix. And then the other thing that affects mix, of course, is the decision around which products to build to go to different markets, that actually go to different customer base, different market segments. And an example would be, an example of that would be, are we going to build more RL product for the networking market as opposed to other product for some other market? And those are -- the prior one, by the way, tends to be pretty consistent. At certain customers, we know we have certain percentages that we're expected to meet in their various categories of that particular customer. But with respect to the market differences, those are pretty hard to predict and that mix changes a lot. In fact when I'm showing the average ASP and the average cost, if you noted on that chart, it was really kind of the commodity spot market part and not reflective of our average ASP because we have a much better portfolio. But I think it that one of the discussions now about how do we help people understand the complexity of our financial model outside of these types of discussions because it just takes more time and we don't have that kind of time to go through it. We might try to look for another forum where Ron can spend time and actually go through how our financial models work and how mix impacts us to help people understand it much better. Because it is -- because we are pretty complex now, and we have enough segmentation, enough portfolio of data that it's just harder to grasp that. Now we'll never be able to tell you with precision ahead of time exactly what the mix is going to be because we're reacting to market. And we can affect that to some degree within a quarter or 2. So anyway, I think that's how we're looking at it.
Unknown Analyst -
And then, Steve, as my follow-up, you've always done a good job of acquiring assets cheaply, especially in bad market environments. But per some of your capacity charts up there, the push to consolidation without a lot of wafer start growth doesn't seem to be as strong. And given your IP-base, given what's going on with the mix strategy, given that it's starting to play out with more stable cash flows, why not look at your own stock as the cheapest asset out there to go acquire during the soft period -- patch of business?
Yes, I think -- let me try to answer that in a couple of different ways. I think, one point you're making is that the industry has been consolidating and there are less and less cheap assets to buy. And that has been a way for Micron to be more effective in our capital deployment primarily compared to Samsung. That's allowed us to grow more cost effectively than we might otherwise be able to do just through organics. And that, and as I noted, if you look at that chart that I showed you, my last slide, that the number of companies that are out there where you can go acquire inexpensive assets that are of any kind of significance of scale are getting less and less. And I think that's true. And that means that it's not completely done, as I said. I think there's another play or 2 to make there, but that also says that those kinds of opportunities are less and less. As those opportunity become less, as we generate more cash flow, then your follow-on comment is, what are you going to do with the cash? Why not, at certain times anyways, why not look at your stock as a cheap asset to buy? I made a comment that as our share price trades below book value, we do tend to look at it in those ways. And as you know, we've done some things historically. Of course, our priority is to make sure the balance sheet is strong, to make sure our cash position is strong and to make sure that we still have the necessary ability to invest in the company in terms of research and whatever the things that we need to do to make sure that we have scale and competitiveness. Now having said that, I also watch with interest the debate that occurs out in Wall Street about -- and in fact that not that anybody here bases what they do on Jim Cramer's comments, but I mean, it was interesting to listen to some people including him just rip CEOs and companies for doing stock buybacks. They need to stick with what they well, et cetera, et cetera. I actually -- my perspective is -- the way I look at that as the same way I look at opportunities for inexpensive assets. That there are times that it makes sense and there are times that it doesn't make sense. And we're going to continue to look at it kind of like that. So I'm not disagreeing with you that there may be times where it will make sense to use your cash in order to reduce your shares outstanding as opposed to other times. But those are kind of game time decisions you have to look at constantly as you're balancing capital needs with balance sheet preferences, with -- reduce our share count. And I've said it before. I mean, I think we're going to continue to look at that, and if you look at the transaction we did most recently, the nature that we did, we did it like we did it because we wanted it to be non-dilutive. Now somebody could've argued that we should've done the stock buyback simultaneously with that transaction I could've waited had I known that our share price is going to be 5.5 instead of 7.5 or whatever. Sure, but that wasn't -- I mean, we did want to make sure that people understood that we were very anti-dilutive in our thinking. But in particular, I wanted to make sure that all of our shareholders understood that in that particular transaction, we are -- there are things and opportunities that may surface. We want to make sure we are a little better positioned than we were then, how to do it, which is why we raised the money. And in addition to that, I wanted to make sure that people understood that we're very sensitive to dilution. And that's why we did the equivalent share buyback at the time so that you didn't have to -- when I say trust me, we're intending to settle it in cash. I didn't want you to have to trust me. I wanted you to know that I was going to take the shares out right then such that in the event we settled in shares, we already covered it because we took those out of the market at the same time. So I hear what you're saying, it's in our thought process all the time, and we'll just continue to evaluate as we move forward. There was a question over here from the room. Okay, go ahead.
Unknown Analyst -
This question is probably for Glen Hawk. You mentioned at the very beginning that Flash was driving the cloud, and in particular cloud storage. But like with the emergence of cloud storage services such as Apple's new iCloud and I think Amazon has competing one. Is it possible that these services could actually have like a detrimental effect on NAND demand as average storage memory would be less for individual devices?
A good question. I love to talk to people about that very question quite a bit. I definitely have an opinion on it. When I made the claim that Flash was driving the cloud, I'm a Flash memory guy, so of course, that's how I'm going to look at it. And I think on this question as to where the storage is going to reside, I have a strong bias that says, not only are you going to need a lot of storage and Flash in the cloud itself, but I think that we're looking at a future where consumers are going to want to have a lot of that available right there at their fingertips because they're going to be able to do different things with that. And I think as time has gone on, we have a lot of fun talking recently about some of the opportunities -- you know, what the future is going to look like and just to call out one example here, there's a great website, if you haven't seen it, davidbergman.net. He's an action photographer and videographer based out of New York, and he has a great website that gives you a glimpse of the kind of things that you're going to see in the future that people are going to do when the next-generation cloud and these next-generation devices are available. And what he did was he was asked in 2009 to photograph the Obama inaugural. So in January. and he thought, "Wow, this is the first time that the swearing in ceremony has been opened to the public." So they opened up in the National Mall, brought in tens of thousands, hundreds of thousands of people, I don't even know how many, and he used the technology called Gigapan to essentially create an incredibly high resolution photograph of that scene. And I think the image that he eventually came up with after using his desktop computer to crank through about 6 hours worth of processing time, I think it ended up being like a 1,400 megabit pixel image. And when you go to this website, you just see nice photos of the inauguration. And you think, okay, what's the big deal. But you have the capability to zoom in and see individuals, individual faces. It is just phenomenal. And even though this took hours to produce and a lot of processing time, I think that's the kind of thing that you're going to see more of in the future. Interestingly enough, he didn't even know some of the moments that he captured with that image. If you zoom in, later on, somebody discovered that one of the people in the audience happened to be Yo-Yo Ma, the famous cellist, who was, of all things, caught taking a photo with his smart phone. And I thought, how cool is that? And I think that's a real vignette into the future. Just imagine when everybody has Gigapan-enabled smartphone walking around the planet. They're going want to record that content, put it in the cloud, but a lot of people are going to want to carry that with them. And so I think that my vision of the future is that the storage is going to be in both places. There's things that we can barely conceive of today that we're seeing people do today that's really exciting. So I'm bullish that there's still going to be a lot of storage in the clients [ph] . Sorry for the long answer, I'm pretty passionate about that.
Unknown Analyst -
Just one on the NAND side for a second. Clearly we've seen a lot of growth in smartphones in the past couple of years, and if we see that continue to grow at that strong rate, one of the things that you need to see is also some ASP declines to kind of get to the $150, $100 price points. Just curious in terms of where you see the bill materials pressure coming from. How much are you seeing in terms of the NAND embedded space? And where do you think -- is there any pressure downward on the embedded content of NAND beyond the 2x per 18-month that you talked about earlier, Mario?
Well, I think that I mentioned before what is partitioning in terms of architecture choice for the different models. I think a couple of years ago, talking about cellphone, when we're talking about other phones without partners, we're starting discussing about $100 maybe going down below $100, $90, $95. Today, in the low end portion, as I mentioned, we have a phone which is in the probably $25, maybe $30. And this is going to continue to go down, not at the same pace as before, but there is still the pressure from the competition in the market to decline that boom. And this is happening for every category. Of course, what do we have to exploit is the parallel growth of the density so that we will compensate the bit cost decline or gigabit the cost decline towards with an increase of density, so that it will be associated to the new application. So overall, we see the kind of stability that each of the management of the mix between gradual, continuous decline, technology migration and density addition in both -- in terms of no volatile memory as well as volatile memory.
Unknown Analyst -
Maybe just a follow-up. A broader question on NAND. You talked about a shift in you focus on NAND to more enterprise applications versus consumer. Can you just talk about if you had your druthers would it be 100% enterprise and 0% consumer or how do you think about the trade-off in that business?
I don't think it would be 100% enterprise. I mean, that's certainly not the only segment that we're going after in a big way. I think that's one that you're starting to see us ramp into in a more direct way. And what I see first in the future is we're looking to assemble a portfolio segment that we're participating in and enterprise is certainly a key part of that. But I think that they're still going to be a lot of great business for a lot of those client devices that require portable Flash storage as well. We're going to go after all that.
It's probably worth noting that whether or not it's in the DRAM space or whether or not it's in the NAND space or actually whether not it's in the NOR space that the margin that gets generated in some of the more attractive segments is a direct derivative of your ability to compete in some of the more commoditized segments. In other words, what drives us for our relentless focus on cost is the fact that we have to be competitive at kind of mid- and lower more commoditized segments, and that drives the margin. Somebody coming in into the marketplace who doesn't have that kind of portfolio who just wants to sell into the enterprise space won't have near the margin that we'll have because of our drive and some of the more commoditized spaces.
I should have added, our entrance into client SSD is a great example. You ran to about 10% of the market. What that enables us to do is to exercise our entire manufacturing chain, our supply line, really make sure that we've got that dialed in. One thing that you want to be careful of with enterprise customers as with automotive and other customers is that, that supply line continuity is something that's very, very important, and that's something that we can provide by having both in our portfolio. I think we can address those really, really well.
Unknown Analyst -
Two questions. First one for Mark. I just wanted to understand how fungible is between DRAM and NAND, given the complexity of scaling going from 3x to 2x up 2x to 1x over the next, say 12 to 18 months. Can you give us the metrics in terms of your past transition versus what do you expect or what do you anticipate?
Sure. Okay, so relative to you, our ability to move capacity back and forth between NAND and DRAM, there is the ability to do that, and we're doing some of that today. We have in the next quarter, you will see from Micron given the relative softness in the DRAM arena. Some of our DRAM capacity in one fab in particular move into the NAND arena. It's not something you do with a snap of the fingers and there can be an efficiency associated with it. But we have flexibility of the margin to move small amounts of capacity. And by small amounts, I mean, thousands of wafers a week, not tens to thousands of wafers a week from one node -- from one technology to the other. There's generally some slight inefficiency associated with that because you have a tool set that's optimized for one mix and as you move to the other, you're going to give up a little bit. But in certain situations, given varying demand profile, that make sense. I think we're not different than our competitors. I think most companies find themselves in the same position that if you want to make a long term transition, yes, a lot of the equipment is fungible. But it's not something that you can do very efficiently on a short-term arena basis.
Unknown Analyst -
And just a follow-up for Brian. How should we think about the specialty DRAM mix by the end of next year versus by the end of this year? Can you give us some kind of the metric there? And what is Micron doing to improve the specialty DRAM mix?
Certainly. So in the DRAM Solutions Group space, even today, I think you're going to find a cross server network and consumer. We are sitting at over 50% of our revenue being specialty DRAM today. That's just for the DRAM Solutions Group. We have room to continue to grow that. Generally, that's coming with improved performance at Inotera , things that we're doing on the supply side to take up the amount of server grade material. Suffice to say, we're at the point where we don't see the demand for those products being the limiter that it once was. So we have room to grow that. Where we end up at the end of fiscal year '12, I'm not going to comment. I'll just say that, that's specifically still our focus to drive it as much into the server and networking space, consumer as well, profitable consumer opportunities like game consoles, DTV as we do. So we're pleased with that percentage.
Unknown Analyst -
For Glen. The DRAM obviously, the demand mix that you've got. Are you still planning on competing or providing TLC in a significant amount? Will that be more managed TLC, Managed NAND versus more in the eMLC area?
Just to clarify, 3-level cell or Managed NAND?
Unknown Analyst -
2-level cell and then it will be incorporated in Managed NAND...
We do have 3-level cell products today on our 25-nanometer. We don't have to run a lot of them today because the segments that we're engaged with, the customers that we're dealing with, that's not a product that's high on their priority list. So we're running multi-level cell and single-level cell Flash for the most part. But we've got 3-level cell there. If the environment changes to where we need run more of that, we certainly will. With regards to the Managed NAND question, I'm not sure that I understood that correctly. But I think what you will definitely see from us in the future is a higher and higher percentage of our bits, our product being shipped with some type of management associated with it, and whether it's something like a ClearNAND which takes care of error corrections or something like embedded MMC, fully managed or ultimately a system level solution like an SSD or a PCIe card, much more than I think will be shipped in that in the future.
Unknown Analyst -
I have a follow-up with the -- I know that Toshiba and SanDisk do a lot with their TLC, and they're playing the commodity market quite a bit and for cars and USB drives and then they're migrating their TLC to some of their managed products like the I NAND would you consider using the TLC with the ClearNAND?
You bet. I mean, I think that there is also an element to what we're doing that we do seek to cram as many bits into the memory cell as we can. Different segments can handle different amounts of that, that's also a function of your error correction, your digital signal processing capability. So we can bring all that to bear in the future.
Unknown Analyst -
Steve, just even since February, there's been an extraordinary amount of movement transactions and so fort in intellectual property, some of the prices being paid for defensive reasons and so forth is kind of extraordinary. And I know you got a portion of your IP with Round Rock and so forth, and we talked a little bit about this last night. But how should we look at really Micron's core intellectual property from a valuation perspective and what you're hoping to achieve in the next few years and maybe sort of discuss through the acquisitions that you've made the last several years. Really what you have that others might find extremely valuable?
I think we look on this IP frenzy, maybe the way you characterize it, with a little bit of interest. But I think also some amusement, really. There are -- I think there are certain companies that have struggled with the lack of IP in their portfolio and are now trying to react to that and do something about that. That's really a lot different scenario for Micron. I think as many people know, we have an incredible IP portfolio. We've been one of the most prolific IP producers in the history of the semiconductor world. And by the way, that all evolved out of a scenario, that in the 1980s when we were very vulnerable, similarly like other companies are vulnerable. We, that time, decided to organically drive our portfolio, which we started doing in really the mid-'80s, late '80s. So when we see companies being acquired for an enormous amount of money to get 5,000 patents or 6,000 patents that's and we, and Micron itself, with our portfolio which is multiples larger than that, we look at it a little bit differently than trying to acquire IP to protect ourselves. I think our intent is, as we've said in the past, we've gone down the path and we want to try to monetize some of our investment in IP and that's what we've been doing. If I had a preference, and a number of us worked on this in Washington, I would -- essentially, the way the patent litigation occurs in the United States has evolved to a point where it doesn't reflect at all the intent of the patent system that we put in place decades ago. And if we had our preference, that will all be restructured and redone. And we think the way that IP litigation occurs in the United States today is ridiculous. But nonetheless, it exists. And so we're subjected to it, and as you know, we're constantly in various battles around that even though we're one of the most prolific IP innovative companies around. And as a result, we decided that we have to assume that the current system is not going to change. And if we're going to be under attack on one side, we need to monetize it on the other side. And that's essentially what we've been moving towards is continuing to try to monetize that IP. And whether it's selling it or whether it's using it for direct royalty payments, et cetera, and we've disclosed a number of things in the past, and if you add it all up, it's actually pretty significant dollars for us. If I had my preference, all that will go away. But it's not going away, so we have to do something with it. I don't think that -- I mean, I guess if someone offered to buy the company for $50 billion for our IP, we have to consider it. But I don't see that for our company in any ways as being significant in that role. And I think it's primarily because the IP aggregation in the semiconductor industry has, for the most part, occurred. And the way I would characterize that is we don't have very many new companies in the semiconductor world that are trying to acquire portfolios in order to protect themselves. We have lots of people and whether it be a lot of the social media or the cloud, I think where people trying to figure out where they've grown very quickly, they've become very large companies and they don't have intellectual property positions. There's a lot of that going on. But I don't see that much of it happening in the semi world. And I think that our focus will continue to be both to build and defend ourselves, but also monetize a lot of investments that we've made over the years.
Unknown Analyst -
And can you just give us an update, may be more qualitative than quantitative. But in terms of the initiatives that you spoke about in February and before about moving into some adjacent categories including LEDs and so forth. Where do you see -- how's that all progressing in this environment where there's just a lot of weak -- a lot of issues surrounding competitive pressures and so forth in some of those categories? How can you use the balance sheet, the company's balance sheet, to accelerate growth in some of those areas?
I'm going to have Mark comment specifically on a couple of those market segments. But let me first, Mark, I'm just going to do backdrop to our thought process around that. First of all, we're a memory company. And that's where the vast majority of our resources are spent. And how do we make sure that we have a broad enough portfolio and we can take advantage of opportunities exist around consolidation industry. As a memory company, we're very focused on that. The reason that we've had an interest in a couple of other categories that Mark will comment on, is because we also view ourselves as a very large silicon producers. And there's a certain expertise that you gain by being a very large silicon producer. And we've looked at some of these opportunities as ways to leverage our expertise and the ability to produce silicon and the ability to have manufacturing facilities to produce silicon as a way to explore whether we have strategic or that we can develop a strategic advantage in other categories that also might be very large consumers of silicon. And that's what we focused on. So Mark?
So along the lines that Steve was talking about, the most significant opportunity that we're interested in is in the solid-state lighting LED arena, and it's because we're convinced it's going to be a very big market and it's an area where we think we can bring core competencies from our memory business to bear to deliver something that is technically very differentiated from other solutions on the marketplace and bring us a potentially sustainable competitive advantage. We're happy with the way that that's going. It's not -- the program is not to deliver a me to LEDs. It's to deliver a superior performance, lower cost, higher quality solution that we can then deliver into value added lighting systems. We're making good progress. We have a business plan for 2012 that we'll start to see -- the first revenue in that market, but it's in the tens of millions of dollars. It's not -- it won't be significant to Micron in 2012. We do think it's a long term, a good opportunity for us and something where we can potentially do something that's quite competitive with other folks in the industry. So that's all that solid-state lighting. The other 2 areas that we talked about publicly that we're interested in -- by the way, on LEDs, the underlying core competency there revolves around substrate engineering and around gallium nitride types of structures. And so there are extensions of that around power devices and things that we may look at as we prove out the differentiated technology we have. The other 2 areas that we talked about having an interest in are the low-power mobile display arena. And again, here, we think we have a great technology that over time will become more important. But it's not significant for Micron today. And again, this is something that we think we have unique core capabilities in, because a lot of these integrated display technologies really leveraged advanced memory, and you can use advanced memory in interesting ways in order to drive enhanced performance. The technology we've selected is very, very low power, and therefore very applicable to portable applications. And we think it dovetails well with the customers we have today. So that, again, in 2012 won't be a significant business for us, but we think it has good long term potential for the company as well. And finally, we've been doing some work in the solar arena. And again, we picked something that dovetails with our core competencies that we thought would potentially deliver a long term competitive advantage. I'll tell you today, the solar business is just a tough, tough business. So we continue to look at what our technology can deliver that's differentiated and add value. And we'll look for ways to monetize that. But that's not something you should think of Micron jumping in a big way into the solar business anytime soon.
Unknown Analyst -
Just a couple of follow up questions. Steve, you talked about all the IT issues. Can I just ask you to update us on the timeline of the Rambus case? I know there's testimony that's going on right now. When is that expected to end and expect a judgment date?
Yes, they are always tough to predict because between the witnesses and there's 3 parties in the courtroom and so forth, I think that timeline wise, at least the last I've heard is that where the schedule is going and with the witnesses and so forth, and that it's probably going to be the next month, maybe a month and a half for that to conclude. And of course, jury deliberations are impossible to predict. And that will affect it. But I think it's -- I think we're getting closer, if you will. And as I said, I think it's sometime that in the next month, 1.5 months that they will probably finish all of the evidence. And then jury deliberation is unpredictable, and that will be whatever it is.
Unknown Analyst -
May I just ask, Glen some questions. In terms of the complexity of the controller technology, can you just talk about how that has evolved as you go into 90 nanometers and beyond? And how do you balance investing your own controller technology with your customers while you also control providers like [indiscernible] How do you manage that balance?
Yes, you bet. To your first question, as we get to each nanolithography, our controller requirements are increasing significantly. In fact there's an order of magnitude increasing the gate count as we go from our 25-nanometer NAND to our 20-nanometer NAND. A tremendous amount of horsepower is required to really get the most out of the Flash as we move forward. In terms of what we do internally versus externally, as I showed earlier, we have a very wide variety of segments of customers that we're working with. A tremendous variation in the form factors that people want. We know that there's no way we're going to be able to do 100% of those controllers ourselves. So it really comes down to, what is it that we're going to choose to do internally versus externally. The example that I had shown, I think is a good example of probably about the hardest controller that you could do, which is our PCIe SSD controller. Again, I think that's a very good proof point for us that shows that we know how to do leading-edge, high-performance controllers and get great performance results out of that. All the firmware, the ASIC, all that stuff is ours, the manufacturing is ours. So there is a good example of where we will absolutely be applying our internal focus going forward. The high-performance drives using the latest NAND lithographies, that's what it will all be about. Now that said, there's a large continuum down to devices like USB and SD cards, that in those cases, obviously, it makes better sense for us to partner with somebody. But even devices like that in the future are going to require some additional enablement that hasn't been there before. So we have to be prepared to work with those third parties as well to create the right controller capability. So everything in between is something that we'll be going after. And you'll see various approaches from us in the future. We're already doing that, basically.
Unknown Analyst -
How much is -- what you ship now is 3 bits per cell and when do you think that gets -- should become used in embedded systems and solid-state drives systems?
A very, very small percentage today is 3 bits per cell, particularly in our SSDs. We're using multilevel cell, 2 bits per cell Flash and single-level cell Flash. Now in the future, it's something that we're looking at. However, keep in mind, every time you go from 1 bit per cell to 2 bits per cell to 3 bit per cell, you're getting a diminishing cost benefit, number one. Number two is the technology scale it's getting a lot harder to use those 3, even some people talked about it and we've looked at 4 bits per cell capabilities. So the amount of effort, it's not obvious to us how big of a focus that should be. We're very comfortable with our current focus, which has been get to the next lithography first, get that into production, get one and 2 bits per cell there. That's our main focus, but we've got that 3-bit per cell and we have more, by the way, if we need to deploy that.
Unknown Analyst -
A follow-on for Mario. Mario, a big part of your presentation was the fact that the complexity in the wireless ecosystem is an advantage to Micron, presumably because you're doing a lot of customers semi custom solutions that drive higher margins. Do you think that complexity stays constant over time or do you think you'll start to see that market need to standardize like the PC market that i.e., you've already seen Nokia move towards Microsoft. You heard HP last week, they love the web OS, but they hate the hardware that's it's going into, and they're getting out of the tablet and the handset business. So does that complexity actually move to more of a standard? And if it does, what does that mean for profitability from a Micron perspective in that vertical?
Yes, I think I showed the example, the PC versus the complexity of the mobile, the wireless market in general because of that. You start with very large fragmentation with that and a huge amount of players, each are trying to position itself in the market, offering differentiated value. And this is going to change through time. Some changes, as you mentioned, are already happening. Some merger happening, some partnership at least like of the Nokia, Microsoft are trying to drive the [indiscernible] differentiation. There is the other effort from the other side, which is trying to protect differentiation as a way to generate value. I think step-by-step, there will be some standardization. There will be definitely standardization. For instance, in the package architecture, what footprint, what number of pullouts [ph] , how many staked [ph] product will stay in the package, are you going to use MCB versus PoP? All of that is becoming more and more a standard solution. But still, there's a lot of differentiation. I think the operating system and the architecture level still the variety is very, very large. They need to happen some consolidation along the way to cut some of this differentiation. Still, we see the value in differentiation because this is our ability to manage the different requirement to get margin and to create opportunity for us to grow in a positive way.
Unknown Analyst -
A follow up for Bryan. What's your view on Win 8 when it comes out next year as far as DRAM content in the notebook market? It is promised to be kind of a slimmed-down OS, more efficient OS, what's the implication for DRAM and I guess, the same question I asked Mario, to the extent that now you're going to open that up, that ecosystem up to ARM, does that at the level of complexity where you can bring value on the memory side as ARM players get into that market?
Well, generally, what we have seen from Win 8 is that the content per box, there is some advertised slimming. I think from our perspective, what we've seen, there's always a little bit of feature creep that comes in as well. And so we're pretty bullish that in the notebook space, content per box will be situated well over the next couple of years between 4 and 8 gigabytes. That's obviously, balanced against some of the tablet in say in light segment, and I think we've commented on some the trends we're seeing there. From a computing platform perspective, what we have seen is that 4 gigabytes is pretty much a good minimum for the kinds of applications that people are looking at running even in the thin and light segment regardless of OS and regardless of processor. So some overall aggregate look with the tablets coming in. Win 8 is not the thing keeping us awake at night in the content space.
Unknown Analyst -
And then relative to ARM being available [indiscernible]
Absolutely. The ARM as a platform frankly has opened up an entire new, I'll call it innovation thread for all of the memory guys, specifically around opportunities to optimize a little bit better for power. Really take advantage of the core reduced power of ARM in general. And as you see, the enable base broadening out with other ARM players out there, the Qualcomms, NVIDIA of the world, the innovation opportunities are going up significantly.
Okay. We probably have the time for one more question if anybody has one.
Unknown Analyst -
Steve, this is a long-term question. But Micron has always been known as the DRAM company, and that is reflected in the share price today. And you've demonstrated in the last 3 years how much more diverse and integrated and the specialty functionality of memory is becoming a big part of your message. Are you going to rebrand Micron a little bit more into being an all-in-one high-end memory company instead of just a DRAM company or a mobile memory company in some sort? Or how do you try to help us convey that message that Micron is different now with these new models?
I think that we -- I think we're impacted by a couple of things and along the lines of that question. One is, just the fact that we have DRAM as part of our portfolio. For those of you that remember TI, until we finally bought their DRAM business, they always traded primarily with the selling price of DRAM in their share price. I know there are companies that I think been reflective of that. And I think we do have an impact from that. But the fact that DRAM is part of our portfolio, until that changes, which is my second point, is that we're also impacted by, I think, a decade of not very good results. A decade of where in the industry was under considerable stress. And until we have greater history, until people can actually see that Micron will go through a cycle and not be as impacted as dramatically as we have been in the past, that I think until that happens, I think we won't get credit for it. In other words, there was a period of time where we were getting credit for the growth rate and the ultimate expansion of the DRAM industry and that's kind of up to the year 2000 and people has always had confidence that no matter what happened with a particular cycle that it would come roaring back and so forth. And really what's happened in the last decade other that -- I mean last year it finally happened where we came roaring back and our results were better than they'd ever been before. But that just 1 year of a decade. I think it's a little hard to expect an investor base to think that 1 year makes a new decade. And so until we have some history, because I actually think that the memory industry will consolidate. There are, as I said, a couple of plays left in the DRAM industry but once that happens -- the reason that DRAM is considered a commodity certainly isn't because of the capability needed to make it. It's not, by the way, because there's also lots of new entrants that are commoditizing the space. I mean, there's really only a few of us left developing the technology. It's still because we have a hangover, a carryover of the way that the supply and demand curve has worked overtime. And I still am a firm believer that ultimately that will mitigate and correct itself. It's not that we won't have cycles like any company does. It's that specific to DRAM, they won't be like they were in the past because the industry will consolidate and be more mature and be more rational. So I think that's coming. But I understand why we don't get credit for that yet. And I think ultimately, some of you might think I'm crazy, but I think ultimately, the fact that we have DRAM as part of our portfolio will be viewed as a plus, not a negative as the industry goes through it's final plays around maturing and that there'll be more volatility in the NAND space by virtue of price elasticity and growth rates and capacity coming into play. Although it won't be as bad, I think, over time as DRAM was historically because the market diversification is already at a point that it took DRAM industry 25 years to get to. So my perspective is that ultimately, we'll get credit for what we're doing by virtue of having a sustainable history that shows that the comment that I made a year ago which was, look, the next 10 years have to be better than the last 10 years. And as that starts to unfold, then I think we will get credit for it in the share price.
Unknown Analyst -
How should we think about production and cost [indiscernible]
Yes. That question is around with Inotera getting better, and IMFS ramping, how should we think about production cost of moving forward? Mark, do you want to comment on that?
Yes, I think, you have to look at Micron's business and the complexity associated with it and try and separate out cost per bit in the stuff going into the commodity space versus cost per bit in the stuff that really is more stable margin and going to drive profitability irrespective of cost reduction. To get some of those products, we don't get to shrink because the customer wants the same product next year and the year after as he's getting today. But if you think about the part of Micron's business that's going into the most commoditized spaces, so the stuff is going into the PCD RAM, the stuff that's MLC NAND going into whatever the more high volume applications are. I think what you'll see is that we'll perform with more on a go-forward basis of that model is not broke and for us yet. Irrespective of the fact that we're getting nearer to the end of the road map on NAND and things are getting tougher and tougher on DRAM, I think we'll continue to drive cost out as we have over the last decade moving forward at least through the next 3 years or so, which is probably enough for the purposes of this conference anyway. Yes, we'll continue to drive cost out in that piece of our business at roughly the same rate the industry has over the preceding decade.
Unknown Analyst -
[indiscernible] for the short term [indiscernible]
You mean, the cost structure? Well, I think clearly, Inotera getting better from where they were, will help a lot. And obviously, IMFS is going fantastic. And that's ramping very quickly. In fact, I was surprised at how many wafers are being shipped now and how quickly it's come online. And there's no question that, that will help us in the short term as compared to where we were a quarter or 2 ago in our cost structure.
We -- let me just follow-up. I think we gave some specific guidance at the last conference call, and we're still on line with that guidance for the current quarter.
Okay. Let's just wrap up. Again, I want to thank everybody for taking the time to come kind of spend some moments with us today, let us talk about the company. We're very appreciative of having the opportunity to do that. Thanks.