Ladies and gentlemen, welcome to Seagate Technology's 2013 Strategic Update. Would you please welcome Kate Scolnick, Vice President of Investor Relations?
Good afternoon, everyone. Thank you for being here. We appreciate you coming over and taking the time to be with us in person. We know there's a lot of events going on today on other coasts, but we think ours is the most important, so thank you for agreeing with us. We have a great program for you today, and I'm just going to give a quick intro here and then hand it over to the real stars of the stage. Of course, no meeting is without a forward-looking statement, so I wanted to let you know we will be making forward-looking statements, as well as referring to non-GAAP financial information, and our risk and uncertainty factors can be found in our SEC filings. Right now I'd like to talk a little bit about the folks that are here today from Seagate.
Many of you know everyone here, but I will go through introductions. Steve Luczo, our Chairman, President, and CEO. Rocky Pimentel, our Executive Vice President and Chief Marketing and Sales Officer. Dave Mosley, Executive Vice President of Operations, and Pat O'Malley, Executive Vice President and Chief Financial Officer. I'd also like to point out to you the folks that are from another aspect of the management team here from Seagate and from our Board of Directors and our Investor Relations team. Chris Merrill and Jingjing Chen from Investor Relations. Chris Oncken is the Chair of our Audit Committee and our Board of Directors. Gary Gentry, Senior Vice President of our SSD business unit. Scott Horn, our Vice President of Marketing, and Dave Morton, Vice President of Finance and Treasurer. A brief overview of the agenda that we'll be following today. We'll be kicking off with some remarks from Steve Luczo.
We'll have a brief CEO Q&A for a few minutes, and then we'll turn things over to Rocky Pimentel. A bit of a break, 15-minute break, then we'll have Dave Mosley and Pat O'Malley on stage, and then we'll open up for a broader Q&A session. After that, we'll be heading downstairs for an executive reception where we'll have some social time to mingle with the executive team here. With that, I would like to turn the stage over to Steve Luczo.
Thanks, Kate. All right. Good morning, everyone. Thanks for coming. As Kate said, it's a pleasure to be here. Again, I think just to put it in context, what we've tried to do in the last couple of years and going forward is to use this as an event to articulate what our strategic view is of where technology is and how storage plays into the transitions that are occurring so that you can get an understanding of then what we're going to do tactically and where some of our capital allocations decisions are made, R&D decisions are made, and also our strategic investments. Consistent with last year, we want to provide a little bit of an overview of a longer-term view of our financial results.
I think what's interesting about this time period of 5 years, last time period of 5 years, incorporated a lot of the impact of the floods. This one's an interesting slice of time because you can almost think of it as the entire recession and post-recession, if we're in a post-recession. It's not clear to me, Fiscal 2008, of course, is our June of 2008 ending quarter, which was certainly the last probably good quarter before the full impact of all the asset bubble took into effect. Really, the 5-year period of time has been a recovery from that. Even against that context, I think, pretty strong results, whether or not you look at it in terms of the technology side, which is, of course, our prime focus, where exabyte ships continue to increase in the 30%-40% range.
Of course, that's variant by application space. Some big backend infrastructures are growing at 70%-100%, others are obviously less to average out at 35%-40%. Average capacity per drive is really interesting, going from a couple of hundred gigabytes to we believe we'll crest 1 terabyte probably sometime in the next 12 months. I think this is a really interesting point that while a lot of investors and industry people focus on unit TAM, which is important, we really view ourselves as being in the business of shipping petabytes. When you take the unit numbers and then take the average capacity per drive, that's obviously the business we feel that we're in because for us, the majority of our investments in heads and technologies and channels of managing big data sets is really around average capacity and how we ship petabytes.
How we package them in terms of individual units is actually probably the least complex thing that we're faced with. We see those trends continue, and I'll talk about that in a couple of slides. Unit growth has been good, although obviously in the last couple of years, the transition between what role notebooks play and what role tablets play has come into effect, and I think that's probably getting into some state of stability with then maybe even, I think, a rotation back to a notebook-type format, whatever it is, when you take a keyboard and plug it into a screen, and then if you put some decent amount of onboard storage, that looks a lot to like a notebook to me, and I think we're going to start seeing that transition back here over the next probably 12-24 months.
The financial performance obviously has been very strong. Cash flow has been good. We think we've been good stewards of capital in terms of returning value to our shareholders in the form of dividends and buybacks, that's a philosophy that we will continue to have. We believe it's a statement of our confidence in the future. I know a lot of people will somehow twist that to say, "Well, that must mean you're not very excited. You don't have anywhere to invest your money." I think that's a very bizarre view of the world.
If you have confidence in how you're making your investment in R&D, capital, and acquisitions to the point that you know you're going to generate enough cash flow to support your businesses and all the opportunities related to your business, and you can still return to your shareholders because, after all, it is your cash. I think that's actually a statement of confidence, not one of not being creative. I actually think the inverse is a lack of confidence. If you think you have to hoard seven or eight years of revenue as your cash, I think it's an interesting statement about what you think your possibilities are in terms of new product refresh that create value in cash flow.
CapEx, I think the team has done a phenomenal job, basically, to deliver this level of technology in terms of advancing areal density and capacity per drive, which for us has been more heads and disks. Yet, we've kept our capital budgets quite tight, we continue to operate in the 6%-8% of revenue range. We continue to believe that's the right target for the next couple of years. Good performance over the last five years. Looking forward, I think just a couple of touch points, then you'll see these themes throughout Rocky and Dave's and Pat's discussions.
To the extent that as investors, you want to have questions about our strategic vision, these to me, the next couple of slides, would be the areas where I think it'd be fun to have the dialogue because this is one where we really would like to have a dialogue because our thesis really is around these assumptions. If you have data points that challenge these assumptions, I think that's where we should have the dialogue. One is around data growth. If you take today's annual shipments of storage, which we measure today in exabytes, the industry will ship something like 450 exabytes this calendar year. I think that was something like 370 last calendar year.
The flash market will probably ship about 50 exabytes, but you've got to remember that about 80%-85% of that is going into phones and tablets as opposed to what we consider mass storage. That 500 exabytes or 450, if you want to just view it from the drive perspective, we believe is growing around 40% per year. Depending on the application set, there's either accelerating rates to that or smaller rates. How do we get there is really a function of our analysis of demographics. We think about the number of people in the world, per capita income, how that translates into ability to buy technology and mostly connected technology. It gets into how many people in the world are connected, and then how rich is the data that they're basically sharing.
We believe the number of people in the world continues to grow. If it doesn't, then I think there's other issues that we'd be concerned about that would be way more important than Seagate stock. We think that the number of people connected grows because we do think per capita income continues to accelerate and the cost of technology continues to come down so that more people can afford devices that basically generate and transfer data. Today, if you think there's 2.5 billion people connected, whether or not you want to include smartphones or not, it's 1.5 billion people on the internet, probably about 2.5 billion people connected if you include smartphones. That number going to 5 billion total connection or 3 billion on the internet, we view that as nothing but a bigger network effect that's going to create data.
We also think the richness of the content increases. Today, sure you have music, you have video, you have some small data sets like spreadsheets, but the video content is getting richer and richer. HD is just starting to enter the world in terms of a format that people start using more and more. 4K is coming right behind that. High frame rates is coming right behind that. When you think about a high frame rate 4K movie, that's 100 gigabytes versus today's 6 gigabytes, and that's compressed.
If you believe that big data analytics is starting to deploy effectively, which I do, I think for the next 10 years, we're finally at that point that the analysis of unstructured data is going to become valuable, then it's all the more important that people use really high-quality, very rich content to store because the ability to analyze it therefore is enhanced. When you take that growth rate and you apply it to the 500 or 450 exabytes that we ship today, that's how we get to our 6 zettabytes of demand annually in 2020. The other big shift that we see is the shift to most of that shifting to cloud as opposed to today, most of it is on direct attached storage or internal devices.
It's about 60/40 in favor of client, if you will, today, maybe more like 65/35, and it's probably going to shift to something like 60/65 in the cloud. The cloud, we actually think there's two clouds. We think there's public clouds and private clouds. We'll talk to that in a second. Again, that's just a big opportunity because it changes the nature of the drive. It changes the nature of the system that we're selling the drive into. What's really interesting is let's not forget that if you take 40% of 6 zettabytes, that's what's going to be on the client. That's still many multiples of what exists today. Even on the client side, there's huge opportunity in terms of the devices that we provide. This is really the chart that I'd like to speak to the most in my opening comments.
The team really will follow up, and Dave in particular has a lot to say about this in terms of what it means about technology and capital deployment. This theme of the demand for storage being the green line, and again, you can argue about is 30% or 40% the right number, and I'm happy to have that dialogue. I want to have that dialogue because it's important to us. The bottom line is, if you take today's financial models and how we invest in our capital and in our R&D and overlay it with what we know about the technology, it kind of lays the basis for storage industry output.
The ability for the industry to basically invest if we're in a 6%-8% of capital mode, and what it means in terms of if our margin structure is at 27-32 points, what does that mean about our R&D budgets, and benchmarking that against available technology that we see. The blue line basically indicates how much capacity the drive industry will be able to provide to users. There's been a lot of focus on unit TAM, as I mentioned before, and a lot of people will say, "Well, the industry had capacity to do 180 million units per quarter," probably eight quarters ago now. We were probably close to that.
Industry's at 140 right now. People say, "Oh, you're at 70% capacity." That's not true, because if you actually take that petabyte average capacity per drive number that I gave you, we're actually probably as an industry capable of delivering about 500 exabytes of data, and we're about 450. From a petabyte perspective, we're actually probably more closer to 90% capacity as opposed to 70% capacity, but the demand is growing almost 2x the supply. The areal density curves right now are under 20%. By the end of the decade, if we think about the next seven years, six and a half years, maybe we can average 25%. I'm starting to think that's aggressive. Today, we still use that number, because historically, we've been able to accelerate the areal density growth.
For sure, for the next few years, it's going to be pretty difficult to get above 20%. When you think about these differences, what it says is sometime between 2015 and 2016, and we can have all sorts of debates of how we slide that intersection, we basically start gapping out our ability to basically deliver petabytes versus what's being demanded. An interesting thing is if you take the peak market of $180 million or whatever it was, 175, and you think about 140, most of that has been a loss of notebook. If you think about the number of petabytes that have been shipped in the notebook market, we actually shipped more petabytes into the notebook market last quarter than we did during the peak quarter.
It just gives you a sense of what's really going on in terms of, quote, "supply and absorption," because again, we think about it in terms of heads, disk, and petabyte capacity, not just unit capacity. The real question is, what do these curves look like? How do we invest? What happens as we get closer to a disconnect between the supply and the demand? As most of you know, one of the difficulties of our industry is if we get behind, it's really hard to catch up because our capital deployment, depending on where your shortages are, can be anywhere from six months to two and a half years in terms of bringing on the right capital in order to basically provide the technology. It depends if you have land. It depends if you have a building ready. It depends if you have clean rooms.
It depends if you have the equipment to put in it. One of the big challenges that we have as a management team is staging appropriately so that when we start to see these shortages, we don't miss too much of the gap. Maybe investors would say that's okay because you're going to have big margin pop, and yes, that's what would happen. On the other hand, you also don't want to miss multiple billions of dollars of revenues that are out there either. It's this balance between making sure that we have things staged where we don't get too far ahead of ourselves in terms of over-providing capacity before it's needed. On the other hand, once the shortage starts coming, that we're not so far behind that we're leaving billions of dollars of revenues behind.
It's one of the things that we look hard at. I think what's instructive here is we firmly believe that this outpacing of demand to supply is real. We watch it every day. We view it as a challenge that we're going to have to meet a couple of years out, and therefore, we are really focused on our core because the opportunity in our core, and whether or not that's HDD or hybrid drives or SSDs, is significant at the device level. In addition, we have all these other things that are going on in terms of mobility, analytics, and cloud, and video. These are creating opportunities because the consumer and the commercial markets are starting to blend. The customer base is changing. Many people are starting to buy technology away from the OEMs who are just trying to go deploy directly.
Seagate's got this interesting position that in our core business, there's a lot of opportunity, but the changing nature of applications and architectures is also creating a lot of opportunity in what I'd call adjacent technologies that actually the drive companies have an advantage to. I think it's going to be interesting over the next few years to see how We and WD and Toshiba deploy not just our capital and R&D dollars, but our strategic dollars to say where do we want to position the companies. It's nothing, in my mind, but good and positive because it's all about growth, but it certainly is a differentiator in terms of where we think some of the real opportunities are and when are they going to show up.
The data trends, I think you're all aware of, the convergence of video being a data set that people care about. Today, it's probably mostly for entertainment or personal. Clearly, again, as unstructured data sets become data sets that are being able to be analyzed to create value, then we think that the video content is going to explode, and especially HD video that people keep for a long period of time. What do I mean by that? Surveillance videos for retail behavior. To the extent that people start being able to mine that data about how an individual operates in a shopping mall becomes very valuable, which means it makes more sense to keep that data longer.
Analytics, obviously, a big shift to people finally being able to say, "Let's take all this great math that we've developed, and a lot of it actually developed in the financial services industry, and let's overlay it into, again, analyzing unstructured data so we can do something valuable with it." Obviously, the trend of continuing to get devices into individuals' hands so that they can quickly create and share content is going to continue. That's nothing but a good thing for Seagate. Architecturally, it's the cloud, it's really the mobile trends that we're positioning ourselves for, we're doing that with areal density investments, deployment of silicon across all of our products, then also making sure that we have a device strategy, a system strategy, a services strategy, and a software strategy that lends itself to the shift to mobile, to client, to cloud.
On areal density, Dave has all operations and all engineering now. Dave's going to talk a fair amount about what we're doing in terms of deploying next-generation areal density and the investments that we're making today. Yesterday, we made an announcement about our shingled magnetic recording. We're very proud of that. We're industry leader there. It allows us to increase areal density significantly. We'll ship 1 million units of shingled drives, these are not little sample things. This is happening, this increases areal density, we're proud of the technology. Flash, both Rocky and Dave are going to talk about. I think it's important to realize that we can enable HDD to have the same performance as an SSD, that performance isn't just the user experience, it's also battery life and everything else around it in terms of shock and performance.
Rocky's going to talk about that. I think you've heard me say for a few years now that I felt that in five or six years, 80% of our drives would have some sort of silicon on them other than DRAM, which of course they all do today, and I still think that's where we're headed. We're probably three years into that. If you ask me two years from now, do I think 80% of our drives will be some form of hybrid, and I think the answer is yes, because we really can deliver disk drive cost and capacity with SSD performance by using a very small amount of flash on top of our drives with some very sophisticated code to make sure that you basically have the experience as if what you're asking for is always in the flash, but it really isn't.
Again, no impact on battery and no impact on environmental. Rocky's going to show you a couple of things, and we have some demos out there. Again, there was an announcement yesterday where we basically are providing a toolkit for our OEMs where they can take our 5-millimeter drives, they can put them in the tablets, and that tablet's glass will break way before that disk drive will stop spinning. We've already done this. We have drives in tablets, and we're excited about the opportunity. Then the computing shift, with a nod to Wenxia at Lenovo, again, we don't think of post-PC. I think that's a very odd way of looking at the world, especially if you've been in the technology business as long as I have. It's a constant evolution.
It's a constant expansion of what technology is touching in terms of at the individual level versus way back in the days when only three people could have a key to the back room, and I was one of them changing tapes, or actually, I hate to say it, cards. To say that it ends at the PC and this is some net zero-sum game between PC and tablets, it's just kind of funny because what's happening is it's just a technology that's getting deployed further and further into individuals' hands. There's been a lot of studies that say that actually tablets haven't cannibalized notebooks per se, maybe at the margin 5%, but that really what it is that people bought tablets.
They may not use their notebooks as much, or they may not refresh their notebooks as much, but the tablet doesn't replace the notebook, in part because it doesn't have storage. So we view it just as on the one end, consumers are getting access to more technology that they can carry with them more and more, which allows them to be better communicators and capture more events that are important to them, which ultimately gets stored on a disk drive, usually more than a couple of times. It's also extending up into a more shared infrastructure on the enterprise side. To me, that's time share 30 years down the road. There's a lot of reasons why those architectures make sense when you think about the management and the cost of deployment.
We do believe that the cloud world is going to split into this thing, that there's some amount of private cloud, so think corporations that are developing their own cloud infrastructure with different characteristics of whether or not they're performance-based or storage-based or network-based, then others, which are the public clouds, which I think are the ones that get more of the press today, the AWSs, the Googles, the Facebooks, the Microsofts. Our belief is that how we're gearing ourselves strategically is, yes, we're engaging with the public cloud providers. They are thought leaders in storage technology. They're not constrained by a legacy view of storage, and in a certain way, they're not that knowledgeable about our devices. The first couple of implementations of those architectures probably haven't taken full advantage of what could happen if you actually worked closer with a drive company.
We are engaging deeply with all those companies I just mentioned, and we believe that'll be an interesting business. But it might take the flavor that what we do for them is somewhat customized because they're all very secretive about how they want to deploy their infrastructure because these are capital budgets now, by the way, that are running $5 billion-$10 billion a year. Those are competitive advantages that they really don't want the other big cloud providers to know about how they're doing it. It creates more of a strategic consulting relationship with them, which again, Rocky's going to talk about, because one of our biggest challenges is how do we develop a go-to-market capability that's a little more technical in that sense.
There's going to be what we think will be the vast majority of the market, which is what we call private clouds, which is that people are going to be standing up anywhere from 20 to 100 to hundreds of petabytes. They're not going to basically put that into someone else's hands, either for security reasons or other reasons. Of course, as a result of the events over the last six months, we actually see a shift in that direction fairly aggressively. An interesting point on that, for all of the focus that people have put onto the cloud, our industry shipped about 20 exabytes into the cloud last year. What's interesting is that if you look at the number of exabytes that have been shipped into the consumer space in terms of DAS, it's also about 20 exabytes.
Now, some of you may have known that, but if you didn't, that's a pretty interesting point that the personal cloud business taken to its extreme, i.e., the individual, is as big as the corporate cloud business. I don't think most people realize that. It's our belief that that's going to continue. However big that corporate cloud business is, we also believe there's a consumer business that is just as important for us to invest in. From an investment perspective, areal density is absolutely key to what we do because our core technology is so important to the growth of data. Silicon is important. We will ship a lot of drives that are either pure silicon or silicon on top of HDD this year, this quarter, and it will continue to grow.
How Seagate takes advantage to the opportunities presented by the shift to mobile, the shift to cloud, and the fact that the network infrastructures are too expensive, so the only way that you can actually provide the user experience that people want is by caching the data closer to the compute device, is really where we're focused. That's what you're going to see mostly in terms of our themes. Overall, you know we're big believers that cloud mobility and open source is what's driving the architectural shifts. We want to position ourselves there. It has big implications for our customers. On the one hand, our existing customers, the OEM, and the distribution base, they're trying to meet that challenge as well.
On the other hand, there's a whole host of customers that have already made the decision to move away from those traditional channels and basically deploy technology on their own. Seagate has to engage with both of those customer sets. There will be winners out of the OEM customer base that successfully transition into being either technology or service providers to the companies that want to deploy technology. There will be losers. There will certainly be a whole host of opportunity around engaging directly, either in terms of what we do in terms of designing our devices or actually reaching into customers that are deploying technology directly. Again, technology, innovation, and leadership, we have to keep investing in our core and advancing our technology. That's what's giving us the opportunity in the other spaces.
We do believe that operationally we are best in class in the world. Dave's going to give a couple of examples. We respect our competition. We think our competition is excellent. I would say in the boxes of technical innovation and leadership versus operational excellence, I think in particular, our competition is very good operationally, and we think that's a good thing. That allows us to be better every day, and we think we are really good, and especially if you look at the advancements that we've made on the reliability and quality of our products over the last few years, it's been pretty remarkable what Seagate's done.
Of course, in terms of our financial model, we're focused on making sure that we continue to have a business model that allows us to fund in our R&D and capital to take advantage of all these opportunities that we see, whether or not they're in the core or adjacencies like SSD or things like mobile or cloud. We also feel that we have an obligation to return value to our shareholders, and we think we can balance that equation. We think our opportunity is big enough, our market position is strong enough that we can continue to do what we've done over the last couple of years, which is not give up the leadership in technology and products, but also return value to our shareholders, and we plan on continuing to do that.
That's kind of the big broad view of things, there'll be a lot of detail to follow through that in the next hour or so, 2 hours, I guess. I would like to just have a kind of a brief Q&A, hopefully around these topics, then we'll turn it over to Rocky.
Over the last couple of years, you and your competitor have been very aligned, driving scale with acquisitions and being smart about adding capacity and smart about chasing the right mix. That has obviously resulted in great financial performance. You now see a divergence in investment. You mentioned that strategic investment will be an area of differentiation with WD spending $1 billion chasing Flash. Just curious to understand, from the outside, it seems like you have different views of that market. Can you just talk about if that's the case-
Yeah.
What your views are on investing in Flash?
Sure. Everybody heard the question? We're good on mics and everything? I think it's not hard to get to the first reaction of, "Oh, it's different." Maybe not. I think my point is, the good thing about the drive industry is we have lots of opportunities, and all those opportunities have fairly significant payoffs in terms of what role we play in cloud infrastructure, what role we play in the software and services that have to actually get deployed with that, what role we play in the number of devices that we make, and what kind of value add is in our devices, what role we play on the consumer side.
It's very unusual for an industry to have this many opportunities, and you say, "Well, why is that?" The answer is because, quote, we'll go back to what people used to bash about the industry, because we are the fundamental commodity. What's being attacked was never the cost of the commodity. It was the value add. It's the software and services on top of that thing that people now are going after. For us, that's a huge advantage for two reasons. To the extent they bring down the cost of computing, which they will, that means more device deployment because we are limited today by budgets that have nothing to do with the cost of a disk drive.
When I read someone that says, "I can deliver" I'm not bashing SSD because I'll get to it, because I think SSD is important for part of an application space. When someone says, "I can deploy SSDs at $4 or $5 per gigabyte, and that's the same as an HDD, therefore, I'm going to take over the world," that person has no idea what they're talking about. The cost of an HDD is $0.06 a gigabyte. They're confusing the cost of the device with the $4 of crud that gets on top of it that's under attack. Guess what? Someone might stack $4 of crud on top of that other device as well. When that thing comes down to people deploy software and services at $0.50 a gigabyte, now you're at $0.56 a gig, not at $4, and guess what?
You probably don't have a big change architecturally, because at the end of the day, HDD is the best thing for mass storage. If you're talking about accelerators, if you're talking about I need to move a huge data set in and out, a little tiny slice of Tier 0 computing, yeah, SSDs are going to be really good for that. We want to play there because our understanding of enterprise workloads is very important to making sure that technology gets deployed correctly. That's how Samsung and Seagate ended up together. Because Samsung clearly has the capability of making an SSD. What they found is that when they went into the enterprise, they didn't understand workloads.
Where we basically merged our competencies, whether or not they were technical or go to market, was to say, let's take what you know about silicon cell structures, let's take what we know about code and interfacing with a host, let's take what we know about making sure that works across 30 years of legacy systems, and go attack this opportunity. That drive starts shipping this year, we think it's going to be a very successful product. I think the WD thing is complicated by, 1, they've clearly shown a near-term preference for investing in silicon. We already have that investment. We're 2 or 3 years into our partnership with SSD with Samsung. The other thing is it may talk also to what's going on with the Hitachi Intel relationship, because that's really been the success of their business up until now.
That relationship basically dissolves sometime soon, and it may just be, you'd have to quiz WD on this, but it may just be that who owns that technology is more reflective of why they think they have to go buy all that technology versus that they view that as a bigger opportunity. By the way, again, WD, like Seagate, has the resources to make multiple bets. The fact that they made that bet this week I don't think means they're going to make that bet for the next 5 weeks. It may. Or maybe it does show a little bit of a difference. We believe that our core opportunity is massive, and we are going to make sure we stay focused on that and deliver those solutions to our customers in an expanding customer base.
Trust me, the cloud is a bunch of disk drives that's connected to a network server and a network switch that gets downloaded. It's a bunch of disk drives. We want to make sure that we own that space technically. That could be at the device level, that could be at a higher integrated level between devices and hybrids and SSD. That could include software, could include services. Yes, there may be a difference, and we'll see how things play out. Did I answer your question? Okay. Yeah.
Steve, you touched on a lot of different talking points, especially around adjacencies and future investments. Seemed like you were talking about maybe a 2 to 4 or 5-year time horizon. If you started prioritizing some of those adjacencies and we're looking at things that could happen maybe in the next 12 months, 24 months, what are the most important maybe additions to the strategy that we should pay attention to?
Well, importance measured by what? You'll probably say dollars, which is a good measure. Dollars is a good measure. In terms of time, it's really hard. It's just what are we all obsessing about at the moment? All of us are fully engaged in our world and where we see our world going, and it's amazing. I can get as excited about some of the opportunities in mobile as I can in cloud, right? I think the answer to the question is first and foremost the core. Things around how do we make sure that we continue to advance the technology? How do we get ourselves to be best positioned so that when the stress between the demand for petabytes and the supply of petabytes starts appearing, that we're better positioned competitively?
There's a lot about supply chain that has to do with that, because while most of us, while both of the three of the remaining companies, two of us are highly vertically integrated, all that really means is of the 200 parts that we put in a disk drive, we make four or five of them, let's say. There's a supply chain behind us that's really, really important. Seagate, even pre-flood, really, my whole pitch when I came the first time and then reiterated the second time was suppliers are customers in reverse. We shouldn't be asking our suppliers to do anything differently than we ask our customers. When we tell our customers, "You should be more strategic, and here's all the reasons why." If we're not doing the same thing with our suppliers, we're hypocrites.
We actually have always kind of had that DNA since 1997. Post-flood, I think in particular, it really caused us all as an industry to reevaluate the relationships that we had with our suppliers and how much risk we were actually placing on everyone as an industry, as a function of our margin structure and our business practices. We've spent a lot of time looking into our supply chain to say, what are the things that we can do differently to make sure that those very important technologies are there when we need them? Because what's going to be interesting about the transition point is velocity. If you have a set of relationships from the supplier through your factories to the customers that creates more velocity than your competitor, that's going to be really important, especially as supply and demand gets tough.
It means every drive I send to someone, the quicker I can get it through the system, it turns into revenue. If you're in a drive-constrained world, what did we learn? Who had drives got revenue. That's what we think is going to happen. We're working hard in terms of the investments that we're making in IT, in business practices with our suppliers and with our customers, some customers get this clear as day, and those are areas of investments that you might not notice, right? They're big investments. A lot of things around the core, both technically and from a business operation perspective. Then I would say, if you forced me to say-- and by the way, in the core, I include HDD, hybrid, and SSD device technology. If you said, what's next, I would have to say cloud.
I think ultimately, the opportunity for a drive company to be a significant voice at the table when people are talking about delivering me 20 or 40 or 50 or 100 petabytes at a time is going to be significant. The metaphor I'll use, it's not one for one, but if you think about what happened in the DAS market, it's going to be similar. The DAS market shifted to the drive companies because at the end of the day, 80% of the BOM was a disk drive. Yes, it mattered what piece of plastic you put over it, and yes, it mattered how good was the software that was on it. Ultimately, we gained that capability, and today probably 85% or 90% of the DAS market is controlled by the drive companies.
I don't know that it's going to be that on the cloud side because there are other technologies that are important, a lot of those other technologies are going to come from open source. I think for the drive companies, how you think about your devices architecturally and how you open up the API so that it works with a wide variety of open source/OpenStack technologies is going to be a real differentiator. How you operationalize it. It's staggering to us that for all the companies that have made great businesses basically operationalizing selling storage, they're not really focused on that anymore. They're either focused on building a cloud and renting it out, or they're focused just on the software.
Look, we're operations people, we actually think a focus on how you actually build and deploy the physical product is going to be really important. By the way, there's a software implication to that. There's a services implication to that. There's a go-to-market implication to that. There's an architectural implication to that. I think that's where you'll see our next kind of area of focus if you force me into priorities. Like I said, the consumer opportunity is as big as the cloud opportunity, we have to balance what we think the consumer opportunity is.
For that, you saw through the announcements yesterday, we do believe there's an ability to advantage OEMs because OEMs are going to play a big role in the consumer device business to the extent that we can make it easy for them to take our technology and deploy in devices so you get the user experience that you want, then that's a great way for us to do it because it leverages off of our business model. I think you'll also see us do things a little more creatively where we actually probably deploy things under our own brands or brands that we own to take advantage of that market. That's the priorities, it's all important. Look, they're all growth markets for us. The good news is we really do think that we can fund opportunity in all of them appropriately.
The follow-on question would be I'll ask it because everyone probably has it on their mind. Your OpEx has gone up a little bit because you've made the investments in the areas. Pat's going to talk about that, and the answer is yes. Yes, that's right. It has, and we should continue to do that. We're not going to go crazy. We don't want to blow the fundamental model of how we can return capital. On the other hand, there are investments that we should be making in market development and technology that relate to these adjacencies that we think have big payoffs in revenues and margin, and not payoffs that are 10 years down the road. We expect returns on those investments in pretty short period of time. I think, again, that's what we view our task to be of managing that balance. Yep.
I'll go to this side of the room, sorry. Must be right eye dominated today.
Thanks. When you talk about the six zettabytes of storage demand that's going to be there, do you have any idea how much that could feasibly go into SSDs versus HDDs?
It's a great question. Today, again, you're at 50 exabytes and 450 exabytes for about 500 exabytes. Of the six zettabytes, what people say is that 500 exabytes of that will be flash. We don't understand how you could remotely get from 50 exabytes total capacity today, which again, 85% of that's going into mobile devices. You're really at 10 exabytes going to 350 or 500 exabytes. That's $1 trillion of fabs that has to be deployed tomorrow. The math doesn't hunt, which is why we think how you get to that is hybrid. If you take 64 gigs and throw that on top of a 10-terabyte drive, you start to have basically 10 terabytes of SSD. That's actually what it looks like if you're trying to solve that tier 0 problem.
When you get behind the math, it's really difficult to say that it's anything more significant than that.
I guess I was trying to think about Facebook recently at the Flash Summit, I think, talked about cold storage, which was their take was a majority of the data they have is written once, read never, and they just want extremely dense, low IOPS, high latency solutions. I'm wondering, is the answer maybe make it bigger and slower versus smaller and faster?
Facebook's unusual. Again, remember my description of how we view the cloud. There's public and private. I would put Facebook in the public cloud business, even though they don't sell their service. My point is public meaning we all reach into and use it either transitorily or more permanently. Private meaning it's more corporate contained. Facebook is an interesting one because they're not actually a data-heavy company. They don't have full motion video. Maybe with Instagram they'll get there. It's a little picture, and it's a thumb up. That's just not a lot of storage. For the user experience they're trying to deliver, response is really important. For them, they are a flash-heavy company for sure. Others aren't. They could see a world, even if, by the way, they do go into more video, however they get there.
They could see a world where they say, "My tier 0 stuff is this, then I've got this huge cold store." What's cold store is the interesting question. We now have solutions that are as cost-effective as tape. The question of cold store, which we actually think the real answer is warm store. There's no value to something that was cold store. What does that mean? Call me, and I'll send it to you in a week? If it means that I have multiple seconds of access time versus milliseconds, that's a rotating drive, too. Yeah, we believe as we customize the solution for Facebook, it may be a 20-terabyte drive that spins really slow or not at all, and it's in a refrigerator and whatever. If it takes 15 seconds to get the data off of it, that's okay.
For Google, who has a lot of video that they need online very quickly, they may say, "That's not the architecture we want at all." Yes, those companies will define architectures that are very different from one another, and in my opinion, very different from what most of the world needs. I think it's dangerous to say Facebook says or Google says or Microsoft says, because there's only about 10 companies in the world, maybe, that can afford that footprint. That footprint is $2 billion-$10 billion a year of investment in capital. That is a big number. Everyone else cannot afford that, right? The target market for us ultimately is really how are the people that are deploying really just a more shared infrastructure going to develop the technology.
We think that is a more of a replicable technology that has to do with business process than it does am I storing videos or pictures or whatever. I don't think it's an inconsistency. I think it's just that's the reality of the marketplace. I'll go to this side for a second. Yeah. Kate, you can throw something at me when my 15 minutes is more than over.
Earlier, you talked about exabyte growth, 36% over the past five years, and your expectation makes all sense in the world that data growth is, let's say, 40% CAGR over the next several years. If you look at the industry, it had been growing exabytes at 50%, 60% a year for several years, and that actually slowed a lot the past year and a half. I guess two parts of the question would be, one, why do you think that's been? Do you think it's just the macroeconomics? Do you think it's been architectural or whatnot? Obviously, you're implying we expect that to rebound or pick back up. The drivers of picking up with mobile and video and analytics would seem like they're already happening now. If that's all happening, why have we actually seen a slowdown rather than that kind of rate?
It's a great question, we think a lot about it, and as you point out, it's almost been Almost the last two years. It seems like depending on the markets, it's kind of slowed down. I do think a lot of it is macro. I think a lot of it is the cost of the systems. I think we're just getting to the point where some of these more affordable, scalable architectures are being able to be deployed outside of places like Facebook and Google and Microsoft. Therefore, it's somewhat constrained the ability for people to basically deploy the technology that they want in order to store things. We can take a look at our EVault system as just a little metaphor. It's a storage services company. We do archiving over the wire into a cloud infrastructure.
The biggest thing that constrains our ability to grow is the cost of the storage, is the cost of the systems. I won't tell what brand name we're using. We use a particular OEM solution, but it's so expensive that we can't actually match it with the decline that people are charging on the cost per gigabyte. Because there are some people who are heavily subsidizing cost per gigabyte, either because they have bigger businesses that are obscenely profitable or because they're being funded by venture capitalists that plow it in because they think they're going to get the value somewhere else. Look, there's a lot of companies that are being successful deploying more users, but they're not making money, and they're basically subsidizing their storage. You can't do that forever, right?
At some point, you're going to have to start getting paid for the infrastructure that you're putting in place. I think part of it is macro. I think part of it is that we're not quite there on the open source solutions taking enough of the margin structure out of the software and services get stacked on top of our drives. It's just starting to happen. We've put up an OpenStack cloud inside of Seagate and, one, it's not easy, and two, it doesn't really have all the attributes that you want at the file system level to run your business the way you do. There's more work to be done there. The other, I think, is just notebooks.
I think given the compression on the notebook market over the last couple of years, when you think about 40 million units per quarter times-- and those are high-capacity drives usually. Those are usually, in today's world, it'd be a terabyte going to two terabytes. That's a lot of petabytes that could be shipped out. I think that's just a function of where the notebook guys were in their product strategies. Until really the last year, I just think the notebooks weren't that compelling. Now, whether or not the new ones have drives in them or not, we can talk about. But today, the thin and light market is still only about 15%. We don't see it that way because all of us are rich compared to the rest of the world.
The reality is it's about 25 million units out of the 170 million units that get shipped. Of that, within a couple of quarters, we'll be shipping 30%-40% of that as hybrid drives. That Ultrabook market that people thought the drive industry wasn't going to touch, we're going to have 30%-40% of that going forward. And it may be growing more quickly because of the fact that we can give you a lot more storage for less money with SSD performance. I think we're in a bit of an inflection point, but we're watching it closely because you're right. That's the big thing. If you say, hey, for some other reason, you're going to stay at sub 30% or 20%, and that's getting close to areal density growth, then we have a whole different equation.
That's exactly what you should be thinking about, and it's exactly what we spend a lot of time thinking about. Other questions on this side?
Yeah.
Yeah, thanks.
Kim Evans from Baird. Just a quick question, kind of dovetailing off Katy's question with a front seat to the Virident acquisition this week. Presumably, you had a view on the opportunity there, and it seems like if we're taking this right, that you feel Seagate has enough with their Samsung partnership, both on the development side and with the sourcing of NAND flash also as part of the agreement. Can you comment on that, if we are reading that correctly? Then secondly, how effectively do you think non-vertically integrated flash storage suppliers can be?
Well, I think the second one, the answer is can be, because if you didn't answer it that way, then all the media companies should have been successful. At the end of the day, flash is a media, but there's a system that gets wrapped around it. We wrap our system around rotating magnetic disks today. We also wrap our system around silicon. That's really where a lot of our value add is. Yes, we have a lot of value add, as it turns out, in making that media, but the value add on that media is nothing compared to the value add on the system and the system-level knowledge.
I think for sure, a non-flash owner can be successful in the delivery of that device because that device is a system, and then especially if you're integrating that device along with a few other devices that you can either call a rack or a big rack or whatever. Yeah, I think there's plenty of opportunity there. In fact, if anything, you might have more of an advantage because from time to time, this silicon provider may be ahead of that silicon provider may be ahead of that silicon provider. That's the trickiest thing about being in this business right now, is that your controller technology is explicitly tied to a silicon technology, and the development time on the controller is long, and the development time on the silicon is long.
You're making these bets on who's your silicon provider and who's your controller tech provider, trying to hit a product release and requirement that's 2 years out. In the meantime, everybody else is doing the same thing. What happens is, if your controller is late or if the silicon is late, your controller was on time, but it turns out Toshiba got to 19 nanometers first, then all of a sudden, now your silicon-- It gets really difficult. For a silicon company, they're making that bet all the time. Of course, for them, depending on who it is, they have other issues that they're trying to solve. Like in Samsung's case, obviously, they're trying to feed a device business. That's really important for them.
We happen to really like Samsung in terms of technically where we see them going, the relationship we have with them. Toshiba clearly is a world-class provider of silicon. We work actively with them. That world is consolidating. I think the answer is yes. I think on the first thing, my view is this, that Virident was a good company. Obviously, we invested in them. Clearly, at that price, we were a seller, not a buyer. We love it when our competitors have to send us checks that have lots of zeros after them. Hopefully, this is the first one this year. Hopefully, the second one is really almost as much as they just paid for that company. Yeah, we were okay with it. Look, it's an important technology going forward. We think the drive companies have a role to play there.
It's software mostly, as Mosley will point out. That's just monkeys typing at a computer, and you just have to tell them what to write. No offense to Microsoft, we don't really believe that. I think there's opportunities to develop that technology, but it doesn't mean it's not important technology, and it doesn't mean what they did didn't make sense for where they're going. It's an opportunity for all of us. Yeah. Yeah.
Part of the spending is slowed down, growth spent by the last year or 2. Anything to do with improvements in using raw storage more efficiently, and what assumptions about efficiency have you made in this longer-term supply-demand curves that you have?
Not a lot. The compression technologies and utilization of infrastructure really happened a few years before that. There hasn't been any big advancements in compression or whether or not you can run your system at 80% capacity or 85% capacity or how often you flush data that hasn't been used. I don't think it's a lot of that. I wouldn't read too much into a couple of quarters trend either. The thing about the cloud deployment, a lot of this is being driven by cloud, a lot of it's being driven by consumer. I do think we have to pay attention to the macro impact on the consumer. On the cloud side, it's really choppy. It really is very choppy.
Again, when you have a handful of companies that are driving so much capital deployment, if one of them shifts their capital deployment strategy by a quarter or 2, it can have this illusion that, oh my God, everything's slowing down. My sense of it really is that if I think back to May, April, thought about the rest of the year, it felt like the cloud deployment was going to be stronger. It seemed when people started getting nervous again about the macro issues over the summer, it's June, July, is when you all of a sudden started hearing company X or company Y is deferring their deployment for 6 months. Not that they're deferring their deployment, for 6 months. I think it's too early, I don't think it is technical. I really don't. We have to watch it.
Yeah, one more, I should let these guys go.
Steve, how the components of the things that you're talking about strategically to become involved in in the coming years well, they all speak to increased value add, providing increased value add to the industry, to the customer. The question is, do you think that the invested capital returns of the company can substantively improve-
Yeah, absolutely
continue to substantively improve?
Yeah.
If that's the case, could that impact what you guys think about cash return in the future, or how might that impact that?
Yeah, I think it does. Obviously, again, the goal of the investment is how do you balance investing enough so you give yourself a fighting chance to be relevant in the space, either with your go-to-market capabilities or your technical capabilities, and then when is that payoff? Chris is here, she can vouch for the fact that the board is really good about making us answer that question of, yes, we support the transformation in terms of taking advantage of the market opportunities, but we also are going to hold you accountable to return on that and when do you think that comes into play. We do believe that while you could have OpEx increase with a stable gross margin for some period of time, when do we expect, basically, to get leverage in revenue growth and gross margin?
Yeah, we do view that as mostly accretive to what we're doing. Some of it, if it isn't accretive, that's okay because it's gross profit dollars that we wouldn't have anyhow. The hybrid discussion gets really interesting there. If you're taking 50% of the ultralight market, and let's just say it's not accretive, let's just say it's neutral. Do you do that or not? Of course, because that's a market people thought we were going to get. Or if we can deploy 5-millimeter drives into 30% of the tablet market, and let's say you only do that at 22% gross margin, do you do that? Of course. That's 50 million units a year that you wouldn't have had. It depends. In general, it's accretive. No, I don't think we have to jeopardize the capital return policies that we have. I don't.
I think the opportunities are that big and that close, and it leverages off what we do. Will we have challenges on the go-to-market side, especially as it relates to end users? Yeah. That would be different for us, and therefore, how do we engage with those OEMs that we think are going to be successful in doing that? That's one of Rocky's big jobs. I actually, if someone said, what do I worry about more, I worry about that more than I do the technical stuff, to be honest with you. Look, back to the Virident question or other companies in that space. Their biggest challenge has been go to market. It hasn't been developing great technology. It's been not being able to afford the infrastructure to actually deploy it. We're at scale when it comes to deploying technology.
It's just, can we shift that competency that we have in sales and marketing to address the new market opportunities? Of course, we think we can, but it's not going to be easy. Okay? All right, I should wrap it up and let Rocky go because I probably went way over my time, and we'll have more time for Q&A later. Thanks. Means you can be quicker.
Oh, yeah. No, that's true. Thanks, Steve. I don't know. Steve pretty much covered a lot of it. First of all, I want to thank everybody for letting me come today, because it's always nice to make a trip. I have a personal little project, which is taking pictures of bathrooms. This is another location I was able to fulfill my personal responsibilities back to the home front. I'm really happy that by now, I think I must have a statistical sample that'll assure me of success with my domestic situation. Thank you very much. I'm going to talk a little bit today about the customer side of things. Steve touched on a lot of these points, so I hope I don't bore you with some more.
First thing I'm going to do is talk a little bit about how we see the environment over the next decade or so, how we've changed our engagement with our customers, also who are some of those emerging new customers, and how are we addressing their needs, and then finally, talk a little bit about some of our product announcement that we just made. One thing that our hypothesis that we're working with is really a view over the next decade or so that the world is a storage fabric. It's not just a big cloud, and it's not just a client. It's things between that whole extreme, the biggest cloud, the smallest client, all of which is going to create data. Steve talked about it.
There's forecasts that the amount of data in the next decade per year could be as much as 20 to 40 zettabytes, depending on which research organization you want to rely on. Of that 20 to 40 zettabytes, 6 to 10 zettabytes is storageable data, and those are all projections. I think the point is it's big and it's huge, and it's a growth opportunity, which you don't expect from an industry technology that's been around as long as storage has. The storage fabric, we need to really focus on how do we service everything from the cloud to the client.
We see clients redefining themselves in many ways, not just the traditional tablet or notebook device that we're used to, but all of these new emerging devices, like cars, intelligent cars, like glasses probably, like watches at some point, and other kinds of communication devices, all of which is going to create data, and some of that data is going to be really important for numerous sources to harvest and mine. We are sensitive that we need to address the storage opportunities there. I think when we get into it more, the announcement of our five-millimeter drive has been a really interesting product announcement.
As we've gone around and talked to key focus customers over the last month about it's pretty amazing where out of nowhere, interest has sparked in non-computing applications in industries where data is important, but they never thought about mass storage being affordable and applicable to what they're doing. Again, the opportunity we see is an incredible growth opportunity as much as any of the digital media growth opportunities that we hear about every day. Let's talk a little bit about how that impacts our customer relationships. Certainly, our traditional customer relationships with the classic OEMs, and everybody knows those names, has been pretty much the core of the business, I'd say, for the first two or three decades.
Now over the last four years and less, there are these emerging relationships which have evolved, which include cloud customers, but not just cloud customers, but client customers, and again, these industrial customers which previously had been serviced one or two touchpoints removed. This creates a whole different dynamic. In the past, the focus was a lot on just selling what we had as products. Now I'd say it's at least equally balanced or more weighted towards discovery of the opportunities with the customers, which is a new phase, I think, in selling storage. We have not been as pulled into idea creation in terms of products and services like we are now.
I think this is a huge opportunity, not just in the ability to sell more storage products, but in creating these value-added longer-term relationships with a whole new set of customers that will shape the experience we have over the next number of decades. There's a lot of traditional things that-- or non-traditional things that they worry about, which is, where is data positioned? How do I harvest this data? How do I then turn it into my valuable economics in my business model? On that, let's just get into transitioning to cloud service providers. It's interesting because we look at the companies that people classify as cloud service providers, like the Google, and the Microsoft, and Amazon, and the Alibaba, and the Baidu. These companies have, at least as an outsider, we perceive they have a certain slant.
Google's a search giant, Alibaba is a commerce giant, Microsoft's a multitude of service provider. What's interesting and really resonated the last couple of weeks of travel is that the really insightful companies think now beyond their services. They think of themselves as a data company. In fact, Jack Ma at Alibaba defines himself as a data company, and his most strategic value is how does he collect the data on his customers and turn around and mine it in terms of economic transactions. It's pretty amazing when you go to that facility and you look at some of the real-time decision support applications and vehicles that they use to run Alibaba. It's something you would see in NASA.
It's really impactful and I think is an indication of how companies will look at data being actually their strategic value and not so much their tactical services and other things that they offer. They look at, how do I get storage? How do I get more and more data? As a result of that, they're focused on a number of things. Certainly, the quality of service to sustain the customer satisfaction. Certainly, the total cost of ownership, which Steve alluded to earlier. This is not a trivial thing. It's easy to think that the architecture to collect and use data is just SSD acceleration. I'm going to spend $10 per gigabyte.
When you think about the scale of some of these companies, even like the Facebooks with 700 million or 800 million users, Alibaba who has a target to be more than 1 billion persistent customers or registered customers over the next 3 to 5 years, and the Googles and stuff, that's a lot of data, and the total cost of ownership of that data is really challenging. The idea that everybody's worried about accelerated data or everybody's worried about just cold data, it's a complex architecture, and we have a role to play in the entire stack, which I'll get into as we get down the discussion I have. What are they worried about? Total cost of ownership, availability of data, quality of service, customer experience.
Manageability is another area because in the past, when you look at storage suppliers' relationship with OEMs, we've been, in many ways, channeled to a very specific purpose within an OEM solution. What's interesting is these new customers that I've mentioned are now technically competent to take storage and use it to their own unique advantage. It's allowed us to start to talk about things that actually the drive can do, like real-time manageability to sustain quality of service, to avoid service outs and things like that is a real eye-opener and has a lot of value to these large service providers and data providers. It's a really exciting time. Another area they worry about, security, environmental management, things that before we were just prohibited from being able to influence because of our prior historical customer relations.
Now, having said that, our historical customers still have a lot of value to us and will still have a role, whether it's in the cloud side or the private side. I'm not here today to say the whole world's going to just cloud service providers. I'm saying there's a whole new set of customers that could definitely change our value proposition as we go forward. Another driving enabler to, certainly, the cloud service environment is the open source initiative. Really, this is to allow those customers that do have that internal competency and can afford to have the technical resources to be able to piece or wed together the key components in the data center themselves. What it does is it's certainly disrupting the traditional enterprise economics, where again, it's only an OEM serviced demand by the cloud service providers.
It gives them the chance to buy elements and do things internally in a unique way that gives them an advantage in how their systems work or how their data is managed. The other thing is defining how their scalability works. We're now able to work directly with them to figure out what's the right configuration in their data centers based on what kind of data matters and where are they trying to move it, and certainly utilizing the external world to be able to contribute to how they're trying to figure out their stuff. As a result of that, we're an active participant in two of the leading organizations in open source, being the OpenStack Software Initiative and then the Open Compute Initiative, which big sponsor is Facebook on the hardware side.
We're very actively trying to develop and work with the emerging ideas that, again, position us in a different way in the value chain. Let's talk a little bit about, Steve mentioned mobile devices are a phenomenal opportunity. We've had that question, what's going to happen with the death of the PC, blah, blah. Yesterday, we made the announcement of our 5-millimeter drive, which is, we think, a significant breakthrough to creating affordability and adaptability in the mobile space. I think out there we have some demos and some show and tells for you to take a look at. We really felt that there was no reason for us to not be relevant in the mobile world, because again, if we believe in the hypothesis of the storage fabric, near-field storage is a really important aspect for the user experience.
What's also interesting, just this morning in having a meeting with the marketing organization, we did a number of focus groups of consumers around storage and how they think about storage. What was interesting is, this was the Mobile Millennial, I think, is what it's called, or the Millennium Focus Group, which is that group between 18 and 32 or I don't know. I'd defer to the marketing organization on this one. What was interesting is, though they recognize that services like Box.net and Dropbox and others are out there, they still expect to have a substantial amount of their information and data within their control. Even though there's things that they will do, they will put in the cloud, they want to have their own personal cloud. Not Direct Attached Storage, but they want to have their own personal cloud.
I thought that was really an interesting insight that actually, to some degree, validates our hypothesis of the storage fabric as it relates to the client or near-field storage expectations or desires. We think that the evolution with 5 millimeter, then we announced our mobile enablement kit, which is really interesting because we've actually developed software that we will make available to OEMs, because many OEMs and others that build tablets build it with an established architecture that has some level of NAND for the onboard memory. Taking that into consideration, what we did is we said, well, some of those tablet manufacturers or mobile device manufacturers will want to buy a hybrid solution, but they may want to just buy an HDD solution because they pre-design with NAND in or memory on board.
We wrote software that allows them to basically buy the five-millimeter drive either as an integrated hybrid device or as a standalone device that can use this firmware to talk to the host design NAND, such that it acts and behaves like a hybrid. We think that this really changes the one, as Steve alluded to, opens up a big new element of the serviceable TAM on the client, as well as enable applications and content providers to create a new experience on mobile clients. We are pretty excited about this. Actually this quarter, I think we will be shipping approximately 100,000 of these devices already, and this will be scaling over the remainder of the current fiscal year for us. We have multiple either production releases or evaluations going on with major OEMs and other major industrial applications.
We are pretty excited about the potential of the client segment based on our five-millimeter solution. I think also some other points which I think are out there and were in the press release. We did this drive with no compromise to what people expect in a tablet. It has matching or better durability from a pure NAND-based device. It has higher performance in terms of system execution. It actually matches the battery life. It doesn't degrade the battery life. Some usage cases would even say it enhances the battery life. Again, there's a lot of redeeming values that have been built into the design of our five-millimeter. We are pretty excited about that. I wanted to move on and talk a little bit about flash technology, because we've already talked about this. Certainly, we see our product portfolio as being a storage solution.
The way that we think about our portfolio of products over the next decade is to be able to provide whatever that customer needs, whether it's very low-cost storage in the form of a traditional HDD or slow spin HDD to the highest performance SSD requirement. Despite the fact that Virident was purchased this week by our competition, it doesn't really change our strategy. We've always been committed to our strategy of having a complete SSD portfolio. I would say that our progress on our SSD unit to date and in this quarter has been really positive.
In fact, right now our forecast would show that as we exit the December quarter, this coming December quarter, at exit, our memory connected or memory related products, that would be hybrid drives and pure enterprise and industrial client SSD class clients or drives, the aggregate of that pool will be over $100 million in our December quarter. We are pretty excited about the building momentum and success that we've had with our flash-based technology. Steve talked about the hybrid drive. This current quarter, September, we will ship over 1 million hybrid drives, which I think meets what we talked to you about back in the June, July timeframe, if not the prior quarter, what we expected.
The other really positive characteristic of the more than 1 million drives we'll ship this quarter is that they're across the client, the mid-range, and the enterprise class drives. We'll be shipping drives in all the key categories with a hybrid solution. I think that's a real exceptional data point in terms of the customer adoption of that technology, which for a long time people said, "I don't get it." Now customers are starting to clearly get it, so we're pretty excited. I think, in summary, the things that hopefully you take away from and can appreciate, we're committed to a deep storage solution to our customers, whether, again, it's from the traditional HDD level to the highest performance flash-based server level, and we'll continue to work to fill that in. World-class go-to-market infrastructure, Steve alluded to that.
I think we have a great opportunity as we see new emerging ideas. We hadn't really built a lot of deep momentum with Virident at the time it was acquired, so it's still fertile ground for us to look at the next set of opportunities that help us complement our overall portfolio strategy. Once we see those opportunities and embrace them, we can make an immediate impact in the marketplace, and we can build substantive economics quickly. We've proven that. We could see that in our work to date, so we're pretty excited about the power of our go-to-market resources to really give us that competitive advantage. Again, the integrated, I kind of talked about this already, the integrated portfolio view of how we'll build a whole stack of storage requirements to our customer base.
I think finally, the opportunity for the data explosion and how key organizations see data as a strategic value is a great opportunity for us to consolidate our importance to them. Thanks very much.
Break time?
Do we have a break?
Yep. We'll just head out to a break.
Ladies and gentlemen, if you could please return to your seats. Our program will begin momentarily. Ladies and gentlemen, please return to your seats. Our program is about to begin. Ladies and gentlemen, please welcome to the stage Dave Mosley, Executive Vice President of Operations.
Thanks very much. Listening to Steve's talk, it reinforced how excited I am about the future that we have in front of us. I hope to impart that upon you with respect to the operations in Seagate, the technology development in Seagate. There's a lot of people that bring these products together. We're quite proud of those products. I hope I give you a piece of that. I can't give you a tour of the factories today. I can't really take you into the minutiae of the technology, but if you don't have a sense of that, please hit me up afterwards, and I'd love to talk to you about it some more. What do we mean by operational excellence in this company? We operate manufacturing at one of the highest scales on Earth. We're one of the most precision automation companies.
We deal with volumes that, frankly, not many other companies ever have to contemplate. We do it with technologies that are some of the most intense, not just silicon technologies that we have to source or help design, but also some of our fundamental recording heads and media technologies, quite impressive. There's a lot of pride coming through in this talk, I think. If I take Steve's opening slide, the six zettabytes that we talked about, what this industry can actually source out of that, I deconvolve from there some of the metrics that we have to go hit inside of our technology development or our operational teams. One of the metrics you could look at is the number of recording surfaces that are going to be in a drive. This reflects move from some PCs to cloud.
What you see is today we ship just over three recording surfaces per drive on average. Some drives only have one recording surface, and some drives could have 10 or more into the future. As you look at the cloud growth, especially some of the big high payload boxes that we'll have, you look at going over five, max again could be 10 or more. That's packing a lot of heads and disks into one of our units, this really gets to the point that Steve made about it's not all about shipping volume of HDDs, it's also about the payload that we have on it. These things we have to translate back in. How do we do this from an operational perspective or from a technology development perspective? That's really the theme of my talk.
The first thing is about building out the specific boxes is almost the last thing we do. We stage the heads of media, we build the box. We really have to look at market segments to determine which box we're going to build. We talked a little earlier about the public cloud and the private cloud. You can see the phenomenal growth that we're expecting in box count here. Remember, this is not the six exabyte case. This is that storage industry perspective on what we think we're going to be able to do. You see mobile PC and desktop PC also on this list. Results in about a 28% CAGR that we have to solution or that we will be able to solution given our current investment levels and current technology roadmaps.
The key thing to point out is that while the units growth in the desktop PC and the mobile PC business may not be big, there still is appreciable exabyte growth happening inside of there. Look at the personal cloud. Sorry, the private cloud. That particular segment is really the one that we need to be focused on making the right designs for and staging the right components for. How do we make sure that we're making the right designs? I think this slide's really about our fundamental componentry, heads and media, our ability to deliver smaller and smaller bits on a disk. We're at the end of a long road transition, which we call Perpendicular recording. There's a lot of sub-technologies underneath that, say, adaptive fly height, for example, that have also aided in this technology S-curve.
We've really reached kind of a doldrums, if you will, in our ability to really push perpendicular technology forward. What these curves represent is our technology demonstrations. Technology demonstration at, say, one terabit per square inch might mean a product that comes two or three years later after we can actually get that into high volume. What do we do with this kind of rolling over of the perpendicular technology? In the past, what we've been talking about is really the next thing to get the bit smaller on the disk, which is Heat-Assisted Magnetic Recording. I'm going to talk about that some more. The interesting kink in this curve has been the fact that it's so hard.
It's not hard to pack more bits there, it's really hard to do it reliably, because what you have to do with Heat-Assisted Magnetic Recording is get a laser down to the business end of the head in order to heat that media up and make sure that the transition happens properly. What we think is we're waiting for this Heat-Assisted Magnetic Recording S-curve, which is the next big S-curve we've invested in for many years. We think we're going to be able to address the interim, if you will, with Shingled Magnetic Recording. I'd like to show you a video on Shingled Magnetic Recording, just so you have a good sense of what it is.
Our planet. 7 billion inhabitants, creating a lot of data. A lot. One problem: where do we store all of that data? We are rapidly approaching the physical limits of how much data can be written on a single hard disk platter. Let's take a closer look. On a conventional hard disk drive, data is written in tracks. The closer the tracks, the more data you can fit on a platter. Using today's perpendicular recording technology, data is written in tracks only about 75 nanometers wide, smaller than a flu virus. We've hit the limit. If we're going to increase the amount of data that can fit on a disk, we need a new approach. Introducing Seagate's Shingled Magnetic Recording, or SMR. SMR is a breakthrough technology that allows for greater capacity, which means more data can be written on a drive without increasing size or form factor.
Instead of laying tracks next to each other, we overlap them, much like shingles on a house. With this technique, we increase capacity by 25%. 25% more capacity means hard drives with the lowest cost per gigabyte and capacities of five terabytes and beyond. Launching a bold new approach to the hard disk drive. Same size, more data. Seagate's Shingled Magnetic Recording Technology, shipping now. Seagate. Storage for life.
With that technology deployed, we think we're going to be able to hit disk drive capacities in the current form factor, not going any thicker, current form factor of 20 terabytes by 2020. It's pretty remarkable. That's even a little bit beyond 25% areal density growth. The key point about shingled magnetic recording is it's shipping now. We've been testing it in various markets. We've shipped over 1 million units with SMR, and we're about ready to be able to broadly deploy it across all of the relevant markets. One of the points I want to make that I'll come back and touch on later is if you remember how those shingles kind of lay down, you need to know the data in a lot of those shingles beforehand. It's very important that you cache and tier properly before you address the media.
That's a very critical point, and it gets to Steve's earlier question or earlier point that he made about how various layers, whether it's DRAM, NAND flash, various layers up the stack from the raw media actually have to be managed properly. Okay. Shingled magnetic recording gets us to about 1.2 terabits per square inch. We've already begun product integration on that. The next technology that we haven't really talked about broadly yet, the industry, there's been various industry forums that are talking about it, is really a read head and data recording channel technology, which is called Two-Dimensional Magnetic Recording. It's really looking across the tracks and getting more information so that you can operate at fundamentally lower signal-to-noise ratios than you were before.
This technology is being developed right now, probably productized about 2016, and may be another one of those speed bumps, 15%, 20%, 25% even, as we go forward. We're actually pretty excited about this one. There's HAMR. HAMR, we've talked about for a long time, which is not read and channel, it's now write head and media technology. It's the next major significant S-curve that gets to the point where we can deliver higher and higher capacity points. We're still looking at product integration. We showed a drive actually in this forum last year, a demonstration drive. Again, working really hard to get that technology reliable. We think we're at that point. That was that little kink in the curve that I talked about before.
Driving forward to prove that we can do the areal density that's 2x or 3x or 4x where we are today. That's certainly possible on the media, to get the whole right process to work is what we're working on. We expect productization in 2016. Okay. How do we think about this development process and then staging of capital, I'll talk about as well, how do we think about this technology development process? We nucleate projects anywhere from four to seven years before they're ready, and we invest in them such that we're going to understand fundamentally from an R&D perspective what is going to be needed in order to productize them someday. We make this portfolio of bets, so to speak, and the bubbles represent different investment sizes depending on how much we need that technology or how speculative it is.
At about four years to two years, we actually go into an integration phase where we're looking at integrating that into componentry that our product teams might ultimately use. Making sure that those components are, again, reliable and can make the requisite yields and quality levels that we're going to need. During the product development process, individual product teams will go off and integrate that with other pieces as well, with other pieces that they might get from various parts of the supply chain. If you look at our relative investment portfolio, it's about 10% is in that four years and beyond timeframe, about 20% is in the integration of the platforms, and then the core products are about 70% of the total investment.
What's instructive, and just to take away, it's not all about the finished disk drive product, it's a lot about the componentry, whether it's heads or media or even channels. About 58% of our total investment is in those fundamental components throughout the course of this, getting them to market. The final drive, we're actually fairly flexible. If somebody came to me and said, "Instead of notebook drives, I need cloud drives," there's a little bit of tooling that we have to change over, and there's maybe some product teams that have to ramp up or down quickly. We can do that all certainly in the last year, and as a relative investment, it's a pretty small investment. We're fairly flexible that way.
There was a lot of talk about flash before, I want to take you through at least the way I see the journey. We do this whole broad line of HDDs and have been for 30 some odd years at Seagate. If I look at about 20 years ago, there was an engineer who came and said, "Hey, let me put some DRAM on the board." There were customers that said no because they didn't want us caching things on our side of the interface. There were probably operations people that said, "No way can you add to the bill of materials." There were all these discussions that went on, and there was probably tension in the system because of it. Those DRAM caches that we have today are very small, 32 megabytes, 64 megabytes, potentially even 128 megabytes.
For those of you who know DRAM, these are tiny pieces of DRAM, specialty DRAM. If you look at about 1% by revenue of all the DRAM that's shipped, it actually ships on a hard drive today. That's not inconsequential in the bigger world. There's a lot of hard drives going out, and they have these DRAM caches on them. If you have multiple terabytes and only a few megabytes of DRAM, we're losing scale. We actually have the capability to use NAND flash as a replacement, eight gigabytes, 16 gigabytes, for a very reasonable cost. We can. That allows us to go manage some of this tiering, caching that I talked about, power, performance, reliability, and even gain more capacity. Witness shingled magnetic recording, you can actually gain capacity by doing some of these tricks, right?
For us, the caching and NAND flash part of our portfolio has been an integral part of how we believe the industry is going to be going forward. Somebody said earlier, well, why hasn't hybrid been adopted enough? I mean, Rocky talked about us shipping 1 million units this quarter across a broad set of products, not just notebook drives. Part of the reason is because we added so much cost when we were using SLC NAND flash 3 or 4 years ago. Now that we've gotten MLC to work and we can get the cost adder to the drive very small, that opens up the whole market for us. I won't talk about exactly what the cost adder is, but I'll tell you that if you're selling a $599 notebook, cost adders that are $20, non-starters in those high volume markets.
That's one of the reasons that hybrid drives have just not taken off until now that we could actually engineer with the NAND vendors costs that were more reasonable for those market segments. With this technology to manage the NAND, we can turn around and apply that to a whole host of different other products that may have different interfaces on them. We're even looking at some interfaces that are very different for hard drives. I mean, we've shipped things like Fibre Channel for many, many years in Seagate. We know a lot about how to make an interface transition. You need a lot of partnership with your customers, but we could go to Ethernet or some other kind of memory bus or something like that fairly easily if we wanted to.
I think that'll be most of the leverage that we get going forward into SSD. I think takeaway, our perspective on this is that we have memory on our drive. We have for a long time. We've had to manage that as part of a tiered architecture for quite a while. The movement to NAND and putting NAND on our drive, we think is logical, especially at the cost and value points that NAND has become. We think it'll improve our many different vectors, whether it's quality, capacity, performance of our fundamental device. I think we can also leverage that, like Steve was talking about, into other market segments as pure NAND play, too. How do I think about this across the portfolio? I mean, on the axis of this pyramid, you see this label that says IOPS and cost per gigabyte.
IOPS is an interesting term. Some people think they know what it mean. What it means to me is 512 bytes that have to get moved from one point to another. The interesting thing about that is it all comes down to these fundamental chunk sizes that really haven't moved as operating systems have not developed for disk drives for probably the last 30 years. We're still getting judged on how do we move this fundamental block around. Disk drives, once upon a time, 512 bytes was a quarter of a revolution on a disk drive. Now it's a little tiny piece. To spread all that, say, if you took a gigabyte, for example, and spread it all over the disk drive, yeah, mechanically, it has to go seek and round and get all that data.
If you could keep it sequential, a lot of operating systems, especially tailor-made operating systems, are very good at this, you get sequential performance that's roughly comparable to NAND flash. Maybe a little bit behind, but technology-wise, we can actually go solve that too, on our core device. The way I look at this is an entire tiered structure of devices that we'll be able to solution. For the high performance, especially in those very small block sizes, it'll be NAND flash. There will be a performance layer below it that'll be very fast data rate, performance HDDs, and there's a layer below that that's just high capacity.
Steve also talked about archive, if we wanted to talk about that later or ask a question about it, we could, archive's a really interesting space about where tape is, how hard drives could be used in that space, whether it's warm data or cold data. Very interesting discussion there. If I look across our whole product portfolio, on the left-hand side of the chart, we have the core HDD products, whether it's cloud and data center products or products that go into laptops or desktops or tablets or so on and so forth. This is what we invest most of the money in, and we get most of the returns from. We also have developed this very good branded storage business that Rocky talked about. We have the DAS that Steve mentioned.
We have consumer NAS and business NAS, you can see some new announcements on that out in the lobby. We have the wireless products and just pure mobile products that I think you're aware of. We also have some services wings. EVault is an archival services, some personal cloud services. We're working very hard on data recovery, a data recovery business that we have as well, and some ways to monitor the disk drive performance in the field and actually recover from certain events that happen in the field. Those are things that we can add value to that I think we can actually get some customers interested in as well. We'll call that whole thing the services business that we're working on. Steve's chart, this is the real big summary chart. The way I feel about it is having to solution this.
Again, I think if we stay on all those areal density vectors that I talked about, if we invest about the same amount of CapEx, Pat's going to talk about what our forecasts are there, that we have been for the money we've been making, we should be on the lower blue line. That 6 zettabyte pent-up demand that's coming, that we believe will come, is that upper teal line. The difference between them is enormous. The big point on this slide for me, from a manufacturing perspective, is what happens if I have to go north of that lower blue line? I think the hard drive business has always been accustomed to going south really quick. As I install capital, I'll have my hand on the handbrake, just like I had to in 2008, right? If we have to stop the capital investments, we will.
Going north is a very expensive proposition, okay? It's going to require a lot of capital install. We won't really be leaning into the wind on it too much. We'll be installing at the same rate, or technology advancements and capital equipment replacements and things like that will be going at about the same rate of capital that we were in the near recent past. We won't really be leaning into this new vector, except I am going to do some automation and metrology investments this year to make the insides of the factories a little bit better. That'll just keep me on that blue curve. Going north of that, getting, let's say, halfway between the blue and the teal line, would require another entire industry. That's how much capacity difference we're talking about. It's not one building. It's not one factory. These are enormous investments.
That's why I think one of the things in the world that people assume is that disk drives are always going to be there. You're always going to be able to pick up the phone and call. Right now, nobody's getting really punished for not having them in their business models, right? Because the industry does have capacity. Soon, if you really subscribe to this kind of growth, and there's a lot of things that you can check to see and make sure it's happening, very soon, the equation's going to change. Then it's going to be lead time for our ability to bring up new factories. Obviously, if we can push areal density a little higher, that means just from our component perspective, we have a little bit more flexibility on having to do that because I don't need four heads.
I only need two to make the same capacity point. Anyway, I think you catch the salient part of the problem. Steve talked about suppliers and answers to one of the questions. I can't remember who asked it. I feel very proud about how we've managed our supply base. When we think about the people in Seagate, it's not just our people, it's also the people in our supply chain. We've been very responsible, I think, making sure that not only that from a business perspective, we have a healthy, sustainable business, but also from our people perspective, upstream in our supply chain and in Seagate, that we have sustainability. Then we've also made sure that we're not damaging the environment in the future.
The way you can read about this is we've signed actually the UN Global Compact, you can go out and read about that on our website. We're also very active in the EICC. Really representing not only our own values, but also our firm belief that this is going to be a good business for a long time to come, and we need to manage it properly. Like I said, a lot of pride in that, I'm also very proud of my supply base as it's been able to morph there. More than 50% of my direct suppliers have been audited by the EICC, an independent body now, we're going to push that all the way up. I think we're out ahead of that trend in general industries. These are things that manufacturing people have to be really cognizant of.
With that, if I didn't, I'd like to just offer up myself later, come by, we can talk about technology, or we can talk about operations and just continue to try to impart that real enthusiasm that I have for the business. Thanks. I'm going to turn it over to Pat.
Thanks, Dave. Always like listening to Dave. Makes me want to be a drive engineer, but they keep on turning me down. I guess I'll just be a finance guy. The other thing I like about coming here, I learn new things. I learned something new about my management team. I learned about Rocky's passion to get pictures for bathrooms. I was just in Russia on a mountain. I got a doozy one for him, so I'll show him that picture if he wants to visit there. Anyway, it's all storage, so it's all good for us. You've heard from Steve, you heard from Rocky, and you heard from Dave, and you really see what's really happening this world of consolidating from last year, I think we showed a chart of 246 drive companies down to three.
That game is pretty much done for the most part. The vertical integrators won. They own the technology, they own the supply chain, they own the capacity, that was the really business model that prevailed. As Steve and Rocky and Dave said, there's a technology that's changing the ecosystem that we're seeing fits today. There's different opportunities because of the way our customers are addressing the market. As Steve highlighted, at the end of the day, we really ship storage. I want to make a point about that in a second that this TAM that we talk about, we get so fixated on may not be the most relevant factor because not all TAMs created equal.
When you own the underlying technology and the capacity and the delivery mechanism and invest it deeply in your supply chain, it's an incredible valuable asset that Seagate and probably only one other company in the world possess, clearly in our space and in many other spaces as well, put us up against many others. Let me give you the financial overview. I'll put the wrapper around what they all said and say, okay, what does that mean to us? How do we think about approaching the business from one investing in just the business, the day-to-day running of the business, what we do with that cash flow, how we deploy our thinking. I don't think there's going to be many surprises here because we've been very open and transparent about how we do this.
It just shows, as Steve Luczo said, we're very convicted on the cash flow of the company. When we get convicted on the cash flow, allows us a point of view that what we should do with that cash flow. Let me walk you through that. I first want to bring up fiscal 2013 results, just for a couple of points. These are some of the key metrics that many of you that follow us already know these numbers, so I'm not giving you anything you don't know. There's two numbers on this chart I really want to talk about. Last year, we had the competitor talk about their business model a couple of weeks before ours, then we came out here, and we sort of had to open it up that maybe the TAM's a little light.
Everyone's like, "Oh my God, the TAM's light. What does that mean for us?" We set a model of 27%-32%, and the reason we said that is because we need that level of profitability, one, to invest in the operations, whether it's the CapEx or the development or the sales channel, whatever it's development, we need that type of development to keep that investment going, that type of margin. What happened was when we were looking at last year, we thought maybe a TAM of an average of about 150 million units. Everyone knows what that means. They have their own point of view what that means. In fact, the whole last year, we didn't have one quarter that was over 140 million, and we still delivered this. It really resonates to the point what Steve said.
We ship capacity, we ship capacity the way our customers want it. It's not about shipping spindles, it's about shipping storage, we're a storage company. That's how we manage our business. How the customers want to packetize it is, that's what we go and do. We were able to manage that business because of these fundamental technology changes that Dave and Rocky talked about, whether it's the cloud, whether it's the engagement with new customers to lower the cost of storage so they could have more storage. There's all these things out there, that's how we look at the business. We need to make that margin to continue these investments. The second one was the operating cash flow. Steve was very clear on this, that we have a very strong point of view.
We don't look to hoard cash on a balance sheet. If you look at the history, that's not what we do. We really don't want ourselves in any challenges running the business. We certainly are open to making strategic investments. We're certainly open to consolidating the industry. We've done all that. We know how to use all the tools and the tool chest provided to us about capital allocation, capital deployment, capital uses. We also say excess cash on the balance sheet, we don't need to have that. We give it back to the shareholders. We're very proud of that because that just shows our level of comfort that we can return that much cash to the shareholders and still feel comfortable about all the opportunities that Rocky and Dave talked about. Let me break this down. I'll start from the top.
Obviously, that's the best place to start from. I'll give you a little look at it. The pictorial of the graph is really the last two years. The reason we looked at the last two years, is that you see the pre-consolidation, pre-flood levels. It was a good business, certainly, we managed the business well there. You saw the bump off of the flood, and then you see the post-flood come down and the consolidation. What you see is an incredibly stable line of the last four quarters. Like I said, not one quarter of the TAM was over 140. We don't get too excited about that.
I'd be the first one to say, I'd love to see that grow, and we do believe it'll grow, but we know how to manage our business and ship capacity in different packages, whether it's five disks, whether it's one disk, whether it's one head, whether it's 10 heads. We know how to do that. That's what we've done for years. What we're here to do is reaffirm the model 27-32. I think someone earlier asked, could that model expand? Clearly, it has the opportunity to expand over time. Right now, as we're in the transition of running the vast majority of our business through the traditional HDD, that model's probably fair, and it gets to a point where we need to update it because some of these opportunities might be accretive to this or even push us further higher, we'd update the model.
For the time being, I think that's a fair model, and I think most folks are seeing some comfort level that we know how to run the business in that model. We do believe we're a revenue growth business, 3%-5%. Some folks may say it's not aggressive enough, some may say it's too aggressive, but we look at that from other pieces. We clearly see opportunities in the traditional HDD. You heard Steve and Rocky talk about getting in tablets. That's largely been a TAM that's been excluded from the disk drive. Whether it's accretive or not, you have to look at it whether the gross profit dollar is going to look at.
We clearly look to go after that market because we believe we have a technology that can scale in there, that can fit in there, and can be a great cost point, reliability point, and a power point. You look at all the metrics, it stacks up very well. We have other parts of the business. Rocky talked about our enterprise SSD. We'll have growth there. As we continue to invest in OpEx and some of these other emerging areas, we expect to see some revenue there. 3%-5%, we feel very comfortable in that revenue growth target and feel comfortable, and that's what we try to drive the team to, and really see that there's an opportunity to do that line of sight. The last one, market share, is more of an outcome than an input.
At the end of the day, you have to have the right portfolio. You don't get market share by just showing up and saying, "Okay, what do I need to price to get?" That's not the way you win this game. That's not the right approach to it. It's about the portfolio. As long as we continue to have time to market, high reliability, and an extensive portfolio, that's probably a fair outcome. Now, if we missed on any of them, we would adjust accordingly. We're not going to push market share because we have any portfolio holes. Our job is to deliver enough for the portfolio so this is a natural outcome given our scale and size, our cost structure, et cetera. That's more of an output, but that's what we look to. We try to measure, that's a statement of health.
That's sort of where we look at the revenue and the gross margin on a go-forward basis. On the non-GAAP operating expenses, you'll see the blue line trend nominally up, and I think Steve highlighted that already. We are making bets. We're getting more convicted on making these bets, number 1, because we've held that margin at 27% fairly consistently, and we're very convicted to hold that. That allows us to put a little more operating leverage into our model about making selected bets organically or partnering with other folks. As you can see, the 12%-14%, with that model, we're fairly comfortable. We'd expect over time that some of these bets come to roost, and we start generating the revenue, as Steve said. You may see that coming down, but those are probably a year to two years of development cycle.
Not long development cycles, not long to market, but it takes some time. We're making those bets very thoughtfully and strategically and metered. We're not trying to flood the farm with all these. We see a tremendous amount of opportunities, and I think someone asked us to prioritize. We do that all the time. Where are these bets? How do we structure them? How do we stage them? What level of investment? That's very normal inside of Seagate. The manufacturing scale that Dave talked about with the magnitude of dollars is kind of funny. We always get hung up on how much CapEx dollars we do, but if you take a look at 6%-8% of revenue, it's a very, very efficient model.
Even Dave had to ramp up his cap spending, and it still wouldn't hit some of the levels where you see semi-cons are doing about 33% of capital to revenue dollar. It's still going to be very, very efficient, but we do feel that we could run 6%-8%. To the second point, it's been very maintenance-focused. We have to invest in test times because density goes up, as you saw from Dave's chart. That means test times go up. We know how to run that, but it's very maintenance-focused. In fact, fiscal 2013 and fiscal 2014, it may be artificially high in the sense that we're making investments in our research facilities.
We have longstanding research facilities for different reasons, whether because of an acquisition in Samsung, we need to have a facility, whether we need to move to a new facility for media folks or our science park design team. These are going to be stay-there facilities that will yield dividends in many, many years. We're taking the opportunity while there's maintenance capital mode to make those investments and fit them in our business model in a very appropriate and metered way, as I said. 70-million-unit capacity is sort of a number that moves a lot. I think Dave and Steve both commented on it. That's a sense of what we do today. There's less client drives and more cloud drives. You may not have that capability, but we have the infrastructure in place.
As Steve even said, when you look at it from how much we're using for capacity of test time, it's probably closer to 90%. That 50 million getting to 70 million in some parts of our supply chain may be sort of challenged if we want to go right to 70 million. We look at that all the time. Dave manages the trade-offs of that on a not just a weekly basis but a daily basis. This is a chart that probably, when we look at the financial metrics, this is the one that we're probably the most proud of because I bring FY 2009 in there for one particular reason. Anyone who's been with us that long knew how challenging FY 2009 was, not just from our technology misstep, but obviously the macro that was very stressed.
Even in the worst time, you see the level of cash flow we generate in the business. Seagate's always been a cash flow generating business, and that's why we feel very comfortable with how much cash we return to shareholders because we always have a good line of sight on cash. You can see the dollars, $10.3 over the last five years. That's an enormous amount of cash flow over the last five years. We're very proud. We get this engine up and running, and it really is a cash machine. I only put the current market valuation up there. I let the market decide what they should put a price tag on it.
When Steve and I and the board look at it's one of the reasons we look at this and say, we think it's very fair to keep on buying back the stock, and we'll talk about that in a second, because we certainly feel it's a good value. If it wasn't a good value, we may not buy it. That might not be the only reason. We might have some other strategic items. We don't have those strategic items. We always feel that we look at certain key metrics and say, "Yeah, it's a good time to buy the stock." Last quarter, we bought four million shares. We're very comfortable with the stock price. We look at it. We're not going to be bad stewards of the money. We do have viewpoints of how do we go about buying the shares.
We shared this chart last year. It's an important chart because some of these are fixed investments we've already made. Once we make the investment, we're very committed to making, we don't want to back off on those, such as dividends and the appropriate debt levels. We have to service our debt. Once we step into an arena of the dividend level, we don't want to retract on that. We'd always like it to grow. I'll go through that in a moment. We really are very disciplined on how we look at this capital deployment. We do have a dividend, an active dividend program. We do have an active share redemption program. I'll talk about those both in a second. Our debt levels, I have a chart on that.
I think we've done a fantastic job on our debt. I'll show you a slide on that as well. We've made investments. We talked about some of the investments. We make investments for strategic elements, for consolidation elements or for whatever. Like I said, the tool chest enables, we know how to do that as well. The one number I put on there is $1 billion. We probably need to run this business as only about $1 billion in cash. We like to have about $1 billion and five to give us that flexibility, whether there's opportunities out there. That's sort of our operating principle, keep it $1 billion and five. The reason I put $1 billion is it only takes about $1 billion to run this business.
When we look at the cash on the balance sheet of $2 billion, we certainly say we have opportunities to continue to invest in the business and return that to the shareholders as well. The dividend program we launched after the fiscal 2009, we relaunched it. I think it's been one of the more progressive dividends. I think part of the dividends is one of the points of the capital deployment where I talked about the principles of. You can see the levels of increases over the years. It's been very healthy. Last year, the 19% raise was a pull-in raise at the same time. We got the Q3 dividend in before the year-end for a lot of U.S. tax considerations. We also did the raise.
If you take a look at the dollar, it's almost $1 billion in the last three years, fiscal years, return in dividends to shareholders. This is a program that will stay with us. We feel very comfortable with it. We're very comfortable with the 10% year-on-year raise. You can take a look. The last time we did was November, December timeframe. Obviously, we're in that mode of we're looking at that again. We're very committed to try to grow that dividend while we're in this state of our capital deployment. Share redemption's given you a picture there. That's about $5 billion we've returned to shareholders through this program. It's been probably the more significant of the capital deployment. Part of that was we wanted to get down to 250 million shares. We put that goal out in 2011.
Share price was 19. We had almost 450 million shares outstanding. We had another 60 million of dilutive shares on top of it. We put that target out there. One, for a lot of reasons, for the type of shareholders we were looking for. That's why we hit the dividend program as well. Then two, to say the type of level of liquidity we thought would be for a company our size. If you take a look at about 25% of the company has been pulled back out of the market. Then 67% is the load of shares. You look where we're at, we've had some limitations on how many shares we can buy, so it's been fairly lumpy, but we've been very transparent what that is. If you look at the fiscal 2013, we've taken that number down to 359 million shares.
By the end of the fiscal year, we should be down to 325. Then we still have the target of 250. With our cash flow, with our capabilities of leverage, and we run our business of investment-grade leverage, so we've never put ourselves in any jeopardy for that. For other sources of funds that Steve talked about that might be being mailed up on a stamp from the south, they certainly could help get us to that level. We feel fairly comfortable, and we don't want to move off that target. At the end of the day, the more important thing on this chart is that we're committed to return our cash flow to the shareholders. Obviously, when you buy the stock, a share price is a consideration, and we have to be mindful of that.
The point of view is that we're going to return a significant amount of cash flow year in, year out, as long as we have that line of sight of cash to the shareholders. I want to put this chart in because a lot of folks think that, oh, Seagate has a bit of debt. Actually, we're not quite that leveraged by any stretch of the imagination for a company our size. In fact, if you look through fiscal 2011, we've actually de-levered the company. As we've been buying back shares, heavily as we've been increasing the dividends, we've also de-levered the company. Not only have we de-levered the company, we've extended the maturities and lowered the cost.
I think that's, Steve and I always work to see where the opportunities are to do that, because our job is really to push that maturity wall out and lower the cost. I think if you look at how that stacks up for the rest of the decade, there's about $600 million left in just principal debt. That is clearly within line of sight of the cash flow you saw the last five years. That's not an issue at all. We will continue to do that as the market allows us to lever up accordingly and to be sort of in that investment-grade banter. That's how we operate ourselves, whether we're there or not. That's how we want to run the business that will continue to look to reshape our debt.
I said that we're very proud of how this has just been pushed out and lowered the cost of it. We'll continue to monitor the markets and where they go and always look at this on a very strategic basis and structural basis for long term. Capital deployment plan, like I said, when we get to 250 million shares, we can talk about how we address this. You can really see how we deployed that cash to shareholders last year. It really doesn't change that much, because we still have a fairly significant share redemption plan in place. That's still going to be a large part of that capital deployment. We still expect to pay out approximately 70% of our cash flow this year to shareholders. Last year, just even the free cash flow was almost one for one.
Almost 100% of our free cash flow went back to shareholders. Again, this conviction on how to run the business, how much cash we need to make these bets. We look at all that, we make those right decisions first, then we come to here. I think it's very thoughtful. Debt retirement, that's opportunistic. We really don't have any structural debt coming due. This year, we wouldn't really plan any, but we always look at our debt and see if it's efficient trading levels and how we manage that. M&A is just opportunity-dependent.
Just in summary, what the model is, to put everything in one page for a snapshot, of how we look at the business, how we want to run the business, and how we get a comfort level on what we want to do, from whether it's investment in just the core, whether it's investing in adjacencies, whether it's investing in the factories in CapEx, whether it's investing in any potential strategic M&A. We go through what we think the model's going to yield. We believe 3%-5% is a fair revenue growth target to put on us. We think we could stand up to that. Non-GAAP gross margin. At this time, we're not changing it, but we did up it last year. We feel very comfortable with that. Like I said, over time, we could address that.
If some of these adjacent bets change the margin structure, we'll certainly be here to update the market. Non-GAAP operating expenses, as I said, for the year two, as we make these bets, we're not going to shy away from these bets. We just think it would be the wrong thing to do. Whether it's Rocky or Dave or other folks on our team making these bets, we're going to continue to make them if they make sense and they can yield a good return in a very short period of time. Capital expenditures, like Dave said, you haven't seen that elbow start lifting yet.
We're going to just run that model until we get the conviction from whether it's the customers, whether it's from the macro, whether we see whatever we believe is sustainable, we can start move that up a little bit, but that's not going to be for this year for sure, probably not going to be for next year until we see that level of conviction. That's the financial model. On the deployment, CapEx, about $850 million. We don't get too excited if it's $800 million or $900 million. That's noise level for us. We'll look at that, because a lot of these things are timing from a quarter to quarter. But we use for planning purposes, about $850 million into our plan. We're very comfortable with that number. That would probably be on the low end of the plan.
Like I said, it was $50 million more. That's not something we really lose sleep over, but we're very thoughtful about it. Dividends, about $550 million, share redemptions, 30 to 35. That's a prescriptive number. We have limitations on what we can buy, as long as we want to protect our NOLs. Today, we're committed to protect them, but we always look at that as well. That would lead us to about, for the course of this year, fiscal year, about that many shares repurchased, you guys could do the math and plug in whatever share price you think. That's the level of shares we'd like to buy this year. With that, I think we've hit all the strategic elements. We've updated the model.
I'm going to hand it back to Steve, we're going to do Q&A up here. I think Steve, or I don't know, Kate's going to be the moderator. We'll just handle that. We'll get the mics out, and you can answer questions.
I'm going to moderate. I'm so nervous that our CFO is doing so many betting. Anyway. I like to think of it as investing.
Yeah.
That's Pat. Let's open up for some questions. Sure.
Pat, just so we can look at the current demand environment, first of all, I think last quarter you had talked about a 135-140 TAM. Are you still standing by that in the current quarter? Also, just so I know the math here, it looks like given the $550 that you've talked about in terms of the dividend, you're really talking about $2.7 billion-$2.8 billion of cash flow from operations. Is that how we should think about the model in the context of that cash flow as you think about the 3%-5% growth, be it relative to revenue? How do you think about the long-term sustainability of that cash flow?
Yes and yes.
Yes.
We feel fine about the long-term sustainability. Next question. Yeah.
Thank you. Sherri Scribner from Deutsche Bank. I just wanted to dig into the 10% dividend growth over the long term. That would assume that either your cash flow is growing at 10% or you're eating into the buyback. I just want to get a sense of what you think your long-term growth profile is for the cash flow.
Well, while you're doing the buyback, obviously, you're reducing your share count so you can keep your dollar cash flow. For a couple of years, we've got plenty of room that way.
Right. Then I wanted to ask a question on the technology side. You, Steve, talked about maybe 20% areal density growth, and if you assume the dollar per gigabyte prices decline at about the rate of areal density for the hard drive industry. I think NAND expectations are a little bit higher, not substantially higher for how quickly they can decrease die sizes. If you assume that dollar per gigabyte decline is similar, potentially, if we're not able to increase areal density, maybe NAND starts to get cheaper. How do you deal with that?
Yeah. We see it the other way. We have pretty deep relationships with the NAND providers. I think if anything-- Look, it's like I said last year, I worry way more about the NAND Flash people not staying on the linear density curve than I do that they do. When you look at the capital that they have to deploy to get to the next generation of capacity per chip, it's tough. Whether or not you're taking the line width down or whether or not you're going to go to 2D or 3D or 2-bit or 3-bit or 3D, the capital costs are just insane. If we actually look at the trends recently, the drive industry is actually gapping out again relative to the NAND cost, where it's still an order of magnitude higher. I actually think overall, the relationship doesn't change.
I worry that it will, but I worry that it's going to change the other way because that impacts us in the network effect. I think it's really important to realize that the boundary of competition between NAND and HDD is pretty thin relative to the advantage that both of us expanding our technology creates. What the NAND folks are doing and then how they deploy that into device-level technology is so important to data creation, data sharing, that I worry about that network effect way more than I worry about what happens at the boundary of some application that SSD is better at. Because the reality is, in the enterprise, in the large data sets, in mass storage, which is what we are for the most part, it's a really thin line. Because a disk drive, Dave talked about IOPS.
The disk drive, when you think about IOPS versus SSDs, there's really no comparison. If you're doing sequential things, there's really no cost comparison on the NAND side. I think that boundary is thin. I worry about it the other way. From everything we see, they're probably more worried about their ability to stay on that linear density curve than we are staying on our areal density curve. By the way, I think we're both really challenged, I think that if you believe in the data growth rates that we're talking about, there will be shortages on both ends of the equation. Yeah. Any more on this side? I can see all these question askers over here. Yeah.
Question on gross margins. Maybe, Pat, you can chime in. Over the last year, you've been at the lower end of the long-term target. What are the areas of flex that in the medium-term, not the longer-term new areas of investment that you mentioned, but in the medium-term, what can take you to the upper half of the gross margin range?
It's cloud growth.
Volume.
Cloud growth.
Mix and growth.
Yeah. Cloud growth. Just because what happens is, again, a unit is in a unit is in a unit. Our ability to respond to upsides on client is a lot easier than our ability to respond to upsides on cloud or enterprise, and that's mostly issues around complexity of the device and test. If someone comes to us in a quarter and says, "We need 1 million unit more of one-terabyte notebook drives," that puts some stress on the factories. We got to make sure we have the parts and all that. Typically, we can respond to that. If someone came and said, "I need another 1 million four-terabyte drives," the answer is, "Great. We'll send them to you next quarter, and by the way, the price is going to be this." It's not the same equation.
To me, what gets us off the 27-28 end of that versus the 31-32 end is what happens with mix relative to overall demand. Now, if macro situations lift, that is a different issue, because now we've got this issue of how do we package our petabytes, and if you create shortages in notebook, guess what? Those price margins are going to go up as well. Right now, the margins in the notebook space are pretty low, but if there was a shortage, they would get high really quickly because it doesn't take a lot of dollar movement to move it up. As I think about the world in reality, I think the notebook recovery won't be so dramatic that there'll be shortages, but I do think it's going to be better than no recovery, so we will get some more absorption.
I think where the event happens that starts to give us more margin opportunity is really as cloud deployment accelerates in terms of these kind of chunky ways that it does, that's where you're going to have immediate impact on pricing that then probably has a couple of quarters of life to it. I think it's the general trend. I'm glad you asked that question because it's back to this, what's the GAAP look like, right? If the GAAP looks like what we drew, obviously, the margin implication is interesting. If it doesn't, I think this is Pat's point is that the industry has tended to operate in this 27%-28% margin range. I don't see any change in that, which I think was maybe part of your earlier question.
I think the industry is very disciplined about what do we need as a margin to invest in our businesses and into the opportunities. I think the answer is as you start getting stressed, and I probably think it's going to happen in the cloud area first, that's when you're going to start to see margins popping up a little bit, just because of the mix-related issues, how it impacts our deployment of capital. Yeah.
Pat, can you provide us with the cash taxes for the next couple of years, your projections as well as the NOL status?
Cash taxes, just plan for about $60 million a year. Just use it as cash, don't view it as a rate. We have about $3 billion of solid NOLs on the books that continue to be looked at, and they're very solid, and they're very recent, too. They're about four or five years old.
Let's run back there. Then we'll come over here. Sorry, I didn't see you.
Thanks. I'm just curious, when you show that shot of capacity versus the drive capacity versus data growth, what do customers typically tell you when they see that gap? Just to get a sense, how much capacity of CapEx would you or the industry need to try to close up maybe half of that gap? Just a quick one. Earlier in the day, you made comments about 140 million TAM. Aaron was talking about your guide was 135 to 140. Does that imply things are maybe towards the high end of the range, or am I sliding this too much?
Again, don't get hung up on units. For the drive industry, 130 to 150, it's all the same. It's all the same. It doesn't impact our cost. It doesn't impact absorption. At the margin, maybe. What we tend to do is go into a quarter, look at demand, unit demand, probably decide that we're going to plan to levels anywhere 5%-10% below what we think the market is going to provide because we can chase the upsides. When you have excess capacity, there's no reason to do that. You go in with a lower amount, you're getting the absorption you want, then chase the upsides, and if you can get them. If the TAM ended up growing from 140 to 150, then people chase the stock up because of that, it's nutty. You're not going to get that much absorption.
If it's 132 instead of 136, are you kidding me? That's like two days of production. It doesn't matter. Don't get hung up on that. The market seems to be right now in a 135 to 140 range. This quarter, could it be more than 140? I don't know. September quarter is a weird quarter. It's the worst quarter in terms of back-end loading. December is usually stronger than September. Sometimes September is better. It's all noise level right now. I think the important thing is watch the petabyte growth. Watch if there is some traction on mobile in terms of notebooks getting a little bit of recovery because the devices start getting more interesting. Watch whether or not five-millimeter drives actually penetrate into the tablet space or some other mobile space. Watch the cloud. Watch the nearline business. Mission-critical, by the way, has been remarkably stable.
What?
I can make a comment about this.
Yeah, go ahead.
If you guys still want to.
Yeah. Make a comment.
It's interesting you asked that question about the two divergent curves or departing curves. With the cloud service customers, almost to an account, the immediate response was, "How can I commit to a long-term supply agreement?
We had a notebook.
They believe in that data curve. That's what they're very fearful of.
Yeah
That data curve is their strategic initiative, and they don't want to be out of pocket on it.
We had notebook customers say the same thing. They said, "We want to extend our LTA through 2016." We have other ones that say, "Can we buy head capacity?" Because they worry about heads being the shortage. We say, "Sure. We'll package them in drives and sell them to you." You get different. EMC would say that that number is off by 4X. EMC says that number-
Five.
Yeah, EMC says that number is 30 zettabytes-35 zettabytes. We talked a little bit, Dave mentioned it, I think Rocky mentioned it, what that number is the annual shipment. When you talk about the corpus of what's out there, today it's about two and a half zettabytes. That number starts to approach 10, 15 zettabytes in terms of what's out there. Then you start getting this machine-to-machine effect, which I think is an area that none of us are really expert at understanding how big the machine-to-machine data learning and data creation is going to be, self-created simulations and things like that. I think that's where EMC gets numbers that are much higher than that. It's not one that people tend to look at and challenge. Sometimes they think we're doing it to promote a pricing discussion, which we're not.
What we do use it for is to promote an end-to-end supply chain discussion between our operations and our go-to-market capabilities to say, therefore, velocity is more important than ever because the faster we can get this stuff through the system, the better. I know Dave or Rocky if you want to add anything.
Yeah. I was actually going to say, another initiative that I think all of you could go pursue is the growing initiative around lighting up the dark data that's out there. This, nobody has really measured, but there is a lot of resources and a lot of early-stage development around how to light up the dark data that resides from years and years ago, which is another initiative that, once it gets lit up and becomes part of the highly sophisticated analytical engines that the future data, who knows? That could change that curve even more.
There's a question over here. Yeah.
My question is on 3D NAND, specifically related to hybrid drives, but you can add more on it if you'd like to elaborate on 3D NAND. Realizing that it's not being produced right now in hybrid drives, and it's developing, can you comment on the technical merits of it relative to planar, the cost competitiveness of it, and are there any compatibility issues in terms of implementing it?
We should probably let our partner Samsung answer all those questions. They seem to be the leader in 3D NAND. We need to be probably a little bit careful about what we say because we probably have a greater exposure to it than maybe the average company. I, for one, would not underestimate Samsung's ability to deliver 3D NAND effectively on the schedules that they're talking about. Dave, maybe, and Gary Gentry here, who also heads our SSD business, may want to weigh in with a couple of comments, but remember our relationship.
I think integration is tough, just like every generation over generation is tough, but we don't really foresee any compatibility issues or anything else. We'll be able to use it when it's ready.
Gary, do you have any thoughts? You want to add anything?
Yeah. 3D is going to come mainstream. Our relationship with Samsung is very deep, even in joint development work. We will launch next year. When it comes to compatibility, as Dave said, there is none. The way we.
No issues.
No issues at all. I agree with Steve, 3D is a big deal and Samsung by far is the leader in deploying it, not just in R&D, but in putting fabs down to support it.
I think the perception I get out there is that people are a little bit doubtful of the ability for Samsung to deliver it on the timetable they're talking about. I, for one, would not bet against that company. It's remarkable. Yep.
Can you elaborate a little bit more on the trade-offs you see between local storage and network costs? You talked about that earlier. Just the metrics you use to compare those two concepts, either from a physics and an economic perspective. Thank you.
Yeah. In general, bandwidth is still an order of magnitude more expensive than storage, they're talking about not coming down the cost curves as fast. They're even more challenged in terms of delivering bandwidth at the rate that data growth rate is growing or that the storage elements are growing. What we believe is that in order to deliver the user experience that people want, especially on a mobile platform, that you're going to have to bring the storage closer to the end device. You can think of that as caching if you want. Therefore, that's why we think these five-millimeter drives have a nice life to live in terms of being on or very close to a mobile device. Rocky, talk a little bit more about it.
Yeah. I would say when you looked at the projected data creation curve that we keep alluding to, when you look underneath the elements, over half the demand is coming from new emerging markets, which is going to be a somewhat cost-sensitive growth dynamic. I think the idea of everything being frictionless from a transport side and being costless is not the right assumption. I think therefore, we believe in the storage fabric because in order to, as Steve said, balance the cost of transport and connectivity as a way to get your content or your data, positioning locally key elements of content or data is going to be important to create the user experience and the cost profile.
Especially if you're talking about professionally created content, the amount of investments that the content creators are putting into high-quality replication of the video experience is enormous. To the extent that that can't get delivered to the consumer, that's a huge disappointment for the content creator because they've spent hundreds of millions of dollars creating this art form that then gets polluted by bad bandwidth.
I think that's one of the reasons why our engagements with the professional content creators, and other device companies that are very interested in providing HD-quality video, which isn't just professionally created, obviously, but thanks to companies like GoPro, we can all go do, is we're getting a lot of traction with them because their business models are based on making sure that that experience is high quality, and it's not going to be high quality if you're sitting there waiting for the thing to download. It's just no fun. We think there's a big opportunity.
Actually, to add onto that's actually confirmed by discussions with communications companies. They're believers in storage as well.
There's another question over here. Yeah, Rich.
Two questions. First, can you just talk a little bit about what you're seeing in the retail market right now in terms of average capacity demanded or any interesting trends from a features, what people are looking for today?
You say in the retail market?
Retail. Secondly, when you look at the SSD space, we constantly get questions about owning the NAND, at what point does it make sense to own that capability even at all in-house. I was just wondering if you look at the history of the drive industry, at what point did Seagate want to be making its own heads? If you could just go back, is there an analogy there? At what volume it made sense?
Yeah, I think I answered that question earlier. I don't know that it's quite the same analogy, but The reason the drive industry tended to get vertically integrated was access to technology, cost, and ability to ramp. The silicon companies certainly seem to have enough technology, so you don't have to worry about whether or not they're investing the right levels of technology relative to what we could do. Having access to it is important, I forget, I think it was you who mentioned that the subtlety of what we have with Samsung in terms of the NAND flash agreement shouldn't be lost on anyone.
We made that very clear when we made the acquisition, that part of it was not just buying the drive business, it was having a NAND flash agreement that basically secures us supply, certain requirements that are easy for us to meet in terms of forecasting. That's not really an issue. Cost, it seems to be a fairly competitive market, I'm not sure we would out-cost anybody. Again, I'm not sure it's the same analogy as much as what do we do with the system on top of it. If you're asking me, are we going to become a fab company, I think the answer is no. What was this other question? Oh, retail. Same trend. They need more and more capacity. You come out with a 2 terabyte drive, and they want that in a small form factor, and then 4 terabytes.
I think the biggest issues around retail at this point now is how do I manage all this content? Easier-to-use interfaces and software that allow you to manage your devices and your content across multiple OSes now, multiple devices, and probably while you're on the road. I think the shift that we see and are playing into with some of our investments is, I think the old view of the world is that your mobile device is somehow an appendage of your home device if you think about where data resides. All the data's back there, and this is the little viewer. I think, like everything else that's happened with respect to mobile, you have to start thinking about the mobile device as the control center itself.
Therefore, that has implications about what's your software strategy and your services strategy about managing content, even if only a sliver of that content is on that device. Even if you only have 64 GB of your 8 terabytes on that device, that device probably is going to take a much more important role in terms of how you manage that content than we probably think of it today. We're pushing that theory in a lot of our retail investments in technology, and a lot of that is software and services, but it also relates to basically device-to-device communication. The other thing is this fascinating world between iOS and Android and Windows you have 3 or 4 different OSes now, plus you have all sorts of hardware packaging around that.
It gets interesting in terms of how do you make sure that all that stuff works together. I think that's probably the biggest opportunity in retail, is if you can solve that problem, you probably can attract a lot of value, and we're working hard on that. Yeah.
First thing you mentioned by December quarter, $100 million in hybrid plus SSDs. I presume that's 90+% will be hybrid or is it a much lower percentage than that? Could you give a little detail there?
No. It's more like 60/40.
Okay. Then, you mentioned the caching software or rather the, I guess it's an API for OEMs to integrate the onboard flash with a 5 mm hard drive in case if they don't want to utilize the hybrid drives. Is there a pricing for that API or is the idea to create lock-in for the OEM to say, "Yeah, we want to keep on using Seagate drives because it integrates better with that caching software"?
Not charging for the toolkit, that's the question?
Yeah.
No, we don't charge for the toolkit.
The logic behind that?
The logic behind it? What do you mean?
That's the firmware.
Yeah, we don't charge for the toolkit.
Okay.
Yeah. If they have flash, we'll give them the toolkit. They can incorporate a non-hybrid drive and get hybrid performance.
Will that toolkit integrate only with Seagate drives or it integrates with all types of drives?
Seagate drives.
Seagate only.
Yeah. Is there any other?
Yeah. Just on that long-term graph of demand versus capacity shift, if that's true and we really have twice the demand than what we can actually output in 2020, we would agree that's a boost to gross margin long term. Would also argue it supports top-line growth, however, and I think if we're right, Pat updated the long-term model with a slight revision down on the top line. Maybe if we could just discuss some of the assumptions by sub-segment, what's giving you pause there longer term?
I think last year we were ±5%, I think that could have been zero. I think I'm tightening the range, I'm not downgrading it. Just a point of clarification. To the point about the OpEx, we certainly think there's opportunity for the revenue line and for the gross margin line. As we make those investments, we'll update the model as that plays through. I think that's a slight subtlety, but we're not really taking the model down because when we did it last year, it could've been viewed as 0%-5%. We're saying we're at a growth company at 3%-5%.
Got it. Thank you.
Yeah.
Quickly, Steve, on the hybrid opportunity, can you elaborate on your thoughts for how the notebook market shapes up? I know you said you still think that there'll be some hybrid type on 80% of the notebook market, but maybe we're already starting to see some books on the high end go all flash, and maybe if you can talk about the thin HDD versus the hybrid versus all flash.
Yeah. If you take just that ultralight netbook, or what do they call the thing? Ultralight and thin and light. Ultrabook, thin, and light, depending on if Intel's paying you or not to say one word or the other. That market today is about 25 million units. Today, that's really where our 7-millimeter hybrid drive is penetrating. That's of the million units that Rocky mentioned, let's just say a big percentage of that is 7-millimeter going into notebook versus going into either enterprise or going into desktop. 750,000 drives growing of 5 or 6 million a quarter is a pretty decent penetration already of that class of product. Your question's a little different. Of the 170 million units, that now is 25 million Ultrabook, how much of that goes? I don't know. It's 15% of the market, which is the high end. You know what I mean?
How big is the high end? By definition, the high end is usually about 15% of the market. The rest of the world isn't Palo Alto, as I like to say. Not a lot of people can afford $1,000 for a notebook or a tablet or whatever it all is when you add everything on top of it. Dave's point's a little different, that when you think about how we've deployed DRAM over the years, and we've really limited ourselves to a small amount of DRAM, we now can hit cost points on using big amounts of flash that allow us then to take liberties in terms of some of the technical things that we can do with that. Therefore, that's where you're going to get this deployment of flash across 80% of our drives, even in enterprise.
By the way, it could only be about reliability because if there's certain things, for example, that you're doing that you're caching all the time into the NAND and therefore you're not causing the head to seek all the time, you might get a lot higher reliability, which means a lot to an enterprise customer not to have to send someone out and replace a drive because a head's failed or because something's broken mechanically. I think there are many reasons why we will deploy flash across all of our devices, one of which also is market-driven in terms of the application set. Do I see that market growing to 40% of the notebook market? I think no, by definition it can't because the price points are just too high.
Unless the world gets really rich, which might happen, in which case, yeah, that's good for all of us. You know what I mean? If you get some big per capita increases in India and China and Africa, and people can start spending $1,000 for a computer, then yeah, you might get really high penetration rates there. Couple more questions?
Yeah.
Hi, thanks. Steve, to what degree are the OEMs in agreement with the hard drive industry now about the need to, I guess, about the gross margin range and the role that they need to play going forward with regards to investment in the technology to keep things moving along as they are?
I think that they understand our business model. I think for them, it's more about getting the devices they need, the quality they need, and hitting their cost targets and hitting their takedowns and things like that, which we get up every morning and figure out how do you basically make this technology more affordable? That's good for all of us. It doesn't mean that you're giving up margin in doing that, though. I think everyone understands that the world post-consolidation, post-flood is a world where supply and demand is in balance more often, and therefore, pricing is more reflective of what the long-term industry needs are and always had been. It's just that what happened before is you had 200 companies go away because they were overtaxed, and I think now it's a more stable industry, and people understand that.
Honestly, I think they think more about it in terms of what it means for their industry. Our upstream suppliers went through it before we did, and then we did, and now you look at that industry, and you probably can make an argument that there's going to be consolidation there. When that consolidation gets resolved, guess what? There's going to be margin expansion there as well because they're operating sub-scale right now in terms of what their financial models require.
Would it be your view that the tenor of the conversation has changed to the extent?
It's way more strategic.
Way strategic, yeah.
It's way more strategic. By the way, where the companies aren't strategic, we don't have the conversation. We really don't. What we tell those companies is, "Fine, put in your order. We'll tell you what we can do. If you come to us with upsides, you're probably not going to be at the front of the list." There's not many like that. The companies that are doing really well, maybe not profitably because, again, maybe there's too many players, if you think about the companies that are winning share right now, Lenovo, Dell. Dell in the server market last quarter did really well. Those guys are so strategic about their supply chain. We just line right up with them. Yeah. All right. One more, or is that it? One more. Kate's saying I can have one more question. Unless there isn't one. Good. Okay. All right.
Well, that's good. We answered everyone's question. We're going to have a reception downstairs. I want to thank all the shareholders here. Thanks for being such great shareholders. For all the industry analysts, thanks for all the hard work you do, and we'll see you next year, and we'll see you downstairs. Thank you.