Altera Corporation (NASDAQ:ALTR)
Analyst Meeting 2013
November 18, 2013 2:00 pm ET
Scott Wylie - Former Vice President of Investor Relations
John P. Daane - Chairman, Chief Executive Officer and President
Danny K. Biran - Senior Vice President of Corporate Strategy and Marketing
Scott A. Bibaud - Senior Vice President of Communications and Broadcast Business Division
Jeffrey W. Waters - Senior Vice President and General Manager of Military, Industrial and Computing Division
Ronald J. Pasek - Chief Financial Officer, Principal Accounting Officer and Senior Vice President
Ambrish Srivastava - BMO Capital Markets U.S.
Arif Nizar Karim - Kilimanjaro Capital, LLC
Romit J. Shah - Nomura Securities Co. Ltd., Research Division
Alex Gauna - JMP Securities LLC, Research Division
Good afternoon. We have a few folks with snacks just coming in the back door here. For those of you who have not met me, I'm Scott Wylie, I'm the Vice President of Investor Relations for Altera. And we're delighted to have you here today.
I have a few tasks to perform at the very beginning of this session this afternoon. First, this presentation contains forward-looking statements, which are made pursuant to the Safe Harbor provisions of the Private Securities Litigation Reform Act of 1995. Forward-looking information generally refers to any information relevant to a future time period. Investors are cautioned that actual results may differ materially from these forward-looking statements and that these statements must be considered in conjunction with the cautionary warnings that appear on our SEC filings available from the company without charge.
Further, this discussion and presentation may repeat elements of prior guidance and in so doing, the company is neither reaffirming or modifying that prior guidance.
Our focus this afternoon will be strategic and therefore, somewhat longer term in nature. While during the course of this presentation, we will offer you our guidance elements for next year as we typically do, gross margin, operating expenses and related. We will not otherwise touch on our existing fourth quarter guidance nor current business conditions, trends and so on.
And in light of this, we ask when we merge into our Q&A session, we encourage your questions. But if you would, please be mindful of those areas where we're not able to comment today. And again, it is a presentation that we think will engender healthy questions and answers, and we invite those questions when we transition to that portion of the event.
We will, after we conclude, adjourn to the room that you all passed by on your way in, [ph] will be an informal reception. All of today's speakers will be in that reception. We'll be delighted to carry on the conversation with you there.
Finally, and once again, we're delighted to see you here today and welcome. John?
John P. Daane
Thank you very much, Scott. This afternoon, we're going to go through a series of presentations. First, Danny Biran is going to present on our products technology, followed by Scott Bibaud, who will talk about the opportunities both within the wireless, as well as the optical markets and communications. Jeff Waters will present on computer, new and exciting market for Altera. Ron Pasek will provide financial guidance. I'll come back at the end for some brief wrap-up remarks. We'll go into a Q&A and then, as Scott mentioned, we'll follow this with a cocktail event for additional questions and answers for everyone.
There are 3 primary messages that we have today. First message is that if we look at what's happening from a cost perspective, and the cost is both from a cost development, as well as the increasing costs of transistors, we find is the ASIC and the ASPP models are under pressure, very difficult for companies offering these products to achieve the expected return on investment from an investor. For many markets, we see that these technologies and products are actually retreating. And as such, have a great opportunity to continue to come in and replace what [ph] our very large markets, much larger than programmable logic, replace them with our products and continue to grow.
Second, as both Scott and Jeff will comment today, we have great opportunities both within existing, as well as new markets, where continued growth, placing a wide variety of products and technology. And then third, as you'll see from Danny, we really have engineered our product lineup to take maximum advantage of ASIC and ASSP replacement. We think with all of these things, we will continue to be able to grow twice as fast as the overall semiconductor industry.
And with that, I'll now turn the presentation over to Danny.
Danny K. Biran
Thank you, John. There are 2 interesting ways to look at our business. Look at it by vertical markets, you can see on your left, this is [ph] Altera revenue by vertical market to the first 3 quarters of the year. And then look at it by product category and we look at high end, midrange, low cost, and what you see on the right-hand side is our view of how that -- it is going to look like over the next generation.
High end is and will continue to be the largest power [ph] [indiscernible] in half of the business, followed by midrange, then low cost. And based on some public comments, we believe that this view is [indiscernible] by our competitors as well.
Obviously, there's [ph] mapping between the market segments and the product categories. And you can see here which are the markets that primarily use the high-end products, which are those that primarily use midrange, low cost. There's a couple of points that I would like to emphasize [indiscernible]. First of all, as you will hear from Scott and Jeff, some of the new emerging opportunities are in the high end. This is why we are definitely certain that, that part of the market will continue to be the largest part.
The other thing that's important to notice, we have been talking for a long time about replacing or displacing ASIC [indiscernible] I'll talk some more about this later. But with the new capabilities that we are introducing, our new products, we now have opportunities to also displace other types of products, microprocessors, microcontrollers, DSP [indiscernible].
And while we don't have time today to talk about all the products and all the market segments, I will emphasize already that with our 28-nanometer SoC products, the FPGAs that embed ARM processor cores, we are already displacing microcontrollers and microprocessors in a lot of different applications. It's not specific to any one vendor, it's not specific to any one architecture. It really is something that we see fairly broadly that customers, mainly those who use them [ph] in the past in FPGA, microcontroller over microprocessor next to it, going into the next generation of their system, would combined them the SoC product.
We talked a lot about ASIC replacements and from time to time, there's a question as to whether this is really happening. And we thought it would be useful to show you this chart that compares revenue in 2008 and 2012 for the custom logic provider. That includes ASICs and FPGAs. And as you can see here, obviously, Samsung, primarily because of the Apple business, is the largest one now, significantly improved [ph] the business between 2008, 2012.
But if you look at all the others, the FPGA companies made significant progress. Both of [indiscernible] climbed 2 places on this ranking. And also, you can see that most of the other vendors have actually seen their business decline, while we saw the business coming up. So certainly, when we talk about ASIC replacement by FPGA, this is something that has been happening, is happening and will continue to happen.
And the reason for that, and you've seen this in different shapes before, this is based on new numbers from IBS, the fundamentals of the semiconductor industry continue to favor the FPGAs. And the reason for that is the increasing cost of creating a semiconductor device as you move from one process technology to the next. And you can see here what is now commonly accepted as numbers from 65, it's $35 million, all the way up to the cost of [ph] 20 nanometers, you see the development cost grew. And if, as a semiconductor company, you don't want your R&D expenses to exceed 20% of revenue, that means that this R&D investment needs to generate 5x revenue, and you see the numbers in red and there's just not enough numbers that can deliver those products.
And therefore, the fundamental trends in the semiconductor industry are against ASICs, against ASSPs. And for those that can build products that go, there [ph] is a lot of different customers in a lot of different markets like FPGA.
So the economic is favoring FPGAs. The question is whether we will be able to take advantage of that opportunity. And for the answer to this question to be yes, we need to address 3 very significant technologies: Performance, power consumption, cost. And I'll say a few words about each one of these.
Starting with performance, and I have here an example from the telco market, but this is happening in a lot of different markets as well, if you look at what happened at the application level, at the system requirements, those systems go from 10 gigabits per second, 40 to 100, 400 gigs per second. So from one generation to the next, they expect the performance improvement of 2.5x [indiscernible]. At the same time, if you look at what traditionally FPGAs would get by going from one process technology to the next, typically, we would get a linear improvement, 20% from one generation to the other. So obviously, that creates a [indiscernible].
As performance goes up, so does power. But many of the applications and systems that we service have a power budget that cannot be [indiscernible]. Because of cooling constraints or thermal management constraints, there's a ceiling that can't be [indiscernible]. So something needs to happen on front as well. And last but not least, as you go from one process node to the next process node, the cost of any transistors going up, but at the same time, customers expect their costs come down.
So something fundamentally needs to change so we can change those slopes and meet the requirements for performance, power consumption, cost. And that is where our partnership with Intel on 14-nanometer tri-gate or FinFET comes into place.
And this, I'm talking about Intel, instead of us curating [ph] Altera charts with the [ph] Intel technology. I'm just going to use a number of Intel charts. So what you see on your left, this is a proxy for transistor performance. You can see the contributors to performance over time.
You can see here in blue that the classical implementation of a transistor, that performance over time not only that it's not going up but it's actually coming down. And because of that, there's always a need to introduce new capabilities and new technologies, such as brain strained silicon, high-K metal gate, those are some examples of the past. But you can see, starting at the 22-nanometer generation, the most significant contributor to the performance, FinFET, and again, what the industry calls FinFET, Intel is calling tri-gate, so use those terms interchangeably.
FinFET doesn't only help with performance. It helps with power consumption as well. Therefore, you can see on the other charts that by using this capability, going from one process node to the next, you don't just get higher performance, but we can actually reduce the power per function.
There's one thing that you make [ph] about Intel's approach to FinFET. When Intel first moved to FinFET in the 22-nanometer generation, they implemented FinFET, but also implemented a full [ph] 22-nanometer processor. [indiscernible] they never missed a generation in terms of die size reduction, therefore, cost reduction.
Other players in the industry used to tackle one problem at a time. So as they move to FinFET, they essentially stay [ph] in the same process [indiscernible]. So you do get the improvements in performance and power consumption, but you miss a generation in die size reduction, therefore, cost reduction.
The partnership that we have with Intel allows us to address performance, power consumption and cost all at the same time. This is what allows us to really penetrate those new opportunities that you'll hear about from Scott [indiscernible] and continue to replace [indiscernible].
The first Altera product that will take advantage of this new capability is Stratix 10, our next-generation high-end FPGA. And you can already see that, as I mentioned earlier, if [ph] traditionally, we would go from one process node to the next, so one FPGA generation to the next, and get a performance improvement of 20%, now we are talking about 2x improvement in performance for the first time in history, bringing FPGAs to [indiscernible].
There are 2 contributors to this performance improvement. One is the Intel 14-nanometer process itself, but also in Stratix 10, we are going to produce 2 high-end architecture that we have developed. The combination of these 2 is what is going to give us the 2x performance [indiscernible].
But Stratix 10 is more than just performance [indiscernible]. I mentioned earlier how important power is. And therefore, one of the important things to look at is not just performance and power, but the ratio between the 2 [ph], often referred to as the [indiscernible]. So efficiencies, performance per power consumption or performance per watt, in Stratix 10, we will get up to 3x improvement in power [indiscernible].
We also announced a couple of weeks ago that on the Stratix 10 FPGAs, we will include [ph] or embed a quad [ph] 64-bit ARM processor subsystem. There is [indiscernible] one of the reasons, as I mentioned earlier, we are very pleased with the traction of the SoC products already, the low cost and midrange, so we decided to expand the offering to high end as well.
And maybe I should mention here one industry analyst who was quoted in Forbes when we made this announcement. He said, "The combination of ARM's most advanced technology on Intel's most advanced technology [indiscernible] that will be very hard to [indiscernible]." So obviously, we are very excited about the ability to combine that ARM capability on the Intel core [ph].
We also had some new DSP capabilities in Stratix 10, as I mentioned earlier. The opportunities [indiscernible] DSP processors. We also will have the highest density, and we can do that because, as I mentioned earlier, the 40-nanometer [ph] process, they do 14-nanometer process and therefore, the transistors are very small, we can get very high density.
In fact, we believe that with the monolithic Stratix 10 FPGA, we'll have higher density than any [indiscernible]. And to complete this, we will also introduce and we'll announce later some interesting products that combine FPGAs, other technologies using 3D capabilities and we'll do this again through a combination of some things we are implementing with Stratix 10 and the access that we have to Intel's packaging technology as part of the partnership [ph] we have with [indiscernible].
Two years ago, we started to talk about the concept of silicon convergence. The fact that you need to embed on the same device, logic fabric, processor, DSP capabilities, Stratix 10 really takes the concept of silicon convergence [indiscernible]. But it's not just about the high end. Even though, as I mentioned earlier, the high end is and will continue to be the largest part of the market, there are opportunities in the midrange, there are also opportunities in the low cost. And with our generation 10, we are going to address all of that.
Arria 10 is our new midrange product family. It is based on TSMC's 20-nanometer process technology. Software for Arria 10 is already publicly available. Many customers are using it. They already paid [indiscernible] the first device. We will start to ship customer samples in the first quarter of [indiscernible].
And we also -- we'll introduce a low-cost family of MAX 10, which combines the best of FPGAs and CPLDs, and that one will be using TSMC's 55-nanometer [ph] embedded [indiscernible].
So by introducing this complete portfolio, we'll be able to take share inside the FPGA world, but also even more importantly, continue to replace other [indiscernible].
I mentioned power a number of times already and you'll hear it from Scott and Jeff as well. There are a number of ways to address power and power consumption. Obviously, the usage of the tri-gate or FinFET process is one of them. There are many things that we are doing in architecture and implementation.
Another very significant step that we took a few months ago was the acquisition of Enpirion power management company. Enpirion is mastering 3 various significant technologies: Switching, packaging, magnetics. This allows them to provide the most integrated power management solution in [indiscernible]. You can see at the bottom right of this foil a comparison, it's an apples-to-apples comparison from a functionality perspective between the Enpirion solution [indiscernible].
That integration and that efficiency provides a number of significant customer benefits. Because of the high efficiency, it contributes to lower system power consumption. Because of the small form factor, obviously, it enables a small factor implementation, but also contribute reliability. And because it is so integrated, it's very [indiscernible], so a number of benefits to our customers.
It also provides us with some very significant benefits. First of all, it allows us to capture a bigger part of the bill of materials of the system and therefore, accelerate our growth. It is not uncommon that a customer would spend between $0.15 and $0.20 for every $1 of FPGA revenue on the power management. So a great way for us to capture a bigger part of that bill of material and we do it without any negative impact on profitability. In fact, the Enpirion business model is very similar to the business model of our core [ph] business.
And what is perhaps even more important going forward, when we design both the FPGA and the power solution, we can also define the communication protocol between them. And that allows us to really provide system-level power optimization so customers can really benefit, combine the 2 devices [indiscernible].
So a very significant strategic acquisition for us and we already [indiscernible] significant traction, Enpirion did quite well before we acquire them. Now that they are part of Altera, we [indiscernible] acceleration adoption [indiscernible].
In addition to Enpirion which will support all our products, there's another thing that supports all our products, and that's the Quartus II design software. There have been also some rumors recently in the industry about what's happening in the software world. So I thought I would share with you a couple of benchmark [indiscernible].
First of all, when we look at the performance of FPGAs, always a function of the hardware and the software. And therefore, one of the important benchmarks of software is what is the performance that you get from your high-end FPGAs. And as you can see here on this foil, Altera's Quartus II is improved. We had a leadership. We lost that leadership for a short while when our competitor introduced a software package. But we regained that leadership and today, if you run a variety of benchmarks representing a lot of different designs and applications, we typically have 20% higher performance of our high-end FPGAs compared to the competitors' high end.
Another very important benchmarks for software, compile time. So compile time directly affects the productivity of the engineering teams of the customer, while in the previous foil that showed performance, higher is better. When we look at compile times, lower is better. And once again, you can see the same pattern. We had the leadership, we lost it for a short period of time, but we regained it and now typically, customers get to those to 30% better productivity [indiscernible] of lower compile times with our FPGAs.
And what's important to emphasize here, this is the same Quartus II package that supports all our product families that customers are very used to using. So there was no disruption to the customer base, they just get it [indiscernible] by using [indiscernible].
When we look historically at FPGAs, they were used by hardware engineers who would program or designed with the FPGAs using what's called HDL, or hardware description [ph] language, such as Verilog [indiscernible]. That was great because by doing it, they got access [indiscernible] in terms of, again, performance [indiscernible]. But the problem is that for every hardware engineer in the world, there are 10 software [ph] developers. And software developers prefer ease-of-use over other efficiencies if they can use can't use languages like [indiscernible] language to take advantage [indiscernible].
So there's this dilemma between -- I'm a software developer. I would like to get access to the most optimized hardware, but there's no way I'm going to let them [ph] Verilog or HDL. So what is the way to bridge between those? The answer to this is OpenCL. OpenCL is a new emerging standard that is not specific to FPGAs. [indiscernible] it with FPGAs, graphics processors, other microprocessors, but it allows customers or software developers developing OpenCL-based language, and then by companies like Altera providing the compiler and libraries, they can take the [indiscernible] and convert it into a hardware optimizing [indiscernible].
Altera is the first FPGA company to support OpenCL as an OpenCL SDK [indiscernible] by many of our customers. And Jeff, in his presentation, will provide some specific examples of markets and segments that take advantage of [indiscernible].
So to summarize this part, we now have access to the Intel 40-nanometer tri-gate process. We are the only major FPGA company that has access to this technology. So that is obviously a significant sustainable advantage for us new. We have a new architecture in Stratix 10, best software [ph] Quartus II. And through the acquisition of Enpirion and our commitment to OpenCL, we are creating even more opportunities to penetrate new markets and accelerate [indiscernible].
I will now hand it over to Scott.
Scott A. Bibaud
Good afternoon. I'm Scott Bibaud. I run our Communications Division here at Altera, and I want to talk to you about a couple of my businesses that will illustrate what we see as the big opportunity for FPGAs going ahead in the future.
The first is wireless. So as many of you know, we're in the beginning of an LTE kind of sustained investment cycle for communications infrastructure to support the rollout of LTE. And some great things have happened to show that, that's successful. In the past year, we've seen 114 new networks to roll out worldwide. We've seen growth of 64% in the amount of equipment spending that's going on for infrastructure equipment for LTE to a level of $113 billion that was spent in the last year. And we've seen that the amount of subscribers for LTE have doubled, more than doubled, and those are all good things.
But the more important thing is that we're set up for this long-term growth, and the drivers are all in place. First of all, although we've grown the number of subscribers, we're only at about 2% of the total worldwide subscribers for cellular using LTE today. And as we all know, one of the big drivers, the biggest driver and will probably remain the biggest driver for a long time for LTE growth is smartphones and tablet devices that encourage people to do video downloads and some of the more multimedia-intensive applications that require the high bandwidth LTE that delivers. So we see that in this market, we're at the very beginning, the early innings of a very long sustained investment cycle for wireless in the world.
And to illustrate that a little bit more, I wanted to show you about -- how a hypothetical wireless network would get built out and how that opportunity -- how it translates into the opportunity for FPGAs.
This is showing a land area that might look very much like the Bay Area, but it's a hypothetical land area with a conventional macrocell layout. So macrocells are placed all over to try to provide broad coverages to the populated areas and this is the way the LTE has been rolling out throughout the world.
But it's -- the LTE that's rolled out to these macros has been a little more complicated. We're supporting multiple versions of the LTE standard in the world. They also have support backward compatibility to a number of other standards. And the frequency bands that are made available all around in different geographies are quite diverse. So it's very complicated environment. But this is the way that traditionally, for the last 15 years, FPGAs have taken advantage of the wireless market.
Macrocells use a lot of FPGAs. And they span -- in particular, they use our high-end FPGAs for switching, traffic management and also future proofing. As the standards evolve, they can make upgrades to the FPGAs, both in the field and by introducing new [indiscernible] that can be brought up with these capabilities quickly. So the start of macrocell is rolling out. It's very good for FPGAs, but it doesn't tell nearly the whole story because these days, the environment for the cellular buildout is much more complicated.
One of the things that we see happening on top of the traditional macro buildout is small cells starting to be rolled out into the market. 2 types of small cells that we see coming out most frequently, first of all, microcells are being deployed today that will provide additional coverage in areas that have difficulty providing coverage. A good example would be here in the city where buildings prevent macrocells from covering wide areas and there'd be microcells that are deployed to the cover other areas.
Microcells are also an excellent FPGA opportunity. They traditionally use the same architecture as a macrocell, just smaller. And so typically, our FPGAs would be slightly smaller. But it's very important to the equipment providers that these are the same architecture to manage the network.
Picocells are also being talked about. They're starting to be deployed out, but it's early days for picocells today. And those will be used more for localized capacity, so like inside buildings or inside a sports stadium. In the future, we think picocells will be served by ASSPs because it's more -- much more cost-sensitive environment. But really today, most of the picocells that have been offered by the bigger cellular equipment OEMs are actually using Altera FPGAs.
So we see a great opportunity for small cells, sitting on top of our traditional market of macrocells, so it is an incremental opportunity.
Another thing we're starting to see roll out around the world is something that's called Cloud-RAN or centralized RAN. RAN is Radio Access Network. And what about Cloud-RAN is, is when the base stations that are typically -- the baseband that's typically in a base station in centralized at [ph] central server rooms and connected over long links to the remote radio heads.
That's definitely another very fertile field for FPGAs. All of the C-RAN implementations that have been going on around the world are using FPGAs. We just did an announcement last month about a strategic partnership between Altera and China Mobile on their Cloud-RAN implementation they're doing in their network.
And finally, all of the diverse elements of this heterogeneous cellular network has to all be connected back into the central core network to be able to deliver all that traffic around. And all of these different message of backhaul are also big users of FPGAs. And they use it for capacity, the different topologies that you need, the different technologies that you need to do the back hauling and for switching.
So you can see that in the more complicated cellular rollout that's happening today, there's a wide variety of opportunities for FPGAs and whereas in the past, it was an important market for us because we had all the macrocells. In this case, it's an even bigger opportunity for us going ahead in the future.
Now all of the pieces here that have to be pulled together and work seamlessly with each other so that when you and I try to download a video and tweet and text while we're 2012 driving, are things that will be -- that have to be supported by our customers, the big equipment OEMs. But the equipment OEMs have got to be dilemma. This new environment of cellular networks is bringing out far more technology than it ever has in the past. And each piece of that technology has got a lot more variables than it has in the past.
So just as a few examples, in baseband and radio, what we see is 2 different versions of LTE rolling out around the world, each of them in different regions and with different carriers having different type of backward compatibility they need to support. I talked about the different types of cell sites that all require different output power levels and different form factors that needed to be supported, and each of those drives for different equipment that need to be developed by our OEM customers.
On the radio side, I'd say one of the more complex piece, and originally when radios were developed for GSM, they had to be about -- they had to support 120 megahertz of bandwidth. But now we can see that out into -- sorry, 200 kilohertz of bandwidth, now we can see it out into LTE Advanced, we're going to have to support up to 100 megahertz of total bandwidth on a single radio head. And we have to support it in many, many different bands all around the world. Each geography being different bands. For some of carriers they have different regions that require different bands. And also some of the bands are contiguous, meaning you can group them together, some noncontiguous, so you can see this is very complicated. One of my big wireless customers recently told me that he can see the day when his team needs to develop a new cellular radio product every working day of the year, that's 250 different products per year.
Likewise in backhaul, huge amounts of variety and what's needed out there for backhaul depending on the type of media that you have available to send the data back into the network, and so this also provides huge amounts of variety of equipment that the OEM have to develop. But the challenge is, as you guys know, looking at my customers, each of them got a lot of pressure on their profitability. So they're trying to cut their R&D budgets, they're trying to improve their gross margin, but they're working in an extremely competitive environment that demands that they hit really cost-sensitive price points, so they have to development individual equipment to meet these items.
And I would like to illustrate just a little bit more, so I'll talk about that radio example I just gave, the 250 radio products a year. I've put together a little illustration here to help you understand what the equipment providers are facing. So a lot of different technologies, they have to develop different products for each of them to try to get to the most cost competitive radio product for each individual market and region that have requirements.
Now ASSP providers have tried to make products to meet the requirements of all these different variations in the past. But so far, we've never seen any that have been really successful doing that. The reason is in ASSP would have to have a huge amount of features supported in that. It would make for extremely large devices that are very hard to verify. And so far, it doesn't seem like any have been economically viable. We've seen a number of players in this industry leaving in the past few years and with others would be -- that they are lowering their investment rates. So really, I don't think it's a market for ASSPs.
What we have seen and you guys have seen us report on this, is that for some of the opportunities in the markets that are big enough, some of our customers have been able to make ASICs be able to address that. But it's important to understand there's only a few players out there who have enough market share and enough R&D resources to be able to develop these ASIC.
Because they have to develop an ASIC for just a few pieces of the market, for the rest of them they have use other solutions to my products to fill those levels. So the other thing that's happening is, as the cost to develop ASIC increases, as Danny was showing on his slide, that little bubble I have a square around will get larger and larger. It has to get larger and larger to justify the additional cost of the ASIC. But the market is actually moving in the other direction where we've seen a much bigger number of smaller circles that have to be covered.
So the right and best long-term way for our customers to develop solutions for this market is by use of FPGAs because they can quickly and easily program both the hardware and the software on the FPGA and they can optimize solution for each of the technology variance.
So today, some people continue to use ASICs. A lot of people -- most people use FPGAs. I think in the long-term, we're going to start that moving more and more towards everyone using FPGAs exclusively. So right here, you can start to get a feel for why I think this will be -- that FPGAs are increasingly essential in the wireless [indiscernible].
As Danny said, this question remains about, okay, FPGAs are essential, but can they meet the need? In the cellular market there's always been a certain performance level that's been required for each of the different new cellular standards that's coming out from kind of 250 megahertz level that was required for 2G, up to today in LTE need about 500 megahertz of performance out of your FPGA to be making products for that.
And in 20-nanometer and prior planer technology, we see the limit of the capabilities of the FPGA getting to about 500 megahertz. But my customers are already starting to talk about 5G cellular and about LTE Advanced and about those 100 megahertz wide radio bandwidth. And we believe that the only technology that will be available in the market meet the needs of that as our Stratix 10 14-nanometer technology that's under development right now.
Just last week, I was traveling and I met with one of my customers who is complaining that LTE Advanced is coming very quickly. He thinks it will start to be fielded within 2 years, and he said there's a huge amount of new technology in there and nobody knows what combinations of that technology will need to be rolled out in the market. So I said the one thing I do know is it's going to require a lot more baseband processing that we're going to have to implement on FPGAs. So you can see as LTE Advanced comes, it will be -- it will need the capabilities that our Stratix 10 products will bring to the market.
So to summarize this wireless area, today, Altera is #1 in wireless radio, #1 in backhaul. By next year, we'll be #1 in basebands as well, so we have a very strong market position here. The market is expanding and will continue to expand for many years. And we believe that FPGAs are bringing very strong flexibility to the market that's going to be necessary as ASICs and ASSPs become less viable. We also have very strong traction in some of the new and emerging technologies. So I think it's an excellent picture of a strong position in a market that has room to grow in the future.
With that, I want to talk about another market that is underneath my division, our optical transport networking space. I know optical transport networking is a lot less familiar to many of you than wireless. And when we're talking about optical transport, we're talking about core of the network technologies, these things that go into Metro, core and long-haul networks. Generally, it's -- the service providers don't need a lot of this equipment, they need small amounts of it, but it's very, very high value. And so the volumes are low, the value is high.
But you can see that as this transition happens to optical, we are forecasting very strong growth going out into the next 5 years, 12%. Infonetics recently released the results on what growth there was for the first half of 2013, it grew at 23%. So it verifies that this is a good growth opportunity as a market as a whole. And what's interesting about this market is, for quite some time now, a big portion of it has been served by FPGAs. So in 2008, about 30% of the market was FPGAs. Altera was #1 in FPGAs. But today, 45% of the market is FPGA, and Altera is #1 overall semiconductor provider into this space. By 2018, we're forecasting that it will be 70% FPGA and we'll still be #1. And I want to give you an idea about why that would be.
So if you look at ASSPs, when -- this market is interesting, each equipment OEM that tries to make a box, has to put lots of standard technology in it. They also got lots of proprietary features, so hundreds and hundreds of features and each one of them wants to put some proprietary ones in there and implement some of the standard ones to make their product unique in the market. There's 10 different OEMs, each of them is making these different mixes of proprietary and standard [indiscernible].
For an ASSP to make one chip that meets the needs of, let's say, the top 10 OEMs, it would have to implement literally hundreds of different features. And for each box that was developed, most of those features wouldn't be used. So it would be filled with dark silicon. Dark silicon still burns power and still has cost, and so it's a very inefficient architecture. And for that reason, and also to develop that huge piece of silicon that has all those features, very expensive, and then verify it is even more expensive. So what we've seen in the past few years is that more and more players that were in this space making ASSPs have either excited or have cut their investment dramatically there. It's really becoming purely an FPGA space.
Now when I -- when one of my customers decides that they want to make a box for OTN based on my PGAs, they take in FPGA, they only put in the features they want, they only put in the proprietary features they want, and they can make something that's really optimized to just what they need. So it gives them better cost, better power. And we think we're very competitive with the best-in-class solutions out there today in terms of cost and power. But we also can give them fast time-to-market because they can change the features at the last minute and support evolving standards.
So you can see the picture here we've been talking for many years about the tipping point when certain markets reach the tipping point. This market is clearly a case where it's more advantageous for customers to use FPGAs than ASSPs. FPGAs are cost and power competitive with ASSPs, and so in the future, it will become more and more [indiscernible] market.
Finally, when people make products in this market, there's many reasons that they'd like to implement everything into a single chip, it's lower cost, lower power, higher reliability. In today's solutions out there in the market, 100G solutions are going into production with single chips. 200G are well into development at this point and we're starting to see people working on 400G line cards that are trying to use a single chip. But what we know is people really would like to get a terabit solution on a single chip. And we think there's only one platform that's under development today that will able to deliver on that promise and that's our Stratix 10 because of the performance and power that it can meet. But with the density that we're delivering on that chip, it also will be able to meet all of the feature requirements that people need in that market. So for OTN, we think that our 14-nanometer product is really providing unmatched density performance for this market.
In 2010, Altera acquired a company called Avalon Micro, because we saw that the OTN was an important market for us and we wanted to acquire the best-in-class IP that was being developed out in the market for this space. That turned out to be an extremely successful strategy for us. This year, we doubled down and we bought another company that was developing great OTN IP top of Altera silicon called TPACK. Now those 2 teams are put together and we have developed what we believe is the most mature and widely adopted 100G OTN IP in the industry. And as proof of that, we have 7 of the top 10 optical OEMs in the world that are using our OTN IP in their development, and 9 of the top 10 optical OEMs are using our Stratix V high-end FPGAs in their optical platforms themselves.
So overall, I think this market means an excellent picture, it's a growing market, is tipping from ASSPs to FPGAs. Altera is the only company that has really gone out and established itself as a true market leader in this space for technology, and so I think it positions us very well for continued growth in this space.
So with that, I'd like to turn it over to Jeff Waters who'll talk about computer.
Jeffrey W. Waters
Great. Thank you very much, Scott. So my name is Jeff Waters, I run the military, industrial and computing division here at Altera. And today, I want to talk to you about what is the fastest-growing business that we have at Altera over $100 million, that's our computer and storage segment.
We're very excited about the underlying fundamentals of this opportunity, primarily because it's based on what is a pretty undeniable growth story, and that is to build out an expansion of data centers and servers to help support all the big data and cloud computing requirements in the world.
When you think about this type of an opportunity, what helps exemplifies it as well is the urgency to try and tackle what is a very big challenge around energy consumption within these data centers. Now FPGAs we think offer a pretty good solution to that, and I'll talk about that here over the next couple of slides.
But if you look at where FPGAs are, it's very early innings. No it's very early kind of in their infancy of adoption, and it represents about a $200 million PLD opportunity for us. With the introduction of OpenCL, with the introduction of our Enpirion power management products and with Stratix 10 on the near horizon, we really think this is a market that has the potential to grow to be a $1 billion opportunity before the end of this decade. And given that Altera is #1 in terms of programmable logic and widening that lead, we think this really has the potential to be the next great growth story at Altera.
You're probably familiar with why data centers are getting built out just as a quick run through of some statistics. 1/3 of every man, woman and child on the planet right now is on the Internet. That Internet traffic has created the need for over 3 million data centers, over 50 million servers, spend is over $105 billion a year just going into those data centers, growing at about a 15% annual clip. And of that spend, about 10% of it is going into CPUs alone. So a very big and growing opportunity.
Now for many of you when you think about data centers, you might think it's dominated by kind of a household name data centers like Facebook and Google and Amazon. The reality of it is that those big marquee data centers represent a very, very small percentage of the overall data center market. The market is made predominantly of a very diverse and wide array of application, a very specialized applications and algorithms for data centers. Everything from oil and gas exploration, to online gaming, to the financial industry, to telecom and also to our own federal government here in the U.S., has gone from 400 data centers in 1998 to over 2,000 here in 2013.
So for those of you that our students have the programmable logic industry, you can probably recognize that when you talk about a fast-growing market with a very diverse set of optimized needs, that presents a very good opportunity for FPGAs.
Now for data center challenges, as I mentioned, there is a very significant energy challenge there. Right now, the world's data centers consume the equivalent of 40 nuclear power plants. One of the things that are really exacerbates this problem is that about 90% of the energy that supply to these data center is wasted. Now as for a variety of different inefficiencies, but from a server perspective, what really stands at the heart of those inefficiencies is the traditional CPU architecture.
For big data and cloud types of applications, a lot of things come into play that create bottlenecks into that CPU. So although you have a very strong and powerful CPU engine that's also assuming a lot of power, it is much more often than not starve for data. Starved over the I/O. Starved going out and coming back from the memory. This basically creates a very low performance.
Now what the compound does, the CPU itself is not architected well to run the very parallel life it needs of most cloud computing and big data types of applications. So you end up with a very inefficient processing and this just adds to more wasted power. So combined, you end up with a very inefficient platform meant to transfer these application. Now what does has driven is a strong push by many to go out and create more of a customized approach for their specific data center and server needs.
Customers like Google are making their own servers and data centers. Other companies like Facebook with their Open Compute platform, IBM with their OpenPower platform and HP with their Moonshot program, are trying to create an infrastructure and ecosystem to make it very easy for even a small player with a very specific data center need to build the data center that can be run efficiently and effectively. So this is a big growing trend in the market, it also favors the usage of FPGAs.
FPGAs, as I said, create a very good solution to the architectural challenges of the traditional server. And it really what comes down to is the parallelized nature of FPGAs. Because of that massively parallel architecture, FPGAs have orders of magnitude gratitude more computational units. And these compositional units can be pipelines in a customized way, they can be configured in a very customized way to do exactly the calculation that you need as quickly as you need it and as efficiently as you need it. This is something that's very unique to programmable logic.
One of the things that I'd like to talk about here is to show you some of these performance per watt benchmarks, talked about them [indiscernible] is one of the key algorithms. That if you look at the algorithms that underlie most search engines, it really comes down to 2 key ones, just both the filtering and also indexing algorithm. This takes all the world's Internet content and makes it easily searchable so that you type a keyword into Google, it comes back to you quickly with a search, a result.
If you take a look at some of the benchmarking that we've done, let's first look at the Intel Xeon platform. This consumes about 130 watts of power and it does the equivalent of about 2 billion transactions per second for these search algorithms, with the performance-to-power ratio of about 16 million transactions per second per watt.
Now if you look at GPU from NVIDIA, you see its power consumption is much higher at about 215 watts. Its performance is also much higher, it does about 3.2 billion transactions per second. But from a performance-to-power ratio, it gives you a very similar result to what you get with CPUs. If you look at the Altera FPGA, not only does it consume much less power than both the GPU and CPU but from a performance perspective, it delivers better raw performance for these algorithms giving you a dramatically better performance-to-power ratio.
Another key algorithm is image scaling. So if you go on to Google, you go into Google images, you find something, you click on it, the very small image it balloons into a very large, high-resolution image. For Google, they don't want to have to store an image of a certain resolution to meet all the various needs of the displays that you have in your life. So they store one and they use a scaling algorithm to very quickly scale it up or down depending on what kind of a device that you use. If you look at the performance across the CPUs and the GPUs as compared to the FPGAs, you see the FPGAs also dramatically outperformed both of those architectures as well.
And then last but not least, financial modeling. At the heart of a lot of financial modeling is Monte Carlo simulations. If you look at Monte Carlo performance, once again, with an FPGA, you get much lower power consumption, you get much better raw performance, almost 3x better raw performance. But then on a performance-to-power basis, a dramatically heightened level of performance at both.
One customer is actually talking publicly about the use of FPGAs in the data center space as IBM. This is our IBM that sees a 1,000 data warehouse appliance. Effectively, what they've done is taken an FPGA-based system and stuck it right at the storage [indiscernible]. So instead of having to go and retrieve data from storage, bring it back into the CPU, process it, put a back out into storage, they do a lot of that processing locally right at the desk. It's not only filters out a lot of data for the CPU to process, but it also handles a lot of it locally. And as IBM points out, using FPGAs we're able to remove a lot of the I/O bottlenecks that exist within the traditional systems and able to achieve a much better energy efficiency with these platforms.
Now talking about kind of some future trends that we have within Altera that are going to impact data center. I'm out on the road talking with a lot of our customers. Just last week, I was meeting with some of our storage customers that sell into the data center space. They tell us a very similar story that I hear from a lot of other customers. And that is that in order for them to really meet the evolving needs of their customer base, they need FPGAs. They need that reprogrammability. And actually, a lot of the value and differentiation they bring for their customers is put into that FPGA.
Now what they have happening though on a daily basis is that bar continues to get raised. But the power consumption needs to get reduced, the performance of the device needs to be able to reduce, the size of the devices that they create need to get reduced. This really puts a lot of pressure on them. And if you look at this trend, though, this really speaks to the strategy that underlies we were doing at Altera. If you look at our Stratix 10 family, our decision to go to 14-nanometer with Intel allows us to get the much higher density. For that same storage customer, we're going to have the ability to take what was 3 FPGAs and integrate it into one larger higher density FPGA.
We're also going to be able to achieve a much better power performance with that move as well. With our acquisition of Enpirion, not only does this help with the power efficiency story, but it also helps us condensed that form factor in the device and allows them to pack more into what they do as a provider of storage by consuming less of that board space with power management. So this really is a great trend. The needs of this space fit in very nicely with what we're doing as a company.
Now you might ask yourself, "Given that FPGAs are such a superior hardware solution, why aren't there being adopted more quickly? Why is it only a $200 million market for FPGAs?" Well, Danny kind of hinted at the answer to that, it really gets down to the programming language for FPGAs. FPGAs are very well suited for hardware engineers. But for software engineer, they're really not accessible because of the programming language.
With that in mind, 4 years ago, we began our investment of the OpenCL, started off as a development program. We've been working hard at that, working very closely with a lot of lead customers out of the data center space, lot of Tier 1 data center and server suppliers. And that culminated in 2013 and us releasing a fully validated OpenCL tool suite for our Altera FPGAs. It has gone through and passed all of the rigorous testing that is put out by that OpenCL standard. For the trade show this week, where were demonstrating both HP and IBM, they have their own people there demonstrating FPGAs being used with OpenCL within their server systems and showing dramatically superior performance with the combination of their CPUs and with our Altera FPGAs.
Given this early release here of OpenCL, we're already seeing a lot of great uptake and enthusiasm for OpenCL. Today, just within the first year of release, 1 out of every 5 of the opportunities that we have within the data center space is leveraging OpenCL. With the rate and the progress that we're seeing, we expect that to be about 1 out of every 2 by 2016. Right now, 4 out of the top 6 largest server suppliers are using OpenCL in their FPGA development, and 3 out of the top 5 largest investment banks are also using OpenCL with their Altera FPGA development. So a lot of great progress.
Now to put kind of the competitive landscape in context here, I want to give you a visual representation of that. Along the y-axis, you see development productivity, so that is how easy is it to develop with a specific platform. Along the x-axis is efficiency, and that's performance per watt, a lot of metrics I've been showing over the last couple of slides
Traditionally, if you're designing a server, this is really what you've had to choose from. On the far upper left-hand corner is multicore CPUs. The nice thing about them, that you have a great well-known infrastructure, you have legions of engineers that know how to develop in C++, right on to those multicore CPUs. The downside, as we've discussed, is the hardware efficiency. GPUs NVIDIA released their CUDA software tools, very C-like in nature, very easy to use. And so they've helped with that development productivity. But as we described, although they can offer some performance gains on CPUs, their power-to-performance ratio is not much better than what you get with a typical [indiscernible].
Now FPGAs, the bottom right-hand corner, a superior hardware solution. But as we've discussed, for the legions of software engineers and algorithm designers, not so easy to develop with. This is really where we see the promise of OpenCL in combination with FPGAs and in working in a heterogeneous computing environment, along with multicore CPUs. We think this is going to be the door opener because it's going to open access to our computing acceleration to a whole new group of engineers.
Right now, the traditional computing methods for servers represents about a $10 billion market. Our feeling and our analysis shows that we think that by the end of this decade, we could have a $1 billion FPGA opportunity, a lot of it really being accelerated by the introduction of OpenCL, along with our Stratix 10 platform.
So in summary, the data center market, driving a lot of server designs, driving the need for server customization to meet all the needs of big data, cloud computing is really getting a great opportunity for us at Altera. It fits nicely with our strategy, we've done a lot of things to augment our strategy to fit with it even better, our Stratix 10 development, our Enpirion power management products, all the investment we've been putting into OpenCL, we think really creates a good potential for us to capitalize on this market.
We're the #1 guys in terms of programmable logic. 5 of the top 6 service guys are using our Altera FPGAs in their systems and developing with them. And although it's a $200 million market today, we do have a line of sight for this being a $1 billion market by the end of the decade and once again, be the next great growth story for Altera going forward.
So with that, I'd like to turn it over to Ron Pasek to review our financial guidance.
Ronald J. Pasek
Thank you, Jeff. By now, you're probably figuring out my section is not in your package, it's not. But we will have copies for you when you leave today on your way out.
So I'm going to talk about 3 things. Number one, I'm going to go through the plan for next year. I'm going to reiterate the change to our business model, which I talked about on our conference call a couple of weeks back. And then at the end, I'm going to, again, reiterate our philosophy and return of cash to shareholders.
Okay. So for FY '14, we see gross margin at 69% plus or minus 1 point. It's actually going to be a slight increase from this year, where we'll probably finish [indiscernible]. By the way, on gross margin, as you know, through the year, it is end market mix dependent, so very [indiscernible].
R&D should be about $433 million. That is about a 10% increase from where we'll finish this year. Large increase [indiscernible] wafers next year and as you know we're annualizing acquisition of TPACK and Enpirion.
SG&A actually is going to be down next year by about 1%. I'll show you a bridge [indiscernible]. We did, as you know, go out and get some debt. We got a net billion dollars in debt with buy backs throughout the year in FY '15. That is going to increase our finance expectancy up $13 million interest expense.
Tax rate, 12% to 13%. This is assuming for now there will be no R&D tax credit, perhaps in [indiscernible] there will be. And again, diluted share count, I'm just giving you kind of where we're going to end this year, I'm not going to predict where that will be next year but obviously, they'll be shares outstanding [indiscernible] will, of course, [indiscernible] buy back opportunistically.
The last piece is capital expense for the [indiscernible]. Just to give you -- that's $40 million. A little history on OpEx, so coming off of 2010, we're very clear that we're going to grow OpEx, particularly R&D, faster than revenue for 2 to 3 years. So we have slowed the OpEx growth. In fact, this year, 2014, 5%. Again, some of that -- 2% of that is the acquisition of [indiscernible].
But we wanted to book a plan where we really were quite sure that we could grow revenue faster than OpEx because we want to really turn that around. Give the investment we needed, particularly in R&D, and now it's time to build expansion and operating margin.
So the bridge from 2013 to 2014 for OpEx is really quite simple. You just [indiscernible] 2 large pieces there that increase next year. One again is the annualization of acquisitions and the other piece is a big chunk, $32 million, in masks and [indiscernible]. Really, all we're spending money [indiscernible]. There is no net new hiring planned for 2014.
So on to our business models. So I've chatted about this on the call a little bit. I think what makes more sense is to talk about the elements of gross margin, R&D, SG&A in terms of ranges. It's never the case that we're going to hit the exact numbers for each of those given year. But we will try and we will overtime be in the range of 67% to 70% for gross margin, and in the next 2 to 3 years, 19% to 21% [ph] for R&D and 14% to 16% for SG&A.
Some of you know that in 2010, we had a really large year in revenue growth [indiscernible] 63%. And that really is because we have good leverage in the business model. So in 2010, what we had was something that you'll see similar to this where we get the high end of the gross margin range, the low end of R&D and SG&A [indiscernible] the leverage in the business model [indiscernible] and that's what we like to see in our business.
So finally, just a little bit on dividends and buybacks. So if you've heard me for the last couple of years, we've steadily increased the dividend. What I've have said is a really want to see that dividend be a regular and return to shareholders. And as well, we'd like to see opportunistic share repurchase as another [indiscernible] shareholders. [indiscernible]. I've committed now for several months to returning over time 60% to 80% of cash flow from operations to shareholders.
Okay, with that, let me turn the podium back over to John.
John P. Daane
Thank you very much, Ron. And again, just so everybody understands the material from Ron's section will be in the back of the room for the people here today as you leave. Again, the reason that we did not include it with your other materials is that at the time we started the presentation, market was still open, did not have a chance to read it into the record of the webcast to make it available to all of our investors. But again, the material that Ron has presented [indiscernible] in the back of the room.
Three foils to wrap it up, and then we'll start the question-and-answer period. What you've seen from Jeff and Scott today are all of the vertical markets on the right-hand side. Starting at the top with our wireless base band market, which is again the base station content, recognize from Scott's presentation that the complexity of new deployments, both on TD and FTD LTE, as well as the need to have backward compatibility and the fact that different network operators are actually requiring different capabilities in terms of backhaul capability or features added to this is increasing the amount of FPGA content within the systems. Additionally as we develop our newer generation Stratix 10, a higher performance, lower power technology, we expect that in many cases what has been implemented in the ASIC will start transitioning into the FPGA.
On the wireless side, which has generally used our mid-range FPGAs, the complexity of the number of frequencies, the bandwidth, the form factor, just the raw number of boards that need to be developed for all of the different technology being deployed around the world means that you need a flexible interface to the analog section that drives the usage of FPGAs. The increasing frequency of such will drive more FPGA use. And again, as Scott mentioned, there's an opportunity here with our Stratix technology in our next-generation with a much lower cost point to displace a number of ASICs that were developed around the technology.
Telecom is an example of a marketplace where the volume simply is no longer high enough. These applications are sort of in the low and medium volumes space. For ASSP companies, to develop ASSPs [indiscernible] a return on investment that investors would expect.
And so they're simply leaving the space or finding that with our FPGAs combined with our own IP content, we can create ASSP-like solutions on top of the FPGA to replace these products. This is a great example of a marketplace where the raw economics or the increasing cost of design are forcing part of the semiconductor industry out of the market, creating opportunity for us to continue to grow.
And then within the computer space that Scott presented, great opportunity for us. This is a newer market, one of our fastest-growing markets. In fact, our fastest-growing market as an industry as we have much higher performance and the flexibility to be able to run data-specific or data-center specific algorithms, much faster performance, much lower power. The benefit of this market is it's very large, obviously, we're replacing not only GPUs, but CPUs in this market. But unlike our other markets, it actually goes through a box refresh or a system refresh at a much faster rate. You see in many cases in telecommunications, that piece of equipment may last 10 years. In the computer data center world, you see the equipment is replaced every 2 to 3 years, which means they're using not only our high-end very expensive FPGAs, but they're replacing it on a very regular basis.
Now that's half of our business that we presented today. But there are growth opportunities in the other half of the business as well. Within networking, the service provider portion of it, that is networking equipment like routers and switches, are sold to service providers, has very similar volume and business characteristics as a telecom space. That is, it's low and medium volume equipment. It means many of the ASSP providers in the past for this industry are also getting out of supplying solution to that space.
Additionally, what's happening, if you've seen in networking, there are a few companies that have very high market share in this space. And I realize now that if they use ASSPs, they will create a me-too product and allow other companies to get into their space and meet with them on price.
Ultimately, since they have such high market share, if they're to develop their own IP and implement on an FPGA, that may mean that the ASSP provider no longer has a large enough business or market available to them to develop that ASSP. And if the ASSP provider exits the business of supplying the solution, then many of the competitors that could grow up in Asia simply have nothing to base their products on.
So increasingly, the leaders are focusing on developing their own IPs. As an example, in a major cable platform, we won a design which uses our FPGA with IP to replace what is a very large ASSP docket.
The automotive business, new applications for us, where our FPGAs are being used in infotainment and in driver assist or things like car navigation system, cameras, recognizing lane departure and other types of applications, using our Cyclone series products with our integrated CPUs. About 3% of revenues but again, a very fast-growing market.
Another example is in industrial, where we've traditionally replaced the number of older ASICs that were used for creating custom bus interfaces for a lot of industrial manufacturing equipment. But now with our SoC products, that is the microprocessors integrate into our Cyclone series, we're now able to replace the ESPs and the microprocessors and lift our overall average selling price quite substantially because we're integrating $20 of content into our FPGAs. That allows us to grow at a much faster rate than the overall industrial industry.
Similarly, within other, which includes markets like Test & Measurement, Medical broadcast, are also a number of applications where we can displace microprocessors, we could displace ASICs and other technology as well.
For those of you that remember our great growth that we had in 2010 will remember it was not just one market that drove the growth. It was in fact all of our markets that were growing very strongly during that period of time. 2010, all of our markets grew between 55% to 75%. The fact that we have very strong growth opportunities across what is all of our business that we believe that we have the capacity to continue to grow at a 2x rate over the semiconductor industry.
From a product perspective, as Danny presented, we've refreshed the MAX series after many years of not developing a product there. We have a very good Cyclone series with embedded microprocessors. Fits wonderfully for the automotive, industrial and Consumer business, which uses the low-end FPGAs in the light blue.
In the right section, we estimated about 25% in the FPGAs industry. We have just taped out a few weeks ago our first 20-nanometer device, which we expect to begin sampling in the first calendar quarter of next year, 2014. And we released our software for that product, very high performance, very full-featured capability, product line for the broadcast, the enterprise section of the networking and also the wireless radio space.
And then also, as Danny presented, for the high end, we really are a generation ahead of our competition. And this is very important because the high-end represents over 50% of the FPGA industry in terms of revenue. And here, you see that we've developed a new product using a more advanced process technology than our composition will have access to that will allow us to do much higher density products, much lower power, much higher performance, all at a lower cost, which will fundamentally allow us to independently and individually really address the computer and storage medical, military, service provider portion of the networking business, telecom, test and measurement and wireless baseband industries, all of which is predominantly [indiscernible] our high-end products.
Finally, there were 3 messages that I started with: Rising cost of design, rising cost of transistors, fundamentally are driving other types of products, particularly ASSPs and ASICs out of many portions of the semiconductor industry because they no longer -- either the customers can longer afford to develop the ASIC or ASSP companies are not achieving a requisite return on investment for the ASSP development as they move out of the industry we move in, provides a great growth opportunity because again, these 2 industries, ASICs and ASSPs, are many times larger than the overall FPGA industry.
Number two, great growth opportunities in wireless, telecom, networking, automotive, computer space and many other businesses.
And then number three, and this is most important, really designed our product line with the idea to take maximum advantage of replacing ASICs and ASSPs across these end markets.
We've created a position where we're well ahead of the competition in terms of technology. We've got the right features. We've got the right capabilities. With that, we believe we can continue to grow twice as fast as the overall semiconductor industry.
With that, what I'd like to do is invite up today's presenters to the stage with me. We'll have Scott Wylie and Cliff Tong with microphones. They will select people who have questions. If you could read your question into the microphone, we will then answer it. Additionally, what we'll do is just kind of raise your hand if you have a question. They'll find, and then we'll start with whoever they pick.
First question would be just relative to the overall operating model, Ron or John. You've spent about 5 quarters below the long-term operating net model now. At what point do you think you're going to be able to get back to that long-term operating model? And what gives you the confidence that you're going to have enough revenue growth whether it's in the next 2 or 4 or 5 quarters to actually get you there? And maybe talk about the top line trajectory [indiscernible] because I think it's about 16% or so growth in current revenue run rate to get back in the model.
Ronald J. Pasek
Yes, I think there's 2 parts. One is what we can control directly, which is spending envelope. And you're going to see spending for a while. And we're at where we need to be from an investment standpoint both for R&D and SG&A. We used some modest revenue growth numbers going out, and it's not very hard to get there. So I don't think it's unreasonable at all. It will be 2 to 3 years.
John P. Daane
What I'd add to that, Jim, is a few years ago, Ron had made a point that we were going to expand our R&D spending for a few years. And that R&D expansion went into a few areas. Number one, to develop microprocessor solutions, which we're using now across our product line, benefited data by including the microprocessors we now get an average selling price uplift for our products because as we pull in that device that used to be on the board next to our FPGA, we now sell at a much higher price than we were before. And that allows us to grow faster. We used the increased expense to acquire OTN companies. In Scott's area, as he presented, to create solutions to replace ASSPs. We used the increase in spending to develop more FPGA families, for instance, doing 20-nanometer today with TSMC. We're also in development with 14-nanometer with Intel simultaneously. And additionally, we use that money to also buy a power company because we view that as really a strategic opportunity.
We have everything we think we need in order to achieve the growth and really differentiate our product [indiscernible] forward. And so as such, we now believe we can flatten out our expenses for a few years and grow our revenues at a faster rate and, therefore, grow our earnings at a faster rate than expense growth much like we did sort of [indiscernible] late 2003 [ph] [indiscernible] 2010 and '11. Next question?
First, actually, [indiscernible]. First is if I look at nearly the last [indiscernible] top line growth has been [indiscernible] but what gives you the confidence that you [indiscernible] get back going above your rate of R&D [indiscernible]. What's going to be different about the next [indiscernible], that's number one. Number two, what's going to be the catalyst [indiscernible] are also [indiscernible] wireless [indiscernible] number one in a number of different [indiscernible]. So to summarize question, what's [indiscernible] faster than the last [indiscernible]
John P. Daane
Well, I think if you start at a high-level, if you think of 2010, 2011, those were very good spend years for infrastructure, industrial, wireless telecom types of equipment, which, obviously, benefit the programmable logic market. If you forward into 2012 and 2013, what you've seen is the opposite. The spend in semiconductor has gone more to the handset side. The infrastructure side has struggled. So you've seen industrial spending over a period of time go down, worldwide economy went through a slowdown, and wireless equipment deployment slowed down during that period of time, and military spending has also gone done, obviously, which affected the spend within the programmable logic sector. What you've seen, I think, in the last year period of time is we've seen a bottoming and a trend towards possible growth in many of those markets. Industrial for us actually bottomed in the first calendar quarter. We had very strong growth in Q2 and Q3. And if you go back to our guidance, referencing again back to our guidance, not updating it today, but again, a reference, we said we expected that this quarter industrial will be up again this quarter. And if that trend rate were to continue, obviously you get growth year-over-year. Wireless, we really have seen deployments for LTE equipment be slow.
In the United States we've seen AT&T and Verizon deploy. Both have stated that they need to spend some more money on their networks, plus additionally, looking forward, we'll get Sprint and T-Mobile spending. There'll be an additional spend cycle in the U.S., combined with the fact that we now have China Mobile starting to deploy LTE. And China, obviously -- combined with the fact that China Telecom has announced the deployment and China Unicom will deploy after that. We have Japan spending. We have South Korea spending. So I think you're going to see more spend across a number of geographies, such that wireless, given a few years, should be a good area of growth opportunity for us. Telecom, again, video starting to saturate a lot of the metro area requiring upgrades of equipment. Every time there's a new standard, Test and Measurement equipment gets refreshed [indiscernible] strong for us very good. And again, there are a number of new technology areas being deployed. Computer, as Scott mentioned is -- excuse me, Jeff mentioned, is an area of great growth which is a new market for us, new applications. So I can't say within any given quarter what we're going to do because visibility is very low for us even for our customers one quarter out. But we see enough opportunity either from an end market doing better or from our opportunity and what we're doing within that end market that we believe that we can grow well. Now vis-à-vis Xilinx, we have 42% market share of FPGAs Xilinx. So anything that we do in any node above 42% means that we're going to take market share. And obviously, if you look at 65-nanometer, if you look at 40-nanometer, now if you look at 28-nanometer, we believe that we have well north of 50% market share. In fact, well over that in new products today is 40 and 28. So that means we will continue to take market share. There is an aberration going on, which you can see [indiscernible] product numbers between us and our competition as they're going through an obsolescence. Obsolescence is effectively pulling in future years of revenue into the -- today. As that transition happens, you expect a normal market share gains that would come from having a better footprint or position than the newer product area to continue to benefit Altera. We think, longer term, we will continue to take market share. Next question?
John P. Daane
Yes. We haven't seen that big of an impact on our business from sequestration. Big part of the business we do in the military part of our business falls into more the high performance computing or data center side or intelligent side of what they do. That said, we are seeing, with many of our suppliers, many of our customers, that there is a bit of a slowdown on [indiscernible]
[indiscernible] a follow-on [indiscernible].
Jeffrey W. Waters
John P. Daane
[indiscernible] Don't have any expectations for forecast. [indiscernible]
Ambrish Srivastava - BMO Capital Markets U.S.
John, Ambrish from BMO. Maybe you can share with us your perspective, if you compare 3G spend versus 4G? So 4G, roughly is about 25% to 30% already of the total wireless equipment spend. And there is -- it's never a linear line, we all know that. It goes in fits and starts. But it seems even though 4G spending is going on, it does not seem to be helping either you or Xilinx. So the question is what do you think is the underlying reason? A, it could be 3G is falling off faster than 2G [indiscernible]? B, it could be even more rapid ASIC replacement? Or third, and that would be healthy to say that in a [indiscernible] conference, is there less content in 4G [indiscernible]?
John P. Daane
So why don't I grab that one. If you look at the sequentials of our business, in the second calendar quarter, we were supplying into the TD-SCDMA build of China Mobile. Remember this was a 6.1. They had done the 6.0 spend in the fourth calender quarter. First calender quarter was in between. Second calender quarter, we did the 6.1 portion shipments of the material for that build. When you get to Q3, TD-SCDMA is done, so we saw our TD-SCDMA business go down. Ultimately, though, what we did is we replaced that with LTE equipment or LTE shipments going to China Mobile. And our wireless business actually grew in the quarter. It shows you the increasing content that comes with the third-generation of equipment. So we've actually seen some good growth within our overall Wireless business. Now if you go back to refer to 2010 as an example, 2010 we were actually deploying equipment not only into China for 2G and 3G, but also North America was going through continued spend on 3G plus an LTE rollout. So what we're seeing today is not all geographies are kind of hitting on all cylinders today. Given time, you could see the potential for that. Because again, as I mentioned, you have more spend that both Verizon and AT&T need to do for coverage build-out, particularly AT&T in North America, you've got T-Mobile and Sprint, which will have to spend -- spending that will come from multiple carriers in China, you could have India eventually start to spend. You can have some of the developing countries like Brazil, which has a need to spend ahead of the World Cup and Olympics do some spend. And then eventually we may see Europe start to spend money on upgrades to their systems. And if that were to happen, you could have more than just one particular operator or carrier in one region spending at a time, and you'll see a pickup in business. Not sure if that answers it, but again, I think what you're seeing now is, yes, we are getting growth on a sequential basis. But what we're really seeing is 4G spend right now in China only 2G and 3G predominantly have ramped down [indiscernible].
When you talked about the 40-nanometer products that you're manufacturing, can you give us an update of when you expect [indiscernible] sampling products or when you expect them to ramp? And also, when you add the slide with the 20-nanometer area, you didn't seem to actually have anything for a lower-end cycle product. Will you have a refresh of your Cyclone line?
Danny K. Biran
So first of all, on the 40-nanometer Stratix 10, we intend to take out the first device for next year. By the way, we already phased out [indiscernible] chip so [indiscernible] process. And software for Stratix 10 will be available in the first half [indiscernible]. About Arria 10, as we said, software is already available, customers are using it. [indiscernible] the first device already a few weeks ago, and we when to [indiscernible] customers [indiscernible] next year, 2014. We called the next generation low-cost primary Max then because as we said, fresh processor -- process, sorry, so it's a combination [indiscernible] FPGA. By the way, historically, if you look at what happens in a Cyclone space, so low-cost space in general. Typically, every company from time-to-time skips a generation because the benefit that you get on the low-cost by moving from one process technology to the next [indiscernible] what you get for mid-range of high-end because the package is the bigger part of [indiscernible]. So we are very happy with where we are with Cyclone with 28. Most Cyclone V and Cyclone V [indiscernible] don't see a need to refresh Cyclone now, but we definitely want to refresh the new [indiscernible]
On Slide 12, Danny's presentation, there was a chart that you showed of areas scaling from Intel versus a competitor chart, which is TSMC. And then the title says competitors may suffer from no chip areas scaling in future generations, which I think would be a pretty important statement. Could you give some color on that statement? Do you think that this is something that just happens at 16-nanometer or does it go further than that? And related question would be do you think your manufacturing advantage of 14 nanometers, does that -- is that something that's like a stopgap that gets maintained over time? Or do you think that's something that grows over time?
Danny K. Biran
So first of all, to explain this one generation of [indiscernible]. Going to FinFET, very, very technically challenging and going to the next process mode, always very technical challenge. As I mentioned, Intel chose to do both when it went to 20-nanometer. So it went to 22-nanometer and they went to [indiscernible] the same time, and therefore, they never lost a generation in terms of scaling. Though to the next process node then [indiscernible] therefore better cost [indiscernible]. Other foundries have chosen to address one challenge at a time. So going to FinFET but not going to [indiscernible] next-generation process. So if you look at [indiscernible] it is essentially the 20-nanometer process [indiscernible]. Therefore, you do get the benefit of FinFET in performance power, but you don't get the [indiscernible] reduction therefore, you don't get [indiscernible]. So for one generation, definitely riding [indiscernible] Intel [indiscernible]
John P. Daane
The reason that's important is remember, wafer prices are going up. And TSMC has discussed this, as well as our major competitor where they have mentioned that 16FF wafer is more expensive than a 20 SoC wafer. So if you have exactly the same die size and you have a more expensive wafer, that means you're going to have a more expensive product. So the benefit that we have of going with Intel is being a generation ahead, you get a 14-nanometer transistor fundamentally, therefore, one generation ahead in terms of performance and power capability. But because the die size is so much smaller, the wafer goes -- price goes up, we're able to get a cost [indiscernible]. So that's why we believe that the high-end -- again, we'll have higher density, higher performance, lower power, and we get that cost benefit. Something our competition can't do across any of those vectors to really compete with the high-end part of [indiscernible]. And the reason that 16FF is not a great mainstream sort of product, the reason that we're doing 20-nanometer SoC as our mainstream products, remember mainstream is cost-sensitive. So if you get the same die size and a higher price wafer and therefore a higher cost solution going to 16FF, that's going to be hard to sell in very high volume wireless application for the customer [indiscernible]. Thank you. Next question?
I wondered if you could talk about the increases in R&D, how much of that is associated with bridging TSMC and Intel products in the same year? And does that mean that this year will be artificially higher than it will be going forward?
Ronald J. Pasek
[indiscernible] you'd see a huge increase because of Intel. As you know, we were going to do a high-end product regardless. I can tell you it's more expensive because of Intel. That's not part of the [indiscernible].
John P. Daane
If you look at the bridge, really, the increase in spending which is fairly moderate for 2014, it's really in masks, which is resulting for the fact that we have a number of products coming out over 2014.
Ronald J. Pasek
To be fair, there is a learning curve for our engineers that are working under 40-nanometer product. Doesn't mean throwing more of them will make it easier, a little bit different.
John P. Daane
Just to go through the one final comment on it, the EDA tools are the same, right? Engineers are fundamentally doing the same implementation of an FPGA architecture from which [indiscernible] leverage the same software system. So essentially, at the end of the day, if you consider that we were going to have to do something at the high-end trading 16FF for 14-nanometer necessarily results in an overall higher spend for our corporation. And as we've mentioned before, because we get that cost reduction, surprised a lot of people early on [indiscernible], working with Intel does not change our overall gross margin production long-term.
Had a question on your data center opportunity. First, it sound to me like you were saying that, that opportunity was potentially a replacement for what is today multicore x86 data center. Can you first confirm if that's accurate? And if so, what does that mean for your foundry relationship with Intel over the long-term given you're relying on their process technologies as you move from a strategic partner to something that much more closely resembles a competitor into their profitable business that seems to fly in the face of that current foundry strategy?
John P. Daane
Yes. So it's definitely not a one-for-one replacement of a multicore CPU. More that you might have multicore CPUs working in tandem with an FPGA. So in the net of it, it may actually reduce the number of multicore CPUs we have. I don't if it's going to put a huge dent into the overall market of what Intel ships. [indiscernible] $10 billion chip into it, and that's since it is going to make for better system. So we think it could also lead to greater refresh rates than we have out there. It could also be a net positive for Intel.
John P. Daane
I think the other thing to remember is Intel is actually advertising, right? So they have papers out that have shown the performance benefit of using an FPGA within a server for acceleration [indiscernible]. So it's effectively something that they've come out and said and discuss and actually shown a lot of data around it. And I think their term is reconfigurable [indiscernible].
Jeffrey W. Waters
Yes, we've had some collaboration with them [indiscernible].
Arif Nizar Karim - Kilimanjaro Capital, LLC
Arif Karim from Kiliminjaro Capital. I've got a couple of related questions on competition that maybe a little basic but I'd like to clear up in my head. So the first is can you just kind of talk about -- we've talked about performance, power being very important metrics in terms of winning business. But can you talk about sort of the 3 most important factors that customers look at when trying to decide whether they go with a new products versus a competitors' [indiscernible]? The second question is how often does a customer, as they migrate from one generation of the product to the next, recompete a slot that you've already won? Just I guess a rough percentage. And then when they do recompete, how easily can they migrate their IP application from one or the other next-gen FPGA?
Danny K. Biran
Okay, so I'll try to answer. First of all, how often will they look at both of us? I would say definitely large customers that typically both of us we look at both of us every time. Having said this, [indiscernible] does matter. So you use somebody who'll go and use other guy only if there's a very good [indiscernible]. And if you go back to the 40-nanometer generation because we have that leadership there, definitely that incumbency [indiscernible]. And so again, it can be done. And to the question of how difficult it is really depends on the specific implementation. Some customers would develop IP that takes advantage of a certain capability that we have, somebody else has, and that would make it perhaps more difficult to go from one to the other. But it can be done. But again, the important thing is if you're the incumbent, either you have to make a big mistake or the other company has to habe something [indiscernible] better [indiscernible]. Again, because of the incumbency we gained [indiscernible]. And that statement is definitely more to true at the high-end and at the low cost. Our low cost, the designs are actually easier and simpler and therefore, [indiscernible] company to the other. To the question of what is more important. Again, highly depends on the application. If we go back to some of the application [indiscernible] Jeff talked about, whether it's OTN or computer and server, customers need the highest performance. So if you can provide them with the highest performance as long as the power is within reason, they will select you. There are a number of military applications [indiscernible]
If we go to the other side on the low-cost, definitely better in case of power consumption better. So again, there's no one universal answer, it varies by market and by application.
John P. Daane
It also gets into tools, and from a productivity perspective, from a feature perspective, it's into intellectual property that we offer, in terms of capabilities toward solutions, more partial solutions, as we implement quickly. There are a large number of factors. So it doesn't become a selling point that we just advertise this across the product. The engagements are all quite different. [indiscernible]
Scott A. Bibaud
Danny, one, I talked about one of the markets, optical transport, where we actually provide a lot of VIP [indiscernible] in that case a customer develops a product is a various [indiscernible] future are obviously much higher because [indiscernible] rely on our IP point [indiscernible] But in each specific market space, as John mentioned [indiscernible] what IP we have to offer, different tools we have to offer, definitely showing a lot about OpenCL, they were the only people offering OpenCL, so those that are getting [indiscernible] are going to put for us.
John P. Daane
Romit has been very patient [indiscernible]
Romit J. Shah - Nomura Securities Co. Ltd., Research Division
Thank you. This, I guess, ties back to next question, which I thought was a good one. The quarter-over-quarter sales volatility has been greater than a lot of your [indiscernible] peers. You guys have been very transparent throughout. But it's hard not to think that structural or paper issue behind it. I guess, looking back, John, are there any issues that you would point to, whether it's execution-related or something specific mix, a headwind that potentially goes away next year?
John P. Daane
I think, if you look at our business over many years, it actually has been more volatile than, for instance, analog companies. On a quarterly basis, that's a very strong growth quarters. We have some of those [indiscernible]. And they are very hard for us to predict because our customers do not have much capability to build. I've always used the example, in this example, I've used are probably 15 years and still holds true today as the basestation manufacturer has a contractual obligation to deliver a piece of equipment, a basestation in 4 weeks. And so, as such, they will basically, they have no backlog, right? And so the waits at the last minute to buy from us, because they all sort of don't want to hold inventory. They can assemble a card and actually test and ship it in 2 weeks. And so, fundamentally, that's why they offer the 4-week guarantee. And so it just provides us very limited visibility beyond a [indiscernible]. I think, if you look at our overall business, there really -- where over time, we've said, for instance, if you go back to 2011, we had some onetime ASIC business because of the [indiscernible] conversion to us temporarily because of the tsunami that impacted Japan. We said that would come out, it did. Talked about the military programs which ramped down they came on and decreased. Really, nothing that we know of today that's necessarily going to decline. But again, we can't predict exactly what's going to happen on a quarterly basis, because of the lack of visibility [indiscernible] have. So at a high level, no, I do not know of any program or major market segment that through a decline. But at the same time, with every quarter, very hard for us to call a quarter plus 1, simply because [indiscernible] forecasts are so inaccurate that it did, every time we tried to do that, we [indiscernible].
Romit J. Shah - Nomura Securities Co. Ltd., Research Division
Okay. I know you guys aren't providing forward-looking guidance. But I just noticed that Q1 had been down in each of, well actually 5 out of the last 7 years, by an average of about 4%. And I was just wondering, is there any reason why Altera's business has typically been down in the first quarter, while Xilinx went up?
John P. Daane
The only difference between the 2 companies that we've noticed over time, is that they've grown in Q1 more strongly than us. So we've grown Q2 more strongly than them. People have surmised that maybe it's because their year-end ends in March. But to be honest, I don't know. I mean, we don't -- it's not like, at the end of December of our year, that we ship a lot of extra product to customers. So we can't really tell exactly why that phenomena has happened over time, and if you really look at our vertical markets, there's no reason per se that any one quarter outperforms any other quarter in aggregate. I mean, if you go back more years, you'd see that we actually grew strongly in Q1. And so, for those reasons, what we've always done is caution people, don't read into these averages, those are given in terms what the average growth would be out of a quarter, because there is volatility. There's a mix dependence. There is what customers are actually doing. There's macroeconomic effects at a larger level, which all pile into this. And so, very hard to figure out what the quarterly sequentials are going to be. They're certainly, from what we've seen, there isn't [indiscernible] cycle in our business.
Thanks. [indiscernible] from [indiscernible] I'm hoping you can talk about business, both in end market level and product level [indiscernible] portfolio perspective? Some of the end markets' structural growth [indiscernible] others seem, perhaps [indiscernible] the growth is [indiscernible]. So when we look longer-term, help us understand which -- what end markets you expect to [indiscernible] from a profile perspective? And also, from a product perspective, [indiscernible] last year [indiscernible] we expect to see more of that over the [indiscernible]?
John P. Daane
Why don't I just start with the acquisition. We don't -- do not comment on acquisitions one way or the other. So we're not going to take that question today. And I think, on the first portion [indiscernible]? Just in terms of the growth of [indiscernible]? I would say the top level and then I'll let maybe Jeff and Scott comment. At a top level, you think of it, Computer and ultimately, Automotive, probably will grow faster than the other portions of the business. But if you were to look at our business over 3 years, you're not going to see necessarily a substantial change. The percentage of the business, just because the other businesses [indiscernible] are growing, and are already a large proportion of the business. Comment?
Jeffrey W. Waters
Yes, if you looked across our businesses, there are some where we're in markets that are growing, in and of themselves. There are others where you have more stable markets, like Industrial, [indiscernible] where we are a relatively new entrant, some growth areas. So they're growth areas for us, even though it's a stable market, like Factory Automation, for example. When we introduced our FPGA families with the integrated ARM micro processors, that opened up a whole new set of opportunities for us in the space. So we look at our industrial growth, it's growing ahead of what we think the rest of the market is. It has less to do with actual market growth, growing more about, at our own growth with that market.
Scott A. Bibaud
When I look at my markets, I would say there are some markets, I think I've highlighted a few of them here today, but there are some that we didn't talk about that so strong, secular [indiscernible] those are certain factors, like for because, in my Broadcast business, we're seeing the move to 4K video that will ultimately drive for growth in that market, but hard to predict exactly when that will come through, but it seems like a growth area. In other markets, it's more driven by the economy. As a whole, I think as a company, we are putting together a mix technology that we will -- put us on a growth trajectory within all of those portfolios together. But each one is different.
Two quick questions. First of all, for Intel, are you in your SOC process or the logic one?
John P. Daane
So we have not commented on what process technology we're using [indiscernible] from Intel. No, I did -- we're not going to today. I think the fundamental thing I'd tell you is the processes themselves are all the same, the same universal curve of the transistors. At the end of the day, just a different, slightly different shift. So all of the technology is fundamentally already done. And as Intel mentioned, they're going into production. I think they're going into production first calendar quarter of 2014. It means by the time we tape out, and start sampling customers, [indiscernible] in production about 1 year ahead of us. And so that's the first time in a long, long time the we've actually used what is a very, very mature technology. It'd help us from a development [indiscernible].
And then the second question, [indiscernible] programmer [indiscernible].
Danny K. Biran
So, what happens is, people write the compiler and the underlying libraries, once you widen open [indiscernible] the complete conversion [indiscernible].
[indiscernible] question you made [indiscernible] in fact last [indiscernible] their numbers have fallen by [indiscernible] around 30, 35 [indiscernible] The first question, earnings volume [indiscernible] you guys talk about, you mentioned the segment far less [indiscernible] are #1, and [indiscernible] 20, 28 now and [indiscernible]
John P. Daane
So maybe I'll start with the market share. I think if you go back in history, the only times that there have been very significant market share shifts [indiscernible] one side failed. So if you kind of go back in time, when we introduced the Quartus design system, and that did not work very well, along with the Apex family, we lost a lot of market share in FPGAs, after gaining a lot during that period of time. And there was -- other than that, it's just -- it's hard to win a majority of a node just because, as Danny discussed incumbent with the IP, other things. I think if you move forward in 28-nanometer, and you look at the size of the Prototyping business, that moves to the next node. And in particular, as we introduce '14, because we are on a more advanced process, we will own the Prototyping business going forward for the next 5 years. So that comes out of 28. Our competition, as mentioned, they expect the high end of 28 to be bigger than the midrange. As we expect as well, Danny presented today, we have higher performance in the high-end. Obviously, pending a majority of the business. If you look at the $30 million plus of business we did last quarter, most of it was in the high end. And so, as that business continues to grow and again, those systems are more complex, it take longer to debug and deploy into the market, as that ramps, we expect our market share to pick up quite substantially. And all we have to do as a company is get over a 42% threshold. I think, right now, we're winning over 50% of the designs out there. I think we're in a really good position year-to-date over the long-term. We're not worried about that at all. Second part of your question?
Scott A. Bibaud
Yes, in wireless in particular, I talked about 3 different areas, baseband, radio and backhaul. There's actually a mix of -- I primarily focus on the high end to the mid range there's also low-end products that sell into that market quite extensively. I think we talked about in the past, we came to market first, then I think we have market share leadership on the high-end, and we see that continuing going forward, John just mentioned it takes a little bit longer to design in the high-end. And so that will cause our market share to grow in 28 there. On the low-end, we kind of see it as a bit of a wash, but the companies in the mid-range, clearly, they came out earlier than us, they've got a little bit of an advantage to market share there. We will gain in market share, but probably I think they have the advantage on the mid-range right now. And that's how I [indiscernible].
John, think we have time for one more question.
John P. Daane
Alex Gauna - JMP Securities LLC, Research Division
[indiscernible] lucky last guy. Alex Gauna of JMP. Thanks very much for taking my question. A 2-part question, if I could. One is on the DRAM side of the business. Part of the argument for moving C-RAN side of the business. Part of the argument for moving is cost saving by consolidating equipment. Confirm that there actually is a, an increased opportunity for you with that, as you take from the basestations and put it into the [indiscernible] type of location, and maybe what you're thinking around market share in C-RAN? I would [indiscernible] work with [indiscernible] advantage there? And then, the second part is, some of the other foundry players have argued that moving to being a foundry service provider is not an easy task, different mindset. Can you maybe talk about how you're finding the experience with Intel? IBM never really [indiscernible] a big player in foundry. Wondering [indiscernible], all around. Thank you.
Scott A. Bibaud
Yes, let me take the first one on C-RAN. So it's early days for C-RAN across various markets. I think where we see that rolling out most aggressively is in markets where they have a lot of Dark Fiber available in the ground, so they can take advantage of that in putting the remote cell sites, quite high data rates [indiscernible] from the radios and the basestations. Actually, I think we see an expanded opportunity to overlay both C-RAN and macrocells out there. We do believe our market share is very strong there. We're engaged with the big players on C-RAN [indiscernible] RAN. And like I said, it's early days, the number of different carriers experimenting with a number different things. In Japan, I'd say that probably, they've been using that type of architecture for their core architecture, there's a -- not as big a departure for them. So it's not really that big a change in our big [indiscernible]. So I think it's a good development for us. It's definitely going to be driving the highest performance-base bands in the market, and Stratix 10 will [indiscernible].
John P. Daane
Yes, what actually happens in C-RAM is there's processing along the network, all the way from the radios, or it gets to a server [indiscernible] as well. So there's more than just server application for FPGAs that we're working with China [indiscernible] on the high-volume spaces [indiscernible] process. On Intel, what I would comment is, for strategic reasons, Intel is very interested being in foundry. I think they've discussed why they're focused on this. Strategy was developed by [indiscernible] Brian, who's Chairman; and Brian Krzanich, who is CEO. Today, and obviously, they're trying to drive this and get other customers. As such, they're very interested in growing. [indiscernible], I think the best of Altera. So far, everything's been good. Still early days, but so far, everything been good. With that, Scott has a few closing remarks, and hopefully, we'll see you in the other room for some, for the Q&A.
As you're packing up today, be sure, as you go out the door, to pick up hard copies of Ron Pasek's presentation. Either door has additional copies. By all means, stop by and spend some time with our speakers, who are in a room off the hallway that you used to come here this afternoon. We'd be delighted to have you as our guest and continue this discussion. For those of you who have need for an archived version of this presentation rather quickly, it will be available later this afternoon through the Internet on our website, as you would expect. Last, thank you so much for coming today. Delighted to see you all, and we look forward to continue the conversation. Thank you.
Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.
THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.
If you have any additional questions about our online transcripts, please contact us at: email@example.com. Thank you!