Xilinx Inc. - Shareholder/Analyst Call

| About: Xilinx, Inc. (XLNX)

Xilinx, Inc. (NASDAQ:XLNX)

February 15, 2012 5:00 pm ET


Rick Muscha -

Moshe N. Gavrielov - Chief Executive Officer, President and Director

Vincent F. Ratford - Senior Vice President of Worldwide Marketing & Business Development

Krishna Rangasayee - Vice President of Corporate Strategic Planning

Jon A. Olson - Chief Financial Officer, Principal Accounting Officer and Senior Vice President of Finance

Victor Peng - Senior Vice President of Programmable Platforms Development

Lori Owen - Investor Relations Professional


Shawn R. Webster - Macquarie Research

Nathan Johnsen - Pacific Crest Securities, Inc., Research Division

Christopher B. Danely - JP Morgan Chase & Co, Research Division

David Wu - GC Research Ltd.

Ambrish Srivastava - BMO Capital Markets U.S.

Auguste Gus Richard - Piper Jaffray Companies, Research Division

John Pitzer - Crédit Suisse AG, Research Division

Rick Muscha

Well, good afternoon, everyone, and welcome to the 2012 Xilinx Analyst Meeting. It is great to see such a tremendous turnout. We actually do appreciate both your time today and your interest in the company.

So my name is Rick Muscha from Investor Relations. And at this time, let me go ahead and introduce the management team that's with us today: Moshe Gavrielov, President and CEO; Krishna Rangasayee, Corporate Vice President Communication and General Manager of the Communication Business Unit; Jon Olson, Senior Vice President and Chief Financial Officer; Steve Glaser, Corporate Vice President of Strategic Planning; Victor Peng, Senior Vice President of Programs Platform Development; and Vin Ratford, Senior Vice President Worldwide Marketing and Business Development. And in the back, of course, is my colleague Lori Owen from Investor Relations.

So to the extent we do make some forward-looking statements today, we'd like you to acknowledge the Safe Harbor disclaimer, and specifically, I'll point you to our most recently filed 10-Q where you'll find all the risk factors that could contain -- that could drive results to be different than what we project.

So our agenda will feature 4 speakers: so Moshe will come up and describe how we are all-in at 28-nanometer; Vin Ratford will then discuss our 28-nanometer leadership in the broader markets as we define them; Krishna will then discuss 28-nanometer leadership and specifically, focused on communications; and then Jon will be wrap up and discuss corporate financials, including FY '13 outlook. We then will have ample time for Q&A, which Jon will be moderating, and we'll have -- Lori and I will be floating around with a couple of handheld mics so that all the questions, your questions can be heard on the webcast.

Following the Q&A, at 5 o'clock, if you kind of just go out the courtyard to the terrace tent, we have a cocktail reception to give you a chance to mingle with the management team and at the same time, take a look at what we think are pretty exciting demos that actually represent functionality from each of our 5 families in the 28-nanometer product generation.

And lastly, I would want to like remind everyone that today, this is intended to be a strategic discussion, so we're not really going to focus on the March quarter or June quarter of business dynamics. And with that, it's my pleasure to introduce Moshe Gavrielov, President and CEO.

Moshe N. Gavrielov

Thank you, Rick, and good afternoon, everyone. It's pleasure to see you all here. The theme of this presentation is all-in, and the book you have was written on this topic related to David Petraeus and what it takes to win. The management team we have here has been around for about 4 years, and we took a page out of the book and identified that what are the things that we needed to do at the 28-nanometer node in order to achieve and deliver to our customers indisputable leadership. So I'll provide a little context on where we are and then there'll be a lot more details from each of the presenters, which will take you towards additional level of granularity.

So from our perspective, this is the result of the journey we started 4 years ago to establish this indisputable leadership at the 28-nanometer node. We've invested very, very heavily in order to achieve this, and what I'd like to do is to highlight some of the elements which we believe are the pillars of this leadership. First and foremost, we delivered the first tape out and the first silicon at 28-nanometer. Since then, we actually have large numbers of tape-outs that have been done, over 10 tape-outs has been done with TSMC.

This highlights that we actually have a phenomenal relationship with them, and this outstanding partnership is enabling us to deliver what we believe is leadership technology. In addition to that, you'll see that we have pioneered an industry-leading 3D IC technology, SSIT, Stacked Silicon Interconnect, which enables us to provide products that's really are breakthrough in terms of their capacity, in terms of the performance, in terms of their cost and power footprint for whole range of applications.

Then we've also delivered -- and this is silicon that we delivered to our customers, actually last year, and now we're shipping to a lot of customers the Zynq product line with the embedded core. That, again is a result of investment we've made over several years. We started it. We recognize this is actually a paradigm change in Xilinx's history. And we announced the architecture several years ago, then talked about the details, and we believe we have a huge lead over the competitor in terms of developing this product.

And in parallel, what we've done is we've invested several hundred man-years in our software system, and we will be releasing a new software system on schedule in the middle of this year, which will be a breakthrough in terms of its capabilities. So all of these elements actually enable us to go from what we believe was our traditional role in programmable logic to enabling our customers to deliver programmable systems integration. That's a big change, and we're at the forefront of that going forward.

In terms of what is driving our business, in my mind there are 3 distinct elements. First and foremost, the programmable imperative. This basically highlights the fact that ASICs and ASSPs, except for extremely high volume consumer markets, are less and less viable, and as a result, FPGAs address a much larger market.

The second element relates to what we used to talk about as insatiable bandwidth, but actually now the twist of that is insatiable intelligent bandwidth. I'll explain why FPGAs are uniquely suited to deliver that and why that is a major driver for our business going forward. Then all of this plays into the relentless drive for higher systems integration, as more and more of our customers' subsystems, hardware and systems hardware gets integrated into the FPGA.

So programmable imperative. This is driven by an economic phenomenon with each generation's technology developing a product from scratch, increases in terms of the cost, the risk and the time. And as a result, ASICs and ASSPs are less and less viable, and this sort of shows numbers and this trend, for better or for worse, is accelerating. It's also driving our costs up in terms of developing the new product. So we are uniquely familiar with what the cost of developing these products are.

And as a result, as we deliver our product, we know that there are very few applications that can justify that level of investment. And you can see that this trend is accelerating when you look at the overall cost of developing the 28-nanometer solution compared to 44 -- 45-nanometer solution, the overall costs are about 2x.

And if you look at that being well over $100 million in terms of development cost per product, then that implies that you need multibillion dollar markets in order to justify that investment. There are fewer and fewer multibillion dollar markets. And as a result, FPGA is becoming the default solution for more and more of these markets that can't afford to design these products.

Now when you look at insatiable bandwidth, this is just a trend that continues to accelerate. And it was the trend that was very prominent in the developed world, in North America, Europe and Japan. Today, we include China in that. It's sort of when you go and look at the infrastructure in China, in certain areas, it's actually better than the infrastructure in the Western world. And that's sort of a given. That needs to continue to improve.

And there's various data points, one of the most-heavily quoted one is the burden that smartphones impose on the systems. When you look at the data bandwidth requirements, it's typically 1 to 2 orders of magnitude more in terms of bandwidth than a normal phone, and that's just sucking up all the bandwidth, requiring a great upgrade of infrastructure.

This same phenomenon is happening in emerging economies and in less developed countries. Latin America, which, in terms of the population, is significantly larger than the U.S. market, is growing at a very fast rate. You can see the growth in Facebook users and mobile subscribers.

Similarly, when you look at India, there are numerous examples of this. And there's a big investment which is starting in India, and it has a long way to go in terms of bandwidth. And then interestingly, Africa is often being ignored. But as the middle class there expand, it's actually is already -- there are already over 700 million subscribers. And as a result, there's a huge investment being done, and it will be necessary in the future in order to increase the bandwidth to support it.

Now the change that is happening is this. If you look at the top left, there was this insatiable demand for bandwidth, and generally, that has grown 5x in 5 years, and it's expected to grow 5x or more over the next 5 years. That is not relenting that's continuing at the same pace and maybe even at an accelerating pace. But what is happening is not only is there a demand for bandwidth, but there's a demand for intelligent bandwidth. There are few examples for that.

When you look at smart vision, more and more of the bandwidth is taking by video. But now there's more and more of a need to do realtime processing of that video as it's actually streaming through. And that generates huge challenges. And FPGAs are actually incredibly well suited to doing that. So the application you can see here is one which relates to the lane detection, and that's something where you do the detection on the video as it's streaming through. You don't do any post-processing in a big way. You have to do it, and that's an example.

Second area we're seeing is a huge focus on embedded security. And again -- and this manifests itself in 2 ways. First, in terms of making sure that the transactions are secure, but then also making sure that the silicon itself is secure and has not been tampered with. And Xilinx in particular has invested very heavily, over several years, and developed technologies which initially were targeted at military applications and now are being declassified and will have broad applicability for a wide range of commercial applications.

And when you add that to the ubiquitous computing, the fact that now there's computing being done -- machine-to-machine computing, expectation of 50 billion connected devices by the year 2020. We actually saw a number of 10 billion in 2016, which implies that if you get to 10 billion in 2016, you'll be -- probably be at even higher than 50 billion in 2020.

And this is a third wave in computing. If you look at the common theme of all of these areas, one thing stands out. FPGAs, which have the highest throughput, the highest bandwidth and have inherently massively parallel processing are the ideal platform to implement these solutions. And because we serve as an umbrella of applications, it's not just one specific one. You can see how all of these trends benefit from that.

The other thing which FPGAs are driving is we can address all of these 4 elements and integrate them into our FPGA. So this enablement, for programmable systems integration and facilitation of the bandwidth, the smart vision, embedded security and ubiquitous computing is inherent to the FPGAs. And with each generation of product, we make sure that we can move forward on all of these 4 axes.

How does this implement -- manifest itself in the market? Take you through a large number of examples here and what I'd like to draw your attention to is the integration needs in each of these markets, so starting with OTN. Integration needs, you can see the bandwidth, ubiquitous computing because there's so many machines that are being used and security is becoming a big deal in this set of application. So FPGAs are actually becoming more and more mainstream in these wired applications. Krishna will take you through that.

Similarly, on the wireless side, the same needs: extreme bandwidth, ubiquitous computing and security. It's a different application. It operates in a different way. But when you break it down, FPGA, inherently, are the ideal tool to address that. These happen to be more cost-sensitive applications, that's why we're delighted to have the best midrange solution, which is Kintex, that can address this specific market, and it will enable us to grow our business in this area going forward.

Another area which is emerging is data center, and again, you can see similar set of applications. This is a new market -- or similar set of needs. This is a new market that we're addressing. It's actually one of our fastest-growing markets.

Then in Mil/Aero. Here, you can see security emerging as a very important part of this market. And the example you have here is of an unmanned air vehicle, but there's more and more military applications that actually require all 4 of these elements.

Then the one, which is near and dear to my heart, as I grow older, is this robotic surgery platform. We actually had the VP of R&D of a leading company in this area who came to our sales conference and just listened to all of the demands, all of our products in the communication space. And then he came back and said, "Well, that's exactly -- those are exactly the problems that we have in the robots that we are developing."

Huge amounts of bandwidth, the ability to analyze that really quickly and identify all sorts of things that are happening in realtime as you're doing the operation because this is an area where there is no room for error, and you can't sort of wait and correct it later. It tends to be a little too late there. And again, it is serviced by our FPGA product offerings. And in this case, they typically use our high-end product just because they just can't get enough computing, parallel computing.

Then in next-generation driver assistance. Very similar, here, focus is on the smart vision, lane detection. There's now legislation, which I believe has already happened in Europe, and it's rapidly evolving in North America, too, which requires all cars to be able to do a significant amount of safety detected processing. And the expectation is that as you do that, and you can look at the sub-bullets here or the quote, the number of accidents and the number of people killed could be brought down drastically as a result of these capabilities.

Then I believe this is the last one, next-generation industrial automation. Again, this is becoming mainstream. I think Vin will talk in more depth and talk about what the needs are. This is a market that benefits from the Zynq product line, and it's a major driver. This Zynq product originally was targeted at this market as one of its major markets going forward.

So if you look at all of these applications, there's a large number of them. We actually had even more examples but we ran out of time. But basically, what we believe that this enables us to do is to increase our SAM very significantly. And as you look at what is available in 2015, as the number approaches, 50 billion -- sorry, 60 billion for ASICs and ASSPs, at that point in time, we believe those PLDs will be able to service 15 billion of that.

The important takeaway here is of that 15 billion, only 6 billion is in what used to be our traditional core FPGA business. Actually, a lot of the growth is in displacing ASICs, ASSPs and increasing our footprint in the embedded market, which is as a very, very large market.

Our technology position at 28-nanometer, we believe or we are convinced, is unparalleled. And when we look at the -- if you look at it in a broadbrush, we have 5 general families; 3 core families: Artix, Kintex and Virtex; and 2 expansion families that enable us to get into the systems integration space. So that's the high-end of the Virtex-6 -- 7 with the SSIT and Zynq products line.

When we look at these 5 areas, we're obviously focused and convinced that we have the best core offering, but we're also delighted that as we catapulted ourselves into a leadership position at 28, that not only do we address the core market, but we actually expanded the market we're addressing.

And today, as of yet, there is no competition in these 2 markets on Zynq and the SSIT. Now imitation being the sincerest form of flattery, we know that the competition is moving as quickly as they can to find competing products of these. But in our mind, we have at least a year to 2-year advantage over any competitor in this space, and we feel incredibly good.

Core business is the largest part of our business, and we're convinced that we absolutely have the best product in terms of price, performance and power in -- for the core business, so it's not that we've invested in these expansion and we have ignored the core. The core is the basis for these expansion plays. And by having the strongest core product, we believe we have a great position.

And our -- my expectation and what we are seeing now is that in the core business, we have a design win rate which is 70% more. And you can talk about it in terms of number of design wins. You can talk about it in terms of dollars, but we're very comfortable and confident in that number.

And then, obviously, in these expansion areas, we believe we have such a high lead that we're really getting all of the design wins in the expansion product. And they have the way to go where we intend to totally exploit our strong competitive position there by continuing to drive that going forward.

When we talk to our customers, this is how we convey the benefit that we bring to them. And when you look at 28-nanometer as really these 5 elements, we enable our customers to build systems with fewer chips, and we enable that to do them faster. As a result, they actually can get to market faster. They can integrate their systems into less and less devices. And more and more is being integrated onto the FPGA, and we enable them to increase the performance, reduce their bill of materials, and in numerous cases, actually significantly reduce the power in their solution. We'll talk a little more about that in particular. Vin will give you quite a few examples there.

The product is available now, and we were the first to tape-out of -- you can go to the tent, and I really recommended that you go there. You'll see it's not foil. We actually have delivered these products. They're working. They have real customer applications running there. So I very strongly recommend that you go and see what we have there. We show all of the products in the most significant demanding application.

We also have started shipping the product in a big way. So in terms of the number of units, several thousands of units have already been shipped for revenue to our customers. In terms of design wins, at the end of the calendar year, we actually believe we exceeded $1 billion in terms of design wins to date. And when we look at the design win rates, we're very confident in our position there, and this will enable us to attack and improve position and address the much larger SAM that the product is capable of.

So to summarize, this is a journey that started 4 years ago. We went for the most aggressive solution in terms of silicon, in terms of tools, in terms of IP, and we've delivered that. There are numerous advantages from having this technology, and now the fun begins as this translates into a significant revenue. Jon will talk a little more about that.

So with that, I'd like to hand it over to Vin, who will talk about our leadership in the broader market.

Vincent F. Ratford

Thank you, Moshe. So I'm going to talk about where we've had traction and adoption with our 28-nanometer family in the broader market. I'm going to talk about why customers are moving to the 28-nanometer platforms as quickly as possible. And then, finally I'm going to give you some real-world examples of where we're winning. You'll also get a chance to see some of that in the tent.

So when we define broader markets, essentially what we're talking about is all of the end markets, except the communications markets. And Krishna will talk about wired and wireless and also, some emerging opportunities we're seeing in the data center as our computing business and our storage business starts to -- become an integration play and give us more opportunities for growth in the data centers.

So we've raised our SAM forecast from last year by roughly about $3 billion, about $1 billion of that is coming in these broader markets. And the markets that are growing the quickest are in the industrial, scientific and medical segment, the aerospace and defense segment, which we report currently as Industrial and Other. That's where we're seeing the largest growth. The automotive market is growing, as on a percentage basis, the fastest, but it's off of a smaller base. And we're really getting this growth as a result of seeing early adoption of the 28-nanometer family and the Zynq platform.

And I'll go into little bit more detail. But specifically, in industrial, scientific and medical, where we're seeing very early traction, particularly with our Virtex and our Kintex and our SSIT technologies in areas like test and measurement, with Zynq in areas like industrial control and automation. And then in Aerospace and Defense, our Artix family is seeing very early adoption in things like handheld radios for the government, in secure communications platforms with our SSIT technology.

And as Moshe talked about, 30% of the Air Force currently is UAVs. And if you're paying attention, the FAA has been asked to open up the airspace so that UAVs can fly in commercial airspace. So that is a massive trend that favors the adoption of FPGAs.

And then finally, as Moshe talked about, these emerging opportunities in active driver assistance, we pretty much become the early de facto platform in those applications with Zynq because largely, not only do you have to do lane detection, but you also had to do blind spot detection, and you have to recognize roadsigns, and you have to control the brakes. And there's a lot of things you have to do that requires tremendous amount of single processing power.

So as Moshe talked about, not only are we all-in from a technology perspective, but we're all-in from a sales and from a marketing perspective. And whether it's with our horizontal portfolio of 28-nanometer products or the Zynq family, which is seeing very significant takeup, I'll talk about some of the stats in a minute, but also, our targeted design platforms strategies.

We rolled this out about 2 years ago, starting with the sixth series generation. And 2 weeks ago, we announced the first 3 targeted design platforms. We're way ahead of the competition in terms of bringing these complete development kits to the marketplace which ease customer adoption. You get a chance to see some of them in the tent. And already, we've had customers adopted, in the high-end, over 500 of those development kits, beyond our expectation at this point in time.

Another reason that we're able to compete more effectively against ASICs and ASSPs is we've invested substantially in bringing into the company system architects and people that have the end market expertise to allow us to be able to go in and show customers how to map their applications into our programmable platforms, and you'll see some examples to that in the design win scenarios that Krishna and I are going to be talking about.

And then, finally, through the ecosystem that we have, not only with their own Xilinx's ecosystem community of over 250 companies, but the tremendous amount of traction with the ARM-connected community, particularly around the Zynq platform. Over 50 of the ARM-connected community partners are currently developing tools and IP and development boards for the Zynq platform. That's a huge head start against the competition.

And then, of course, finally we have a world-class distribution partner in Avnet that is rolling out a 22-city, thousands of people tour to roll out the 28-nanometer family, starting in April. And they're going to be leveraging all of these targeted design platforms. So we're going to be taking it to the broad markets in a big way.

And here's some of the early scorecards. Moshe talked about some of the high-level numbers. I'm going to parse them a little bit for you. Obviously, with over $1 billion worth of design wins at this point in time, about $500 million of that is in the broader markets. And that's coming from roughly about 250 customers that are designing with the 7 series at this point.

More interestingly, over 200 companies are developing for the Zynq platform. And as Moshe indicated, we pretty much have open field running when it comes to opportunities in that space relative to our competition. So you might ask why is the value proposition so compelling in this broader markets. And again, as Moshe pointed out, in almost every one of these end markets, ASICs and ASSPs are too expensive to bring to market.

They are increasingly under-servicing the applications. Companies have a long life cycle with their technology, and even older generation ASICs are having to be replaced and integrated into our platform. The net result is the customers get a tremendous amount of capability through that integration by delivering either higher performance, lowering the cost or reducing the bill of materials. I'm going to share some examples with you.

And of course, FPGAs have always delivered higher-level system performance, and companies are finding interesting ways to be able to take advantage of that as we take them into the next generation. And then of course, ultimately, what's really important is to have world-class tools and the complete solutions stack for the customers. We've made, as Moshe indicated, tremendous investments in the last 4 years with our ISE software and our TDPs.

And this summer, we will roll out what we believe to be a major breakthrough in the next generation of software, with over 100 companies already designing for that software platform. We think this will allow us to take our productivity to the next level for customers for multiple generations of technology.

So when we talk about performance, it's important not to fixate on any one single metric. Oftentimes, people tend to fixate on one particular number like 30s line rate. But the reality is, if you look at what the FPGA is doing from a processing perspective, you need to be able to look at the I/O bandwidth, you need to look at how quickly and how broadly can you access memory. Moshe mentioned some of those single-processing applications. DSP performance is really, really critical.

And then finally, you need more capacity in the FPGA in the form of bandwidth and in terms of logic cells. And in every key figure of merit for every family that we have, we have a competitive advantage relative to the competitions. And in some cases, it's as much is 2x, and in other cases, 1.5x.

The net result is, across the board, family-per-family, we can deliver higher-performance. But that's not the only reason why customers are moving to the platform. You have to balance the performance with cost and with power. And what that means, fundamentally, is if you can map system performance per watt on the y-axis and cost on the x-axis. And you compare what we're able to deliver versus what the competition can deliver.

If you start with the Virtex family and in particular, the SSIT devices, as Moshe pointed out, we delivered substantially higher bandwidth and substantially higher capacity, albeit, at a higher cost per device. But when you're able to integrate multiple devices into a single device, we lower the BOM cost. And that's true across the entire Virtex family relative to the competitive offerings.

The mid-range family, the Kintex family has been very, very successful particularly in the wireless market, as we're able to offer a higher level of performance for this next generation of wireless applications, while also bringing the power down in a meaningful way. And Krishna will talk about where we've been able to have very early market success with that mid-range Kintex family.

And then finally, the Artix family, which is really -- brackets our competition very nicely, both at the mid-range and at the low end, offers, again, substantially higher performance. Or for people that don't need the performance, we can offer lower power or lower cost.

How do we do that? Well, as the graph on the left basically shows a series of engineering investments that were made across the entire product portfolio, that which results in us being able to reduce the total system power by 1/2 from one generation to the next. And you'll actually be able to see that running in a demonstration in the tent where we can show you the 2000T running at substantially lower power than equivalent to a discrete number of devices. And that's done through a whole host of architectural process, circuit and software investments.

One of the benefits of having that is that you can then, through integration, start to reduce the system-level power, either by integrating multiple devices, as in the Zynq example, in the upper right-hand corner, or through the SSIT technology in the lower right-hand corner. And oftentimes, people are surprised to hear that the Zynq family will actually fit within a 2-watt envelope, which is a key metric for embedded application.

If you can fit within that 2-watt envelope. You can fit up in the rearview mirror on that automobile, not need a heat sink, take very, very small space. And you can start to fit in a lot of embedded applications that, historically, FPGAs could not reach.

Another key pillar of the value proposition is being able to reduce the BOM costs. And again, through the integration that we're able to offer, we're able to deliver a much lower total systems cost by reducing the number of discrete components. In this case, with Zynq we're able to take discrete processors and DSPs off the board and integrate them into the Zynq platform. With our analog mix signal capability, which you'll also see demonstrated in the tent, we're able to take a number of discrete converters and filters and passive components off the board and be able to reduce the cost in that way.

Of course, what our customers always see is our software. And what they measure, once they are done looking at the silicon specifications, is how quickly can they get their design up and running and get their product into the marketplace. The tremendous value proposition of the FPGA is its rapid time-to-market.

Over the last 25 years, as the FPGAs have become more complex and are now approaching the gate count of previous-generation ASICs or even current-generation ASICs, there's a need to dramatically up the game in terms of overall investment and the overall way that you attack design productivities. So we just had a very comprehensive program where we're attacking it in all of these key areas.

If you follow the ASIC and ASSP world, you'll know that the integration of IP is one of the key barriers to being able to get an ASIC or an ASSP into the marketplace. And we've got a tremendous amount of investment in what's called design reuse and the ability to integrate all that IP. Our current platform and our next-generation platform are going to make that a lot easier for people by adopting industry standards, by making the IP scalable across different product families, so you don't have to re-architecture IP when you move from an Artix to a Kintex. And then finally, by integrating all that into the targeted design platform, we give the customer a reference design that gives them a good starting point right out of the box.

So in every key figure of merit that customer measures around productivity, design reuse, ease of use, turns per day and their ability to take advantage of the platform through our support organizations, we have a significant competitive advantage with our current IC platforms, and we're dramatically going to up to game this summer when we roll out the next-generation software platform, 4 years in the making, hundreds of people, a whole new paradigm shift for productivity.

So how are we doing in the design wins. Well, if we focus in at this point maybe on just Zynq, because Krishna is going to talk about some of our 28-nanometer families, what you'll see is there is roughly about $150 million in design wins today. I mentioned that there are over 200 companies that are developing applications. And at about 150 of those are customers in those end markets, about 50 of them are the ARM-connected community and significant ecosystem partners.

But what's interesting about this is how pervasive the applications are. In north America, we got a major design win in a next-generation point-of-sale scanning system from a company that you would recognize. They've been in the business for a long time. They've moved to the Zynq platform.

And what's interesting is the initial win at that particular customer is for a single socket. But that customer is planning to use Zynq in all of its point-of-sale terminals, both the portable ones, the midsized ones and the high-end ones. So that initial socket win of $18 million is just the tip of the iceberg of the future revenue that potentially could come as that company drives Zynq pervasively across all of its own platforms.

And you might ask, why would a company do that? Well, the reality is if you look at the value proposition of Zynq, you've got the de facto standard in instruction set architecture from ARM, you have a tremendous ecosystem to support it, and you've got the flexibility and extensibility of the FPGA platform. For companies that are investing a tremendous amount in software and IP across multiple processors and multiple platforms bringing -- being able to bring that altogether provides a tremendous opportunity for them to take advantage of the Zynq platform.

So again, whether it's in imaging or whether it's in test and measurement or whether it's in surveillance, it really is a platform play, not a socket sale. And then, what I'm going to do now is I'm going to talk a little bit about industrial networking, medical imaging and surveillance and show you the value of the integration play that we are bringing to our customers.

The other one that I thought was kind of cool is I've got a personal interest, I go to Palm Springs periodically. Anybody that's been down there knows all those wind turbines that are twirling around. We've just been designed into the world's leading manufacturer of wind turbine. Now if somebody had told me 5 years ago, 4 years ago that Zynq would've been used in wind turbines, I would have said -- never would have thought that would happen.

So I think that reflects the kind of innovation and the opportunity.

So let's take a look at medical handheld ultrasound. Again, this is the platform sell. You've got companies that are trying to build portable platforms. They're trying build car-based platforms. They're trying build high-end platforms. They want the reuse of a single software architecture. They have the ongoing ability to improve the resolution and the image processing capability between the transducer and the display and give the doctor more information quicker, so the patient doesn't have to wait for analysis.

In order to do this, they have to take advantage of the integration of the FPGA to handle the beam-forming data coming off the transducers and then be able to take advantage of the high bandwidth between the FPGA and the processing platform to be able to do realtime single processing much, much quicker. And if you look at the value proposition, we're basically reducing the number of chips in a meaningful way. We're dramatically increasing the capability through increased performance, and we're lowering cost and power at the same period of time.

Now what's interesting about this particular market is from a development perspective, it requires moving to tools like MATLAB and higher level synthesis where the algorithms are developed in high-level languages. And again, our acquisition of AutoESL and our next-generation silicon platform has been developed specifically to be able to take advantage of these high-level languages. And our ecosystem partners, like the MathWorks, are investing heavily in our design flow for the Zynq platform that will be able to take us into these marketplaces.

So medical and handheld is an opportunity for the expanding SAM. We've got a number of design wins in it already. Surveillance is the next one. Whether you talk about China and their deployment of many, many surveillance cameras for their China safe program or whether you look at the need to upgrade the technology in London to higher resolution and being able to do more advanced analytics in the camera, that requires more signal processing power. It requires more computational power.

And one of the beauties of the platform is it can be used with a wide variety of sensor technology, whether it's 2D or 3D or any other kind of high-resolution sensor. And what you're looking at here is the ability to integrate all of the discrete capability in the image processing ASIC, the video acceleration in the FPGA including the scalar, et cetera.

So this is, again, 4 chip to 1 chip integration play where you're able to take advantage of the integration to deliver much higher levels of performance, being able to deliver 1080p, 60 frames a second and be able to do facial recognition on multiple faces in a crowd in realtime is the requirement. You cannot get there unless you have a platform like the Zynq. And again, we see that as a significant new market and in expanding our SAM opportunities.

Finally, in the motor control area, again, this is a bit of the greenfield opportunity for us. We've historically not been in motor control. It's typically been more or less the kind of a microcontroller market. And increasingly, we're being drawn into this because several things are going on. First, the motors themselves are becoming more sophisticated, and customers are looking for ways to make the motors run more efficiently from a power perspective, more accurately and be able to turn on and turn off much more quickly so that they can essentially manufacture more quickly.

And in this particular example, we've got a significant design win in a major European motor control manufacturer. Interestingly enough, the customers's equipment is used in processing 28-nanometer semiconductors. So it's kind of interesting that their major customer is building semiconductor processing equipment to manufacture 28-nanometers, and now they're moving to the 28-nanometer Zynq platform.

So again, the value proposition here is compelling by being able to reduce the cost, lower the power, and most importantly, give added performance. The other thing they told me, which I thought was kind of interesting, is if you look at a cellphone and you look at the enclosure, this piece of metal that's been machined is manufactured by one of their pieces of equipment.

And so what's important is being able to manufacture this very quickly with very, very tight tolerances. Anyone that knows what a maniacal person Steve Jobs was about detail knows that being able to manufacture this with very tight tolerances was critical. And all of this kind of technology requires these advanced motion control systems. So I thought that was kind of interesting.

So wrapping up, as indicated to you, and Moshe pointed out to you earlier, we really are seeing tremendous adoption of our 28-nanometer family across all of our broad markets. Again, whether it's in ISM or aerospace and defense, we're seeing very early takeup in the applications. And we have a tremendous value proposition that translates well based on our integration story and our ability to deliver value to the customer. And I think as I pointed out to you, the numbers speak for themselves, and they're only going to get better. Krishna?

Krishna Rangasayee

Thank you, Vin. Good afternoon, everybody. I'm going to talk you through the 28-nanometer leadership we have in communications. Moshe talked about the team being all-in, and all-in for communications really starts with building the best portfolio for communications.

So this is silicon, software and IP. The journey started about 3.5 to 4 years ago, and the initial market success validates to us that we're headed in the right direction, and the numbers will probably highlight that story better than my words.

We created a business unit to maximize our subsystems about 1.5 years ago, and it brought a lot of the pieces of the company together and gave us a very sharp focus in building very strong relationships to customers. It gave us an opportunity to align our internal roadmaps with their roadmaps, and that is a critical foundation that, really, has given us an elevated viewpoint of why we could be a bigger player in the communications market more than where we are today.

A combination of the 3 acquisitions we made over the past year, combined with the dedicated design services team really enabling us to accelerate our ASIC, ASSP SAM expansion. That is the cornerstone for our growth strategy in communications. It's a growth horizon outside of the FPGA play. It's really a growth horizon that lands in the mainstream logic IC play. And we are in early days, but we have made significant forays to the ASIC and ASSP market, and that is the growth horizon for us in communications.

In terms of numbers, quite a few numbers here. There are 3 that I'd like to bring to attention. Number one, we have won, to date, the 28-nanometer technology, more than $600 million in design wins. And majority of these are with the top 10 customers. This is a number that, to a certain degree was, we had high expectations of the 28-nanometer technology in the process nodes. The numbers are living up to higher than our original expectations. And we continue to raise the bar in ourselves.

The initial market success validates that we're headed in the right direction, and obviously, it's early days. The other number that I'd like to highlight to you is the design win rate against our immediate competition. So we are winning 7 out of 10 platforms, and we have the highest visibility into our top tier customers. And the validation from, not only from our internal data, but validation from data from key customers highlights to us that we are winning at a rate that is outside of the norm for us as an FPGA industry.

The other more important number is that more than 1/2 of design wins that we have to date, as a part of $600 million design wins that we have, have really come at displacing an incumbent ASIC and ASSP. Early days again, but the strong growth that we have really has to come from eating into ASIC and ASSP SAM.

Moshe has talked to you about the program for quite a few years. We are beginning to live it today. And more than 1/2 of our revenue and the outlook for us continues to be the same. It's going to come in through growing into the ASIC and ASSP SAM, and that's the rationale for why are SAM outlook is really be a much larger SAM outlook from the traditional FPGA SAM outlook.

The wired communications market by the 2015 time horizon is a $5.2 billion SAM for us, 3 subcategories that we put a lot of emphasis in: number one, the broadest market that we played actively at significant role for many years is the service provider market. That's a $2.6 billion SAM; the enterprise market continues to grow at the significant pace, and that represents the second-largest opportunity space for us. It's $1.5 billion; and emerging new area, which is the convergence of the networking storage in the server areas of the data center market.

The submarkets that I've highlighted in green are the areas where we see the highest promise of growth. These are 3 key areas where we expect to grow more than 10% CAGR over the next 5 years, those on the metro, the enterprise switching and the data center market. And over the course of the next few slides, I'll walk you through some examples. Our customer representation is the value that they're continuing to bring to these 3 marketplaces.

Wired communications and wireless have both grown for us as a company. Wireless has outpaced wired communications growth's for the past 4 to 5 years. As we look for the next 5 years, we do believe that both the wireless and the wired markets will continue to grow, but we do think that wired will outpace wireless in our growth outlook over the next 5 years.

Currently, across our top 10 OEMs, our revenue as a percentage of a large ASIC consumption of the BOM cost consumption for key customers is 12%. And if you abstract, if you will, a service provider line card, and this is a block diagram of an abstracted view of it, our roll, essentially, to the revenue we are shipping today, is mostly bridging. So it's coexisting with ASIC and ASSP functionality and bridging them together. And we are one of the leading suppliers in the space today. But strategically, our fit has really been one of fitting in with the ASICs and ASSPs.

What really is changing for us, and speaks to some of the numbers that I gave you before and also speaks to the growth horizon for us if we look out to the 2015 timeframe, is we are going to be increasing or we project to be increasing our logic IC BOM consumption of the top tier customers from 12% to 25%.

Three key reasons why this is going to be happening. Number one, 28-nanometer represents a giant leap for us, from both technology competency and from a total cost of ownership from our key customers. As the number of ASIC and ASSP suppliers become lesser of an option for wired communications applications where most of the volumes is really mid-range and as we fill in the gap, in terms of technology and the business context, the third element of it is really the IP acquisitions that we've made over the last 3 -- over the past year.

Combination of all 3 factors, increases our opportunity for us to go from being the bridging functionality essentially to really be in the heart of the system. That is the transition that we stand today. And the benefit of it early, thanks to the $600 million in design wins that we have to date. But we hold a strong outlook toward where we could be in wired communications going forward.

As part of my role, I spend a lot of time visiting with top tier customers, and what we get constant validation from our customers, both on our 28-nanometer technology but broadly our systems expertise and competency is one of capability. We are providing a game-changer of an opportunity with 28-nanometers. We are providing capability systems for the leading best-in-class system performance capacity and features.

Integration is a cornerstone of our success, and we'll continue to be a foundation for our growth outlook going forward. Integration and taking a leadership on that enables us to translate the lower system power and BOM cost and to a large degree, delight our customers with what we can bring to the table. And this spans, as I've told you, way outside of the normal FPGA context. We are today a mainstream logic IC context, and that is the growth horizon for us as a company.

As the technology and the business competency gets addressed, we are one the few companies with a broad scale to really compete in a very effective way. A key gap that we are beginning to address through our acquisitions, but also with internal capability, is our IP in services. The combination of all of this, again, allows us to meaningfully expand and compete against the ASIC and ASSP landscape in the wired communications market.

This is an example of the service provider metro application. We have provided solutions, and we are one of the leading incumbent suppliers in this market today. Broadly again, as I told you, if you abstract, we are fundamentally providing capability to stitch and build systems together where the core in the heart of the system, to date, has essentially been ASIC and ASSP solutions.

The transition for us is really to become the heart of the system. And at this SAM is around $260 million in the 2015 time frame. The root cost enablement that we are providing is a story of simply integration. We are uniquely capable of stitching new capability we're in there is dramatic integrations, that we are offering through the 28-nanometer technology that leads to improvements in performance, dramatic reduction in BOM cost and power.

And we acquired a company recently, that is also the best-in-class in terms of Ethernet and Interlaken solutions. And that value proposition in the combined outlook gives us a strong growth outlook for our service provider markets.

The metro optical transport area, if you roll back 3 to 5 years ago, there are many, many ASSP suppliers in this landscape. This was essentially an ASSP/ASIC market. As we look at it today, and as we look at the revenue that's going to happen over the next 2 to 3 to 4, 5 years, this is essentially switching from being an ASIC, ASSP market to almost a FPGA-only market.

That's a broad transition on a large SAM, and there's very many simple reasons why this is happening. One is the technology needs are very difficult. The number of suppliers that have access to this technology that can credibly deliver the scale to Tier 1 customers is reducing.

The third element is really around the IP is really, really very hard to go get. Again, we acquired a company in this landscape, combined them together, that is the value proposition we are driving here today.

Data centers, I mentioned to you, to us is a relatively new growth opportunity, a greenfield opportunity. The combined outlook on the networking, the storage and the server markets come in together, and we are seeing very significant activity, and we expect to grow on a smaller revenue basis, very significantly in this area.

The story is very similar to the value proposition that you've seen before: the integration, it's performance. And in this regard, it's not just bandwidth alone. Latency is an extremely critical aspect of the data center market. There isn't a technology provider that brings the value proposition we bring in at the bandwidth, the latency and the cost and integration capabilities that we are pulling together.

The SSIT or TSP technologies that we enable is really not only just a prototyping element, but hitting a home run with regards to both the OTN market and also, with the data center market. And we expect to see that technology proliferate beyond just a prototyping into really be wired communications mainstream.

The summary on the wired communications market, we are seeing our SAM expand from 6 billion -- 65-nanometer, $1 billion to $2.4 billion in the 40-nanometer node and increasing significantly to $5.2 billion in the 28-nanometer nodes. And we also expect simultaneously our penetration into the logic IC BOM consumption of the top-tier customers grow from 12% to 25%. And the combination of that gives us a fairly bullish outlook with regards to our growth horizon in the wired communications market.

Switching gears into the wireless segment. By 2015, the SAM here is $2 billion. We have seen a lot of growth in wireless over the past few years. We continue to see a strong bullish outlook. We are today actively engaged on every major program globally with every major top-tier vendor.

3G is really still 27% penetrated the global population. LTE is in a very, very nascent stage, and it's less than 3% consumed today in the market. We do believe that the shipments for wireless will continue over the next 5 to 7 years, and they're going to be predominantly oriented around 3G and LTE. We are in advanced stages of the LTE-Advanced and future technologies so 4G-plus technologies are key customers. And we have a strong growth outlook for all aspects of this market on the radio, on the baseband and the mobile backhaul.

If you look at our fit today in the market, we are one of the leading suppliers in this landscape. We have been for quite a few years. We have had our challenges with our 40-nanometer technology with regards to radio. Radio has really represented a lot of our revenue basis. And I'll walk you through where we are at the 28-nanometer on the next slide.

The radio footprint, we're one of the leading suppliers. And on the baseband side, it's a growth opportunity for us. We fundamentally provide connectivity and L1 acceleration. And today, the backhaul market has really been a low-performance market, and it's essentially been an ASIC or an ASSP market. If you look at the consumption for the top 10 customers, our penetration in the wireless is fairly significant even with that nascent footprint today. It represents around 24%.

A few key things have changed, and over the next few slides, I'd like to highlight to you why we really think we have come back in a very big way and why wireless is going to be strong growing market for us. Number one, the radio, I mean, as I mentioned to you, we have had our challenges with the 40-nanometer technology at the radio platforms of the top key OEMs. We have almost won every major radio platform in the world today.

And that is fundamentally riding on the back of the Kintex product line. So Kintex today has become the de facto wireless solution radio of choice. And we have strong hopes for the product. The success to date is exceeding our expectations, and we do have a longer-term outlook to really building upon that momentum as we look at the next few quarters.

On baseband, there's a growth opportunity. There are many challenges the industry is facing today with regards to baseband. The baseband is evolving from essentially a DSP-oriented architecture to a DSP/packet-oriented architecture. The wired and the wireless roles are beginning to collide in the baseband. We are one of the unique companies that actually has a strong foothold in both areas, and we are leveraging that value proposition into our key customers. And we do think that we have a very compelling value proposition we're going to provide to the baseband opportunities going forward.

I'll talk you through this slide on this to amplify the story more. As the backhaul market moves from 1 gig to 10 gig, as customer architectures evolve, as the wired and the wireless worlds collide most heavily in the backhaul side of it, a lot of new growth opportunities for us. And performance is now becoming an extremely important criteria, while cost and power continue to remain significant problems.

We are, again, bringing a very unique value proposition there, both against the incumbent ASICs, but also against the new competition for the multi-core vendors, and we are very winning significantly against both constituents.

Our value, and this is the simplest message that our customers communicate to us, they see very little to no competition to Kintex. It is really becoming a strong foothold in wireless. And in the past 9 months, we have actually had an amazing success, and we really have a very strong outlook to what Kintex is going to represent to wireless market.

We architected this product line specifically for the needs of this market, and that's beginning to pay off in very strong dividends for us today. It brings 2 simple value propositions, which are easier to put on PowerPoint but very hard to really make happen in silicon. One is we are providing the lowest total power solution today. There is no competitive offering to that position that we have today. Second, we are bringing a disruptive price-performance per watt solution. This is not just against our immediate competition, but this is against incumbent -- the ASSP-oriented architectures is our competitors and also, against new multi-core architectures.

On the radio head, this is an area where there's a tremendous amount of innovation, tremendous amount of cost pressure and tremendous amount of churn with regards to key customers. We are becoming the de facto leader in this area with regards to radio platforms. And as we go through more time, I do think that Kintex value proposition will be heard loud throughout the wireless industry.

The story is one of integration, dramatic integration story on 8 chips to 2, along with that comes the benefit of the performance that we brought to the table and out of cycle, very significant cost and power reduction. This is an iteration that actually is going to really make a very big difference to our customers as they manage through the CapEx issues and their wireless challenges. There is no answer to Kintex. And that's something that we are very keen on ensuring that we continue to deliver the value to key customers.

With regards to baseband, our fit today is really one of connectivity and L1 acceleration or an FPGA as the core processor to the DSP farms that our customers use, another area where the wired and wireless markets are colliding. We are -- with our acquisitions, with our technology making a very significant value add, and as key customers are re-architecting the system, we are actively invited to be advisers and actually be an extended part of their R&D team as we really elevate ourselves from a supplier to a strategic partner and really enabling this key industry transformation and helping our customers through a very, very difficult phase of enabling the success for them.

The story is again a very consistent theme, and you almost get the repeat of about the same value proposition. The value is the same, the numbers and the details and the context are different. It's integration, it's performance, its cost and power. The baseband market is also heavily a software-centric outlook. And one of our key acquisitions with a company called AutoESL is really bringing in, again, a unique value proposition on how we enable a high-level abstraction language and enable a software community to touch FPGAs and derive the benefit of what our technology brings to them.

The mobile backhaul space is an area of tremendous growth, and much like the data center market, a greenfield opportunity for us. As this market gets into a performance-oriented outlook and as the 1 gig transitions goes to 10 gig, there is many, many different unique proprietary architectures on those. Every OEMs architecture is quite different than the others.

The incumbents have been mostly an ASIC or an ASSP, and we are today winning very heavily against the ASIC and ASSP incumbents. We are also actively competing with multi-core SOCs that are actually actively trying to get into this space, and our value proposition, primarily led by 2 product families Kintex and Zynq, are really creating a unique value proposition that our customers really like.

One key value that we bring to the company is really one of we are an enabler of our key customers. And unlike ASSP competition, wherein customers worry if their roadmap is really going to be held captive to an ASSP company, we are in the active business of enabling and helping our customers retain their IP internally. That is a strong value proposition that we continue to drive and help our key customers through driving very significant challenge in the wireless market, through our technology and our economic investment in this area to drive value to them.

A summary on the wireless market. Again, as the SAM from 65-nanometer has gone from $1 billion to $1.5 billion in the 40-nanometer and moving up to $2 billion in the 28-nanometer node. And our outlook is to really grow our penetration of the top-tier customers from 24% to 35%. Again, a strong growth outlook both wired and wireless. And as I mentioned to you, we do think that wired will continue to grow faster than the wireless market.

In summary, we are all-in from a communications context. We do think that the initial market success validates to us that we have the best portfolio in the industry, early days, and we do think that there's a lot of growth ahead with the 28-nanometer technology in the portfolio that we have. A dedicated business unit is really helping us a lot today in enabling key customer partnerships and accelerating our opportunity to grow.

The IP and the design services are enabling us to actively penetrate and increase our SAM expansion opportunities with ASICs and ASSPs. And the value advantage is that I've highlighted to you and the numbers that I've shown you leave us to believe that we have a clearly leadership with 28-nanometer today, and we continue to want to build upon this momentum.

Thank you. And with that, I'll transition to Jon who'll walk you through the financials and the outlook.

Jon A. Olson

Thank you, Krishna. Up to this point in time, you've heard us talk about the opportunity, about our early success against that opportunity and the confidence we have in how well the 28-nanometer, 7 Series and Zynq families are doing in the early days. My objective today is to talk to you about a little bit about the business model, and I'll get into the some high level and some details around that, talk to you about end market growth and also, about our view of market segment share shift that's about to happen.

So we're running the company today much like we have all along here for the last 3 or 4 years, and that we have an intense focus on developing products that can expand the SAM and drive the top line growth level, while still maintaining a level of financial discipline. And I'm going to talk a little bit about that as we walk through some of the slides here today.

We do believe that we'll outgrow the semiconductor market in the neighborhood of 2x. So just a data point between 2007 and 2011, if you look at the WFTF data for ICs without memory, that's grown about 4%, and PLDs have grown just about 8%, just a little under 8%. So clearly, the opportunity has been monetized at some point. But we think we're at a platform now to grow even more out in time and monetize that in a way that hasn't been seen before in the PLD industry.

We continue to keep the same basic business -- fundamental business model around gross margin and spending. We do believe in operating leverage, and I'll talk little bit more about that, that will grow spending over time at a lower rate than revenue. We continue to focus on cash generation. I've got some good data to share with you about that. And lastly, we do believe in a capital allocation that allows us to generate lots of cash, but also return much of that cash to shareholders.

So going back to one of Moshe's slides around the opportunity for us and how much we think ASIC and ASSP sockets that were traditionally there in the last generation, we're going to be moving to PLDs over the next generation. It's really around how much we've been able to enable that growth through this integration play, as well as adding new features to our product family that allow us to step up and do things we haven't been -- haven't done before. And I hope the examples that Krishna and Vin have provided to you, which are just a very small number on a grand scale that we really have really communicate to you how well we're doing in that particular market expansion capability.

So what does that really mean from us from an overall end market growth perspective? So at this point in time, we're not declaring our growth any greater than we have in last year or so on, in total an 8% to 12% 5-year CAGR top line growth. But there are some changes versus what we told you last year, and those are indicated by the arrows on the right-hand side.

So in communications, we've moved that up slightly, and that's really on the basis of much about what Krishna talked about. The addition of our play in the data center and then more broadly, on the technology to move towards 100 gig and then 400 gig, and the unique opportunity we can play with our very high-bandwidth capability of SSIT-based products at the high-end of Virtex. So these are the things that move us into a space because of that bandwidth that we haven't been there before.

Industrial and Other continues to grow, in our mind, in the 8% to 12%. This is still an underserved market. There's a lot of legs left in this particular market in the broad-based medical area, surveillance and security and also, in aerospace and defense. Moshe showed the drone example. We're very penetrated into avionics and the visual capability required by those kinds of surveillance and drone capabilities, and we do believe that aerospace and defense will continue to grow for us despite what's going on today with a view of cutting the defense department budgets.

In the consumer side, we scale that down a little bit, and that's really from -- on the basis of the pure consumer play. Now it's not a tremendously large part of our business, and we still are very well penetrated into displays and the high end -- high resolution technology televisions. We were the pioneer in terms of allowing the 3D TVs to get introduced at the rate that they have been.

But unfortunately, that market hasn't grown. And I think as much as our expectations -- the end market hasn't grown as much as our expectations were a year ago.

On the other side of that, automotive has a higher growth rate than we thought last year. And the 2 of those don't quite balance each other out as automotive is small going to be very large. But net-net, we're down a little bit in consumer and automotive. Data processing, pretty much the same, the same kind of that about where we play in the pure computing perspective, as well as storage.

Now let me switch to market share. And what this slide depicts is a relative comparison of share between Xilinx and our nearest competitor over that time period that's listed here actual through '11 and then estimate using our published guidance from both companies. So this is a relative set of numbers, and as you can see here, in 2009, '10 and '11, you see the decline of our share in the PLD market on a relative basis, and that's driven by our absence in the high-volume 65-nanometer nodes -- that we do not have a product family there. That node is extremely successful node and then being late to the 40-nanometer product offering.

Now having said that, we do have a very strong 45-nanometer product offering, and I'll talk a little bit more about that a little bit later. But we believe 2012 is the inflection point for us. And I want to explain to you why I believe that or why we believe that. So I'll spend a little time on this page and make sure that I can make it clear as possible.

This is Xilinx's share relative to our nearest competitor by node for calendar year 2011. This is our estimates based on information that's -- public information, as we track it from a market research perspective. The reason I chose to use the calendar year is because, when you're talking about shifts of relative share percentages, it's about the change in the run rate that matters. It's not about what the cumulative numbers are, it's change in the run rate. So I want you to think about that as we walk through here.

So 90-nanometer, very successful, approximately 60% market segment share. At 65-nanometer, we were approximately 55% in calendar year '11, and again, we didn't have this high-volume based products.

But this generation of technology was kind of the first breakout generation, if you will, against accelerating ASICs and ASSPs. And there's still a lot of leg left that in particular generation for us. There's a tremendous number of designs that are just coming to market in the wired technology space, as well as aerospace and defense, which is usually a lagging adopter of technology by the time that they get to production.

Those still have a tremendous amount of design wins yet to materialize and be monetized into revenue. And typically, what happens is in the tail portion of the node, it's usually the high-performance products are the ones that generally fuel some of that cash flow because of aerospace and defense.

Then at 40-nanometer in calendar year '11, we estimate we're at 35%, approximately 35% market segment share at this point in time.

Now we were late to market, we have a tremendous number of design wins, particularly with the higher bandwidth capability, high-end 30s based products that are just now starting to get into production mode. And our belief is, on a run rate basis, we will start moving the 35% towards the 50% range.

Now the one thing that's going to really help us get there is the fact that Spartan-6, high-volume, 45-nanometer product family, there's really not anything there from our immediate, nearest competitor there. And those designs are just now starting to get real traction and mounting up in terms of revenue.

So we clearly believe that, that we're kind of at a low point, if you will, and as these things kick in, the run rate share will move towards 50%.

Now, as you've heard our level of confidence around 28-nanometer, it's like well, okay, greater than 70%. We've seen -- we've shown you the numbers. But also, I want you to focus on the fact that there are categories of products where our nearest competitor do not have a product or they're not playing at this point in time. The very high end of Virtex-7, high-bandwidth, where there's lots of logic cells needed, where we service the competition against large ASICs, in terms of the SSIT technology and also, then in Zynq, where the competition is behind us, we believe in the neighborhood of a year in that particular area.

So I have one other proof point. It's a little bit of an ugly slide for us, but it does show that we've lost significant amount of share in the overall high volumes node, so this is all high-volume products shipped in these particular years.

Going back to our original Spartan, going back to the competition, the original cyclones, et cetera. And in 2011, it was clear that we have gained market segment share back from them. So if you're looking for proof points of this inflection point and why we feel so strongly is it's -- some of it's already happening, or has already happened and moving in that direction. And other things are right on the edge of happening. So we're very convinced that 2012 is our inflection point for share.

So now let me talk a little bit about some of the business metrics. Over the last 5 years, we have continued to increase our cash flow as a company, or our annual cash flow generation. In fact, in 2012, we're forecasting that this will be a record operating cash flow number for us. So despite the revenue being down in the neighborhood of 4%, we believe our operating cash flow will be up plus 5% or 10% on a comparison to our fiscal year '11.

And that's really a testament to our drive towards making an efficient business model, but also really focusing on working capital.

And then over that same period, we've returned greater than $2.5 billion of cash to our shareholders through a progressively increasing dividend program, as well as an opportunistic buyback.

If you look at the -- if you counted the dollars on the previous slide, we've generated an operating cash a little over $3 billion during that 5-year timeframe, and we've returned to $2.5 billion to shareholders.

So we have a strong commitment to continue to, as we generate cash, to return cash to shareholders.

Now let me talk a little bit about our gross margin. Over this time period, we've also improved our gross margin. Through this cycle, we've been able to maintain 2 business cycles, which actually just kind of 2 represented in there. We continue to be able to generate gross margins at a very high level. And in FY '13, or fiscal year '13, we estimate that our gross margin will average 65%, about the middle of our model. And that's really because of a variety of factors.

But we've been launching and taping-out products at a very rapid rate. We've been working through the introduction and the early ramp with results, although some yield penalties, et cetera, but we've been very rapidly moving up -- to up the yield curve, to improve yields as well.

We also believe, because of our competitive advantage and time-to-market, that we have some level of pricing power. So we've pointed out several times today about between Zynq and the high-end Virtex family, there really isn't any competition from an FPGA perspective, and that gives us an opportunity to hold value, if you will, more easily than the past. So we're very positive about that. And that should generate, overall, a more profitable portfolio.

And then, lastly, we continue to have an intense cost focus around the company to look at older generations of products and figure out how we can take out dollars and how we can lean out our model. And we've been quite successful in doing that. We have a very organized program in order to continue to ring those things out.

So going back to the yield comment a little bit. Just let me give you an example, or show you how well we think we're doing with respect to our lead product, which is the Kintex 325. That was the first product that we birthed in March, and it's the lead product. And this product is our -- will have initial production status this quarter and then broad production in the June quarter. So we're moving quite well along the progression of getting products to production.

So let me explain this slide to you. The dotted line is our internal yield forecast for the 325T. So this is what our cost models indicated at the beginning of the generation that we've said this is what we're trying this is where we're trying to get to. And this is in conjunction of working with our foundry partner and -- in creating this particular yield roadmap.

As you can see by the blue line, that we're significantly ahead of that yield curve. And that's obviously, a great thing for cost, and it's great thing that our lead product, which has a lot of the fabric that shows up in other Kintex parts, et cetera, that we really have worked out the issues there and we're moving along very well.

So even though the 28-nanometer generation is not significant dollars at this point in time to move the needle on gross margin, this is one of the reasons that we think we'll be able to have gross margin move back towards the mid and then, beyond that, to the, more towards the upper end of our model.

So now let me talk a little bit about operating expense. So I have said before that we're focused on leverage and that we're, over the time period of running the company, that we want to make sure that we do spend less on average, in terms of a growth rate than our revenue growth rate. And if you look at the 4-year CAGR with, again, using these 5 years of the, as the proof points, but the base period being FY '08, we've grown operating expense approximately 2.5% during that time period.

At the same time, revenue has grown 5%. So in any given year, the numbers could look different and there are trends from technology that make it different, but we measure these things over a longer time period. Because as you know, we invest now and the payoff is several years in the future because there's a very long gestation period to get a return strictly because of the dynamics of our business and the markets that we serve.

For FY '11, this is our estimate for OpEx, $870 million, a 9% growth year-on-year. R&D going up, on high, a little more or more than SG&A. SG&A being held pretty much to, I would say, it's more like normal inflation and we're certainly not structuring to add a tremendous amount of headcount and resources in the SG&A area.

R&D is going up at that rate, primarily for 2 reasons. One is we had a number of assets that move from one year to the next year. So again, we're in the, driving to the peak of our tape-out ramp as we go through this quarter and the next quarter or 2. And some of those moved out, as we realign some of those to more easily or that should be better compete, if you will, against some customer desires of when they needed certain products with certain features. So that changed the calendarization. The second thing that's driving this is our move towards the next generation product family and generations.

So if I move to a page where I try to show you the balance here of what generation and what nodes and where we're spending our money, as R&D being a 100% and these percentages showing the percent by generation of what we're spending. As you can see here, a tremendous amount of development and then execution against that development roll out in tape-outs in FY '12 for us on the 28-nanometer nodes. And then we finish that up in FY '13, which is essentially rolling out the rest of our family members.

We're also putting a significant amount of additional resources and time and focusing on our next-generation product family. So we're clearly starting to roll off after the first half of the year, and then go into the next generation in the middle to the second half of the year.

So then looking at, say, general summary of our financial metrics. Gross margin, again, I said we expect gross margin to be in the middle of the range for next year. We've given R&D and SG&A ranges here again, that tie to that $870 million number that we talked about. Nothing particularly different in intangibles or other expense in this particular case because we have interest to expense and the rates for rate of return for our treasury investments are very low, as everybody knows. No change in the tax rate. And CapEx in the $60 million to $70 million range.

So for the grand summary of today, we are very convinced that we're on a path to have new levels of displacements against ASICS and ASSPs. And I hope you feel the same level of excitement and confidence that we do at our company about all these successes. We try to be as transparent as we possibly can by showing you the information that we see today.

And one of the biggest advantages that we have is this: A capability for integration. And integration that then allows power and cost reductions for the end solution for our customers. Because that's what it's all about. Device to device, you can have differences. But if you're really solving the problem for your customers, that's what matters. And we think we have a generation of technology that really accelerates that.

We continue to have a strong commitment to shareholder value, to drive that value through cash generation and invest in ourselves in order to grow the business and then return excess cash to shareholders.

And lastly, we talked about this all-in mentality that we have at the company. It's really a moniker for us to understand and communicate how aggressive we think we've been in terms of designing product families that have a real value to the customer, but our own personal commitment to make sure that those value advantages get monetized and we have the business success that goes along with the technology successes.

So with that, we'd like to move into the Q&A session. And so I'd like to invite my friends and colleagues up to the podium. We do have roving microphones. We have lots of time for questions, so please don't get frustrated if I don't call on you. And we're going to split the room.

Question-and-Answer Session

Shawn R. Webster - Macquarie Research

Shawn Webster from Macquarie. So you spend a lot of time kind of talking again about how PLDs are displacing other chips within the systems. You reiterated your long-term 8% to 12% growth rate. One of the things, just doing some simple math, you can correct me if I'm wrong here, is you talked about content increases, like, say, in the wireless base station, for example moving from 25% to 34%. But the BOM was dropping 30%. So if I do simple caveman math, that would suggest that your dollar content is roughly stable in the system over time. So what I'm trying to reconcile, and then the other piece of that is your revenues have grown roughly 5% over the last 5 years. So how do I walk from what's been happening in that math to superior revenue growth rates for the industry because of PLD displacement? Or is it the case that you're simply just destroying the ASICS market?

Moshe N. Gavrielov

Krishna, why don't you start?

Krishna Rangasayee

Can I get an easier question? It's an excellent question. And I almost think there's 3 parts it. And if I sort of paraphrase. There's one element of, historically our gross being 5%. That's one element of it. The second one is we are increasing our share, the BOM cost. But customers BOM cost itself is dropping. And where the math add up in terms of our growth and what the trends are in revenue versus BOM cost reduction. And the third one is really around why do we believe we have a strong outlook on our -- are we really competing against the ASICS and taking them off? So I will try to answer 3 parts of it. The material change is really around 28-nanometer being a very big step function in terms of transferring to the Delta and what they bring in total cost of ownership. An overly simplified outlook is we are seeing dramatic drop off in ASICS and ASIC design starts with our key customers. So that aspect stuff that we see. There is different connotations to what it means in the wired communications market, different connotations in the wireless market. So we clearly are growing at the expense of ASICS. We also are beginning to eek into the ASSP aspect of it. So those 2 represent growth horizons. What we have today demonstrated to you is really our early days in this -- winning against the ASICS and ASSP competition. And the winnings we have today will translate to revenue. And that will materially represent, if you will, a higher revenue growth rate than our traditional 5% growth. So that's one aspect of it. The third element is we continue to actively monitor customers BOM cost reductions and their operating expenses and CapEx reduction requirements. We believe that we are, yes, eating into our immediate competition. But there's also growth in the market. And while the overall market suffers from macroeconomic constraints, individual key customers are winning at the expense of somebody. Some other of their peers, if you will. So we are actively keen on partnering with the right leaders in the market and we want to enable them with the technology, and yes, we are growing at the ASICS and the ASPs. And yes, we do plan to grow at a higher rate than 5%. So hopefully, that answers your question.

Victor Peng

And wireless is the area where there are the highest pressures, right? In other markets they tend to have less of a challenge? So if you look at wireless, of all of our major markets it's actually the one where in order to keep stable, we actually have to run faster. And we are running faster, and we're running at Kintex with design with that in mind in order to provide solutions, right? And Kintex is replacing parts that in the past were implemented the high-end of the product software. In most of the other markets the dynamics are not quite as significant as that. Wireless is about 25% of our business. Now and it's the one where -- which clearly is one of the more challenging one.

Jon A. Olson

I think, the, on the wired side, I would say that the penetration we're getting because of our, of the tightness we have now with the IP acquisitions that we made, is driving more value for us, more dollars from that. Because I'm very -- watching Krishna on monetizing these acquisitions that were actually just not, just buying them and giving it away. We're actually monetizing that, and that's helping us to get that penetration, particularly in the wired segment.

Nathan Johnsen - Pacific Crest Securities, Inc., Research Division

Nathan Johnsen of Pacific Crest. I have 2 questions. One the software platform. You guys highlighted a new software platform coming out this year. I was wondering if you could talk about what the incremental feature or the changes there are and how that compares to competition? And then secondly, just on the acquisition front. You guys have talked about a few of the key acquisitions just recently. Are there acquisitions out there that you guys are looking or any key strategic holes that you see going forward?

Rick Muscha

I'll see on the software question and probably Moshe wants to handle the acquisition questions. So I think your question was what are the key technology and/or value propositions that next generation software platform, relative to where you are today. And we're not doing a product announcement today. So I'm not going to go into excruciating detail. But at a pretty high level, if you think about the evolution of the software for FPGAs, one of the key metrics is how quickly can you compile the design? And as the designs get larger, the compilation time goes up. Both us and our competition face that increasing complexity and that increasing compilation time. And what customers have told us, over the last couple of years, is that we've caught up to our competition with our base platform and that our compiled times are no worse than the other guy, but that we both need a dramatic step improvement. And this new software platform offers significant increases in compiling -- decreases in compiled time. And it's architected for dramatically larger designs. So whether it's a large design or whether it's a small design, the design will compile much quicker. And we're talking about numbers of 2 to 5x faster. So instead of having a turn per day and waiting around for the compilation to complete, a customer could have 2, 3, 4 turns per day. And we've seen many examples already where customers have benefited from that. And these are customers that are using the competitions platform. So if you can provide that kind of dramatic productivity improvement, you can have a tremendous impact on the customer's ability to get their product into the marketplace. So that's one -- turns per day. I talked a little bit about that, and the underlying technology that achieves that, we'll be talking about as we roll that out into the marketplace. And as I said, there are over 100 customers using it. And they're using it on large designs. They're using it on medium-sized designs. And they're benefiting enormously. And those that have made the transition are very happy with that. The other major thrust is around the ability to integrate all these elements, the IP and the, all of the traffic management IP and all of the third-party customer IP, being able to integrate that very, very quickly into the platform, be able to verify it, be able to connect it together, be able to reuse it, that's a huge productivity improvement. And in order to do that, you have to architect your database and your data model from the ground up to be able to handle that kind of integration. Finally, you need to adopt a set of industry standards. So we'll have more detail. We'll start talking to that. We'll start talking to the press and analysts starting in the April timeframe. And that product, as I said, will be rolled out this summer. So I think, I hope I answered the software question.

Nathan Johnsen - Pacific Crest Securities, Inc., Research Division

Without announcing the product first.

Moshe N. Gavrielov

And this is more and more important because the level of integration is just going through the roof. So whereas in the past, the FPGA was used for a small portion of the design. Today, the FPGA is becoming mainstream and sometimes we need design, sometimes, the separate set for or ASP. But if you look at that, it requires a totally new system, which was architected from the ground up hence the huge investments we've made. Plus, we need to make it faster to do. With regards to our acquisitions, we have done, over the past 15 months, maybe less, 4 acquisitions. And I think the total is under $100 million.

Jon A. Olson

Less than $100 million.

Moshe N. Gavrielov

Yes, we don't break it out. So we report cap, right. So they're hidden there. So we've consumed them. We haven't taken any special expenses. So some expense, the run rate actually, to some extent, relates to those acquisitions in terms of the initial cost to -- one of them was to provide breakthrough capabilities on the software side and actually will be mainstreamed as part of the new, as yet to be publicly divulged product. That's probably the worst kept secret in the world. And the other 3 were wired communications significant differentiators, which are key to enabling us to replace ASICS and ASSPs in the wide comms market to the -- I'm delighted with the results there. We actually, I think, have 100% retention of all of the people, which is important because these are a little technology tuck ins, to be totally honest. We have the capacity to do more, if you look at our balance sheet, and to the extent that we can find more, we will do that. The results so far as being good. If they exist, we're interested. In some cases, they don't exist. And the integration is being smooth and all of the people are around and they're helping our business in a big way. I think Krishna might just sort of highlight some of the benefits there.

Krishna Rangasayee

Yes. Is the mic on? I think, as Moshe mentioned, the big transition for us in communications from 2, 3 years ago, where we are now, the customer's architecture are ASICS and ASSPs first, and then in FPGA. And today the big transition, based on all the data that we've provided you with today is FPGAs increasingly first. And if I cannot get it done in FPGA, then I'll look at it in ASIC or in ASSP. And in that transition, as we solve the technology complexity and the business complexity and cost and power complexity, the increasing gap is around system low IP. And that's the key area, where we acquired these 3 companies. And almost every one of the key design wins we have today leverages that investment. And we are beginning to see early signs of that helping us a lot more going forward.

Lori Owen

In the back.

Christopher B. Danely - JP Morgan Chase & Co, Research Division

Yes, it's Chris Danely. Just a question on OpEx. Jon, I think you talked about some masked costs being pushed from '12 to '13. So was that $40 million in savings that you guys have from basically last year this time to this year? Was that all masked costs? And was that all a pushout? And then also, I think in one of your slides, you had OpEx peaking somewhere around the middle of fiscal '13. Should we imply that OpEx for fiscal '14 could be flat to down, versus fiscal '13?

Jon A. Olson

So the first, the amount of, the amount of shift, really not going to talk exactly how much. But that 50 or 40-ish or so million dollars that were declining, a good portion of that was the masked shifted over the second half of the year, moving that. Not its entirety, but I would say, a reasonable portion of that. And I'm really not -- I'm reticent to make any prediction about '14 because last year, I sat here and made a prediction about '13 and things came out different. And I think, really, it's around when the technology, the timing of R&D may turn out to be more when the technology is ready for the next generation because we're committed to being there first, and where that falls with respect to our fiscal year kind of thing. So it's really hard to answer the FY '14 question, just because so much of that is when does the startup of next-generation 20-nanometers, et cetera, really happen? So yes, I have nothing.

Moshe N. Gavrielov

If I can just add some color because Jon and I obviously have had significant discussions. We're ahead. We're way ahead. We want to exploit that. We could artificially dampen that, but that would be the wrong decision. And we had a heated discussion. We shared it with the board. We believe, not only are we ahead in terms of having technology earlier, we believe we have the strategic advantage due to the decisions that were made. We want to make sure that we capitalize on that as a result of actually some expansion of the product offering, the 28-nanometer, which is reflected in these numbers. And there were no -- it would be a shame, I would say, it would be criminal not to bank on that. So that's sort of the rationale.

Jon A. Olson

And if you go back, for those of you who are here and from last year and when we talked at out Analyst Day and we talked about the profile of R&D and the little house thing that I built or whatever to show you was going to go up in '12 and down in 13. I went back and looked at the basis for that in terms of the 2 years in total. And we're only a few 2-ish or so percentage points higher in total. It's just a matter of where the distribution is between the 2 years. Despite the fact that Moshe did assert that we decided to really capitalize with some additional things in 28-nanometer during this timeframe. So I still think we're being fiscally responsible, as you can tell, we have vigorous discussions about this, back-and-forth. You want the last word?

Moshe N. Gavrielov

Well, this is what we had Valentines Day together. Jon and I, in the restaurant yesterday. So it's like a discussion with my wife. She always has the last word. So yes, dear.

Jon A. Olson

Thank you. That's a visual for everybody.

Christopher B. Danely - JP Morgan Chase & Co, Research Division

Just a quick follow-up on what Moshe said. So you guys, keep talking about we're ahead of 28, ahead of 28 in 70% of the wins. Your competitors put out a revenue number for them at 28. Would you care to share us, share with us your revenue number at 28?

Jon A. Olson

It's going to be greater than 70% on a run rate basis. No, we're not talking about any actual numbers now, et cetera. The numbers from both companies are small. And I think you're going to see some acceleration on our part. We talked about greater than $10 million in June. And that's the number we're giving you. 28-nanometer greater than $10 million in the June quarter.

Christopher B. Danely - JP Morgan Chase & Co, Research Division

In terms of the market share projection you outlined, giving 200 bps of share this year and then another 100 bps of share next year. I'm assuming you did a lot on math to kind of come to that conclusion. I was wondering if you could maybe bucketize for us over the couple of years how much share you expect to gain in terms of your share improvement at the 40, 45-nanometer. They kind of bump up from 35, 50. You talked about as one bucket. Second bucket being how much you expect in terms of your leadership on 28-nanometer on the conventional platforms, and third bucket being how much you expect from your new unique products, like Zynq and SSIT, where you competitor doesn't have an offering. And I have a follow-up.

Jon A. Olson

Okay. So first, so to make sure that you're -- we're clear on that -- on what the basis for the graph was. It's not a revenue forecast for us. It's not a market growth statement. It's really on a relative basis. And we think, in most environments, that whatever the market growth is and whatever the revenue for us might be, that we'll end up with that share. So I'm trying to make sure that people aren't out there backsolving for the revenue forecast that we had because we've looked at it in a number of different environments. So first on the 40-nanometer, as I said before, we are just getting into, we'd say, is a significant ramp in the high-performance 40-nanometer capability. And if you look at the growth rates of 40, the change between our 2 companies over the last several quarters, on just, on that note, there really hasn't been much. So -- and I am very confident of where we have our upswing coming in the next several quarters. That's just on the high-performance side. And then when you layer on that high-volume category, these are the things that move us towards that direction and get us to 50% on a run rate basis. I don't believe that we're going to get to that 50% immediately in the next 3 quarters, et cetera. This is a, as the generation matures, in terms of as it gets to its peak. That -- those are -- that's the timeframe, I think, we were, we've been thinking about. So it's not quite as influenced in that next year as maybe you come to that conclusion. And then on 28-nanometer, I think we'd be pretty disappointed if we didn't see out of the shoot as our products become production worthy, a pretty significant movement up in terms of revenue on a relative basis than in previous generations. And trying to parse it between how much is in the SSIT plus Zynq versus the other, I don't know that I have a very good way to do that. It is pretty clear that one of the categories for SSIT is emulation, and that typically is a leading kind of thing versus a lagging. It is continuous, but it will jump up faster. So that might indicate that, at least in the early part, we might get a little more from those areas. Zynq, on the other hand, is going to be a longer, slower growth because we're targeting a certain set of markets and applications. Ecosystems have to be developed. And this is a long play towards the integrated SSC technology for us.

Victor Peng

Maybe if I can jump in the SSIT. Jon did allude to of course, in the very early stages, emulation prototyping. But it's also the platform for which we get the highest 30s bandwidth, total I/O bandwidth in general. So it's also definitely very strong in communications. In Aerospace and defense, as well as test and measurements. So it's actually pretty broad, but agree with Jon [indiscernible]. And in fact, many of those customers in prototype and emulation are building the next generation communications devices, right? And these are massively complex designs that they're in process of development.

Christopher B. Danely - JP Morgan Chase & Co, Research Division

I just have a quick follow-up for Krishna. I hope, I promise I'll be quick. On the wireline exceeding wireless groups you talked about, are you making a call on wireline equipment? Your customer's equipment exceeding wireless equipment? Or are you making call on the FPGA portion of that?

Krishna Rangasayee

More of the latter. The simple answer is more of the latter. But we do think that there might be some macrolevel changes. But to a certain degree, I think there are plenty of pluses and minuses. And we aren't necessarily banking on the macroeconomic alone for our growth horizon. While that will help and they're leveraging that or taking advantage of that area of growth, the majority for growth is probably going to be coming from displacements.

David Wu - GC Research Ltd.

A few little ones. Some years ago, you guys had the power PC hardcore and a lot of us got very excited about that, and your competitor didn't have a hard-core. Now you have Zynq. What's different now that Zynq seems to be taking off whereas before the hard-core got quietly dropped?

Moshe N. Gavrielov

Okay. Let me answer that. That was the easiest decision I had to make. Basically, the power PC is an attractive architecture, but this is an orphaned one, and doesn't have broad applicability. And so we decided to make 2 changes, to move to ARM, which by far is a, the most popular architecture, has the best ecosystem around. And then we also decided, as opposed to the previous generation, where it was embedded in the high-performance fabric. This is now embedded in the low and midrange product offerings. So we believe it has much broader applicability, both because of the architecture and the choice of the fabric that it's embedded in. And actually, all of the previous generation customers who have adopted it are enthusiastically making the transition. We expected for there to be a stronger backlash. Actually, there was almost no backlash at all. And so this was a very easy decision to make. And with regards to the number of design wins, I doubt that we have that many of design wins in the previous generations. And we have them now even before we have the product.

Rick Muscha

Yes, I think to Moshe's point, the first generation, first and second generation were really for people that wanted to build their own microprocessor subsystem. And it -- that was a very interesting thing for people in the wired communications market and the aerospace and defense market. But there's a limited number of people that can pragmatically do that. If look at the Zynq platform, it's designed to be programmed from day one by the software development team. And it's very, very interesting, we have seen quite a number of examples now where customers have been developing their application on our emulation platform and the software teams are well along in terms of migrating all of their PowerPC and all of their non-ARM-based software onto that emulation platform. And actually had a conversation with one of our industrial companies yesterday about where they are in that transition, and they are well along. The other interesting statistic is how quickly, when people get the silicon, they're bringing up their development system. So we shipped our first silicon to our first customer. In 41 minutes, they had the application up and running. And the software team was developing their application even further. So the fact that it's a processor-based system, and allows the software team to get started on day one, that's a dramatic shift from the architecture we had before. And it's opening up a broad range of markets and applications.

David Wu - GC Research Ltd.

And just another one, for Jon, if I may. Jon, your long-term gross margin, 64% to 66%, I believe. You also talked about how things are going nicely at TSMC and your yields are better than you expected. Your competitor has better than expected yields. They're 70 something percent. They expect it to come down intact about 67%, you're at 65%. What do we expect, given that both of you at the same place, essentially same cost structure, that those gross margins will be the same? I -- you'll do better than you expect toward there, or come down further than they expect?

Jon A. Olson

Well, I can't speak for what's going on at our competitor and what their trends are. I can only hear the same thing you're heard. But from our perspective, we're certainly happy with where we are in the yields and we would anticipate that 28-nanometer gets to be a bigger portion of our business to have more uplift than drag. I'm not raising my number. I don't, at this point in time, I guess we'll just have to see. But I'm still confident that we'll end up operating in the middle upper end of our model when all said and done throughout this generation of products.

David Wu - GC Research Ltd.

But to your knowledge, there is nothing special about Xilinx that would have a higher cost in the long run?

Jon A. Olson

I mean, our -- the design of our products could be different and the cost points could be different, et cetera. But the yield is not going to be necessarily the one thing or the other in terms of the maturity. Consistency is pretty much gets to maturity equivalently for everybody.

Moshe N. Gavrielov

So just one data point there. HPL is a subset of HP. And it's broadly used. HPL is broadly used. But any learning from HP, which is the more complex process, will most likely benefit the HPL yields, which at this point, are ahead of where we expect it to be. So I don't think that strategically, we have any disadvantage, I would say. The opposite is true, at least on 28-nanometer.

Ambrish Srivastava - BMO Capital Markets U.S.

Ambrish with BMO. Moshe, a question for you maybe. When you talk about the $1 billion opportunity, how does that compare with past transitions? And then talked a little bit about the broader market? And I get that Zynq is penetrating newer markets. So when you look at the $1 billion: A, how does that compare with past transitions. And then, at this point, at a point where you are right now. And then how do you compare that with what is new versus where you're replacing your existing or PLD-existing tenant. And then I have a follow-up for Krishna.

Moshe N. Gavrielov

Okay. So from the sales, effective to $1 billion is about 60% to 70% higher than where we were on the previous generation of product. So it's way ahead where it was and it's growing at a about much faster rate. And it's one of those things which actually think to accelerate to the offering. Now we have leverage to make range. We have sampled the high end. We sampled the Zynq, this is where the sales force is going crazy and just bringing in the design win. So I would actually, the 60%, I'm not impressed with. I think it should be, it's easy for me with the VP of Sales isn't here. I would expect to see more than double. And just, generally speaking, I would be disappointed if the 28 node isn't twice as large as any previous node in the FPGA world. And that's actually a pretty tall order if you look at the 65-nanometer between us and the competition, that was a fabulous node for the FPGA companies. But I expect 28 to be at least twice that node. So that's -- those are the numbers that I expect to see.

Krishna Rangasayee

Just to add on to your last comment on how much of it is eating into ASIC ASSP SAM? At least from a communication compass, more than 1/2 of your design wins are against ASIC and ASSPs. So the growth there I'm seeing isn't necessary cannibalization alone, if you will. I think we are seeing growth into incumbents that I think we have traditionally not be able to compete well with, And today, combination of technology and IP is really opening up the door for us.

Ambrish Srivastava - BMO Capital Markets U.S.

A quick follow-up for you, Krishna. We've spoken about this in the past. And now they have a term for it, head net. It's to call heterogenous networks. Now the CAGRs are talking more about it. Obviously, chip guys are talking more about it. The ASSP guys, it's to their benefit to talk more about it, because PLDs can't compete there. Can you just share your views on what -- you sell into that market. What are your current views on that?

Krishna Rangasayee

Okay. So heterogeneous networks and are a equal monikers small cells, right? And so -- and I think, unfortunately, I think these 2 terms are so overly used that there's quite a bit of confusion on what one means versus the other. For my perspective, cellular networks, or base station macro networks today, fundamentally are coverage mechanisms. So they enable coverage. There are many areas, particularly like urban areas, like where we are right now, wherein there's issues with regards to access to data capacity with the coverage. And so, what -- and as the number of iPad users increases, the network is stretched and pushed beyond its limits in terms of their ability to manage coverage to data capacity. So heterogeneous networks, and our small cell infrastructure will coexist with macro cells. And I, for one, do not see one competing and displacing the other. In the long run, or the next 10 to 15 years, perhaps, there are macro trends that affect it. But for the next immediate future, we do not see them competing against each other. In fact, there's a coexistence of the 2 that are going to be evolving. Every customer strategy is different. As I articulated, even 3G is only 30% accepted in the world. LT is only 2% to 3% accepted in the world. And so there's a lot of legroom left that carriers and the customers that we have to monetize the investments and continue to drive growth. Small cells are unequaled in phenomenon that extend and bolster that capacity and the investment that they were made. It's going to take years before this plays out. So from our participation in small cells, there is broadly 2 delineations to the market. There are small cell networks or head nets where are fundamentally one small cell network can support more than 64 users. They have a very strong compelling value per option, both with Kintex and the Zynq, they continue to enable and drive that and they're driving a design win activity. And I think less than 64 is probably going to remain a consumer-centric or a very high-volume ASP play. And we do not actively choose to play in that both from our business model, but also at fit complexes. So I hope that answers your question.

Auguste Gus Richard - Piper Jaffray Companies, Research Division

Gus Richard, Piper Jaffrey. It appears that you're using Kintex-7 against Virtex-4, the 40-nanometer part at the base station, guys who're under extreme cost pressure. And you're using aggressive pricing at cost down? Makes all sense to me, I think I've seen this in the past. My question is, does that serve to reduce the size of the overall market, driving market share gains for you guys? And historically, what I've seen when you've had a price war like this, there's been elasticity 3 to 6 months, 6 to 9 months out. Would we expect that in this environment?

Moshe N. Gavrielov

Okay. All right. I think maybe I should request easier questions going forward. Very good question, again. So one interesting aspect of it is we took a step back on, and we understand the macro complexity you are laying out, and you hit the nail on the head in terms of the customer BOM cost reductions, the income and supplier challenges, the race towards releasing the BOM cost-reduction significantly. We took a huge step back and built Kintex to fit in that environment and not only enable customers with their challenges, but also enable us to retain a long-term value proposition and a profitable long-term value proposition. And one key area why we are winning so much against the incumbent, against our immediate competition, is really almost a question of you need to hit a certain performance threshold of 361 megahertz to really compete in LTE platforms. And the data we have, been led to believe, based on customer interaction, is that, that's an area where we have a very significant strength and our immediate competition does not. So it is not a question of a, I mean, from your vernacular, a price war, or a window of opportunity where we're going to win, and perhaps through another model, because it's an opportunity we're pushing. So long-term value proposition. We built a product that's built for success, from a variety of different dimensions, and we are seeing the signs of that benefit.

Vincent F. Ratford

So can I just -- I can't answer from a business perspective, but maybe from technology, you seem to be very focused on the BOM cost, right, aspect of it. But there's also the aspect of, from the end system perspective, the cost for port or the cost per gigabit of alloys, right, that they can deliver. So that's where the performance side of the equation comes in, also from a power perspective of the equation comes in. So it's not just at reducing system BOM cost, we're also raising the performance and capacity, right? And when you look at that and translate that into the economics of people needing to lower the cost per gigabit, lowering the cost per user per port in their systems, now that's actually quite compelling, more than this 30%, right? Because you saw those factors of, X factors of increasing performance. So you really need to look at the combination of both. And if I could, maybe just from a technology perspective, hit in the head net in general, I mean, one broadbrush thing to think about is that FPGAs are excellent technology vehicles when they are new architectures. When the standards are changing and they're changing fast, right? That is basically where programmable technology is a home run, right? When you have stability and everybody know HDOC 264 isn't going to change and it's stable and nobody needs anything more or less, that's fine, okay, you can fix it and it gets time for other people to integrate it in. Right now, all these explorations on architecture is when there's fast, as Moshe talked about, moving market, moving applications, and people are experimenting with set of sales, all the compute at the remote, at the radio head or the down here. This is when we provide this platform that lets people explore all these architectures and get to market very fast. So that's broad channel, I think, it's is something that I think you should keep in mind. Plus I think the constituents say, hey, ASPs are better for that and so fourth. But think about it, this is where we shine in terms of having a flexible platform.

Auguste Gus Richard - Piper Jaffray Companies, Research Division

One follow-on. In terms of the maturity of the interposer technology for the stacked silicon. How is that coming along? Is that meeting your expectations in terms manufacturability yield, et cetera?

Vincent F. Ratford

So I could take that. So again, as we said, we're really excited about the customer demand. And as I mentioned earlier, it's not just an ASIC prototyping and emulation, we're seeing it in many different segments. And so to some extent, I guess what I'd say is that, since that demand is so great, right, we're working through that. We've delivered to customers. It is leading technology. We're ahead of the competition by at least 1 to 2 years, right? They're not even offering that. So I feel that we're on track with this leading technology. And you heard Jon saying we're very comfortable of raising the sort of what we'll be delivering in the next quarter or so. So I think we're in very good shape as far as that's concerned. And I think the other thing that you keep in mind is that this is not a one node thing, right? This is a strategic technology, right, that we have and we have a significant advantage of and we'll be leveraging that pretty significantly in multiple markets. So yes, we're comfortable with that.

Moshe N. Gavrielov

And just one follow-on. In the tent, everyone kept saying in the tent, In the tent, you can go and see that 2 different applications, which demonstrate the value proposition. And our expectation is, based on customer demand, that we're going to ratchet up the production and in the second quarter, in the June quarter for it to start being significant. At this point in time, we are shipping to customers -- not huge amount. In the second quarter, those -- that should grow. And the demand is bigger than what we have expected. And we're going to need to work through that. It's the best sort of problem to have.

John Pitzer - Crédit Suisse AG, Research Division

John Pitzer with Credit Suisse. I have a couple of questions. First on the 28-nanometer ramp. I know you're a little bit reluctant to give out absolute revenue targets. But I'm curious, relative to 40, 45, relative to 65, how do you see the rate of the ramp at 28? And given the considerable incremental investment at 28, do we think about it as an ROI neutral node, accretive, dilutive node, as you think about the investment versus time to ramp? Then I have a follow-up.

Jon A. Olson

Yes, so clearly, the ramp that we expect on 28-nanometer is much faster than either the last 2 technologies for us. We're seeing a -- and it really comes from 2 things. One is we had pipelined many tape-outs. So we have, we are moving along the maturity curve of getting things to production faster than we ever did in either one of those 2 generations. That's one thing. And the second thing is the level of adoption, as we've talked about the pull from customers. The hundreds of customers who want samples and they are already converting some of their software designs into prototypes, et cetera. Those things are just much, much more significant than they were in the last 2 generations.

John Pitzer - Crédit Suisse AG, Research Division

Moshe, you used the 2x market opportunities. Should we think about 2x faster ramp as well?

Moshe N. Gavrielov

Yes. And as Jon said, we're already seeing that happening. And the amazing thing is in the past, we led with our complex, very expensive products. Now what you're seeing at this point, most of the revenue is actually from the midrange product offering, which is for better or for worse, it's significantly less expensive. And once the Virtex product, and the SSIT products start ramping up and the Zynq will also ramp up pretty quickly, you'll start seeing that accelerate. So the 2x, it's actually, if you compare apples-to-apples, it's a faster, it's more than a 2x faster ramp rate because what we're looking at now is Kintex, but we'll start seeing SSIT has ASSPs, which are 2 orders of magnitude higher.

John Pitzer - Crédit Suisse AG, Research Division

And then, as my follow-up. I know in this industry, it's dangerous to extrapolate a single quarter. But when you look at your comms business, it actually peaked well before the cycle did, semi cycle did. I think just under $300 million back in September of 2010, it's also the segment that's down the most. How I reconcile that with your optimism around growth going forward? Is that just a difficult comparison both in China back then, constrained CapEx now because of the post merger in the U.S. or is there something else going on? How do I reconcile the fact that the business you're most optimistic about hasn't grown in a couple of years and is down the most of all the segments?

Vincent F. Ratford

Okay. I'll take that one. All true. Everything you talked about is true. I think we were late on the 40-nanometer node. And that has 2 connotations. One is since we didn't have a strong 40-nanometer wireless offering, that really hurt us with some of the key customers. And as I've articulated to you, those key customers have come back to us in a very big way in 28-nanometer. And so we are very bullish in our prospects on wireless. And at 40-nanometer, we had a challenge with wireless. The second aspect to it, and since we were late on the 40-nanometer node, our key customers, to a large degree, held on to Virtex-5, and we are seeing the benefit of that growth on the wireline perspective. And to a large degree, I think they are yet to see the ramp on the Virtex-6 product line. And so, to a certain degree, we are driven by the market elasticity of the macro. The Marco is pushing us more. But our product momentum, really, is regrouping on 28-nanometer. So what you're seeing today is a -- in effect of the macro combined with the 40-nanometer portfolio challenges we've had. To a large degree, we are quite aggressive about our outlook on where we are on the 40-nanometer portfolio, like Jon mentioned in the slide that you have. And bolster that with the 28-nanometer horizon. To a large degree, that's a rational wire growth outlook looking forward is a lot more aggressive than the past 2 years.

Vincent F. Ratford

And the other thing that drove the comparison is not having a good, not having a 65-nanometer high-volume product. So all of your data points are correct. And the reasons are very evident. And you're right, communications took it on the chin harder than any other business.

Jon A. Olson

And the delta from that time period, as you noted, to now, is you can really just monitor the wireless CapEx spend and deployments and see those changes for us. And I think there was, I believe, that by our customers who were very strong, ramping up to the right forever kind of thing, and then they kind of leveled off, they did inventory corrections, which is probably going through now. And I think as that wave of CapEx returns, or whenever it does return, but, I think we're thinking it's coming, that, that will help bolster the short-term while the 28-nanometer parts are getting the traction in terms of moving to production.

Unknown Attendee

All right. You can't see me, but I have to stand up. It's Arnold Chan [ph] at Avian [ph]. I have a couple of questions. Maybe a little bit provocative. So first question for Vin, Victor and Krishna. So historically, FPGAs, you've used it for prototyping, for bridging. What you're showing is that you are going to subsume the functionality of the MCUs or those multi-core CPUs. If you look at what these guys would say the exact opposite and the growth rates in the past are multiples of yours. Could you describe why that's going to be the case for you in the future?

Krishna Rangasayee

Okay. I'll take a first stab at it. So #1, I think multi-cores are not a new trend. In fact, from my perspective, this is the third wave of industry investment in multi-core, right? So this is a 15-year-old trend. If you look at the aggregate revenue that the multi-core companies make in lieu of the broader logic IC SAM, it's a very small number. So it's no surprise that they are seeing a CAGR, that's a growth CAGR, that's a much bigger number than our revenue basis or maybe on the broader logic IC landscape. The value proposition that we bring is a very substantial one. And to a large degree, the largest focus for us is really on the ASICs, ASSPs SAM expansion. Clearly, we are very cognizant of where the multi-core companies are, we understand our value proposition, we understand theirs. And as we believe that we increase our portfolio strength, not only in silicon capability, but productivity is one dimension, the other dimension is in IP. We do think that there are very strong performance, cost and power equation that I think you need to hit all 3 simultaneously well for you to achieve the high value that you need to bring to key customers. Multi cores have their strength, but in the net sum, we are increasingly becoming a better option for more customers than anybody else. So it doesn't surprise me that the multi-core industry has a lot of growth CAGR. But if you look at the effective dollars multiplied to the CAGR growth, we do believe that they have a stronger value proposition. So that my outlook.

Vincent F. Ratford

And if you look at the end markets where Zynq is getting traction and where our 28-nanometer products are getting traction, they're not traditionally been highly penetrated by the multi-core guys. They're still very, very difficult to program, the multi-core platforms. And our customers look at the application that they have, whether it's industrial automation, or aerospace or defense, and see that we have the right programming model. And now that we have a processor-centric development platform, our customers can get up and running in their application in a matter of hours and days, which is, by the way, a significant part of the value attraction for the multi-core guys. And as we continue to refine our ability to make it easier for customers to integrate their algorithms and integrate their applications, our value proposition gets stronger and stronger. So to Krishna's point, I'm not worried about the multi-core guys in the broad markets. They all have some penetration in certain applications. But at the end of the day, I think our value proposition holds. And fundamentally, there's still an element of the platform where the customer wants to insert their own IP and their own algorithms, which is something they cannot do in if they're buying the same multi-core platform as every other customer.

Victor Peng

I'll just add one brief thing, is that from a technology perspective, power and performance has to be optimized together. And if you look at having worked in the graphics area, if you look at the power performance share out there, it doesn't work in the embedded spaces in a lot of areas, right? So again, multi-core is just trying to exploit parallelism. FPGAs have historically always been an excellent parallel solution, right? And in fact, it's also, can be optimized to be way better from a performance perspective. The thing that we have to improve, and we are, is the productivity aspect of it, right? And so more on that later.

Unknown Attendee

And if I can just add a question. This is for Moshe and Jon. Again, be a little provocative. You have a military analogy here all in. General Petraeus is not winning the war he's fighting, right? Obviously, I'm sure that's not the message here. I'm looking at Jon's slide, and this is taken out of Jon's slide. Your revenue CAGR is 5% for last 5 years. It's a tech company, it requires a lot of changed investment risk. Why would someone be interested in a company that's grown 5% a year? I know your position is dominant, but is the market going to be a lot faster? And could you talk a lot about that a little bit?

Moshe N. Gavrielov

Okay. [indiscernible] Now let me just share with everyone that I've known Arnold [ph] for probably 15 years. And so I would have expected a softball from him like where did Krishna buy his tie or something like that. But you sort of -- so much for old friendships. But I know where you live, Arnold, nevermind. But it's a very good question. And deal with the military analogy and the all-in reflects what it took to actually -- you have to invest a lot in order, and this was what we've done in the third. And you can argue, and depending on which side of the political spectrum of whether the war in Iraq and Afghanistan are our finest hours. And I won't go there. I will talk about Xilinx. When I joined, I joined after 10 years in the DA industry. And the reason I joined was, it was absolutely clear to me that designing ASICs and ASSPs is becoming difficult, to close to impossible. And the investment is huge. At the time, they talked about the programmable imperative, and they said, okay, a new fool on the hill, who sort of predicts that FPGAs are going to kill ASICs. And I think that what happened to change that, is that FPGAs actually crossed the threshold in terms of capabilities where you can actually do interesting things with them and they really can provide a complete solution. As a result, you're seeing accelerated growth, and the ASICs, the FPGA market has grown from under $3 billion to $5 billion. And the expectation is that it will continue to grow at a certain rate, but there are 2 things that needed to be done in order to exploit that. You need to move to more advanced technology. And we have done that. And that opens a bigger market. And in addition, we have addressed all of the competitive challenges we've had. So I agree with you, 5% a year is not something to run victory laps around. But we strongly believe that now we have the portfolio that enables us to achieve the 8% to 12%. And it's obviously incumbent upon us to demonstrate that. I am absolutely confident in our technology and market position at 28, and that will enable us to do that. So I think your question's backward looking, are very legit, right? And we need to prove to you why it's different, and the difference is in quality of execution, plus passing this threshold, where we can actually provide a solution to more and more markets.

Jon A. Olson

Is that the end of provocation?

Unknown Attendee


Jon A. Olson

Okay. We're going to move to David.

Unknown Attendee

A soft one and less soft one. The soft one, really, is we're now into sort of the end of the correction period for customer demand and inventory levels and we can look forward to at least a rising trend. Can you compare and contrast this time around versus the last cycle, which was, I think, early calendar '09? And I remember in those days, nobody really dreamt that the recovery were quite as rapid as it turned out to be. So can you compare it this period to last period? And the not so easy question, actually, is that when I look at you and your competition, it, statistically, it seems that one company says one note very well, and the following note, the other guy does better. And when I look back at why that is the case, it seems to be, mistakes were made. But human beings being human beings, I assume mistakes will always be made. So is that possible that you can do superior offering in more than one note?

Moshe N. Gavrielov

Two excellent questions. I'll let Jon answer the easy one, and I will take the hard one.

Jon A. Olson

So the -- to me, the characteristics are a little different in terms of what happened in '08/'09 and what's happening now. And I don't know that it says that there isn't going to be a period of rapid growth some time later this year. There very well could be. But it seems that, right now, there's a reticence than, particularly in communications, is building your infrastructure out of reticence to go put CapEx spending in there right now. At that time, in '08 and '09, you saw stimulus packages going on everywhere around the world, China, U.S., big time in Europe, et cetera. And that really helped us from an infrastructure build, and we came out of it very, very rapidly. This time, I still see, my perception is, well, things are a lot better in Europe, but there's a lot of concern over the macro issues in Europe. And then there's the wireless infrastructure around the world. Are they really, is CapEx really going to start in China? Is AT&T and Verizon really going to start spending later this year? And while it is possible, I don't feel that there is as much pressure before for stimulus where that typically ends up being infrastructure stimulus that nations do. It's not all electronic stimulus, by the way, but there, some of it did go in that area. So I do feel a little bit different about it. One other difference, I think, is that back in '08 and '09, we saw lots of customers decommitting programs and projects. And that has both maybe a good effect for PLDs because when they turn things back on, they can only use the PLD, and that's really helpful. But it also killed some things along the way. And therefore, made it a little harder for us to grow, because there's so many projects that just laid dead or companies laid dead. I don't see that many dead companies in this last cycle. So therefore, that's a little different. So are we going to come out really, really faster. It's there going to be slow climb out? It's difficult to say. I would say slow is probably my take versus a very rapid like it was. I mean, very rapid and pronounced between '09 and '10, last time.

Moshe N. Gavrielov

And so with regards to the philosophical question, this is the stuff that Greek tragedies are written about. And hubris and of all that as soon as you are successful is the seed for destruction. So and because you sort of assume you're infallible. So you could see the numbers, we're already investing heavily in the next generation, which is the 20-nanometer node. And we expect the competition to be as fierce, if not fiercer there. And we take the competition very, very seriously. What I will say is the person we have running engineering was Victor. I knew him when he was 25 years old. He would run 6 marathons a year on average. And now it's 20 years later, and instead running 6 marathons a year, he runs 4 ultra marathons a year. So we have a person who is, I think, can, in terms of his ability to push forward, we have the right mindset there to keep pushing as hard. Because it requires that level of commitment. This is really difficult stuff. I think SSIT having the highest performance, the lowest power, a very good footprint and delivering it on time on this breadth of product requires tremendous tenacity. And I think we've built up an engineering organization which is second to none. That is exactly what PSMC tells us. We use those words. It's the highest level throughout their organization. So hopefully, we'll disprove all of the Greek analogies and deliver at 20.

Vincent F. Ratford

So if I could just add and appreciate the fact that despite knowing me since I was 25, he hired me. But I would like to say also that not only, as Jon and Moshe planned the investments that we're doing, but if you look at the things that we did that really are different at 28, they are not one node wonders, right? This is SSIT. That's strategic. That's not just about getting large FPGAs done, right? There'll be directions on that for a long time. T&C did a joint announcement with us when we announced this technology. They don't -- that's not their usual habit, right? They don't do that because they have to be neutral to their partners. But we truly did a leadership technology there. Zynq, extensible processing platform. That is a breakthrough technology. It isn't the same old 440 redone. I mean, the process's subsystem, in its own right, is a very powerful SOC, and now it's closely coupled with a very high bandwidth interface with the ASI interface of the fabrics. That's not a one-note thing. That's a whole new product class, okay? And then we haven't announced it. So I just as an engineer, I just love to talk about this. But there's a whole thing we're going to do in terms of that ease-of-use productivity, right? Reducing the -- some of the barriers, and this is going to go on for multiple generations. And the thing about software tools is it rises all boats, right. It's across the entire product portfolio. So none of these things are single node wonders. The fun fact that Moshe makes sure I don't get too full of myself, and I keep burning the midnight oil. But really, I think that's the sort of thing that you look at. And just look at what's on the public record in terms of nearest competitor, right. I mean, their equivalent product is, by their own statement, a year out. Silicon, we've already shipped it to customers, right? They don't have anything like SSIT, you can go out there and see the SSIT technology. So I think we've done that, and we're knocking with the foot off the gas pedal and I'm excited for when we can talk about the other thing. And back to the provocativeness, I would say is that a software programmable multi-core, you can get to a sub-optimal solution fast, right? But if you want a really good solution, FPGAs are really pretty awesome. And again, from our perspective, we're going to make sure that people can get to that solution faster also.

Lori Owen

Okay. We're going to cut off the question. One more. Apparently, we're not going to cut off the question.

Unknown Attendee

Jon, just a couple of model questions. First, on the OpEx. How should we model it in terms of the trajectory? When do you expect OpEx to peak on a quarterly basis? And the second question is, is 28-nanometer accretive to gross margins as yet? Or I guess my question is a 28 increase as a percent of sale. Should we expect gross margins to kind of improve?

Jon A. Olson

So, on the first one, there will be, in terms of the tape-out for 28-nanometer, they're definitely are first half loaded, and the peak will be, I would say, likely the June quarter. But in terms of the first half, second half distribution of the R&D spending, it's only slightly more oriented towards the first, sort of the first half than the second half. So I wouldn't get too, from a modeling perspective, I wouldn't put too much of the R&D very front end loaded and have the back of the year be really light, because I actually think it's almost fairly balanced throughout the year. As the tape-outs decline, some of the other things that we're doing start to ramp from an OpEx perspective. So a slight bias towards the first half in terms of spending. And the second question was around 28-nanometer accretion. So clearly, we are in a good position on yield on a relative basis. There's a lot of products we're birthing. We do believe that, as we get towards the second half of FY '13, definitely, I think there will be some accretion going on there in the second half of our fiscal year.

Unknown Attendee

Okay. Then one for Krishna, if I could. Krishna, I'm just wondering if you look at globally, the CapEx spending, we look at that number at a very high level. It doesn't seem like it's changing a whole lot. But we keep hearing from all of you guys that wireless spending is horrible, especially in the U.S. and China has slowed significantly, and then India, we've been talking about 3G network for 2 years, and we haven't really seen that translate into revenues for you guys. I'm just curious, what should we look forward to? I mean, if you're thinking about a recovery in your business, what are some of the things that we need to kind of keep an eye, on as we look out to the next 6 to 12 months?

Krishna Rangasayee

So I think the wired and the wireless markets have different cadences in terms of leading indicators of market fluxes, right? So CapEx is one element of it. The second one is RFP tenders that are floated and who wins what portion of it. And from our vantage point, it really comes down to leading indicators in terms of design wins and traction and increasingly how much of the market share are we really gaining from the ASICS and ASSPs. So we track that internally. And I'm sure some of those public indicators on the RFPs, on the tenders that are floated, and also the CapEx is obvious to you, the nuances is not one at the reading the highest level relative to change, it's really understanding the shades of gray in terms of not only flat, flattish CapEx structure or macroeconomic constraint, there is still a lot of movement within submarkets and leaders within that market. And we spend a lot of time looking into it. And that, perhaps, is the only way to really rationalize through this. All of us hope that the macroeconomic structure will lift the picture up. But clearly, with wireless we have challenges, different challenges with wired. But over the long-term horizon, we've continued to be bullish on the outlook because the demand's clearly there, and it's a question of does the demand add up to the macroeconomic, add up to the business model, add up to the right deployments. And that chain needs to be fruitious. So wireless spends, we have positive cycle. It's going through lots of challenges. Wireline, in my mind, has been in a funk for 5 years. And in the next 5 years, we see that really firming up. So we need to stay close with those leading indicators. And we have our internal lens to it. But perhaps there are external lens that could be put into. We look at it as a 5 to 7-year outlook. It's clearly got a lot of growth [indiscernible]

Rick Muscha

All right. So thank you very much for the great questions. You're going to have an opportunity now to mingle with some of our, with the executives here over in the area to my left, to your right, where the tent is, where we have a variety of demos that we've talked about, some of those key areas here today. So we invite you to go over there and have a little beverage and snack and look at those demos. I think you'll be able to see real things that we've talked about today. So thanks very much, and have a great day, and look forward to talking to you over there. 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.


If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!