Seeking Alpha
We cover over 5K calls/quarter
Profile| Send Message|
( followers)  

Xilinx, Inc. (NASDAQ:XLNX)

Barclays Global Technology, Media and Telecommunications Conference

May 22, 2013 12:25 pm ET

Executives

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

Analysts

Blayne Curtis - Barclays Capital, Research Division

Blayne Curtis - Barclays Capital, Research Division

All right. Thanks for joining. I'm Blayne Curtis, semiconductor analyst here at Barclays. Thanks for joining the first day of the Barclays TMT Conference. Happy to have Xilinx with us today. We have Moshe Gavrielov, President and CEO. My thought here was -- my kind of view is the semiconductor industry is of one of more modest growth. I think Xilinx is interesting as the -- or one of both caps, and he'll walk you through why they have quite a secular story. He's also been in semiconductor, as well as DDA, for, what, 30 years. So it has a very -- a long history, as well as being a business that is broad, as well as very focused on process technology, has a very good view of the overall world. And maybe just to start off with that, I thought it would be good, given this is a broader audience, not all semiconductor people, Moshe, whether maybe you could just talk about a little bit of Xilinx story, give them a backdrop, and we'll start there.

Moshe N. Gavrielov

Sure. Well, thank you for joining, and it's been a privilege to share with you the Xilinx story. Semiconductor industry is 60 years old, and as such, is maturing. And there are transitions that are happening and opportunities in a maturing industry that actually Xilinx is benefiting from.

So a little background on Xilinx. The company will be 30 years old next February. We invented the FPGA, Field Programmable Gate Array. One of our 3 founders actually had the patent on it. And we've grown over these years to service a very broad set of customers. So at any point in time, we typically have 15,000 to 20,000 active customers in over hundreds and hundreds of applications.

And so if you look at the world, we're in every single products you can think of, electronic products, except ultra high volume consumer devices. But barring that, everything else is powered by FPGAs. And FPGAs from generation to generation have come to the point where they're implemented in the most advanced process technology available. So we are actually the first in 28-nanometer. We're going to phase out the first 20-nanometer products, and we're already working on the generation beyond that. What that enables us to do is to benefit from being at the leading edge of technology, and to broaden the market we service.

So in the past, our customers have thought of FPGAs as having a peripheral role in their design. More and more FPGAs have moved to becoming the center of the customer's design. And as such, FPGAs today have, and Xilinx love this, have high performance embedded microprocessors, high-speed serial interface, a lot of memory, a lot of gates. And inherently, the most important differentiation we have is that FPGAs are flexible. They can be programmed both before they -- as they're being designed or actually as they're being used, they can change their functionality. And that's why they have a very broad applicability.

So think about it as a very broad player, which addresses the center of the customer's design. Some of the key applications we're in is in the backbone of the Internet, both on wired and wireless networks in Mil/Aero applications, we're moving into automotive applications, high-end consumer, et cetera.

So that's the end of the paid advertising here.

Blayne Curtis - Barclays Capital, Research Division

Thanks. And as you mentioned, Field is -- are not new entrants to the market. I think maybe the difference is that there's a widening gap, and that's been the story, and it's taken a while. You probably even admit to play out, but I think we're getting to the point now where the focus is on just how expensive it is to be a player at the leading edge. And maybe you can just give people a perspective as to where you are, where are the chips that you replaced and where do you see that going?

Moshe N. Gavrielov

Yes. So as Blayne kindly pointed out, I think in the industry, for just over 30 years now, we started back in the '70s. And for the longest period of time, there was a large number of semiconductor companies, all of which could afford to move with, every 2 years, to a new process node and achieve all of the benefits of that. Now what has happened is over the past 10 years, the cost of developing products, the new process nodes, has sky rocketed. So in the past, you could pull hundreds of thousands or millions of dollars developing new products from scratch. Now at the most advanced process node, depending on how you look at it, it could be as expensive as $100 million to develop a product from the ground up to address a certain application. Very few, if it takes $100 million, just to develop the product, then obviously, the market needs to be very large to support that. You typically are looking at a market which is a minimum 10x to 20x larger than the development cost in order to defray those expenses. There's very few markets, except high-end consumer, that can address that. So in the past, the customers could design unique products called ASICs, which was specific for their application or there were dedicated semiconductor companies that had niche vertical products targeted at the application. That has gone away. To a large extent, that is no longer economically feasible. And in more and more applications, FPGAs are the only viable solution. And as a result, as we move forward, and we provide more attractive silicon with a higher level of integration, then we can address more and more market. And that's the secular growth. As Blayne pointed out to, this has been counted ahead of its time, so you might have heard this story 20 years ago, 10 years ago, and you might wonder what's different now. Well, what's different now is that the alternatives have become prohibitively expensive. And so FPGAs are indeed servicing a much larger available market.

Blayne Curtis - Barclays Capital, Research Division

And I think it's expensive for others, it's also expensive for yourself. Obviously, last year, there were some headwinds as far as top line, which was out of your control. When you look at, I guess, to the one change that semis have been on is Moore's Law. When you got the 20-nanometer, it becomes more expensive for a transistor. And that's kind of the -- not the demo too far in the lead, but that's kind of the last generation on a particular type of transistor. There's a transition to a newer one. You may get better performance gains. Maybe you can talk about how do you manage the, one, the cost? And then two, do you see the semi industry moving, continuing to move down Moore's Law? And when you -- you're already moving forward with 20 nanometers, some are skipping it. Can you talk about the decisions that are being made by semiconductors?

Moshe N. Gavrielov

Yes. Well, I'm an optimist because when I started back in the '70s, there was a general belief that the semiconductor world would end at one micron, which 25 years ago, we hit that, and there were expectations that you couldn't move forward beyond that. And any time, over those past 30 years, any time you've gotten to what were perceived as a theoretical limit, then there was a lot of ingenuity and investments made, and there were breakthroughs, and Moore's Law continued at that certain clip. What we're seeing now, as you're pointing out, is that both the design costs are accelerating and the manufacturing costs are accelerating. And so the dynamic, technically, it's still possible, but it's a lot more expensive. The way we address that is we have to be a lot more deliberate in and thoughtful about how we spend our R&D dollars. And whereas in the past, we'd have a process which I would call spring and pay, where we'd come with a whole host of products, and we could afford to do that, and maybe 20% was successful in the market. Now we have to be a lot more different deliberate. We have to work a lot more with our key customers a lot more closely, and we actually need the higher and larger portion of experts in the target applications that we're addressing, so we can make sure that although we're doing less products that we have a higher likelihood of succeeding. And that whole deliberate approach and more thoughtful approach is something that we're taking. And so far, so good. It has been successful with 28-nanometer, which we started seeing this trend. It's the fastest-growing new technology for us. It's probably growing at twice the rate of the previous generation, if not more than that. And we're starting to see the benefit of that. And it's just incumbent upon us to continue with that approach for as long as possible. We're focused on 20-nanometer now. We expect to tape out actually this quarter. We've announced that. We expect to be one of the first to tape out the 20-nanometer, and we're already working on the generation ahead and beyond that. And we can afford to do it.

Blayne Curtis - Barclays Capital, Research Division

Do you want to talk about the generation ahead? But maybe just taking it back a little bit, and you do have a broad exposure, networking, industrial, consumer. What's your just overall sense of the business environment? And we've seen a bit of a mixed bag. As people look into the summer here, for a guy who does see a good slop of the overall demand, what's your general sense outside of any Xilinx's specific driver?

Moshe N. Gavrielov

Generally, it's incrementally better. I wouldn't say it's great. People aren't dancing in the street, but I promise that the general feeling is somewhat better than it has been for the past 6 to 9 months. So the level of design activity is the length [ph] of trade. And just generally, the orders in the shorter business seem to be holding [ph] up. And when we talk to the executives and our customers, they seem to be quite positive without being too optimistic. So it's -- but first, [ph] it has been a topic of time, not sure what is driving that, but I continue to believe that it's going to be a good second half of the year.

Blayne Curtis - Barclays Capital, Research Division

And when you look back, last year, it was similar optimism, obviously, we thought slowdown as an industry into the back half of the year. What do you think this year that's any different than last year?

Moshe N. Gavrielov

Well, Xilinx specifically, we're in benefiting from a virtuous product cycle. So we led the market at 28-nanometer. The design wins are now moving into volume production. It's accelerating as quickly as we've ever seen or faster than we've ever seen any node. So that is driving our optimism. It's a product cycle. In our case, because the market is expanding, the new products can address markets that we only had peripheral success in the past. For example, automotive, with our new Zynq-based product, which has an embedded ARM core. We're getting a lot of design wins, which are now starting to transition to revenue and driver assistance. There's a whole market expansion that we are seeing. So both of which was product cycle and market expansion is what's driving our business. The third element is, since 50% of our business is in wired and wireless communication, there's an expectation of an upgrade to LTE at a global level, and it's happening country-by-country. Still expecting China to go through a significant upgrade towards the end of our fiscal year. There's undoubtedly a need for all of the other carriers to catch up with Verizon in terms of the LTE coverage. There's big investment happening in Latin America. Brazil starts to host the World Cup and the Olympics, and they have to upgrade their networks. Korea and Japan are doing some upgrades. So that wireless infrastructure should be an additional tailwind that helps drive our business forward.

Blayne Curtis - Barclays Capital, Research Division

That segment seems to dominate [ph] industry. You get the quality of the questions you get as far as when's that going to pick up. So I guess I'll ask you the same question as far as have you seen any indications, and it doesn't have to be just China Mobile. I mean, as far as you did mention these other geographies, it seems like China Mobile seems like we keep waiting. But have you seen any better indications in recent months as far as that you are getting the orders and you will see that pickup?

Moshe N. Gavrielov

So we've gotten the design wins where you have a 28-nanometer Kintex product, which was tuned specifically for wireless. And this we can tell, shipments up to now, it actually has 90% market share in that segment. So we're hoping that this will actually continue to do that well. Design wins, they're broad, they're global, and the deployments will happen at the rate that either economics will enable in the North America, the economics actually are driving towards that because as soon as you have Verizon with leadership, it pushes AT&T and the others to have to catch up, and they're way behind. So there needs to be an upgrade there. China Mobile is waiting for the spectrum to be granted by the Chinese government. And it's any one guess -- anyone's guess as to when that will happen, but the expectation is it'll happen soon in the next couple of months. And once that happens, then they go to their numerous suppliers and they try to roll it out. They've been public. They were talking about the 200,000 base station rollout. That doesn't sound like a huge number, but it is. In that business, that's a huge number and back to the -- expected to be a multi-quarter rollout. So that is not an if, it's only a when, and I don't see it getting pushed out, the start of that beyond this year. We think that it's going to start this year for sure.

Blayne Curtis - Barclays Capital, Research Division

It was a hard -- that particular market is always hard to gauge. There is a government interaction element as well and it seems that obviously, the spectrum hasn't been allocated. There's also a discussion about whether some local suppliers we get some of the bids. And as far as your exposure, is there any way you can quantify your exposure to the potential vendors, that without naming obviously particular names, but whether domestic suppliers versus your in suppliers versus smaller domestic suppliers, how broad is your exposure?

Moshe N. Gavrielov

So there's 3 Chinese suppliers, and there's about 3 or 4 international suppliers, and we expect the market to be split where the Chinese companies probably get 70% of it, and the others share the other 20% to 30%. We have design wins with all of them, but in some cases, we have a higher penetration of our -- of these potential customers' products. So we do have a preference that all we can do is just win as much as possible, and hope that the ones where we have a higher percentage of the silicon value actually get the larger parts of it. Fortunately, we don't get to pose on that.

Blayne Curtis - Barclays Capital, Research Division

Right. I do want to ask you about the wired business. At your Analyst Day, you were actually almost more positive about the growth there, and it does open up a couple of questions as far as some of the drivers there and also some of the strategy to bring in an outside IP, offering more than just a programmable chip and bringing actually specific IP into your chips. And maybe you can just talk about where you're most positive in wired and then we can delve in further.

Moshe N. Gavrielov

Yes. Well, we've been bullish on the wired for 3 years now and unfortunately, haven't quite seen the growth yet. But the reason we believe it this time it's different is there's 2 elements that are happening. One, there's a movement to the cloud, and data centers are becoming more and more essential. And the data center in wired technologies and companies that are in those spaces largely overlap. So we're expecting some growth there. The other thing, which is driving it, is actually wireless smartphone requirements. There's a lot of video there. There's a lot of what we call tuition-oriented [ph] applications, and that's generating tremendous stress on the underlying wired infrastructure, too. So we believe that the 2 of those are going to drive the upgrade. Our biggest customer in this area is Cisco, and they're quite positive about it, so we're keeping our fingers crossed. That looks good so far, but I'll believe it when I see it.

Blayne Curtis - Barclays Capital, Research Division

As far as -- I mean, a competitor of yours, your only real major competitor, bought an OTN asset. You had already purchased one prior, but maybe if you can just talk about some of the IP and where there's other areas to bring PLDs to really replace application-specific devices that...

Moshe N. Gavrielov

So as I mentioned when I started, the semiconductor industry is maturing at 60 years old. There used to be a place for small communication semiconductor companies. And for those of you with 10-year-old memories, you might remember there's a whole plethora in the early 2000, very successful smaller communication semiconductor companies. What has happened is the cost of playing has gone up to the point that these companies, which are typically $500 million and below, they just cannot participate. So they don't have roadmaps. That opens some opportunity for us. But in order to service that market, it's not just with the FPGA, there are actually some protocol-specific IP that is essential to address their requirements in order to displace these ASSPs. And what we have done is, over the past 2 years, we've actually bought 4 small IP companies, and with -- who have state-of-the-art industry-leading technology for those 4 protocols. And that is enabling us to displace these communications, ASSPs at an accelerated level.

Blayne Curtis - Barclays Capital, Research Division

And the area you're expanding your TAM is integrating microcontrollers or processors with your PLD, if maybe you can talk about what you've announced. And then is there any -- where could that roadmap go as far as Altera just purchased the power management company. Can you talk about other adjacent -- a piece of silicon and expanding your channel in general?

Moshe N. Gavrielov

So we have a new product line, which is implemented in 28-nanometer, called Zynq, and it's a product line which has a dual core A9 processor. And it's a very powerful processor. It's the high end of the midrange of the processor space, and we've actually come out with a whole family of Zynq products that has that embedded processor there. This enables us to move into a whole host of new markets, which realistically we couldn't touch in the past. And so if you look at automotive, industrial and the high level of integration in communication, they all can benefit from this product. So this is actually, in terms of design wins, it's the fastest-growing segment of our business. It tends to take a little longer for that to transition to revenue, so we're still at early days there. But this is an integration play. We're now -- we're integrating more of the customers' bill of materials into the FPGA. And so we're replacing controllers, microcontrollers, microprocessors from companies like Renesas, ST, TI, Freescale as they're being integrated into the FPGA. And this is an expansion play for us. In order to do that, we -- it's tightly knit. It's a processor which is tightly knit with the FPGA. It provides higher performance. There's a whole host of applications which we call Vision applications, which benefit from this. And when you think about that, think about surveillance cameras, right? They used to be dumb surveillance cameras, which were monitoring -- or they weren't monitoring, they were just filming. But now think about surveillance cameras that actually have detection and can keep looking for people with certain features in real time, right? And for better or for worse, that is becoming a reality in the world we've been, and there's more and more need for that. So the Zynq product actually has had great success in those sorts of applications.

Blayne Curtis - Barclays Capital, Research Division

And that's an expansion of TAM. There's also share within TAM. And I think you'd admit that at 40, there were some areas where maybe you didn't do as best as good as you would have hoped. How have you right at that with 28? And then as you look at 20, where do you see your advantages?

Moshe N. Gavrielov

So at 40, we clearly had fallen behind our competition. And as a result, we played catch-up there. We were resolute that we would not repeat that at 28. So we're actually the first to market at 28. We have the broadest product offering. We have a very differentiated product offering, so it's not just a me too there -- on the products there a little earlier. It actually is significantly different. If you look at the high end of the product, we invented and implemented an approach, which enabled, we call SSIT or 3D ICs, which enable us to have a product which is twice as large as anything the competition has today. And so that gives us a differentiation at the high end of the product offering. The Zynq product offering, we were in the market at least a year, and probably more than a year ahead of the competition with a broader range of products. And on top of all of that, what we did is we implemented that entirely new software suite called Vivado, which we implemented from scratch. This is something which enables our customers to get to market faster. It's a much more effective product suite, and actually to achieve better quality of results, in other words, the FPGAs operate faster at the lower power point. So there's a lot of differentiation at the high end and the Zynq product offering with the software. We actually have a low power footprint, generally speaking, vis-à-vis the competition. So we think we have a great position. If you look at the last quarter, we shipped 65% of the node, right, and almost twice as large in terms of shipments as our competitor. So, so far, so good.

Blayne Curtis - Barclays Capital, Research Division

Well, you said the node, just to clarify, it's 28-nanometer node base, you have 65% share in...

Moshe N. Gavrielov

This past quarter, yes.

Blayne Curtis - Barclays Capital, Research Division

And when you look at 20, it sounds like you feel confident that you can -- I'm not going to pin you down in a specific number, but you feel very good about it?

Moshe N. Gavrielov

We -- everything indicates that we're going to be ahead, and we're going to use all of the techniques that enabled us to have a differentiated product at 28 going to replicate and drive them at the 20-nanometer node. So I'm pretty confident that we're going to have a very good footprint at 20-nanometer, too.

Blayne Curtis - Barclays Capital, Research Division

I did want to go back to the process technology and this segment. As far as 20-nanometer, I think a lot of people are going to skip it, given as far as the mobile guys, a lot of the companies are going to see a bit of [indiscernible] To big chip, you'll get yields and improvements at least it seems to make sense. When you look at 14-nanometer and 16-nanometer, the new transistor technology at least at 16, can you talk about the advantages there? Do you -- is the cost for transistor still going to be higher? And as far as performance advantages and cost savings, if you can quantify any of that, that will be helpful.

Moshe N. Gavrielov

So the 20-nanometer node and the node beyond that, be it 14 or 16, and there are different marketing terms for products which are actually quite close in those -- in that nomenclature are coming to market very closely together. They used to be a 2-year delta. This is unlikely to be anything like that between generation to generation. The benefits that the 16-nanometer will have with these 3-dimensional transistors, or tri-sets they're called, is it will have higher performance and lower power. In order to achieve that, it actually requires a somewhat more sophisticated manufacturing technology. So inherently, the technology, the manufacturing technology is going to go up in price, but the good news is that there appears to be more and more foundry, viable foundry partners at that node. So I think the market dynamics may compensate for the fact that, that node will be a little more expensive inherently if you just look at the manufacturing process. So with regards to our position, our products tend to have very long life cycles. They have a minimum of 5-year design life cycle and a 15-year manufacturing life cycle. So each of these can last a very long time. And our expectation is that they will last a long time because the investment required to get to that node is very significant. If that does not happen and there isn't longevity, then it actually compromises the return, and we're very focused on making sure that each of these nodes are successful in of its own right. Just to give you a handle on what it cost us on the 28-nanometer, that node is probably an $800 million R&D investment. That's a big investment. So you could multiply it by 10, and you can guess what sort of revenue we would like to see from that node. That's a lot of money, yes.

Blayne Curtis - Barclays Capital, Research Division

Great. And being at the leading edge, you spend a lot of your time actually existing in process development. Is there any challenges that the semi manufacturing industry has to kind of overcome to make it to 14 and 16? There's been a lot of talk at UV, it seems like there's still throughput's not there. I mean, is there anything that would snag the timing of the rollout and/or impact cost?

Moshe N. Gavrielov

The expectation is that the earliest that, that will be viable is 10-nanometers, which is the node beyond that. And there's a lot of work that needs to be done in order for it to be viable. There's also the transition to even larger wafers, which again, in of itself, is a big investment. So I think there are a lot of things that are going to be defined over the next 2 years before that next node is indeed viable.

Blayne Curtis - Barclays Capital, Research Division

Right. And you probably don't want to answer this, I think you got asked a ton at your Analyst Day on this as far as your choice there and I mean, I guess, at some point, you'll talk about it. But there's clearly, your incumbent supplier, who's been a foundry, there's obviously Intel making strides to become -- lever their leading edge technology. Can you talk about the benefits of being with a guy who has operated the foundry, as well as do you think there'll be any -- not saying which way you're going to go, but just maybe if you can just contract the 2 options as far as timing?

Moshe N. Gavrielov

Yes. So we haven't announced who we will be using. We did announce we used TSMC at 28. We have used them at 20-nanometer. We have not gone public with who we will be using at the node beyond that. At every point in time, we have -- we look at the viable players. And the more players there are, the better off generally the industry is. And we measure them based on the availability of technology, the availability of capacity and their ability to support the foundry model. And our business tends to be a very long-lived, low volume high mix. In other words, we have -- we ship 10, 15, sometimes over 20 years to the customers. We tend to have small volumes that accumulate over those 20 years or 15 years to a very large scale of business. But the ability to support that is very different from that of a traditional semiconductor, one product, ultra high volume. And so we make those decisions. The proof is in the execution. If you look at 28-nanometer, it was the first time we used TSMC. We started last, we finished first. We have the best product and we have the highest revenue. It's a combination of the software, the silicon technology, the design technology and integrating them well together. And that this our differentiation, our ability to do that. We've laid the foundation to take it to 20 and the generation beyond that. And so I'm very comfortable with our competitive position, and we've proven it on 28. We're about to prove it on 20, and I believe that we'll prove it at the node beyond that, too.

Blayne Curtis - Barclays Capital, Research Division

Obviously, that's still a decent way is out there as far as timing. I think the concern on, still in the industry is that this -- you mentioned $800 million to 28. Where does that go for 20? I don't know if you would get -- something else there, but how does that relate back to your OpEx line? And then maybe if you could talk about, as far as gross margin, is as you -- the different drivers as far as mix, as well as, as you start adding all the different additional content, do you actually get some benefit from that as well?

Moshe N. Gavrielov

So our model is 64% to 66% gross margin. We're actually operating just above that. And then R&D is 17% to 18%, and SG&A, I think, 14% to 15%. We're slightly above that on both the R&D and the SG&A. Because of -- now most of that is driven by the fact that this past year has been a slow year, so the growth wasn't there to support the investment we have made. What we're doing at this point is, we're, in dollar terms, we're keeping the investments in R&D at the level where it has been, and we are not continuing to grow that. And we believe that we can retain leadership at this level of investment, and we're waiting patiently for the 28-nanometer and the broader markets that it services to facilitate the top line growth. And in reality, by being the first to the node and servicing a larger market, that's what is going to drive it. And this year and next year and the year after, the projections we gave was for a 3-year compounded annual growth of 8% to 12%, right? So that should -- as we achieve that, even if you use the lower number, we will very quickly get back to model.

Blayne Curtis - Barclays Capital, Research Division

Maybe you can talk about the speed of ramp of 28. It's still small as an overall dollar number of the overall pie. But if you talk about how that has tracked towards your expectations and then really, as you look forward, when does 20 layer to the mix? And what would -- if you did gain back the share, how would that translate to revenue?

Moshe N. Gavrielov

Yes. So it takes 2 years from when we bring the product out. And so we brought it out in March, silicon, first silicon was March of 2011. In the past quarter, we exceeded $40 million in shipment for the whole year. Our year is in April to March here, we shipped over 100. We're projecting 250 million for the next year, which will end in March. So it will more than double over this fiscal year, and I would expect it to continue to grow at an accelerated rate. So it will be at about 10% of our revenue 3 years after we introduce it. If you look at the 20-nanometer, that implies, if you take that same -- those same parameters, then if we're taking out now, then likely towards the end of 2016 is when it will be at 10%. There's long delays. This is a business for the patient. The good news is that there's an overlay of all of the past technologies that we continue to benefit from, so it's not very volatile. It actually grows slowly and then continues for a very, very long time.

Blayne Curtis - Barclays Capital, Research Division

So there's been -- a perspective all the focus on 14, but 14 revenue would be 2018 or...

Moshe N. Gavrielov

If that's -- if you take these past numbers, that's likely to be the case. I have not seen it change any faster than that. 28 is coming out faster than any other generations. So this is not a slow introduction, it's a fast introduction.

Blayne Curtis - Barclays Capital, Research Division

I did want to ask you, we talked about semi industry being more mature, slower growth. You're seeing better return of shareholder value. ADI last night raised their plans and took out debt. You're seeing lever of the balance sheet. Maybe you can talk about, both you and Altera, you sit on a lot of cash as a relative portion of your market cap. You've done some M&As, but not big M&As. Maybe you can talk about the strategy there. If you do pay out about 70% of your cash flow, if you include the buyback probably sort of last year, but what about the strategy that the dividend is probably, if you look at the Analog group, maybe 3% to 4% is kind of what a div yield is you've probably has gone amended [ph] that range. What's the kind of plans, and would you be receptive to doing more leverage on the balance sheet, as well as returning more cash?

Moshe N. Gavrielov

So we're committed to both dividends and buybacks, and we have significantly increased our dividend, I think, for 7 or 8 years in a row. The last time we increased it by 16% to 17%, it was by $0.03. It went up from $0.22 to $0.25. And we have typically -- so we lead with dividends, and we're committed to growing the dividend as long as we can grow the dividend. Clearly, there's headroom there. And then we do buybacks at the minimum to address dilution, but we've been pretty committed to doing buybacks, too. At the right price, we have done them. And that's part of our strategy going forward. We have, if you look at our balance sheet, then we have quite a bit of a very low interest debt on the balance sheet, and that has helped us do that. But we actually have the capacity to do it out of operation. The company has traditionally, over the past several years, every quarter, delivered tremendously positive cash from operations, and that's what we're doing. That's where probably you're 70%. So it's been fluctuating between 70% to 80%, then I don't see us changing from that norm in any big way.

Blayne Curtis - Barclays Capital, Research Division

One struggle that tech companies have is that a lot of their cash is offshore, and cash flow is offshore as well. I mean, as far as restrictions, they continue to raise that. What portion of your cash is onshore? What you'll say maybe down in Washington explaining...

Moshe N. Gavrielov

Yes. So a significant part is onshore, but it's less than 50% because actually most of our business today is offshore, right? We have a large business in the U.S. and North America, but it's still nowhere near to 50% of our business. Most of it is offshore. And we just make sure that we always have enough to run our business to pay for dividends, and we'll continue to do that.

Blayne Curtis - Barclays Capital, Research Division

Right. Well, I really appreciate your time. Thanks for doing this. And like I said, I think you do offer, for the market cap that you do have, a really interesting growth profile, and I wish you all the best.

Moshe N. Gavrielov

Thank you.

Blayne Curtis - Barclays Capital, Research Division

Thanks.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

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

Source: Xilinx's CEO Presents at Barclays Global Technology, Media and Telecommunications Conference (Transcript)
This Transcript
All Transcripts