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

Executives

Moshe Gavrielov – CEO & President

Xilinx, Inc. (XLNX) BMO 2013 Technology, Media & Entertainment Conference Call December 10, 2013 9:30 AM ET

Unidentified Analyst

Well, welcome everybody to our second meeting this morning and a great pleasure to welcome Moshe, CEO of Xilinx. And Moshe, welcome.

Moshe Gavrielov

Thank you.

Unidentified Analyst

Appreciate your coming here. Before we go – same format fireside Q&A with a little bit time at the end and I apologize for the last meeting, I didn’t have enough time. But before we go into my Q&A, Moshe, if you could walk us through, just help us for those who may not be as familiar with the Xilinx story, what are some of the key factors driving the business, the operating model and where do you see Xilinx going?

Moshe Gavrielov

Okay. Well, it’s opportune and thank you for giving us the speech [ph] here. Xilinx is the technology leader in the FPGA space. We actually this morning announced our 20 nanometer technology portfolio. If you look at our position at 28 nanometer we were the first to tape out the TSMC. 20 nanometer we were also the first to tape out with TSMC. We actually have over four months advantage over any competition. We announced the details of our product offering at 20 nanometer. It’s a very broad, it encompasses the mid range and the high end of our product offering, much higher performance, lower power, better software and as part of that we announced the high end product, which is the 440, which enables our customers to use 4.4 million logic cells that gives us a factor of four advantage over the competition. So our competition’s largest device at this node is around 1 million, we are at 4.4 million. So basically we had a generation ahead with 28 nanometer. Last few quarters we had over 70% market share at 28 nanometer. We believe that we can continue our hegemony on the high end 20 nanometer too and delighted to be here and to answer any questions you may have.

Unidentified Analyst

Okay. So you brought up 20 nanometer and we were at the Intel analyst day where they also talked about their foundry business and they brought up Altera’s 14-nanometer. So can you help us understand, at least clear some of the – whether it’s the truth or misperception that at 14-nanometer Altera will have a meaningful die size advantage? And assuming they do, what are some of the things we should be looking at? Is it just the die size advantage? What are the architectural differences? And you can also go back and kind of walk us through what you did in 28 nanometer along the three categories?

Moshe Gavrielov

Sure. When you look at the product, there are numerous elements that go into it and if you look at 28 nanometer, though it’s similar disinformation – competitive disinformation about Xilinx’s position. So it sort of – I smile because I see the same thing repeated. So if you look at the results of 28 nanometers and we have over 70% market share, and that’s a result of having a well-rounded solution. There are five elements to that. One is the foundry. The second is the architecture where we have clear leadership with our fifth generation columnar architecture, it’s a technical detail but it gives us huge advantage. The circuits that we have used, that’s the third – actually enable higher performance, lower power. If you look at our 30s, they have much better characteristics than anything from the competition.

The execution is the fourth. It really has been spectacular. We had a lead on the first tape-out of 28, but that lead was just a month lead and yet we have 70% market share. At 20 nanometer we have over four months lead. So it’s an even bigger lead but we actually got into production. We are now ramping very rapidly over $50 million in the June quarter, over $80 million, well over $80 million in the September quarter and we projected hitting and exceeding $90 million in the December quarter. So it’s the fastest ramping product.

When you add all of those four elements, the last one – the fifth is the software, and that’s where we invested – about five years ago we started a big investment in redesigning our software suite. Our software suite is what our customers use in order to implement their solutions and our silicon [ph] – we designed it from scratch. As a result, what our customers are telling us is that the utilization of our silicon is typically 20% higher than what they can achieve on any competitive product. The turnaround time can be as much as four times faster. So when you add that, that’s the fifth element – the foundry, the architecture, the circuits, the execution, and the software, that has given us a major lead.

If you look at comparing our die size to the competitive die size, from the competition at 28 nanometer we are typically 35% smaller. So on the face of it, it should have been the same, right? Because they were using TSMC, we are using TSMC. So I am very confident and we’ve already proven it at 28, we are just about proving it at 20 nanometer where we have silicon four months ahead of the competition. The silicon works, it’s very competitive and we are working on the next node with TSMC, the 60 nanometer node. And I am convinced that not only in terms of schedule will be there but in terms of all of the attributes of the product, you need to look at the current generation and just see where we are, with so much better than the competition, that’s why our market share is there. So I am extremely confident in our ability to repeat this at the future nodes. And I believe that this noise from the competition is an affirmation of how far behind they are on the 28 and 20 nanometer nodes, and we expect to demonstrate and deliver leadership at this future node too.

Unidentified Analyst

So maybe we can go into market share. In 40, you came from behind and now at least per our estimates you are very close to 50:50 in terms of market share. 28, very clear leadership, 70:30, and when we go back and look at overall market share, for Xilinx versus Altera, you’ve stayed at around 55, 60 number. So what is the correct way to think about market share? Is it the Intel AMB or is it all in normalizing to 50:50? Or is it somewhere in between, and why do you believe it’s somewhere in between?

Moshe Gavrielov

Well, our products are moving into the center of our customers’ designs and as such they like to have alternatives. And at the leading edge, these new nodes 28, 20, 14, 16, whatever it is, the customers like to have two options. So the bigger customers tend to have a balance between the two companies. That balance tends to swing over time based on technology superiority and being the first to market. So after – since the 28 nanometer node, we really have changed that balance and the superiority on the silicon and on the software has really enabled us to grow market share. We expect that at the end of our fiscal year, it ends in March, we will have gained for the fifth [ph] year in a row significant market share. 28, we expect that 28 nanometer to drive our market share gains for another three years easily because it is now just starting to move into high volume, it might even last longer than that.

So if you look at our market share and the fact that we are over 70% now, that should grow and then the 20 nanometer restarts and then [indiscernible] to that. Typically achieving more than 60% market share at a node is really an incredible achievement. It just sort of highlights you have a major advantage and that starts to accumulate over time.

Unidentified Analyst

So just as a reference in 50 – in 40, what allowed you to get back?

Moshe Gavrielov

Well, in 40, we were late to the market but we had a broader product offering than our competitor. So that broader product offering has – enabling us even though we started a little late to catch up. So there is Spartan-6 product which is growing very rapidly and that is helping drive that at between Virtex 6 at the high end, Spartan-6 at the low end, between the two. Those were – we are starting to approach 50%. If you look at that a couple of years ago, we were probably significantly lower than that. So we have been catching up in terms of where we are and it is driven by Spartan-6 in this case. And there is – it’s probably a superior product and at least at this point in time there isn’t direct competitor to it. So I expect it to continue to thrive going forward.

Unidentified Analyst

And then you mentioned software and I think it’s probably not as well understood in the investor community, what has that allowed you to do? Can you help us – just kind of give us some frame of reference of when you came in, you looked at what capabilities were lacking and then what were you able to bring with your ADA background and then what is measurable in terms of – just design feature is not probably the right answer. But is it time to design or is it the different number of designs that you have been able to populate?

Moshe Gavrielov

Okay. So let me just explain in the very high level of business we are in. We monetize our technology by selling chips to our customers. When our customers buy those chips, think of them as a high-tech fan [ph] book which they then can customize to their specific requirements. So it’s not like an Intel processor where it’s designed and in terms of its product, we provide a general set of capabilities which is then customized by each of our customers. Typically 10,000 to 20,000 customers a year develop a whole host of applications, and in order to do that, they need a software suite. And we provide that software suite to do that.

Now when I joined, which it will be six years in January, and during the first three months I spent a lot of time with customers, and I got uniformly less than positive – significantly less than positive feedback on our software suite. And when I came back we recognized as our software which was – had been designed 10 years prior to that was incapable of addressing the requirement of the new technology. So we started from the ground up investment in the software suite and it was over thousand menus [ph] that was a Herculean effort. We released it starting from two years ago and now that is the mainstream software suite that we have. The result of that is that our customers can do their design much more effectively, typically four times faster in terms of implementation than they could with using the competitive software.

They can also achieve higher density. In other words, they can pack for into the design and that’s where they can have a more cost effective solution or lower power solution. So the software is really the portal into our silicon. Everything in terms of exploiting the silicon goes through the software and having a very efficient and superior software suite is like having better silicon. And now we have that and that big investment is paying off in terms of customer satisfaction, in terms of scalability, in terms of competitive valuation where our devices can achieve higher performance, lower power, high levels of density. So it actually is the biggest lever we have in these tens of thousands of designs we do every year and it’s a result of a big investment that was made over five years, and we believe that it basically gives us as equivalent to the generation of technology having the superior software.

Unidentified Analyst

Let me just switch a little bit over to what we – I know I have been wrong on it on trying to predict when LTE will have an impact. And I think collectively the industry has been not accurate in predicting when that will happen. But before I go into what we learned and what’s now all over the news in terms of the licenses. If you look at 4G versus 3G and we go back to the transition a few years ago when 3G was transitioning over 2G. If we calculate the – 4G spend versus 3G, it seems 4G is roughly 20% in terms of overall wireless equipment spend. But yet it has not really translated into, at least has not shown up in the revenues for either Altera or Xilinx and I know you will give me a different answer for the impact on Xilinx. But what I am struggling, I think most of us are struggling with, is it because maybe content is lower or 3G is going down faster than what we thought? Or is pricing having a big impact? So if you could help us understand in the framework of 4G has kicked in to some extent but yet it is not materialized into?

Moshe Gavrielov

So I mean there was a lot of sub-questions there. So first of all, if you look at it, then like politics, wireless deployment is local, right? So it’s different in each geography and 4G generation was initially done in Korea and Japan and then later in North America and Verizon was clearly the leader. But now what is happening is it’s accelerating in North America and AT&T is catching up and the smaller carriers are catching up. Verizon largely has finished its deployment which is at the core of their advertising, because they are ahead, they already highlight how good of a network they have. So there is opportunity for growth in North America with AT&T completing its deployment.

Europe has fallen behind. Europe used to be a leader in wireless infrastructure with the GSM, which is a 2G-ish to an 0.5G-ish solution. They’ve fallen behind and now with Vodafone I think they are going to invest very heavily to upgrade the networks in Europe. And what is happening and this is the one which is a big question mark was when will China start full fledge. So it has started, we actually saw parts in the September quarter, we expect to fill a lot of 28 nanometer silicon in the December quarter into 4G. The licenses were just granted last week. Over the next few years we expect for there to be an upgrade in terms of base stations of well over a million base stations. About two-thirds of that would be China Mobile and one-third would be the other players. The other players also have other protocols that they support. So those are going to be given out probably in 2014. So we expect a surge in 2014 in terms of wireless deployment in Japan.

Now if the sub-plot that you referred to was okay, so this is starting, it’s happening at different rates in different geographies, but why are we not seeing a huge surge is because the older technologies, actually the investments in those is dropping at a more rapid rate, in particular 3G we think it’s dropping very rapidly and basically the investment in 2G except in some emerging economies is very low, and that was going on for a long period of time. So 4G is growing. The content – the FPGA content in 4G is higher than it was in other applications. So we do expect that over the next three years, it will be a three-year deployment. It will help drive the wireless to higher numbers.

Unidentified Analyst

So just shifting over to the one – the other segments that don’t get a lot of attention, industrial, aerospace and others. Last three quarters, these segments have really grown both on a Q-over-Q basis as well as year-over-year basis, and I know the feedback I get from based on you is well that’s because of last minute buying. But can you help us understand – have we hit a main [ph] the curve that we should expect to see as 28 starts to also deploy in these end markets, because these are smaller end markets for you – they have not – they have outgrown the overall ASICs and ASSPs but not to the extent the concept [ph]. So what has driven this growth and what should we expect?

Moshe Gavrielov

Well, our technology portfolio now enables us to participate in lot of different markets and at 28 nanometer we have a product line which is called Zynq which has an embedded ARM core and that has been very successful in markets which have been secondary and tertiary for us in the past. So it’s a great expansion play in industrial, in automotive. It enables us to strengthen our position in Melero [ph] which is a pretty big market for us, about 15% of our business. And so the role of the FPGA is changing and that’s why we are replacing ASICs and the ASSPs in some of these additional markets. And it’s starting to happen now on an accelerated rate. If you look at our 28 nanometer shipments to date they have been – the fastest growing segment was communications but now the other segments are catching up and we expect that to continue to drive our business going forward. So I actually believe that these markets which ebb and flow. I mean they have a very unique characteristic and industrial is really an umbrella for – anywhere from 20 to 50 different segments. They tend to have a different dynamic and they grow slowly, but the best if you had to hedge in terms of these markets, automotive is a nascent market for us. We are at the early stages there but we expect things to help drive growth there. So we actually expect 28 nanometer growth in these other markets to be more or less equivalent to where we are seeing it on wireless and wired.

Unidentified Analyst

Okay. With that, I will see if there are questions in the audience. And as a reminder, if you would like to – we’ve got a question here.

Question-and-Answer Session

Unidentified Analyst

I guess, I will ask [indiscernible] first, you might have seen the slides from Intel’s analyst day, where they showed – basically Altera as a new foundry customer for them achieving, what they believe as lower die size and lower cost over time. What do you think about Intel as a foundry sort of enabling your competitor and what is it that you’d say?

Moshe Gavrielov

So Intel has great technology. TSMC is absolutely the best foundry in the world and the best foundry in the world manifests itself through the process technology, the service levels, the business terms, capability. We had options and we looked at all of the options in each of the generations that we decided to continue with TSMC that has definitely proven itself with 28, it’s proving itself 20, and we expect to have an incredibly competitive product offering at the next node. I believe that this Altera quoting Intel and then Intel quoting Altera and this vicious cycle of patting yourself in the back is very amusing but there was a similar set of claims at 28 which we’ve basically blown out of the water, 20 we feel very confident.

If you look at our die sizes of 28 using similar process node to that Altera used, we were both using the TSMC at that point in time, our average die size was about 30% to 35% smaller. So there is a lot that goes into the die size. TSMC is focused on having a very competitive product and for us through the architecture, the circuits, the software, et cetera I am very confident that at this next node we will have a great technology and it should enable us to continue to expand our market share.

Unidentified Analyst

I had one more, Moshe. What are the puts and takes on gross margin at the operating model? You guys have been at the higher end, is foundry getting more competitive and now that TSMC would love to keep you as a customer much longer than – as long as they can, is pricing getting better for foundry?

Moshe Gavrielov

Well, our gross margin has grown from the low 60s, mid 60s, it’s grown to 69%, it’s actually above our target. That’s a result of a lot of efforts. It’s more focus on customer design in terms of our foundry relationships. We are helping them drive defect density which reduces the cost of the die. It definitely helps if there are several foundries at these advanced nodes, that was not the case in the past. And so it’s all of these elements and a singular focus in terms of our operational execution which has enabled us to do that. So we continue that focus. It’s a result of probably hundred different initiatives in various areas, reducing the packaging costs, reducing the test costs, optimizing defect density making sure that we have our inventories done as efficiently as possible. And so all of those have enabled us to raise the gross margin to 69% which as you point out is the very high end of the range and the best we can tell it’s going to stay there for few more quarters. We haven’t commitments yet and haven’t made projections but all of the work to keep it high is definitely there.

Unidentified Analyst

Very good. Thank you very much, Moshe. It’s pleasure to have you.

Moshe Gavrielov

Thank you.

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

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

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

Source: Xilinx's CEO Presents at BMO 2013 Technology, Media & Entertainment Conference (Transcript)
This Transcript
All Transcripts