Rick Muscha - IR Moshe Gavrielov - CEO Lorenzo Flores - CFO

William Stein - SunTrust Ambrish Srivastava - BMO Capital Markets Blayne Curtis - Barclays Tristan Gerra - Robert W. Baird Ross Seymore - Deutsche Bank C.J. Muse - Evercore David Wong - Wells Fargo Securities Joseph Moore - Morgan Stanley Vivek Arya - Bank of America Toshiya Hari - Goldman Sachs John Pitzer - Credit Suisse Chris Danely - Citigroup Srini Pajjuri - Macquarie Research Chris Caso - Raymond James Ruben Roy - MKM Partners Vijay Rakesh - Mizuho

Good afternoon. My name is Skinner, and I will be your conference operator. I would like to welcome everyone to the Xilinx Second Quarter Fiscal Year 2018 Earnings Release Conference Call. All lines have been placed on mute to prevent any background noise. After the speakers’ remarks, there will be a question-and-answer session. [Operator Instructions] I would now like to turn the call over to Rick Muscha. Thank you. Mr. Muscha, you may begin your conference.

Thank you, and good afternoon. With me are Moshe Gavrielov, CEO; and Lorenzo Flores, CFO. We’ll provide a financial and business review of the September quarter, and then we’ll open the call for questions. Let me remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the Company. We wish to caution you that such statements are predictions based on information that is currently available, and actual results may differ materially. We refer you to documents the Company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important risk factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements. This conference call is open to all and is being webcast live. It can be accessed from our Xilinx Investor Relations website. Let me now turn the call over to Lorenzo.

Thank you, Rick, and hello, everybody. Sales in the September quarter increased for the eighth consecutive quarter to $620 million, up 1% sequentially and up 7% on a year-over-year basis. Sales for the first six months of the fiscal year were also up 7% versus the same period of the prior year. Our growth continues to be driven by our advanced products which grew 21% year-over-year, supported by the overall strength of our product portfolio. Gross margin was 70.2%, in line with our guidance. Operating expense was $249 million, less than guided as R&D came in slightly lower than expected. Operating income for the quarter increased 2.6% sequentially to $185 million and our operating margin was 29.9%. Tax rate for the quarter was 10% including a discrete item related to the accounting standard for the treatment of stock-based compensation. Our net income for Q2 was $168 million, or $0.65 per share. Some key points on the balance sheet and cash flows. We ended the quarter with $3.7 billion in gross cash and $1.9 billion net cash after our debt. Operating cash flow was $202 million. We returned $257 million to shareholders during the quarter in the form of $87 million in dividend and the repurchase of approximately 2.6 million shares for $170 million, an average price of $65.04 per share. We continue to execute on our share repurchase program with the intention of exhausting our authorization over the next several quarters. We currently have $445 million remaining on that authorization. We ended the quarter with diluted shares of 258.2 million shares, which included the impact of 7.6 million shares from the warrant associated with convertible debt. The convertible note itself had no impact on shares for the quarter, as the note was completely settled in mid June. For a complete explanation of the impact of the outstanding warrants on share count, please refer to our convertible FAQ on our Investor Relations website. With our stock price higher since the time we established our FY18 share count guidance, we are now expecting ending share count in Q4 FY18 to be in the range of 255 million to 260 million shares. Now, onto guidance. In the December quarter, we are expecting sales to be between $615 million and $645 million. Our backlog is up significantly heading into the quarter and we expect continued growth in our advanced products. From an end market perspective, we expect industrial and A&D to increase again. Communications is expected to be approximately flat while broadcast, consumer and automotive is expected to decline. Our gross margin will be approximately 69% to 71%. We expect operating expense to increase to approximately $260 million, with R&D driving the increase. We continue to invest in both our technology and our customer relationships as we expand our leadership and our markets. Finally, our tax rate is expected to be between 11% and 14%. Our financial performance year-to-date and our guidance for the third quarter indicate to us that we are on track to our annual guidance that we provided at our analyst day. Let me now turn the call over to Moshe.

Thank you, Lorenzo. I’m pleased with our consistent execution as demonstrated in the September quarter financial results. We delivered our eighth consecutive growth quarter with sales increasing sequentially to $260 million. Our operating profit increased approximately 30%, enabled by gross margin and with slightly better than expected and tightly managed operating expenses. We continue to benefit from the combination of our diversified multi-market portfolio coupled with clear technology leadership over the three most recent nodes. Five of our eight end markets increased in the quarter. Our industrial, aerospace and defense primary end market set the new quarterly record with sales of $278 million an increase of 17% from the year ago quarter. The broadcast consumer and automotive segment also delivered significant sequential and year-on-year growth. On the product front, our advanced products category delivered record revenue. This category, which already represents 52% of our overall revenue, will continue to be our growth driver in the foreseeable future. Our Zynq SoC platform, critical element of our multimarket expansion effort, delivered another sales record of growing 65% from the year ago quarter. The Zynq SoC platform now comprises 12% of our overall revenue. Expanding on this theme, our industry disrupting RFSoC family targeted at 5G cable and wireless backhaul applications is already shipping to numerous customers. 20-nanometer sales increased very significantly and also attained the new sales record, driven by strength in multiple end markets. 16-nanometer portfolio continued its accelerated sales ramp, again setting a new record. We’re now shipping 34 unique products to more than 930 broad market based customers. This is a very substantial sequential increase of 50% in the number of unique devices, 75% in the number of customers, both are significant harbinger of future success and demonstrate our technology leadership as noted. I’m gratified by the recent significant announcements highlighting our strong customer momentum in cloud computing. This is one of our key medium to long term market expansion opportunities. As projected in our previous earnings release, Amazon Web Services has made a uniquely differentiated SDAccel development environment available to F1 users while also expanding their FPGA as a service F1 deployment to multiple region. In addition, Huawei announced their FPGA as a service instance based on our Virtex UltraScale+ FPGA. Finally, Alibaba, the largest cloud service provider in China recently announced their next generation FPGA as a service F2 and F3 instances, both based on Xilinx technology. The significant type of scale customer adoption momentum is enabled by the combination of silicon leadership coupled with advancements we’ve made with improving ease of use. This will enable a much larger community of software and application developers to derive full benefit of FPGA acceleration. Finally, as we look forward to the December quarter, our guidance of 615 to 645 is expected to be driven by the accelerated growth of our advanced products portfolio delivered across numerous end markets and all three of its technology components. Achieving the midpoint of this guidance will establish a very significant new sales record for Xilinx. Consequently, we remain confident that we’ll deliver on our fiscal year 2018 revenue target of approximately $2.5 billion. In addition, while we continue to invest in our technology leadership, share gains and market expansion efforts for sustained revenue growth, we remain committed to delivering 30% plus operating margin exiting this fiscal year, as well as preserving our commitment to returning cash to shareholders. Now I’ll turn the call back to the operator for Q&A.

[Operator Instructions] And our first question comes from William Stein from SunTrust.

Moshe, I’m hoping you can help us sort of align what we hear as very good traction in the early stages of these deep learning projects by the web scale providers on the one hand versus sales coming in a little bit lower than expected in that end market for the quarter, and perhaps some effervescent optimism about that end market? I’m wondering if you can help us sort of think about how we should expect that end market to trend over the next few quarters and beyond.

Actually, sales didn’t come less than expected; they came where they were expected. If you look at communications on a whole, wireless was down but data center business was up. But, we are at the very, very early stages of deployment of this technology. And it takes series of time, typically years from when the deployment starts to when the potential is realized. As such, what we are directing everyone is to $200 million to $300 million in calendar year 2020, and that’s still a plan and we are on plan to deliver that number. It just takes this length of time. The biggest impediment we had was expanding the potential customer base. And programmable logic in the past has been directed only at hardware designers. And in order to make it available to a broader audience, we have developed this SDAccel capability that is now being deployed. That should enable the adopters beyond the traditional existing users of our technology and that will -- that gives us confidence in the potential of achieving the $200 million to $300 million. But patience is important on this because it just takes time until it has that viral impact. And the broad acceptance by, by far the biggest things we think is huge stepping stone towards that. And as the -- that together with the SDAccel are the enablers. And over time, we do expect to see this translate into revenue at a faster rate, but we have been very clear with regards to setting expectations as to when that will happen and we continue to project the $200 million to $300 million in the 2020 timeframe. It’s the right thing to plan [ph] at this point.

One more, if I can, related to the same topic. Investors are seeing some outstanding growth from NVIDIA in GPU for similar applications. Maybe you could give us a very summarized version on how you compete with that technology among the web-scale providers and all the engineers and companies that want to develop AI applications, what’s the sort of pitch for FPGA relative to GPU today?

Sure. So, the biggest historic advantage, and NVIDIA had two significant advantages, one is they recognized this market very early and pursued it several years ago. And as a result of that what they’re seeing now is the fast growth but that similarly to our predictions, it took several years until it manifested itself in revenue. The other advantage they had is that inherently a software -- it’s a software program technology and now we’ve gotten to the point where our technology can be addressed in the same way. The advantages we have largely in the area of providing overtime superior performance in a broad set of applications, typically at lower, significantly lower power point. And over time, we expect that to be important in a broad set of applications. So, that’s the value proposition and that’s why you’re seeing these deployments that these -- obviously very significant hyper scale companies are one of the biggest issues when you look at these big compute firms is power is the big limiter. And we can provide them with either higher performance in some applications, not in all of them or definitely on a broad basis we can provide similar performance at a much lower power points. And the combination of those two are prevalent and are likely to provide over the next few years the opportunity for growth for us. But that’s the biggest difference is the delivering the software accessibility, which we now have and also catching up, which we’re executing with vigor on that front too.

And our next question comes from Ambrish Srivastava from BMO Capital Markets.

I’ll be courteous to my fellow analysts and ask only one question. It may not be your favorite one, Lorenzo, but certainly one of my favorite ones, accounts receivables. Going up two quarters in a row, they went up last I think three or four quarters ago and then you brought them down. But, I’m just looking at Q-over-Q and Y-over-Y, receivables are going up higher than sales, both on Y-over-Y and Q-over-Q comparison. Could you please help us understand why that’s the case?

Sure. I think in general, the same phenomena is happening over the past couple of quarters that happened when we had the relatively high numbers in the past, and that’s just timing of shipments in the quarter. As you may recall, we had a quarter where our receivables got very high; next quarter, we reported very low receivables, and that’s about a year ago. And the key thing we monitor on that is the quality of the credits underneath that, -- are these customers going to pay? And the answer is in general, it’s a very, very, very high rate; we have minimal, negligible write-offs on the receivables. And so, we’re highly confident, this is all collectible and will show up in cash flow. And we’ll see different patterns maybe in the future based on timing of shipments during the quarter. But that’s really all what happens. Shipments in the last months tend to be collected the next month or the first month of the next quarter.

Our next question comes from the line of Blayne Curtis from Barclays.

Moshe, I was wondering if you could just go back to the com business. You mentioned in the September wireless was down. Just curious what trends you are seeing in wireline. And then, I know visibility has been kind of low for people, just curious your perspective as to when wireline can come back?

Well, if you look back at the previous quarter, the June one, wireless had a huge surge, and that tends to volatile. And indeed this past quarter, the September one, it trended down quite a bit actually. And our expectation is that wireless in December will go up again, maybe not as high as it was in June, but we should see growth there. Wired, generally speaking has been flattish with somewhat of a downturn, downward trend, it sort of -- not a huge downward trend but it is trending down, and we are not at this point predicting any huge surge; it’s likely to go sideways or maybe continue on a very slow downward trend this coming quarter. If you look at our overall results and Lorenzo broker it down into the three categories, then, our expectation overall is for communications to be flattish and very significant growth on the industrial; and the third market, the broadcast, auto and consumer to be slightly down. And that should net significant upside overall.

And our next question comes from Tristan Gerra from Robert W. Baird.

Looking at data center, are you seeing interest from anybody for a Xilinx FPGA becoming the first access point in the data center, basically delegating the CPU as a second integral? [Ph] And how big is the opportunity for FPGA to become more of a central point in some data center application in the medium term?

If it become any time soon, the first one, then my $200 million to $300 million would be ridiculously low number. And so, what is happening is a transition, which is now accelerating in the industry and it’s a transition where CPUs have basically exhausted their ability or have only minimal ability to provide higher performance to these very parallel applications. As a result, there is this transition to heterogeneous architectures, which have a CPU and have other forms of acceleration. GPUs have been pursuing that several years now and now it’s very significant business for the GPU vendors. And programmable logic has emerged as a third alternative. But, we don’t expect it to supplant CPUs totally. And that’s -- but it will provide a very significant enhancement for this heterogeneous computing and will provide an alternative and that it’s not that CPUs will go away but there’ll be a need for less CPUs for some of these applications and they will be replaced by other modes of acceleration in particular programmable logic over time.

And our next question comes from Ross Seymore from Deutsche Bank.

Moshe, I wondered if you could just go a little bit into the broadcast consumer and automotive segment. That had a nice pop sequentially. So, could you just describe what that driver was behind that? And then, with the guidance for that to be down a bit in the quarter, I know the automotive is the sexiest part about that. But, could you give us a little more detail on what’s driving that sequential decline as well, please?

Well, I believe that all three of the sub markets actually grew quarter that we just reported. And we expect automotive actually to continue growing but the other two to shrink somewhat. And overall, it’ll be flattish to slightly down with -- I mean, that’s our current prediction with automotive continuing to grow. And these markets are actually quite desperate, you can call them the other markets to some extent, and they’ve different dynamics, which drive them, automotive what is driving our growth is continued deployment of numerous generations of driver assistance technology and that’s sort of the biggest driver of the technology there. In consumer, it’s a whole plethora of applications, which from time-to-time have surges and we benefitted from one of these surges this past quarter. And broadcast has been doing well but it’s a small market for us and that’s a market which is going through structural change where as a market generally, it’s not growing in a big way and there’s a transition to the cloud in some of the broadcast applications. So, that’s one of the inhibitors to its growth. That’s about as much color as I can share with you at this point.

And our next question comes from C.J. Muse from Evercore. C.J. Muse: I guess the follow-up to Ross’s question on ADAS. I believe this was roughly a 5% business for you guys in fiscal 2017, curious if you think this could be a 10 plus percent business over the next two to three years. And then, specific to today, can you provide an update on customers transitioning to 16-nanometer and how this is impacting both unit demand as well as your ASPs? [Ph]

So, let me start from the last question, or the last sub question that you asked. We are seeing in ADAS transition from our original Zynq product to our MPSoC, which is the 16-nanometer generation, pickup there has been quite substantial. And quite a lot of the customers are now deploying the second or third generation, which is in 16-nanometer and we expect that to be a big growth driver for us going forward. We do expect this market to provide double-digit growth per year with the potential for that continuing for next few years. And that’s just with the -- for the deployment of ADAS and the new generation of ADAS, and then early versions of automated -- autonomous driving, which will have very powerful, centralized control unit, but then will continue to have a lot of additional auxiliary technology required to implement the solution. And we will likely play significant role in the second part. We are not differing the first ultra-high performance processing portion of it. And so, if you look at the arithmetic, could it get to 10%? Yes, it’s sort of moving that direction but I think it has a few years ahead of it to deliver on that potential.

Just to clarify, C.J., was your question on 16-nanometer related to the broader adoption or just in ADAS? C.J. Muse: It was just within ADAS.

Okay.

And then next question comes from David Wong from Wells Fargo Securities.

Further on the thoughts products for ADAS. In the future, will there be any specific types of hard course or other specialized types of circuitry that you might want to incorporate in your offerings for ADAS or autonomous driving applications?

I’ll let Lorenzo, go first because Lorenzo drives much fancier car than I do and that you can take that one for the bank. So, I think he will be more of a true consumer advocate of this. And then, I might add color later.

My car is 10 years old and the driver assistance is in the form of mirrors. But the serious answer to your question, David, is embedded in the products we have today, particularly the Zynq-based SoC products are hard in IPs that are well-targeted to the needs of our customers who are developing ADAS applications, not the least of which our features that enable safety and security at the device level. So, yes, I think we continue to evolve our platforms in that area with -- again, overall, it drives a greater integration of capabilities that really provide that longevity and the platform design for our customers. So that’s the overall view of what’s going in our products that will be suited for ADAS. Moshe, do you want to add anything?

Yes. And specifically, if you look at the MPSoC and I’m tying this to previous question, if you look at that that’s a 16-nanometer product, then there is hardened safety, hardened security, there is hardened [ph] graphics, there is a lot of hardened cords. You just can look at that. A lot of those were added there based on request from our automotive segment, and that will continue to be the trend going forward.

Our next question comes from Joseph Moore from Morgan Stanley.

I guess one clarification, I think you said, Lorenzo, that backlog was up significantly. Is that quarter-on-quarter? And if so, how does this extend backlog and then sort of flattish guidance or slightly up guidance, are you just seeing longer lead times, can you talk about what’s driving that?

So, let me talk. Yes, backlog is up significantly in the quarter. And this actually relates back to Ambrish’s question on AR. We do have an unpredictable, if you will, pattern in any given quarter of what goes in the book, when the customers want to ship. And, we work with our customers obviously to hit their schedules and their requirements and drive our business at the same time. So, there are occasions when we’ll have the book fill up toward the end of the quarter and that kind of carries forward into the next quarter, and that’s the phenomenon we’re seeing. So, no, it’s not a function of lead time or anything structural like that; it’s a timing thing.

And then, for industrial, you’re obviously growing a lot there. I mean, can you talk a little bit the sustainability of that? And how much of that is sort of the share gains you guys have been talking about, whether it’s Zynq related or process node related, versus just a stronger industrial environment?

I believe, it’s a combination of the two. And the environment is stronger and we’ve been waiting a long time for this to happen. But now, we had a very, very, very broad set of design wins in industrial, which was Zynq based. And we were getting a little antsy because they were taking a long time to translate into revenue. Good news is that that is happening now, happening very broadly, and it’s where we believe we have a very strong position vis-à-vis a traditional competition but we’re also seeing fewer and fewer dedicated ASSPs which are targeted this extremely fragmented market. It’s just the market which has built up a lot of very small niches and it’s very difficult to come up with an ASSP that addresses them all. And so, our position there we believe is very strong. And what we’re seeing now is the Zynq design wins, 28-nanometer move to production. If you look at Zynq, then the first generation was primarily wireless oriented and there was a lot of that and that’s happened relatively early. Then the second generation, which is continuing now is in driver assistance in automotive. And now, we’re seeing the expansion into a lot of different markets and the industrial also -- aerospace and defense is also, we’re starting to see those translate to revenue now. So, both the markets we believe are doing reasonably well or actually better than they have been in long period of time. And our competitive position is stronger than it has been as this market continues to be very fragmented and there’s less, very targeted ASSPs that can pick off these small niche applications.

Joe, I want to go back to something I actually didn’t address in your question that I picked up, which is you said, a flattish guide. I think I would want to point out that the midpoint of our guide is $630 million, which is meaningfully up from where we are today, so strong backlog is consistent with that.

Our next question comes from Vivek Arya from Bank of America.

Moshe, I just wanted to revisit the assumptions underlying the $200 million to $300 million opportunity that you have set out of calendar 2020. Is that opportunity just for Xilinx or do you think that’s for the entire FPGA market? And if it is for all the FPGA makes, then, how does the competitive landscape impact that number? Because even if it may not be all the way there today, surely in the next three years, they can use their incumbency to add more competition in that market. So, any color there would very helpful.

Sure. So, we -- I agree with your observation. When we look at the 200 million to 300 million, this is our number. We expect the overall number for programmable logic to be significantly higher, at least twice as much. Intel does have a very significant design win in this area with Microsoft that’s being very well documented in its public domain information. And this is just starting. And there is seven major hyper scale companies and then there is the two additional ones who are pursuing this with vigor, at least two that we’re involved in, which is Huawei where have the design wind and IBM where we have a very type relationship. So, if you look at our overall position together with the ones we’ve already announced and we’re engaging with all of the others. The 200 to 300 is what we believe we can do in data center in its various forms in that time frame. And the market in of itself we expect to be quite a bit larger than that, at least twice as big. And obviously Intel has its incumbency that is true but we have superior product on the acceleration side. And we believe it will get a significant share of the market based on that. Market overall should be significantly more than that.

And as a very quick follow-up, maybe for Lorenzo on OpEx growth and I know you’re not providing the guidance for next year yet. But, at what point should we start to see your sales growth do better than OpEx growth? Thank you.

So, you’re right. We have not provided any FY19 guidance yet and we will do that in the course of time. Just on OpEx, I will -- want to repeat something I’ve said in previous calls with respect to the flow this year; we are expecting it to decline in the fourth quarter of this year. We have some local highs here in this current due to take out expenses primarily. So, we expect it to decline in the fourth quarter.

And our next question comes from Toshiya Hari from Goldman Sachs.

I had a follow-up question on your cloud computing business. Last month, at your developer forum, I think in the keynote, the gentleman from the AWS gave a very constructive and positive commentary on your offerings and how they would expand that business going forward? And obviously, from a customer diversification perspective, you have announced couple of partnerships. So, I guess the question is, I appreciate this is going to take a long time, but what would you need to see to gain incremental confidence then for you to raise that long-term number?

We would need to see broad adoption, and that tends to be viral, so you get the early adopters, we start using it, some of them start seeing superior results. It generates some excitement, and then you start getting more and more of these application developers. And I think, the key metric will be when will we see the next acceleration in the application developer. Realistically that couldn’t happen up to now because it was limited to FPGA designers with SDAccel and all of the additional things we are doing. We now have provided the foundation for that. And expectation is that it will feed upon itself, once they start seeing superior results in terms of acceleration. And so that will be the thing to watch. And I think the fact that you are seeing Huawei and Alibaba within less than a year jumping in and following in Amazon’s footsteps that is a proof that there is a lot of potential here, because you now have three very significant players providing this as a service. Amazon obviously started a year ahead of everyone and is still way ahead of everyone, at this point, but the others are going to pursue them with vigor, as you can tell from their announcement. And the expectation is that the applications, which will benefit from this -- from parallelism, which is inherent and available, and programmable logic will be the driver for expansion of this, both in the cloud and over time in embedded applications too.

Our next question, we will retry the line of John Pitzer from Credit Suisse.

Quickly, Lorenzo, going back to the OpEx and correct me if I’m wrong, I thought that this year was going to be a relatively tan [ph] year for tape-outs and that would have been a tailwind for OpEx reductions and what sort of offset that was some of the software development work you needed to go after new application acceleration. One, was that wrong? And I guess, two, on the second side of software OpEx to go after new opportunities. Has that expense peaked or is that something that will continue to grow out over the next couple of quarters and how do we think about OpEx around these new applications?

I think the best way to think about OpEx with respect to those new applications and in general is, go back to the guidance we have provided for the full year, which is going to be between 990 and 1010 millions of dollars for the year in terms of OpEx. Take the guidance I just gave and add to it the first couple of quarters, we are right around $750 million of OpEx for the year. We are going to be -- going to be in our range. We’ve accommodated both the new investments that you talked about with respect to software development and our tape-out schedule. The comment I made around tape-out was explaining why this particular quarter we’re going to see OpEx go up and then go back down in Q4.

And the current tape outs are due to 16-nanometer going to production and happening faster than we had expected, plus a few additional expansion products like the RFSoC, which was sort of new. And so, it was never expected to be the low tape-out there. And the reason that it’s a high tape-out here is the technology is doing so well or relatively high tape-out, the technology is doing so well and it’s in such pristine condition that we are now rolling out the entire product offering, and that’s why we have 34 unique products already shipping. And that’s been -- you multiple that 34 number by several million dollars for each tape out. That’s sort of an expensive occasion. [Multiple Speakers]

My second question is just a follow-up, Moshe. Last week when TSMC reported, I was a little bit surprised after the first time ever they called out sort of crypto currency ASICs as being a meaningful driver of their business in the quarter. And if you start to look at the breakdown of some of these crypto currency hardware devices, often times you are finding an FPJ sitting next to these ASICs. And so, I’m kind of curious, is this the sizeable market that you guys think about, how big today under what sort of end market category are you putting this or how do we think about the size of this market and I guess importantly the sustainability?

Actually, you’ve framed the question quite well, John in that. So, maybe, we are little too elliptical when we are talking about our -- the strength in our consumer business. We do see -- we do have meaningful business with some of the crypto currency providers, like you said the role our product plays, which is typically a Zynq-based product is as a controller, if you will for these ASIC-based mining platforms. We don’t -- at the same time, we say that we acknowledge this meaningful business in this quarter. We obviously want to see some longevity in that business and breadth of that business before we point it out as a key driver. We are really happy with it right now but we don’t want to get a little bit overheated in talking about the potential for this. We do, as I said, I’ll go back to same words, some longevity in that business and some breadth across multiple customers for that business for us.

And our next question comes from Chris Danely from Citigroup.

On the communications and data center business, how much these days is comp versus data center? And then, where do you expect that to be in say year or two?

As we continue to talk about the nascent state of the data center business, or at least emerging hyper scale data center business. And as you know, Chris, from following us, we’ve always had storage and network-based business in the data center. It’s about the same size as it ever was. The vast majority of the data center, I mean, the communications and data center business is wireless and wired. And then the true data center piece has been around 5%. And that’s again transitioning from the older businesses that we had their into this newer hyper scale data center opportunity.

Any projections for year two?

No, not yet

Well, 200 million to 300 million, and it’s not there yet. So, it’s significantly less than that, but that’s the projection.

Our next question comes from Srini Pajjuri from Macquarie Research.

Hi, Moshe, just a question on competitive landscape. It looks like Intel is finally shipping their 14-nanometer part. Just curious as to what you’re seeing out there and which end market, if any do you think that they’ll be more competitive versus less competitive?

Well, from everything we’re hearing, there continue to substantially issues on performance, power and software. And to the extent that they have shifted with relatively selectively to their very, very largest customers, which makes a lot of sense. We, to contrast, shipping it now, it’s close to a 1,000 customers and we have 34 distinct products and counting, so already shipping. So, we believe that the leadership we have in terms of the breadth of applications that we can support, quality, the lower power, the better service performance are still very significant. And what we’ve heard from customers is that the lead we have in terms of the majority of the technology is a year and a half or more, which is very significant, it’s probably more than we’ve had any point in time in the past. So, not saying they’re not around but so far our breadth and quality are helping us prevail. And if you look at the results on the 16-nanometer, it’s growing at an incredibly fast rate, it’s actually growing much faster than we had expected it to grow, at this point in time. So, we continue to do well with this technology. And we never rest on our laurels as we have a very worthy competitor, but we’re raising to exploit the huge lead we have.

And then, just one clarification, Moshe. I think you said wireless will grow in December quarter. I know it’s a lumpy business but just curious as to what’s driving that growth in December, whether it’s China or India or something? Any color would be helpful.

Well, we think it’s all of the above. We think that the past quarter was a substantial low quarter -- substantially low quarter. And that what you’re seeing is we saw some inventory reduction at our largest customers and we think that now they’re going to be shipping again based -- and we’ll benefit from their lower levels of inventory which they had built up before, and that should help us across the board but in China and India in particular too.

And our next question comes from Chris Caso from Raymond James.

My question is on automotive. And I guess what we’ve seen presently in ADAS system, there is generally some content from PGA, some purpose-built silicon and general purpose. As we go forward and that industry matures, what’s the reason that FPGA stays on the board and what’s the reason that that content stays with FPGA and wouldn’t necessarily consumed into ASIC or general purpose?

Okay. That’s the great question. What is happening is this market is evolving very, very, very quickly and much faster than it typically has in the past for traditional automotive. And there is huge race to provide higher content and to address new sets of sensors and things of that nature. And those are evolving very, very rapidly. So, if you look at where we’re seeing success is we are seeing success not with our traditional programmable logic, but actually with our Zynq product line. And as the Zynq think product line is actually very powerful embedded processing, don’t think it’s just this programmable logic; it’s actually a very powerful embedded processor with a lot of hardware built accelerators and tightly linked with programmable logic. Now, the value of the programmable logic there is that it enables the customers to have a platform-based solution where they can adapt the platform very quickly to address numerous models and numerous changes. And as this market is evolving very, very quickly and there is new sensors and new algorithms and there is some elements of machine learning which now being integrated into it, then the programmable logic is actually very well-suited medium to address those. So, that’s why we believe that there will -- as it continues evolve rapidly, we can actually provide a fine tuned solution with the control, with a lot of the embedded hardware, accelerators and with the programmable logic. And so, it’s sort of a different from the traditional ways of thinking of programmable logic but it’s just used for prototyping. Here it’s actually -- the Zynq product offering covers huge portion of what the customer requires; the embedded hardware addresses other areas; and then the programmable logic and its flexibility is actually well attuned to rapidly changing market. Where I do agree with you is saying 2030, 15 years from now, if the whole market becomes commoditized and doesn’t move, that’s when it is likely to be less advantageous to have these capabilities. But, there is a long way between where we’re now and when it reaches that stable point. And as -- even though there is a tremendous level of excitement on autonomous driving, it’s going to take several years before it becomes the prevalent form of transportation. And that’s likely to take tens of years in reality. And as it evolves over this period of time, it will change, and we are likely to continue to have attractive value proposition for all of the areas which are dynamic. Now, just to sort of highlight, so I’m not overselling, we are not addressing the ultra-high performance central processing element, the ones which NVIDIA is pursuing and Intel is likely pursue and potentially other players, we are looking at the rest of the market even when it evolves to autonomous driving; there’s going to be a lot of additional auxiliary functions which we will be very well adaptive.

And our next question comes from Ruben Roy from MKM Partners.

Moshe, I had a follow-up on the competitive discussion you were just having a couple of minutes ago. It seems like you are competitor, Intel, has recently increased their focus on 10-nanometer, and it seems a little surprising given the issues, the well-known issues they’re having with 14, the conversation has shifted to 10 at this point. But just wondering from your perspective and now that you have some product in hyper scale data center customers and various new applications. Are you hearing any feedback from customers around acceleration potentially of your own roadmap or how is that all working out?

Well, we have accelerated 16-nanometer very significantly, and that the significant number tape-outs this year were to move it all into production. That’s where we believe we have the broad leadership at the current nodes. And so, I think we’re in violent agreement with you on that front, as our customers. We are working feverously on our 7-nanometer solution with TSMC. We expect to tape that out next year. And we are very pleased with the progress that TSMC has made. And I was just at their 30-year anniversary and they’re full steam ahead. We are -- very, very, very big mobile customers of sort of standing in line, they are all there and they are targeting their next generation 7-nanometer and we will be immediately after them, taping out our devices and benefiting from the huge learning that that volume will run through the system. So, in this business, if you ever stand still, you are likely to have the tortoise and hare scenarios. So, we are running as quickly as we can to make sure that we continue to exploit our leadership and prolong it into the next generation. For us that is taping out next year and it is going to be a phenomenally attractive generation of technology.

Just a very quick follow-up, just point of clarification on the commentary on data center and percentage of revenue et cetera. When we think about traditional products that you are selling into whether it’s networking, storage et cetera. How do you think about that? Is that sort of a steady state type of business that’s going to grow at low single digit percentage type of growth rates? And the new stuff is really just starting out from a very small numbers today, obviously with the expectations for fast growth or as the traditional stuff sort of maturing in and following off while the new stuff comes up?

Well, that entire industry is going through a transition and significant portions of it are getting cloudified. And when they get cloudified, there is opportunities on the networking and storage as well as the acceleration side. So, what we expect to see is portions of what we called wired communications, reemerge as our data center business but they will still be used for storage and networking. And it’s sort of somewhat similar with regards to the technology requirement. So, that’s likely to be a significant transition for us. But, it will probably offset each other without having tremendous impact either away. The growth area is acceleration and the acceleration is the growth area because it really provides an alternative to the traditional CPU computing. So, I think your read is correct that the transition is happening and the classification is going to change and to some extent, it might at some point become a little more difficult to figure out what is wired business and what is data center business. And at this point, we look at them as adjacent and transitioning from one to the other, but the technology requirements are similar on the networking and storage side, and they are unique and different on the acceleration side. And that’s the growth opportunity. So, hopefully that provides some color to your question.

Our last question comes from Vijay Rakesh from Mizuho.

I was just wondering when you look at, you talked about incenting [ph] and how you see Intel performing down the road. Have you seen NVIDIA coming into incenting market as well? And today the GPUs are probably bit tied but do you see them being a viable competitor in that market too?

They are definitely focused on influence too [ph] and they are coming out with new versions of their technology, which are quite departure from their traditional GPU architectures. And if you blur your eyes, they actually look a lot like more technology that we have, a much better match for than the traditional GPUs. And I think what they are doing makes sense. They have a good position and good starting point. But for broad applications and influencing, different architectures than traditional GPUs or GPU derivatives are required. And as you look at some of their new solutions, they actually are moving in that direction. And we believe we are ahead of them in the flexibility we can provide actually as the advantage, the inherent advantage we will continue to have over time.

Okay. Thanks for joining us today. We’ll have a playback of this call beginning at 5 p.m. Pacific Time, 8 p.m. Eastern Time today. For a copy of our earnings release, please visit our Investor Relations website. Our next earnings release date for the third quarter of fiscal year 2018 will be Wednesday, January 24th after the market close. We’ll be attending the following conferences this quarter: The Credit Suisse Technology Conference on November 28th in Scottsdale; The NASDAQ Conference in London on December 6th; and The Barclays Technology Conference in San Francisco also on December 6th. This completes our call. Thank you very much for your participation.

Once again, this does conclude today’s call. You may now disconnect. Thank you for your participation.