Advanced Micro Devices Inc (AMD) — Q2 2019 Earnings Call Transcript

Thank you, Erica, and good afternoon, everyone. With me are Victor Peng, CEO; and Lorenzo Flores, CFO. We'll provide a financial and business review of the September quarter, the business outlook for the December quarter and the revised outlook for fiscal year 2019. We will then 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 that actual results may differ materially. We refer you to the 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. As we discussed in last quarter’s earnings call, in addition to GAAP financial measures, we will also be disclosing certain supplemental non-GAAP financial measures used by management to evaluate the company's financial results. We provide these measures to facilitate period-to-period comparability for purposes of evaluating continuing business operations, by excluding the effects of non-recurring and unusual items, such as amortization of intangibles and certain one-time items related to acquisitions. We believe that sharing these non-GAAP measures will be helpful for analysts and investors in analyzing the company's ongoing core business. A reconciliation of non-GAAP financial information to the closest GAAP measure is included in our earnings release and has been posted on our Investor Relations website. 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 Victor.

Thanks, Matt and good afternoon, everyone. I'm very excited to report that executing on our new strategy is returning results in FY19 that are well ahead of our original plan. We have broad strength in our business in Q2 with growth in six out of nine of our end markets. Establishing ourselves in the data center and accelerating growth in our core markets resulted in a new quarterly revenue record of 746 million, which is up 19% year-over-year and drove the non-GAAP EPS, up 30% year-over-year. Given our first half FY19 results and the continued near-term strength we see in multiple markets, we're raising our guidance for all of FY19 to be between $2.95 billion to $3 billion. Now to be clear, we are watching the macro environment very carefully. Nonetheless, we feel this is the appropriate forward guidance. Let me share some recent business highlights. Communications was very strong in Q2, driven by LTE upgrades, pre-5G and some early 5G deployments. Our wireless business grew very significantly with broad-based strength in both radio and baseband applications with major OEMs across multiple geographies, and we resumed shipping to ZTE. Our wired business grew due to customer transitions to next-generation products in several applications, including OTN Metro, access, and data center interconnect. In the data center segment, we continue to build momentum with hyperscale companies and acceleration deployment. For example, Twitch, a subsidiary of Amazon, shared at our Developer’s Forum how they achieved the industry’s first broadcast quality live streaming platform using a new video and coding format, VV9, all accelerated on our FPGAs. Samsung announced at their Tech Day their smart SSD product, which was the result of a year-long collaboration between our companies. Samsung’s smart SSD combines our 60-nanometer ultrascale plus FPGA with their leading storage technology for accelerating computing near the stored data. With regard to FAS, our leadership position in the China market expanded as Alibaba changed their FPGA as a service from beta access to general access. This falls on Huawei’s general access that occurred with their FAS program this past June. Recently, Amazon Web Services doubled their FAS availability zones from 4 to 8, adding London, Sydney, Frankfurt, and most recently, China. Continuing with our data center platforms, earlier this month, we announced Alveo, a portfolio of powerful adaptable accelerated cards that will increase performance in industry standard servers, both in the cloud and in our on-prem data centers. This covers a broad range of applications, including AI inference. Alveo did ship for revenue very late in Q2, but the revenue ramp really begins in the second half of FY19. We expect revenue to grow at a good pace, but in absolute terms, will be relatively modest for FY19. We do expect Alveo to contribute meaningful revenue in FY20. Additionally, we made excellent progress in our data center partnerships and ecosystem. Working with AMD, we achieved an industry record for the highest AI inference throughput performance at 30,000 images per second, with less than 2 milliseconds latency, all packed in a 4U server form factor with epic CPUs and 8 Alveo accelerator cards. Huawei has announced that they are integrating and deploying Alveo acceleration cards in their server and product portfolio, and we are collaborating to foster an applications ecosystem in China. Inspur, a leading global data center and cloud computing solutions provider, announced that they'll be qualifying Alveo cards together with their server platforms. On the investment side, we made multiple investments last quarter in applications ecosystems, both in the data center and in other markets for a breadth of applications, including database acceleration, mobility, and next-generation technology. With respect to acquisitions, we completed our acquisition of DeePhi Tech, which is an AI technology company, and integration is going quite well. Their products and technology significantly strengthen our AI capabilities and what we can offer for use cases in the cloud, at the edge, and at endpoints. Now moving on to our progress overall as a platform company, our 28 and 60-nanometer Zync product families continue to grow very robustly. Zync sales achieved a new record, growing 70% from a year ago quarter, representing 18% of sales. Growth was across a broad set of applications in communications, automotive, primarily advanced driver-assist systems, and industrial end markets. Our Zync RFSoC products had a revenue growth of approximately 4x versus the prior quarter’s revenue, and we have well over 100 unique customers at various stages of engagement, seeing our opportunity pipeline growing in double-digit percentages. The most profound recent milestone by far was our official announcement of our 7-nanometer Versal product family, the industry's first ACAP. ACAP to refresh your memory stands for adaptive compute acceleration platform and this is a new product category that goes far beyond the capabilities of FPGAs. ACAPs are adaptable, scalable, and are heterogeneous compute platforms that are both hardware and software programmable. The Versal family will accelerate a range of applications, including AI inference and applications that have embedded AI across multiple end markets and from use cases that span the cloud to the edge and endpoints, including communications infrastructure that connects them. Versal delivers effectively the power of full custom silicon without the high cost and the long development cycles of silicon, and they can be optimized to accelerate applications on the fly and in actual running systems. We’re on track to tape out the first 7-nanometer Versal product at TSMC later this quarter, and we hope the silicon is available next year. In closing, I'm extremely proud of our team’s excellence, innovation, and consistent execution. I'm excited to see the fruits of our efforts in our Q2 and our expectations for the full fiscal year. We continue to be deeply focused on executing our strategy for sustained, robust long-term revenue growth and shareholder value. So thank you and now, I’ll turn it over to Lorenzo.

Great. Thank you, Victor and good afternoon, everybody. As Victor said, we are very pleased with our business performance this quarter and our outlook for the rest of the fiscal year. Xilinx established several financial records and we are again in the fortunate position of increasing our guidance for FY19. I will elaborate on that guidance after reviewing the quarter. Now for Q2. Revenue was at an all-time high of $746 million, growing 9% from last quarter and 19% year-over-year. We exceeded the high end of our guidance on the strength of our advanced products, which grew 25% over last quarter and 43% year-over-year. Victor pointed out the strength of Zync. It grew 70% year-on-year and highlights our progress as a platform company. From an end market perspective, we saw strengths from three of our four primary end markets. Data center and TME grew significantly, with both businesses contributing to the increase in sales. In communications, wired grew more than expected, and wireless was particularly strong. Automotive, broadcasting, and consumer were stronger than expected, as automotive and broadcast growth more than offset a slight decline in consumer. Industrial and A&D declined as expected, with a decrease mostly in A&D. Finally, channel revenue was in line with expectations at $19 million. Note, we remain below target levels of channel inventory. Gross margin was 69%, slightly below our guidance, due to the strength of wireless and lower A&D in our end market mix. GAAP operating expense was $282 million, and non-GAAP operating expense was $279 million, both in line with our guidance. GAAP operating income was $233 million, or 31.2%, and non-GAAP operating income was $236 million, or 31.6%. Our tax rate was 10% for the quarter. Our non-GAAP rate was 6%. The primary difference between these rates consists of discrete items related to tax reform. Our GAAP net income was $216 million, equating to $0.84 a share of earnings. Non-GAAP net income was $221 million, or $0.87 of earnings per share. This is a record level of earnings for Xilinx. Diluted shares for the quarter were flat at 256 million shares. There are a few key points on the balance sheet and cash flow I would like to highlight. We ended the quarter with $3.4 billion in gross cash, down slightly quarter-on-quarter and $1.7 billion in net debt. We generated operating cash flow of $313 million, due in part to reducing accounts receivable to $372 million, approximately 45 days. On capital allocation, we returned $114 million to shareholders with $91 million in dividends and $23 million in share repurchase at an average price of $66.08 a share. As Victor discussed, we completed the acquisition of DeePhi in the quarter. This acquisition, intended to accelerate our strategy, the dividend, and our buyback, reflects the application of our capital allocation strategy as discussed at our Analyst Day. On to Q3 and our updated FY19 guidance. We expect revenue to continue to grow with Q3 between $760 million and $780 million. We are forecasting growth in communications, data center, TME, and industrial and aerospace and defense. Particular strength is expected in aerospace and defense, wireless, and TME. Automotive, broadcasting, and consumer looks to come in approximately flat. Channel revenue is expected to be between $10 million and $20 million. In Q3, our gross margin is expected to be approximately 69%, and GAAP operating expense in the range of $295 million. We expect non-GAAP operating expense to be approximately $290 million. The increase quarter-to-quarter in operating expense is related to increases in employee compensation, including profit sharing and sales incentive, along with the full integration of the DeePhi acquisition. Other income will be approximately $5 million, and our tax rate is expected to be between 10% and 12%. For the full year FY19 forecast, we now expect our revenue to be between $2.950 billion and $3 billion. At the midpoint, this would be 20% year-over-year revenue growth. Gross margin for the year is expected to be between 69% and 70%. GAAP operating expense for the year is expected to be approximately $1.155 billion, and non-GAAP operating expense is expected to be approximately $1.140 million. GAAP other income for the year will be approximately $15 million of income, and non-GAAP other income will be approximately $5 million of income. Our GAAP tax rate for the year is expected to be between 10% and 12%, and the non-GAAP rate between 9% and 11%. Share count is expected to be flat to very slightly up for the rest of the year. In summary, we expect the year to show exceptional growth in revenue and profits. We are at the initial stages of realizing the benefits of our strategy and execution, and we are excited to see our opportunities developing ahead of our expectations. Let me now turn the call back to the operator for Q&A.

Thank you for taking my question. I’d like to inquire about your commerce business, specifically regarding any geographic concentration. Can you quantify your exposure to the China market and discuss growth excluding China? Should we anticipate this to be the new baseline for the communications business, and how sustainable are these trends as you look ahead to 2019? Thank you.

Well, China is an important market, but as I said in my comments, we saw growth actually with all the top customers in that segment worldwide. So I would say that we're seeing growth in a number of areas, although China is important. I think we have mentioned in our forums that, in any event, one of the areas where there's perhaps an acceleration of 5G deployments happening in Korea. So again while China is important, we're seeing growth actually in multiple geographies. We have consistently, in the past, said that we expected the really large ramp to happen in 2020. It does appear that it’s starting a bit earlier. We still believe that it will continue to build and carry over indeed, of course, past 2020, but it seems to be starting a bit earlier. So I think from that perspective, that's a new recent trend I suppose.

You talked about the growth in Zync. That was pretty impressive growth there. How do you see it? Are you expanding the TAM or these sockets that you might not have had if you had not had Zync integration and what's your visibility on continuing this very strong ramp that you've seen in Zync?

I believe Zync has a very broad appeal, which is what makes it exciting. We were pioneers in this area some time ago, and while it took longer to develop initially due to the extensive software work needed, we are now witnessing strong growth across the board. We believe we're taking market share from other solutions, such as ASSPs and embedded controllers, and possibly even FPGAs. We are expanding Zync's presence and have made significant investments in 28-nanometer technology. Although most of our revenue still comes from 28-nanometer Zync, we are seeing a strong increase in design wins for the 16-nanometer version, with most of the revenue from that still yet to come.

And I guess I’m out with several of your customers at your Developers Forum and we kind of heard about the use of Zync and specifically we heard the comment that if Xilinx had more libraries, then we’d be able to use more applications. How do you think about that from an investment standpoint? How do you balance? It seems like there is a fairly direct tie to developing more of the Xilinx generated IP and revenue, but it obviously takes a lot. So how do you balance that in this growth environment? Should we expect R&D to maybe continue to move up, as you look at those opportunities?

It's a very good point, and again, we're really trying to work that balance of continuing to invest. We can keep on this very strong growth, but also return to the investors. And so, you will see some growth and on a percentage basis, we're increasing our headcount in software and IP and things above the silicon level if you will, much more. Having said that, we're also putting a big focus that's why you hear me in the highlights and I think hopefully you took away from our Developers Forum that we're really driving the ecosystem. So it’s not just the Xilinx R&D budget, but it's the collective budget of the ecosystem, as well as partners like AMD, Samsung, and others. So it’s a matter of trying to do all those things. But it is part of our balance on how we deliver returns as well as continue to invest, so that we can get higher leverage and more leverage in the model.

I guess first question, as it relates to wireless and early 5G deployment, can you speak to whether this is just prototyping today or other and then importantly, what you've learned now on a content basis and how that translates into sustained growth into 2020 and beyond?

Yes. So the first part of it, what we're now seeing is not prototyping. I mean, I think we've been saying for some quarters that we're inversely on the early proof of concepts and prototyping. Now, what you're sort of seeing is some of that going into production and I would say and then you heard me make the comment of radio and baseband. So, we do have more content than we had traditionally. RFSoC is another element and that is actually still emerging and in terms of what we’ll see there, that is absolutely production. It is the most cost-effective, power-efficient, and size-weight form factor based solution out there. So, yeah, it's pretty strong. Sorry, second part of the question regarding.

What you’re looking from the early deployments today as to what the revenue contributions can look like as we go into deployment?

We continue to see wireless performing very strongly for the year. While we anticipate some challenges at times, overall, 5G is set to be the largest deployment the industry has experienced. We are well positioned for this. In particular, with RFSoC technology, we have a very strong presence in the radio sector. I believe this will develop further, although we should expect some fluctuations as we proceed.

Great. I guess as a follow-up here, at your Developer’s Forum, you launched Versal. I think one of the key questions was software to follow up on Joe's question, so curious how we should be thinking about key milestones there in terms of developer’s training instances, what should we be focused on?

Yeah. We're trying to share each of those elements, right. I mean, as you just said, expansion of FAS, but also non-FAS. It’s certainly not all about that. Indeed, there tends to be a little bit of focus on us in data center compute, but the Samsung SSD product, I mean, we accelerate in storage. We also have positions in smart mix and converse mix. I think there is a very strong trend in the networking side of the business. There's the ecosystem and then our board business, right, which I think we gave a little bit of the foreshadowing, all the way back to our Analyst Day and we would execute that. So I think, watch to see how many people are developing and offering that board and how that board business more meaningfully contributes to revenue in the next fiscal year, that will be another sign of things. On the investment level, if we can keep up this great growth, we're going to return some to shareholders and we're going to reinvest to keep this firewall spinning.

Victor, I was wondering if you could elaborate a little bit more, in your prepared comments, you said that you have your eye on the macro and if I look at the businesses that are probably most macro-influenced for you, it’s industrial and A&D. And yet despite kind of a macro backdrop that we’re all worried about, you’re guiding pretty strong growth for the December quarter. Curious if you could kind of square that circle for me and as you answer the question, maybe talk about what might be share gains that you're picking up in those end markets?

A&D was already down in the current quarter, but it followed a very strong quarter, and we believe it will strengthen again moving forward. We do not see any signs of weakness in the A&D area. One specific area to highlight is our basket that includes testing, measurement, emulation, and prototyping. Within that, semiconductor testing is showing some weakness as expected. However, emulation and prototyping are demonstrating great strength. Considering these mixed factors, TME is still on track for an excellent year. Additionally, in testing, we have a notable situation with 5G test equipment where we are seeing strong demand for clear reasons. The strength of our diverse offerings, along with high value and innovation, makes us less sensitive compared to other products that have more options for customers to switch between or that are not as high value or central to the system. This is one aspect at play. Furthermore, we entered this quarter with a strong backlog, and we feel good about our guidance. As we look towards FY20, we remain aware of the market environment but are seeing positive signals for our business.

That's helpful. The other follow-up, just relative to. Go ahead, Lorenzo.

No, I wasn't sure if you were referring to the overall growth we're expecting from quarter to quarter or just the industrial and A&D sectors. One area we didn't cover is the ongoing strength in communications, especially in wireless. We believe this is connected to the continued rollout of 4G and LTE, as well as the growth in 5G. So that's the final aspect of the overall growth narrative. I'm sorry, please continue.

And then I guess, Victor, as a follow-up, I was intrigued by your comments about growth in datacenter in FY20. Your confidence level seems pretty high. Seems like you’ve got a good line of sight. I know,there is probably a lot you can't talk about, but I wanted to give you the opportunity to elaborate on that comment and what gives you that confidence, is this new customers, is this silicon to board? Maybe if you can help us just kind of parse that out.

It's a combination of factors, as I mentioned. We did generate some revenue at the end of Q2 from Alveo boards, and we currently have two models, the U200 and U250, with plans to launch additional boards in the coming quarters. While the growth rate isn't very high, it's starting from a small base, so it's not significant yet; however, FY20 is expected to be more impactful. Previously, we didn't have any revenue from boards, which is certainly a factor. Additionally, there are many projects I can't discuss in detail right now, but I can mention Samsung’s smart SSD launch and that we have secured a design win for a smart memory DIM with a major memory supplier, though I can't disclose their name. Many of our customers prefer to keep their plans confidential for competitive reasons, but we believe that in the next fiscal year, we will start to see more substantial revenue. We have been experiencing good percentage growth from a smaller base in the datacenter segment, and particularly in the latter half of next fiscal year, we anticipate that everything will start to come together.

I just wanted to make sure I understood the confidence for the full year. Victor, is it and Lorenzo, is it dependent on comps, 5G, 4G as well as LVO beginning to kick in and what about the other segments, so I'm talking beyond the quarter that you're guiding for?

Let me give a briefing and I’ll let Lorenzo add colors to this as well. I mean, I think, one of the reasons and I mentioned also while we're very dialed in to watching the macro situation, we still feel that the appropriate guide is because of the breadth of the strength that we saw in Q2, which is carrying over into this quarter and also from what we can see for the fourth. So, and clearly as the first half came in stronger than our initial expectations for the year, so I would say it's the breadth, yes, so that means generalizing your statement, you just pointed out communication and then data center boards, but it goes beyond that, because we're seeing strength in all the other markets, even though within some of those markets, there are weaknesses. As I said, TME, there are weaknesses in some of the segments within that bucket, but the overall larger end market that we report out looks really good for the year.

I don't have much to add. It's quite evident, considering the relative size of the end markets and the dynamics we discussed regarding the growth in communications. This area continues to be a significant revenue contributor, but our portfolio of end markets is much broader than that.

In communications, you mentioned it increased 35% year-over-year, but it must be significantly below the peak levels we observed a few years back. Could you clarify how far we are from that peak? With the addition of more content, should we anticipate that this number will exceed what was achieved in the last build as we move into the broader 5G rollout over the next six to eight quarters? Thank you.

Give me just one second to make sure I'm not misaligning things. It is still, from both of the end markets in communications, it is still meaningfully below the highs we've seen, for both wired and wireless in the past, three and five years ago. But I think also the other thing to think about is, given the growth of our other end markets, it's a smaller percentage of our overall business. Our overall exposure to that is smaller, but we do expect that we have headroom, as we’ve discussed in both wired and wireless to get past the historic levels at some point in the future.

Victor, obviously, there is quite a bit of concern around end markets, like industrials, automotive, semi test, as you sort of addressed them. Can you give us an idea of what’s sort of embedded in your back half guide for some of these end markets where we do have concerns?

Well, I think I talked a bit about the TME, right. Yes, there are some weakness in some areas, but then some areas are up. So again, from what we can see, we think that for the year, TME will be up quite nicely overall and certainly in the first half, everything is strong, right, including semiconductor test. On automotive, we're seeing growth, and we do still expect it to continue, but just to put that in context, I think we sort of talked about how automotive is sort of around 7% of our business. So, I mean, even if there's some fluctuation on that, just to put that in context, right, so I would say again, yes, we watch it very carefully in some segments that will feel it, but just not to be repetitive, but since we have such broad strength in some of these areas we know the dynamics are such that they won't be affected like, since the operators decided they want to pull in 5G, they are not going to back off. Now, they may modulate the ramp, but since we have always said that we really were planning more of a ramp in the later timeframe to some extent, just the fact that accelerated has been upside, right, and then RFSoC, which we've never had before, that continues. We would have liked three more orders just this last quarter, because this is still the product like that. And so, we will have more value in the radio side and there is more radio that’s going to be deployed. So, yes, there are some areas that are weaker, but again because of the broad strength and some of the leadership products that we have, I don't think we're going to be as affected, but of course we’re going to watch this very carefully.

Yeah. We're in a good position obviously in areas like wireless and even in auto, where the platforms that we've been designed into with our products, our high value products are tending to be in the ramp phase. So while macro factors may impact it, they're still in some ways gaining share versus alternatives in multiple end markets, but it puts us in a pretty good position.

Great. And then as a quick follow up, Victor, obviously, you've been on the road quite a bit hosting a lot of sessions and I'm sure you've had a conversation with customers, but for Alveo specifically, what's been the feedback so far? You talked about the ramp into December and more so into calendar ’19, how should we think about the magnitude of that ramp again for Alveo specifically?

Yes. For FY19, again, it’s starting from effectively zero, right, just at the very end of last quarter. So, it will grow on a percentage basis, but we're not saying that to be very material. But through the course of FY20, as I said in my comments, it will be meaningful revenue. I think the interest is high. I do think that how much more upside versus it maybe being a little bit moderate, there is certainly a degree of around some of the things that the early questions around application development, in fact also just education and so forth and that's why we did the Developer’s Forum and hopefully you saw some of the momentum that we are getting in strong interest there, right. So it's early days, but we do feel like the signals are, there's a lot of interest and revenue being meaningful in the FY20 timeframe, not FY19.

You’re now starting to go after semiconductor content that traditionally has been discrete chips, adjacent to FPGA such as DACs and ADCs universal RF series. What are the other content adjacency opportunities that you are seeing and can you talk about the candidate, if any, in addition to that could be play as part of the FPGA integration going forward?

One point to highlight about Versal is that we want to ensure people understand it is not just an FPGA. This is due to the diverse range of compute engines and infrastructure we have developed. It includes multi-core SOCs, a network on a chip, next-generation programmable adaptive hardware, fabric, and distributed memory. Some products within the AI core series will feature a new architecture referred to as the AI engine. The level of integration is remarkable; it retains multi-MAC capabilities and high-speed components and can interface with integrated HVM on a silicon interposer. Overall, it's a highly complete and powerful platform. This aligns with our strategy to grow our Serviceable Available Market (SAM) as we face competition that extends well beyond traditional FPGAs. We are committed to developing a comprehensive software stack and fostering an ecosystem around it. Therefore, we are incorporating a significant amount of capability into our Versal offerings.

And then just going back on the, while the opportunity, even if we assume that the ASPs are half the less price that you’ve mentioned, trying to reconcile this with your targeted incremental revenue coming from data center suggest that you’re probably making very conservative market share or adoption rate assumptions, any color you could give us in terms of assumption of market share or any type of adoption rate that you see for that new revenue stream to get to your guidance.

What I would say is that we are trying to be measured in our approach since this is new for us. However, we are seeing strong interest, even though it's still early days. We are actively developing our go-to-market channels. I would again emphasize that we plan to be competitively priced while delivering significant value, and we expect to capture market share. The acceleration segment is relatively new, but we believe this represents a new product revenue stream that will help expedite the process for people to market their applications. Our pricing reflects the value we're offering, and we anticipate capturing that value in this emerging area.

I just wanted to follow-up on kind of the data center question that was asked earlier. It seems like you've got very good proof points right now with several of your customers like Huawei, WS, and Baba adapting FAS, but it seems like there's a much more meaningful opportunity, if you can convince one of your customers to move forward with an internal acceleration architecture, similar to what we've seen at Catapult. What do you think it will take for one of your customers or potential customers to move forward with that use case and is that something that we can think about happening potentially next year, that's potentially baked into your expectations for more meaningful growth in data center?

Well, look, I mean I agree to that. The world is not just that fast and indeed, like I said earlier, for us, it’s unlike many other suppliers, it's not even just about data center compute, right, like in storage; just heard, we're working on memory and we're in smart and conversion mix. But now back to your point, I agree and I'm sure you heard in my opening comments that the Twitch group, right, and they did acceleration on video, right. Just goes to show again, we do do that and we are working with others. Again, it's just always a bit challenging in terms of when things can be shared, but yes, we have many engagements for internal acceleration, some which involve machine learning, some which do not. I think Alveo is going to also help that because again, it lowers the barrier significantly for people to develop applications and also to bring that on-prem. So, it’s not everything that’s around in public cloud, right. So I agree with you and I hope that I can share more of that going forward.

And then my follow-up is on 5G. It seems like your customers are approaching 5G with a combination of using RFSoC and MPSoC and some other FPGAs. Can you just talk about your expectations for 5G? Do you expect the majority of your 5G customers to move to RFSoC and because of the integration and performance in the BOM cost savings, does that give you more staying power late in the cycle as your customers consider kind of ASIC reversion as an opportunity?

Absolutely. I think I couldn’t have made the same statement better than you just articulated. I mean, look, the RFSoC really is like, it’s not hype. I mean, there's nobody that has a product like that and by the way, it's not just 5G. I mean, it's any kind of massive MIMO or antenna array application. We see it in cable; we see it in other kinds of radar applications. And people are changing their architecture, the radio architecture based on this. And so, in that sense, it’s certainly pretty sticky. I don't really see what in that particular instance, people being able to disrupt that and it's just getting started, right. We just recently went into production and people are deploying. So I think RFSoC will certainly be very, very strong. And, but you're right, it was not just RFSoC. We’re seeing usage of other MPSoCs as well as just our FPGAs, at the 16-nanometer node. Some of that over the course of time could go to that, but right now, things are so dynamic and moving so quickly. That isn't happening and I think that's going to actually 5G is so ambitious and there's so many things happening that that will probably happen for a bit longer, but I'm not necessarily suggesting that our position in baseband will be as durable as our position in radio, for instance, right. I mean, I think, radio is and we already have a roadmap, right. So it's not just the first generation RFSoC. We already have the product, the next generation as well as in 7-nanometer.

Hi. This is Jay for Ross Seymore. Thanks for letting me ask a question. You touched on this previously, but how do you view Xilinx’s benefit from the 5G transition as compared to the 4G transition, perhaps from a magnitude and duration perspective?

Yeah. I mean to expand on it, I think 5G is going to clearly overtime significantly past what 4G was as the industry and that's for us, it's a bigger opportunity as well and it's both because of the 5G from a technology perspective is so much more disruptive. It's much broader and it really is not just a communication standard, but it's really being used as a term for a basket of technology, including things like massive MIMO, which isn’t inherently necessarily 5G. There is IoT being associated, also automotive. So, it's very broad, very ambitious. And then it's not just the segment is bigger, it's also like we're innovative; we added value that we're adding more value like again, in RFSoC for instance. For that matter, MPSoC, but back in the 4G, we didn't have integrated multi-core RFSoCs and we didn't have really high-performance ADCs and DACs monolithically integrated. So yes, I think, the general market will be bigger and broader and I think our opportunity over time will be significantly larger.

As a follow up on the OpEx side, how far along are Xilinx’s software investments in support of the data center efforts? Do you expect those investments to slow or do they need to persist on a structural or customer-by-customer engagement basis?

Well, I think, in broad strokes, it's certainly not going to slow. I mean we need to, in fact, within what’s reasonable and affordable, so that we could return values to shareholders and still continue to invest. As I said, we’ll probably invest from a headcount perspective, fortunately more in software and things about silicon than we would at the silicon level. Having said that, we are investing in the ecosystem and also doing partnerships. So we're really trying to expand the footprint. So it's not just us. Now, the other part that you alluded to is also true, right, some of the really, really big customers, there's a lot of customization. There's things that we work very closely with to make sure that we deliver to their needs and we are investing and doing that as well for obviously the select top customers. So yeah for both of those reasons, just the broad horizontal investment as well as some key major customers, we are and we will be investing more in software and IT and related things.

Just curious into the September quarter, you mentioned ZTE was back, just curious how much that was and maybe versus what you were planning when you guided? And then just longer term, just curious, you talked about earlier 5G. If you can just relate it to the ramp of 4G. That’s all two sharp quarters in the beginning of ’14, and then kind of fell. It sounds like 5G may be a little longer lived, just kind of curious, as you look out over the next few quarters or even years, the trajectory of 5G versus 4G.

The first part regarding ZTE is important to note. As we mentioned when the initial de novo order was issued, we do not have any customers at or near the 10% mark. That being said, ZTE is a significant customer and has contributed to our growth, but it’s worth emphasizing that our performance is not solely dependent on ZTE. We are also experiencing strong early deployments and activity from other key wireless customers. Since this situation arose at the beginning of fiscal year ’19, I want to clarify that ZTE's contribution is not exceeding our original projections. They did have one last quarter to fulfill requirements, but their impact is in line with our original estimates. We did remove ZTE from our plans at the start of the fiscal year, but once we confirmed their reinstatement, we made the necessary adjustments. Regarding the ramp-up of 5G, I believe it will be more extensive and last longer than previous generations due to its ambitious nature and the multitude of innovations, especially in the business model surrounding it. While we are maintaining our 2020 outlook, there are indications that companies are eager to initiate deployments next year. I am hesitant to predict the exact trajectory, but I anticipate significant opportunities as data demands escalate and the need for low latency and high bandwidth grows. The infrastructure build-out will be considerable, not only in terms of wires but in enhancing the entire network.

Great. Well, thanks for joining us today, everyone, and we’ll have a playback of this call, beginning at 5 PM Pacific, 8 PM Eastern 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 2019 will be on Wednesday, January 23 after the market closed. As far as conference participation this quarter, we will be attending the Credit Suisse Technology Conference on November 27. This completes our call and thank you very much for your participation.