Intel Corporation (INTC) Management Presents at Bank of America Merrill Lynch 2018 Global Technology Conference (Transcript)

| About: Intel Corporation (INTC)

Intel Corporation (NASDAQ:INTC) Bank of America Merrill Lynch 2018 Global Technology Conference Call June 6, 2018 11:45 AM ET

Executives

Dan McNamara - General Manager, Programmable Solutions Group

Analysts

Vivek Arya - Bank of America, Merrill Lynch

Vivek Arya

Good morning, everyone. Thanks for joining us. I'm Vivek Arya, Senior Semiconductor Analyst at BofA, Merrill Lynch. I’m delighted to have Dan McNamara, General Manager of Intel's Programmable Systems Division, what the artist - formerly known as Altera, join us this morning. I'll start with a few questions from my side. But please feel free to raise your hand if you have a question. We’ll leave some room at the end for talk for more Q&A. Dan, welcome.

Dan McNamara

Thanks.

Vivek Arya

Thanks for being here. Maybe as a start, if you could take us a step back, which is Altera has now been with Intel for about two, two and a half years or so. How has been the journey like so far, how has been the integration? Give us a report card of what has happened so far.

Dan McNamara

Sure. So we are coming up on two and half years, and it has been quite a journey. So I would say that it's going very well and there's a number of reasons for that. So let me go through them. So first and foremost, Intel set us up as an independent business group. And that was really for two main reasons. One is to maintain our focus on our customers, because in the FPGA business, your customer relationship and the customer support model is very important. And that was a little bit of a concern from our customer base when the announcement was made back in 2015 that Intel would acquire us, customers were concerned about that support model. So we maintained that.

The other area was really maintaining our focus in our core competencies, which is FPGAs and software. So that has gone well. But we’ve also been able to leverage the breadth of Intel across platform engineering group, across TMG manufacturing. But I think the big wins have been across the sales team. So very broad sales team. There’s really broad customer relationships that we really just didn’t have, so that’s gone extremely well. And then lastly on that front is really the independent other business groups where we’re partnering with them on different solutions.

And then if you look at second [sweet area] [ph] is really about transformation, so we’re integrating and transforming. If you recall, in 2015, we stumbled a little bit on the 20-nanometer node. And one of my most important focus areas was getting our engineering team back on a repeatable cycle of innovation. So we made a bunch of changes in terms of work structure, brought in some new leaders and that we really have that team back really getting in their groove and delivering products on a regular cadence. So that’s gone well.

And lastly, we came in with a very strict three strategic imperatives that we’re very focused on. First is the best silicon and [service] [ph] technology and that hasn’t changed, that’s really important for our business. Secondly, it’s about growing the user base for FPGAs. And what I mean by that is really building enabling software flows with an ecosystem of partners and IP to enable software developers to leverage the benefits of FPGAs. For a software developer, FPGAs can be a little bit scary at times, right, because you’re actually developing a chip. So we’re trying to abstract the hardware away and really develop more libraries and IP to really enable them to get to market and realize benefits more quickly.

And last and I would argue is probably the most important is what I call unmatched value to the customer and the shareholder. When we work together with the cloud business group and the network business group and the ITG business group, we’re putting together complete solutions, whether it’d be an SoC plus FPGA, or a Xeon plus FPGA, or even an ASIC, we come to the customer with a complete solution with an overarching software stack and IP libraries. So that, I believe is -- I always talk about that being a multiplier in value, so that’s gone well.

And if you just look at it, at this point two and half years in, we’re really -- I believe the flywheel is really moving well. Last quarter, we grew 17%, I believe. We also grew our advanced products 43%. And we also hit, which is actually the biggest early in -- the biggest and best early indicator is, we had our largest design win quarter ever in the history of Altera in Q1, and I think Brian mentioned that during our earnings call. So I would say that the first two and half years have gone extremely well and we’re really starting to move forward and accelerate the transformation and integration.

Vivek Arya

If we listen to your competitor then it’s like -- they thought they would take a lot of share, right, because the integration would go through stumbles. But what we have seen is you guys have actually outgrown them in the last -- at least the last year, perhaps the last two years. Even recently, you mentioned advanced products grew 43%. I think, they mentioned only 28%. But the place where lot of investors are sort of concerned is about some next generation nodes. So give us a sense of where you are with Stratix 10 and the 14-nanometer, and then we’ll talk about 10-nanometer later.

Dan McNamara

So 14-nanometer, we are in production. But we’re also, just like we always do with FPGA families, you get out with a couple of early devices and you proliferate out derivatives and you fill out your family. So that’s going extremely well. We’re continuing to deliver the new products this year on Stratix 10 with different tiles. We’re going after that business in a completely different way than our competitor is.

We’re building out, what I would call, a chiplet strategy, which has a base FPGA with tiles that are integrated via, what we call the embedded multi-chip interconnect bridge, which is an Intel substrate technology, which think about is a silicon bridge in the substrate. So that gives us very low cost, low power opportunity to integrate heterogeneous components. And we’re doing that and you’re going to see more of that over the next year in terms of different tiles and different flavors of tiles coming out. So we're very focused there.

We are also focused on 10-nanometer. And we’re on schedule to deliver 10-nanometer next year. We believe we’re going to come out with a very compelling performance per watt and solution set, again, built off of tile architectures and deliver into our main focus areas which is the cloud data center, the infrastructure and there is -- the infrastructure is a big or longer conversation and then the edge, and then our traditional embedded businesses.

So right now, while we were a little bit late on Stratix 10 and 14, again, the lateness for 14 was really about our stumbled pre-acquisition. If anything, since we’ve got to Intel, we’ve accelerated that schedule.

Vivek Arya

Got it. Give us a sense from a customer perspective I know on Wall Street, we fixate on who is ahead who is behind in these node wars. But how do the customers look at this. So for example, from the outside the perception is Xilinx was -- TSMC they are ahead on 7-nanometer. Intel has said that they will be somewhat delayed on 10, which is equivalent to what they are doing. Does it actually put you at a competitive disadvantage, or is this just perception?

Dan McNamara

I think it's mostly perception. If you look at where we are on 10-nanometer, we as the FPGA group were later in the cycle for 10 at Intel. And I know Brian has talked about this a lot and Marty has talked about a lot that we’re working on yields for Cannon Lake and we’re wrapping, and we’ll ramp the full production next year. For us, our tape out was late in the cycle for 10 anyway. So we believe that in ’19 when we really start shipping products, we’re going to be fine.

When you think about customers, customers want performance per watt and solutions. And what I’m focused on and what we are focused on at Intel is bringing that maximum value. So whether it's 14-nanometer or 10-nanometer device, it could be sitting next to a Xeon or some new SoC technology and that solution, if it meets the need, that’s what's important to the customer. Now, are they waiting and excited about 10-nanometer? Sure, we have many engagements that we have customers that are waiting for our first devices.

So I think from -- the big difference, if you look at the two companies, it's really about the architecture, I talked about, and it's really about the solutions that we bring to bear from Intel standpoint. I think it's really an unmatched situation in terms of what we bring to bear across IP, silicon and software and then obviously, bringing that to different vertical segments also.

Vivek Arya

Does it mean that historically what we saw was, for example, they might have done a better job at 60, you did a better job at 40. So this thing has pendulum has sort of swung back and forth between Xilinx and you guys over a number of years. Is it -- do those nodes matter as much as they used to in the past, those node transition and who is going to be first. Or is it more a case of, look, we are exposed to these end markets. We are better exposed to these end markets that are going faster. So just being plus minus three months here or there, it doesn’t make as much of a difference as it used to in the past.

Dan McNamara

Yes, I would agree with you. That's exactly what I was going to say. If you’re in a window of a quarter or so, it really doesn’t matter. Okay, there is a design or two that you’re trying to hit and that may go one way or the other. But it really is from what I'm seeing out there, especially in the newer and the fastest growing markets, like the cloud and now the edge and even now the infrastructure, it's really about the solutions that you bring. Is it hitting the performance per watt? Is it hitting the compute need, is it scalable and things like that? So that’s what we’re really focused on. And I truly believe, if you go back in time, 40-nanometer we had a very, very large lead by probably a year. We stumbled at 20, which made us late for 14 and our competitor was out with a very broad offering at 60.

Going to the next node, I see us very close in timing. And as I mentioned, there's some different strategies that are being employed between the two companies. But the reality is I think the timing is going to be less important as much, and it is going to be more focused on the solution that we’re brining.

Vivek Arya

Now, going over to growth, before Altera became a part of Intel, I think the FPGA industry went through a period of very soft growth, partly because I think there was the hangover from the 3G build out back in 2010 when I remembered, Altera, grew over 60% I think that year but then there was that hangover for a number of years. But when Intel acquired Altera, I think the growth rates was 6% or 7%, you guys have actually outgrown that. Has this just been a datacenter story or has the growth been broader than that?

Dan McNamara

So I would say that the majority is datacenter, but we are seeing -- so datacenter, I think Brian mentioned this in the Q1 earnings. We grew 150% in Q1 year-on-year. Auto, we grew over 50% in Q1. We're also seeing a broad base of military embedded growth. So it's -- I would argue, it's datacenter and what we would traditionally call embedded businesses, comms are still a little sluggish in terms of revenue. But the design activity is very, very strong. So if you look at the recent -- last couple of years and even this year, it's really about datacenter, auto, and some of the more broader embedded markets.

Vivek Arya

Why is there a contrast between your comms growth versus the comms growth in DCG. How are they able to grow faster?

Dan McNamara

Okay, so that's an interesting question and I wanted to talk about that, because I believe that most people are missing what Intel is doing in the infrastructure. So if you think about 5G, and that's the way it's today and if you think about what that brings to the market, obviously, very low latency, high bandwidth, interactive services in a very impatient society.

The catalyst behind 5G is really the transformation of the network from the edge all the way through to the cloud, and that's we’re moving or replacing traditional customized router, switches, load balancers and things like that, that have been distributed for many, many years into more of a [cloud] [ph] server type architecture. So we're seeing that type of growth already in the infrastructure.

And when you look at FPGAs, FPGAs have a play in that, because FPGAs can provide really good accelerator story or application space for this, the virtualization. What I mean by that is when you have a carrier, so we're doing some trial with China Telecom. And they are looking at traffic shaping, quality of service, deeper level packet processing. And they would like to do that with an FPGA and then leverage the rest of the workloads for the Xeon. And the reality is it’s still software defined because an FPGA is software defined. So I would say that what we’re seeing is growth from the transformation of the wired network, if you will and then 5G starting to happen also. But I think whatever it needs to be cognizant of this there’s a transformation and that’s the real catalyst for 5G. And I think our comms folks are seeing today.

Vivek Arya

I think 5G is actually -- it can be a very interesting and unique story for Intel, because you are able to attack it from so many different sides, whether it’s within the handset that modems write, whether it’s in the base station with you guys or back in the datacenter, the number of -- and IoT, autonomous. I think almost every business of Intel comes together in 5G like nothing else. But if you look at your business focus on comms, is it as intense as it used to be in the past, because it seems like most of the discussion has now shifted to the data center.

Dan McNamara

Yes, I would say that comms -- I have this question a lot, by the way. But so for us, comms, 5G is a really big opportunity for FPGAs. There is much more complex processing in the base band of the actual base station. So FPGAs are playing a very important role there. And by the way, to your point we’re grooming up next to Intel SoC technology and Xeon. So again that complete solution helps us there.

The radio, there is going to be more radios. And FPGAs play very well in radios and traditionally have, so we see a pretty big opportunity there. And then as you go back in through the edge and what I talked about the virtualization of the network, we see continued growth there, it’s different though. We used to be, we used to be the guts of a router in the data plain, driving very high speed serial lanes. I am not saying we’re not doing that anymore. But it’s going to be more of a solution set for the virtualization world, where we’re building specific solutions for different carriers to do, whether it’s a virtual broadband network gateway or virtual edge packet core, there is real solutions here, instead of just delivering a device and having the 10s -- the company deliver the actual IP. We’re doing a lot more of it.

So we see a big opportunity here. We don’t see it coming back really to a probably ’19 and beyond. This year, we’ll probably see a little bit of growth in communications, but overall, it’s really a big design win phase for us in comms.

Vivek Arya

And if you look at the content opportunity as you move into 5G, we saw FPGAs benefit a lot in the 2G to 3G cycle, perhaps not as much from 3G to 4G, I guess probably because 3G CapEx was coming down at the same time 4G was growing. So net-net there wasn’t much capital. But there was also a lot of move from FPGA to certain ASICs and other merchant silicon. What give you the confidence that 5G would be different than thus far? Do you think 5G would look more like the 2G to 3G transition, or the 3G to 4G transition?

Dan McNamara

So if you look at where ASICs come into play, typically in the history with 5G, it’s really the radio. And they spend -- I not saying that people don’t convert FPGAs, ASICs in the baseband. But overall, we would see convergence in 3G and 4G from the radio standpoint. And what we see though in 5G is through so much flexibility needed, and we see the com OEMs, really looking at how do they harden certain things, but they also still need some FPGA capability. And what you'll see ultimately longer term is it’ll still go ASIC on some of the high runners. But I believe that there's just going to be more breadth of solution in the radio space that will allow us to maintain FPGAs in there for the long haul.

Vivek Arya

So AI is the buzzword of the day, and with Intel, there are different ways of going after that market. Something your competitor has done has put a lot of focus into getting their own developer ecosystem started et cetera. I know Intel has many ways of going after the AI market. But if you just look at your business, then what has been the progress so far, what are the milestones. Is it as important to you guys as it is to the rest of the company?

Dan McNamara

So let me address -- there is couple of -- maybe I'll talk AI specifically and then maybe talk about the developer community, because it’s important so AI, clearly, the fastest growing workload. What's interesting is that what you're seeing is so many different applications. Whether it's image, text, speech -- there's an underlying architecture and topology that our customers are looking to solve. And the beauty is that with Intel is that we're going after this with a broad portfolio and it goes from Xeon FPGA to Movidius to the Crest family to potentially GPU. And we're doing that under a one overarching software stack. And we believe that clearly one size will not fit all. We have competitors that are going after with point solutions and maybe a software flow.

And then if you translate that back to FPGAs, FPGA is a really good interface. You probably heard of Microsoft Brainwave, leveraging FPGA's to be even competitive with anything out in the market today in terms of interference. Very high throughput very low latency, and that's what FPGAs are really good at, because we have a very, a parallel architecture with embedded memory and embedded DSP, and that really enables customers to make decisions, which inferences on a frame by frame or packet by packet basis.

We are building out a complete flow just like many, but it's going to be all encapsulated under N graph or OpenVINO, which we've announced recently. And we believe that customers, some very savvy chip level customers, will do the usual way and develop it completely on their own and with our help, but there's other customers that want that overarching software, so to get to 95% there and then we can tweak to get that performance.

So we're seeing very good results so far. If you think about it, Microsoft's a very public. We've announced [indiscernible], which is a Chinese manufacturer of surveillance cameras. We've announced a win there. We've announced NEC in Japan with facial recognition. We've announced JD.com, which is Chinese retailer. So those are the ones we can announce. But we have many, many engagements even in the automotive space where people are looking to leverage this really core inference capability with very, very low latency and low performance per watt. So we feel very, very good, as one piece of the puzzle for Intel.

I truly believe that Intel, overall, as we go forward here in the next year or so, will become a very well known name in the AI space if we aren't already. But I think that people were -- it'll be a de facto going forward. As you look at all of the competing solutions and the applications, if you just take one level down from an application, whether it's speech or image, the choices and the different needs of the customer, you have different network topologies. You have different precision and accuracy needs for instance. You have different latency and throughput. And then all of this has to be typically done under some performance or cost model. You can’t do that with one device or one family of devices. You have to go after this as a complete solution solving the customers’ problems. And that’s what we’re doing today is we going to our customer with representatives from all groups that have a solution. And we sit there and we listen and then we come back together and we go back to the customer and say, here’s what we think.

So it’s really that I believe, that’s why I am so excited to be part of Intel is we are not just an FPGA company. We can integrate Xeon in package. We can integrate SoC technology from the communications group in package. We have this heterogeneous capability built-in to the company that we can very, very quickly iterate and deliver the best solutions. So from that part, we feel really good.

On the developer front, that’s an interesting conversation too. So we’ve seen in lot of conversations about developers, by a number of customers out there, or competitors out there. And I would say that our approach here is edge to cloud. So let me start with the cloud. If you look at what Microsoft has done, they’ve talked about Brainwave. And Brainwave has been used for being searched for some time now. But they’re actually bringing it out to developers today, both from the cloud standpoint and bringing it to the edge, which will drive more developers. We have already Alibaba has FPGA as a service. We have OVH, which is a cloud service provider in Europe with FPGA as a service. We’re working with a number of other folks to drive this ecosystem from the cloud standpoint with FPGA as a service.

And if you look at the enterprise, we’re partnering with folks like Dell and Fujitsu and Inspur in China to develop what we call our pack and stack strategy, which is a programmable accelerator card with the Xeon acceleration stack, with over 30 companies today delivering solutions on top of it, and that number will grow to probably double that by the end of the year. And we’re going to market for on-prem analytics and things like that, whether it’s AI, or financial risk analysis, or database acceleration. We’re building out that suite.

And then if you go to the edge, which is a very fast growing area, we just announced OpenVINO, which is an open visual interference and neural network optimization kit. And it’s really just a complete open software kit that enables developers to quickly create a solution, whether it’s on FPGA or a core, or Movidius, they can download this quickly, get all the libraries and really ramp quickly to really go get that visual interference application, which is very, very important. So we just see it as not just boy, we’re partnered with one cloud service provider, but it’s the whole path from edge to the cloud from developers.

Vivek Arya

So think that’s the important question, because again the perception is that, look, the large GPU player has -- I think, they’ve announced 850,000 developers. I think Xilinx had an Analyst Day recently. They said they will get to 400 by the end of the year or something. Is it -- so the difficultly in programming FPGA as an obstacle for AI adoption. Do you think we are passed that point or is there still a lot of software work and software tools that you guys need to produce from your side to make it easier for the developer communities to use FPGAs. So it just becomes -- out of -- not even really care whether they’re using an FPGA or a GPU?

Dan McNamara

I would say, we’re getting there, right. We’re beyond the 50% mark, but we’re not there yet. And here’s the challenge, we can build out these software and abstract their hardware, and this is just an FPGA conversation at this point. When you really want to eke out that last 10% where an FPGA really gets a lot more competitive than say GPU, there is some hand tweaking. And what we’re trying to do is build out the software so people can quickly get there and then look at the performance and then if needed, we can take them the rest of the way, because at the end of the day when you abstract the hardware away, you do take away some of that granularity and that final green architecture. So it's a little of both. What we’re trying to do is enable the masses and then in specific areas really dig down with our customers and our quick support model to really optimize to that finer green level.

Vivek Arya

Maybe a similar conversation on the automotive side. Again, Intel has a number of tools, right, and the tools to approach back, whether it’s Atom, or whether it’s -- what you got from Mobileye and obviously, the FPGA solution. How do you decide which is the right solution to pitch to the customer?

Dan McNamara

So in auto, it’s very interesting for us. Again, like we said, we have Xeon. We have Mobileye. We have FPGAs. We have Crest devices coming. What we’re looking at is what is the architecture of the future. And when you do that, the answers come out very cleanly, right. FPGAs become sensor fusion and a sensor aggregator to take in all the numbers of cameras and all the other senses that are going to be coming out of the car. And take that preprocess and then send it into, whether it’s a Xeon or what have you.

We look at it very similar to what we -- the way we look at AI, it’s architecturally what is the customer trying to solve at the top level, and then we go in and put the solutions. But sometime we just get a call and say look, we’re interested in an FPGA to do e-vehicle inverter control things like that where we're seeing a lot of growth for the electric vehicle. So it’s not always that way, because you still have a large number of ECUs that people are trying to solve problems with. And we go with at that as needed.

But I think at the end of the day, we’re trying to raise the conversation up to the top level architecture and then we propose, from an Intel standpoint, what we believe is the best solution we have. So it’s very similar to the AI conversation.

Vivek Arya

I think, historically, your competitor had always had a stronger footprint. I think you guys always did better in the data sector. They somehow did better on the automotive side. Why was that? Was auto not a focus as much, or was that back in the bay? Is it become a bigger focus now? Do you think you’re going to catch up in terms of share?

Dan McNamara

So there is no question that our competitor was more focused on auto earlier than us. And we were keeping it very focused on low cost low power families and we’re doing okay with that. But our competitor really aggressively went at 28 with some products there, and started winning some on the cameras that are out there [multiple speakers] exactly. And we win some camera business too. I think one of the public announcements we made is with DENSO in Japan, which is a stereo camera.

So I think it was just more time and we saw the compete -- we used to call it completely storage based. We saw that and we thought it was a very big opportunity and we’ve been talking about that since 2014. They saw auto earlier and they have been investing for a while. We’re seeing very good growth in auto today. But again, if you think about the total solution from Intel, we believe we’re extremely well positioned from an auto standpoint overall with Intel and even within FPGAs. I think our competitor has a single solution that they are going to go target against a broad array, and a portfolio that I think is unmatched in the industry.

Vivek Arya

And then back from your experience at Altera a lot in embedded products. Is it fair to think of the FPGA industry? Like if we take out the half of the business that are datacenter and comms and so forth, the other half is sort of analog in micro controller like, that it has the similar advantages of selling to a diversified market, it goes through channels. So do you think that's a fair competitive or there are differences, and how we should look at that part of the business?

Dan McNamara

I think in the embedded side of the house, you should - it has the same characteristics around very long life expectations. And you design in into your box for medical whether it's ultrasound or CT scanner, you have to support that for a long life. It's traditionally similar to what we see in the other segments it's typically around digital signal processing, which we’re good at from an FPGA standpoint, and it's glue logic. So tying a lot of the interface between the analog, or if you think about being forming for ultrasound and things like that in medical, military is very similar. Military it's all about taking the radar inputs from the RF and processing it into the main processor. But that’s very similar to comp. So there’s some similarities to the traditional micro analog. But what we do well is really process large amounts of data very, very efficiently, and we do that across the board in the embedded market, even in the industrial space.

We had a lot of growth in motor drives and programmable logic controllers, and motor drives, it was really just doing algorithm acceleration off of DSP or a microcontroller. So this acceleration piece, which we talk about today, is we want to be the accelerator of choice from the edge to the datacenter to the cloud. That’s being going on for 20 years, that accelerator piece. So I would say that there is some similarity, but it's more tied to large amounts of the data being processed simultaneously.

Vivek Arya

So for example, we had Texas Instruments and analog device there yesterday. And very excited about their industrial growth right, which for -- I think, ADR has been over 20% over the last number of quarters. Do you see similar rates of growth, so again back for the prior question, should we expect if the analog micro industry and grows whatever say 6%, 7%, 8% a year. Is that the right way to look at your opportunity outside datacenter and comms?

Dan McNamara

Yes, I would say there is some puts in there. If you have military and then you have industrial and then you have what we call broadcast, if you think of the video switchers and things like that. So there is probably some higher growth and there is some lower growth. But in general, I would argue yes high-single digits is probably the way to think about our embedded business over the next four to five years.

Vivek Arya

And one last one there is when Intel had acquired Altera, I think some of the flood again that was there well a lot of these embedded markets are very small, people will not pay attention to them to the extent that datacenter, and comms will get the attention and focus. But the one counter to that was Intel had a very strong IoT business, which I imagine, right goes through a similar set of very diverse customers. So talk to us about -- is there collaboration between the two groups, like there is pull through between the two groups.

Dan McNamara

Yes, so our partnership with the IoTG group is actually going extremely well. And if you think about the areas where they’re focused smart retail, smart surveillance industry 4.0 really for compute, all of that we’re putting together complete solutions. If you look at the OpenVINO work, that came out of IoTG, and they worked with us in a number of other product groups. So that has been a really good story for us. That is where I came from in terms of that side. So I really see very big opportunity to really partner with the IoTG group to build these complete solutions. And it’s going extremely well, whether it's in industrial and military also, the broadcast sector and then the really fast growing, what I would call vision at the edge, that's a big focus for us. And that's what that what OpenVINO was created for. So that relationship and partnership and more importantly the solutions we’re bringing to market for our customers is really exciting, and the customer uptake is really strong.

Vivek Arya

Got it. Maybe one question on the financials. And as a standalone company, Altera would routinely get very strong low 30s operating margins. Now at least to the segment level that is disclosed, it’s lower than that. Why they have been that even though you guys have grown so well? Why have the margins come down? Or it is just accounting?

Dan McNamara

No, it’s not accounting. What we -- I mentioned earlier, we stumbled at 20. And we really needed to invest, over the last couple years, to get back to the pace that the market and our competitors were at. So we really feel we’re approaching that today. We’re getting really back on that, repeatable cycle of innovation that I talked about. What you'll see from us -- so we’re on the lower range, historically.

And what you’ll see from us and you probably see this from the most of the Intel is we want to accelerate the top line, in by the way very highly expansive TAM, right now. We see the TAM for FPGA is growing pretty aggressively over the next five years. So we believe we’ve got some good top line acceleration there, but we need to be disciplined on the spend. I believe we got back to where we wanted to be. Now, it's about efficiency and also leveraging the best of Intel.

There is opportunity there for synergies within the company where PSG -- at Altera we don't need to invest in all of these other things, we can leverage it. And that's why I keep coming back to our architecture with tiles. It's a lot less expensive, from an R&D standpoint to build out that ecosystem then build up everything monolithic and go after every market with a monolithic device. So I think you can think about this as profitable growth managed spend in a very expanding TAM from that standpoint.

Question-and-Answer Session

Q - Vivek Arya

Let me pause there and see if there is any question from the audience. Okay, let me then just ask this last question, which is diversification that in your data center business. So Microsoft has been a customer, a strong customer for a long period of time. Talk to us about diversification outside of Microsoft, where you are and what’s been the progress there.

Dan McNamara

In order to talk about data center, you just touch on where we see FPGA is playing in the cloud, first of all and then we’ll talk. There is a lot I cannot say in terms of engagements. But we do see a tremendous pent-up demand for understanding and enabling with FPGA since some of the challenges in the cloud. So first and foremost, we call it infrastructure. And this is removing the network stack and accelerating the network security storage off of Xeon. And if you think about it for a cloud service provider, anytime we're using Xeon to run all their workloads is not good. So there's a big networking and overarching infrastructure off-load goal. And we're doing very well with that, and we had engagements across, I could -- I don't want to name them, but we all know the customers. And this is a really interesting area for FPGAs, but here's the other piece of that.

Vivek Arya

Could you give us an [indiscernible] about the customer?

Dan McNamara

Just in general, think of the top ten in cloud where we have engagements across all of it. But here's the interesting part again, and I want to bring this back to why the acquisition is so important. We have a group called the Connectivity Group at Intel. They have a large cache of IP for network and storage, and compute offload. So not only we’re bringing an FPGA to market, we're bringing this IP and implementing that in our FPGA. And that is a clear differentiator for Intel FPGAs.

So again, this total solution set really brings to bear really strong solutions for our customers. So that's the infrastructure. The other one is what we call these predictable high cycle applications. So this is the deep learning, the Brainwave. This is a video transcode. If you think about what's happening with these device with our handhelds is, all the video in different formats need to get rationalized in the data center. So video trans code is a big one.

Again with our parallel architectures and our solution sets, we can manage and process that data effectively. And then there's the third level, which we call these higher level apps, which think of those genomics acceleration, financial risk analysis, database acceleration. And we're going after that with again the pack and stack story, which is complete solution with the partner ecosystem. So we believe that our traction here is going extremely well. And by the way, we are still very bullish, we were bullish in '14, we were bullish when the acquisition was announced. We're also, I would say, we're even more bullish about our traction today than when the announcement -- when we closed the deal based on the uptick. So we see those three areas. And it's hard, because you can't -- only certain customers will allow you to talk about them publically.

So we've announced [indiscernible] -- Microsoft, obviously, that's a very good partnership and again we're working across the board. And I would say why do I believe we'll win, it's really about that end solution with Intel.

Vivek Arya

Thank you so much, Dan for giving us your time. Thanks everyone.

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!