Intel Corporation's (INTC) Management Presents at Nasdaq 39th Investor Conference (Transcript)

|
About: Intel Corporation (INTC)
by: SA Transcripts
Subscribers Only
Earning Call Audio

Intel Corporation (NASDAQ:INTC) Nasdaq 39th Investor Conference Call December 4, 2018 3:15 AM ET

Executives

Murthy Renduchintala - Chief Engineering Officer and Group President, Technology, Systems Architecture and Client Group

Analysts

Joe Moore - Morgan Stanley

Joe Moore

Hey. Good morning. Before we get started, I am Joe Moore, by the way, from Morgan Stanley. Before we get started, Intel has asked me to read the Safe Harbor. Today’s discussion may contain forward-looking statements. All statements made that are not historical facts are subject to a number of risks and uncertainties and actual results may differ materially. Please refer to Intel’s most recent Earnings Releases Form 10-Q and 10-K for more information on the specific risk factors that could cause actual results to differ.

And with that, I am happy to introduce Murthy Renduchintala. Murthy is the Group President of Technology Systems Architecture and Client and the Chief Engineering Officer at Intel. The Group is responsible for aligning technologies, engineering, product design and business direction to extend Intel’s strategy and speed execution across the client device and IoT segments. So I am thrilled to have you back. I think it’s your third time here, if I am not mistaken.

Question-and-Answer Session

Q - Joe Moore

Let me just put up a big picture, you guys had a great 2018 in terms of numbers. You have had some challenges on the manufacturing side. So we will go into the businesses in detail. But can you maybe just give us a view of how did you guys grow so much this year across all of the segments. Can you just give us a general big picture of what we are seeing?

Murthy Renduchintala

Sure. Well, good morning, everyone. It’s great to be back in London. I think in many ways 2018 was a banner year for Intel, not without our challenges, as you described. From the expectations we set, with respect to revenue in January of $65 billion, we have raised that over $6 billion over the year. We have just recorded a record quarter in Q3 $19.3 billion. Year-to-date, our data-centric business is growing like 24%. Our client businesses have grown 9% year over year. So, all in all, I think, it’s been a very good year by the numbers.

As you pointed out we have had our challenges. I think one of the consequences of getting $6 billion of upside in revenue is we have been chasing supply and some of our businesses are somewhat constrained in terms of equalizing demand and supply. We have invested about $1.5 billion of extra CapEx since our January guidance, bringing our CapEx investment for 2018 to about $15.5 billion.

And as you pointed out, we have had some challenges on the manufacturing side in terms of getting our 10-nanometers out into the market per our original schedule. So, all in all, I think, it’s been a great year by the numbers but not without challenges.

Joe Moore

Yeah. That makes sense. I think you guys talked -- you talked about data-centric just now. You guys have sort of used that as a guide post for investment. But your client business has grown a lot as well. So, how do you think about the trade-offs between those businesses and how do you balance investing in client…

Murthy Renduchintala

Yeah.

Joe Moore

… which is still a good growth engine and enormous profit pool versus the growth in these data-centric markets?

Murthy Renduchintala

Yeah. Great question. I think that both businesses are very complementary in many, many regards. I mean the overarching strategy for Intel is really driving the transformation of the intelligent digital world as we see and the continued explosion of data in every sphere of our lives.

We believe that the data-centric business is the growth engine of the company and over the next few years will ultimately become more than 50% of the company’s revenues. However, I think the PC client business or the client business I should call have been particularly resilient and they provide us a profit and the scale engine that allows us to make deep investments in leading-edge technologies on the manufacturing side, on the IP side, on the architecture side and on the platform development side. So you essentially have a growth engine for the business, as well as a foundation that gives us scale and the ability to invest, to drive those leading-edge technologies.

We are playing in a TAM that we have characterized north of $300 billion, and therefore, there’s a lot of room for us to grow. A lot of our growth is going to come from those data-centric businesses, about two-thirds of our projected growth is going to come from those client-centric businesses. But, again, the IP synergy between those two businesses is very, very large.

Joe Moore

Great. And I did want to ask also on a more big picture side, because you are relatively new to Intel, I think, three years or so, with a history at Qualcomm. You come from an organization that was fabulous that certainly differentiated on process technology in different ways, but versus one that we have a fab and you think about process. Can you talk about the differences between the two companies and I think you brought a lot new to Intel in terms of thinking about alternative ecosystems and the right way to approach those things. Can you just talk about how…

Murthy Renduchintala

Yeah.

Joe Moore

…you marry those two experiences?

Murthy Renduchintala

Sure. Well, in many ways, I think, what I have really tried to do at Intel is to fuse the best of the fabless ecosystem in my 20-plus years of experience with what’s really great about being an IDM. I am really trying to get that explosive chemistry together.

When you are in the fabless environment, however influential you are and Qualcomm was quite influential, you are one of many voices in the ears of the foundries in terms of understanding and implementing the requirements that you feel are important for your technology roadmap going forward. You are constantly having to have your requirements arbitrated with others, so that the foundry can get a definition that can host a large number of customers and you make compromises as you go through that.

In an IDM situation such as we have at Intel, you have a much more dedicated mindset in the definition of process technology towards the optimum point for your power and performance of your leading-edge products, and therefore, you have a very sophisticated dialog that allows you to drive for in-segment leadership in areas where power and performance are really key.

The other thing that’s really been revealing to me is beyond just processing technology, how important advanced packaging and agile assembly and test capability are. They are actually as important for product leadership as even the fundamental process technology.

So what you learn in an IDM is when you are able to have a high correlation between process, definition and architecture requirements, and you can fuse that with advanced packaging technology and agile assembly and test, that’s a pretty powerful mixture.

And the real challenge for me and for others in Intel is how do you keep that innovative flair across the entire panorama of skills in Intel in balance so that you can then translate that to schedule predictability across our roadmap.

So it’s one of the challenges we’re harnessing and putting a degree of boundaries on the innovation spirit of the company. It’s one of the challenges to make sure we translate that into timeline accuracy.

Joe Moore

And I think one of the changes that you have made to the client business in particular, you used to have the sort of more rigid tick-tock methodology and you have shifted towards more waves of innovation. Can you can you just talk about that trade-off, and obviously, part of that is, if the timeline takes a little longer in process, it’s hard to do it every other year. But it seems like there’s more to it than that. It seems like there’s more to it, thinking about bringing the best products out at any point in time.

Murthy Renduchintala

Yeah. I mean, that’s really the point, Joe. I mean, the entire focus of our business units across all of our roadmaps is really to drive product leadership. And process technology is important to the degree that it underpins product leadership. It’s necessary but not a sufficient condition for product leadership.

And one of the things that we have been really ramping up over the last few years is flexing other muscles inside the company beyond just process technology to really drive product leadership to its absolute apex. So, for example, IP excellence, architecture excellence, in-package memory, interconnect, advanced packaging, platform innovation, software, security. Product leadership comes from combining all of those ingredients, together with process technology, in a judicious manner and what you are able to do is continue to maintain a product cadence with meaningful generation-to-generation improvements even if one or two of those vectors aren’t firing on all cylinders. So, I think, that’s one of the things we have really created, a resiliency in the chemistry of our approach towards product leadership.

Joe Moore

Okay. Great. And then on that topic of 10-nanometer, you have talked about it a couple of times. This year you announced a delay in the timeline. Can you talk about that? What -- give people a little bit of an overview of what drove that delay and are we still on track for 10-nanometer next year?

Murthy Renduchintala

Sure. So our timeline for 10 nanometers remains as we stated in April. We are aligned with our plan to have client systems on the shelves by the holiday season of 2019 and then the rest of the portfolio product shortly thereafter, including our Xeon Scalable processing capabilities. So no change there, we reiterated that on our last earnings call.

And again, just to go back to the root causes of the challenges we have had on 10 nanometers and it goes back to the early definitions of 10 nanometers in 2014, that even predated my time with the company, where we signed up or we had the ambition of delivering a 2.7x scaling factor beyond 14 nanometers as we approached 10-nanometer and bear in mind, this was in an environment where we still had no EUV.

And that required a tremendous amount of process and technology innovation, many, many revolutionary experiments to be bought together at the time that we thought were bounded and calibrated, but ultimately when you put all of that chemistry together, we clearly underestimated the challenge ahead of us. So we have decomposed the root cause of that risk profile and put in place a plan that we think is now measured towards the timeline that I just spoke about.

Joe Moore

And what will 10-nanometer bring you, I mean, you did a Technology Day in 2017, where you sort of talked about the benefits of…

Murthy Renduchintala

Yeah.

Joe Moore

… higher transistor budget. It doesn’t seem like it actually moves the clock speed higher necessarily. So it’s more about other dynamics. But, obviously, you have been -- competitively, you have done very well with the 14-nanometer portfolio. But can you just talk about the benefit…

Murthy Renduchintala

Yeah.

Joe Moore

… of 10 to make sure?

Murthy Renduchintala

The ambitions for 10 haven’t changed. I mean even though we have had the trials and tribulations with 10, the power and performance and transistor density targets that we set in 2014 remain the same.

We could have taken two approaches, we could have slackened off on the tech specs of 10 nanometers and move towards a faster schedule or we could have maintained the tech specs and basically done the work to get it out on the timeline we talked about. And that’s really because of the resiliency that we have been able to engineer in 14 nanometers.

Intel doesn’t really do a great job in terms of speaking about its manufacturing technology, I think, in that scenario we can become sharper on. But we have had three iterations of 14 nanometers within the last few years. I mean if you look at the performance of the current generation of 14 nanometers shipping products compared to the very first generation from a process-only perspective we have improved that 30% to 40%, in terms of transistor improvement. I mean that’s kind of like a [more] lower and a half in many respects. So, yeah, I think, one thing we should be clear is that we have had multiple iterations of 10 -- 14 nanometers…

Joe Moore

Just maybe give different numbers?

Murthy Renduchintala

Well, I will leave that to my marketing colleagues. But what it has done is basically allowed us to continue with the very aggressive tech specs of 10. We haven’t had to abate them. And again for us 10 is really all about continuing to drive the power, performance, equation, as well as delivering on the transistor density.

Now, we are thinking much more about the data-centric requirements of our process technology, as well as our PC-centric technologies and they are slightly different. PCs are very much about burst performance, peak transistor speed and frequency, servers are much more perf-per-watt, perf-per-core, perf-per-dollar, and therefore, you are trading off those requirements.

And what we have really tried to do is to make a process that is a sweet spot of serving our data-centric products, as well as our PC products. And I think that’s been something that we continue to define as being important and in 10, we maintained that challenge even though it takes us a little bit more time.

Joe Moore

I think that’s pretty interesting because I think one of the advantages Intel has had historically is having a process technology that so optimized for microprocessors. Can you talk about even with the microprocessors now there’s these different requirements and as you start thinking about foundry and thinking about other things, it seems like you have to change the process to alter those things. So can you talk about the benefits of focus on CPUs and having a processor focus on that versus sort of and I know you don’t have aspirations to be a broad-based foundry, but that process technology that can do other things, how do you think about those trends?

Murthy Renduchintala

Yeah. I mean again process technology is one of those absolutely vital element ingredients for everything that we do and it’s not just about scale of processing CPUs. We are now moving into a world where parallel computation is becoming extremely important, whether it’s general purpose GPUs, whether it’s FPGAs, whether it’s custom silicon for AI.

The ability to do vector, scalar and parallel compute now become really, really important and they all have requirements in terms of power versus performance versus area. And I think what’s really happening is that the general roadmap for excellence in computation, whether it be scalar or vector, or matrix is now interacting with the requirements for high-speed interconnect.

One of the things, the challenges we have is it’s great having this massive processing engines but you have to feed them with the ability to get data on to and off those platforms whether it be through memory controllers, high-speed IO or other interconnect mechanisms, and sometimes the physics of scaling logic don’t necessarily correlate too well with the physics of scaling a high-speed IO. So, mixing those together has been a really interesting technical challenge for us.

And we’re taking some extremely novel and disruptive approaches to solve those. I can’t say too much about that here, but, will be talking about that in the future. But I think those are kind of like really exciting challenges that lead the spirit of innovation inside the company.

Joe Moore

Great. And then just last question on process technology, what happens now with 7-nanometer? Do you -- when does the clock start on 7, do you need to get 10 kind of up and running and stable before you can focus on 7 or can you do that in parallel?

Murthy Renduchintala

Well, 7 nanometers for us is a separate team and a largely separate effort. And we are quite pleased with our progress on 7, in fact very pleased with our progress on 7, and I think that we have taken a lot of lessons out of the 10-nanometer experience as we defined that and defined a different optimization point between transistor density, power and performance and schedule predictability.

So, we are going back to more like a 2x scaling factor when we get back to 7 and then really moving forward with that goal. So we are very, very focused on getting 7 out according to our original internal plans.

One thing I will say is that as you look at 7-nanometer, for us this is really now a point in time where we will get EUV back into the manufacturing matrix, and therefore, I think, that will give us a degree of back to the traditional Moore’s Law cadence that we were really talking about. 14 and 10 were really about double patterning and quad patterning in the absence of EUV.

Joe Moore

Yeah.

Murthy Renduchintala

And I think that was one of the key things overcoming those challenges while still driving for the same generation-over- generation transistor density that was really the technical challenges.

Joe Moore

Yeah. Okay. That makes sense. In terms of the PC business that you oversee still over half of the company, it’s growing really nicely this year. You have continued to, I think, each year you have come here we have talked about average selling prices going up and everything else sort of flattened out but they keep going up. What do you think is happening in that business and are we at the point now of where we can look for stabilization we are sort of on a refreshed pace for the PC market?

Murthy Renduchintala

That’s a great question. I mean, first of all, let me say, absolutely great year for our client businesses. We grew, as I said, 9% year-over-year. In Q3, we delivered our first ever $10 billion quarter. And the growth in the business, the health in the business has really been driven particularly from the commercial and the gaming environments, and even in the client businesses, we have also been gaining share in our modem business as well.

So across the board, I think, it’s been a great year for the PC business and that reflects the payback of a strategy that has made our business so resilient, which has been based on real segmentation and investing in areas where we see real growth and being tremendously operationally disciplined in the way we invest.

So if you look at -- what I am really pleased about the business is the operating margins of those businesses year-on-year have improved in a really great clip, as well as being able to keep revenues growing even in a scenario where from 2011 to today, the TAM has declined by 25% to 30%. So it’s been a great example of operational excellence and focused R&D and identifying the areas of growth in what some oversimplifies a monolithic market.

Joe Moore

And do you still talked about it as a unit that, I think, is in mild unit decline, a market that’s in unit decline. Should we still think about that being the case or it just seems like we are getting to?

Murthy Renduchintala

Yeah. We take a relatively cautious approach over how we manage the business, focused on really defining areas of growth and then specific R&D to go after those areas of growth and fiscal discipline and so we won’t change our mindset.

However, I do think that as you look at the future, I think, you are seeing a degree of settling into specific use cases for things like the PC, the 2 in 1, the tablet and the smartphone, that I think all are identifying specific optimized use cases for each of these platforms.

Joe Moore

Okay. Great. Maybe shifting gears a little. Talk about wireless. You guys have had some success in LTE around thin modems. I know there’s a big push within Intel around 5G. Can you talk about the status of that and how much are you investing? How should we think about the trajectory of 5G?

Murthy Renduchintala

Well, 5G, we regard as a very important accelerant to our overall focus on transformation and it’s all about the massive growth of data that I think 5G catalyzes. I think it’s very important to also understand that 5G is much more than just an air interface. It’s just -- it’s much more about data rates and bandwidth and I think it’s a significant transformation on the network side as it is on the air interface side. I think the network environment that we see today is as we go through 5G is going to go through, probably, a significant a transformation as well we are from analog to digital in the late 1980s, early 1990s.

I think with the advent of software-defined networking and network function virtualization, you are really moving towards a scenario where the network is a virtualized cloud type environment and I think that in-and-of-itself provides a tremendous degree of agility and extensibility of the network architectures going forward.

On the air interface side, what also becomes really important is now latency as a third coordinate of importance in addition to bandwidth and data rate. The services and offerings we see in the future for the kind of new classes of machine-to-machine services are going to be extremely dependent on ultralow latency, and that’s going to be a key feature of 5G.

And I think Intel looks at 5G as an area where both across the cloud, the network, and the client, we’re a player that is unique in being able to see an end-to-end play across that entire environment.

In terms of specific timelines, Joe, we announced in November that we pulled in the launch dates for our first multimode 5G and our LTE modem, the XMM 8160. We will be able to provide samples of that device in the second half of ‘19 with commercial devices in the first half of ‘20. That’s kind of the timelines we are talking about.

Joe Moore

Okay. But given where you start out, I mean, 5G is first of all both a new infrastructure and a new modem technology, and it’s also a bunch of different technologies, a lot of different pieces. Intel’s aspiration is to be broad across the entire scope of those technologies, both the infrastructure and …

Murthy Renduchintala

Very much so. I think we see 5G drawing more and more technology reuse out of the Data Center into the network. The network will cloudify effectively where you will have compute , storage, and data distribution embedded in the fabric of the network, and what it really means is you are pulling the Data Center technologies into the network, and then I think mastery of the radio or interface to be able to link the client-based data explosion onboard that on to the network is really where wireless becomes a critical control point IP.

Joe Moore

I guess your aspirations on baseband market share would be higher, and I mean, is -- if you -- do you see this as an integrated applications processor being an important part of that?

Murthy Renduchintala

Yeah. I mean, I think, that the modem -- the modem product line is an interestingly important product line for us, but as an IP wireless is going to be even more explosively important and not just cellular IP, I think it’s also going to be WiFi as well, so wireless connectivity in its broadest definition I think are going to be really important IPs.

I mean if you think about it, Joe, Intel’s strategy is really all about processing, storing, and distributing data, right. So, wireless technologies in that scenario of distributing data is a really important fundamental technology.

Joe Moore

Okay. Great. Maybe I know you don’t oversee the Data Center Group directly, but it’s a question that a lot of people have. The strength of that business has been really, really impressive and particularly the cloud piece, also the enterprise piece of the cloud last quarter growing 50%. We have certainly seen some indications that we are going into -- entering into a digestion period for cloud, we are seeing it from memory and hard drives. Intel doesn’t seem to have seen that, you’ve talked about Q3, you’ve talked about Q4, so – but why the Data Center business has been so robust for you guys?

Murthy Renduchintala

Yeah. Well, first of all, it’s been a great year for us. I mean, I think, year-to-date we have grown 24% which has been a very, very pleasant surprise. And as we you said, grown by excellent growth in the cloud, and we have been really pleased with how that’s gone this year.

I think as you -- as we’ve tried to size up what’s happening as we go into 2019, a lot of third-party reports that I have read, I am sure most of the audience have read, have talked about some degree of moderation in the growth, still growing next year but maybe at a slightly moderated rate, and it will be interesting to see how that plays out.

But for us, the transformation of the cloud and the cloudification of the networks is a multi-year journey, massive technical transformation and one that we think is going to be a long-term growth vector for our DCG business.

And one of the things that we are needing to make sure we are positioned for is, as the cloud part of our business becomes the largest part of our DCG business, then the lure of large numbers comes into play, and therefore we take a little bit more of a moderated perspective on the CAGR of that business over the long term.

Joe Moore

Yeah.

Murthy Renduchintala

Yeah. The other thing that’s really surprised us in the Data Center business has really been the strength of the enterprise. I think it’s a mixture of strong macroeconomic conditions together with an increasingly sophisticated IT deployment of workload management, have really led to growth in the enterprise business in 2018, and that’s been a really good eventuality for us, so it’s not just been the cloud, even the enterprise has performed well in 2018.

Joe Moore

Yeah. Makes sense. On the topic of competition across both of the markets across PCs and in Data Center, AMD has, obviously, got a more competitive roadmap than they have had for a while, but you guys have talked about still having better per core performance and the advantages that that brings. So, can you just talk a little bit, generally, obviously, so far the competitive environment hasn’t affected Intel that you are ASPs are going up in every segment, your market share has been holding in nicely at every segment. Can you talk about how you think about that for 2019?

Murthy Renduchintala

Yeah. Sure. Joe, I think, it goes back to this whole focus of how do you drive product leadership and it being a result of excelling on a very wide number of parameters, really making sure that the elements of product leadership are as broad as possible so that we can excel in as many vectors as possible.

I think for us, we have got a really clear roadmap. We have excellent support for our customers in terms of being able to deploy solutions very, very rapidly. We have taken a great deal of focus in evolving both our 14-nanometer and defining our 10-nanometer program, and we have really talked about how we go way beyond just thinking of our product leadership being a CPU-centric phenomenon.

What’s becoming more and more important is essentially the power of the platform. For example, if you like, take our PC business, you now have not only the PC, but what we are doing with memory, what we do with high-speed interconnect, what we are doing with connectivity, and then as you go forward what we will be doing in the discrete graphics space.

And then how do you bind all of that together with a support system and development ecosystem that allows people to innovate on top of our platform, so we think that product leadership comes from being able to deliver on a multitude of different vectors in a way that customers can continue to differentiate.

Joe Moore

Okay. Great. And then, just final question on the Internet of Things business, about 5% of your business has been growing nicely. Can you talk about that opportunity longer term and maybe a little bit shorter term as you talked about some declines this quarter around the silicon constraints, can you just talk about what you are seeing in IoT?

Murthy Renduchintala

Yeah. So, the IoT business is a really, really important business for us. And I think as we have moved into defining the future strategic landscape for that business for us, we are really focused on particular verticals and the particular targeted set of workloads, so for us the real key verticals we think are important are in the industrial retail smart city space.

And the real sweet spot for us in terms of IoT growth is the edge of network, where I think workloads such as edge inference, vision, and video, as well as the aggregation of compute are going to become really, really important.

And the IoT business is a really cool business, because it leverages so much R&D from our other businesses, so it is a R&D and capitally efficient business that leverages a lot of R&D, but is also supportive of some specific R&D in those areas where we believe we can get real differentiation, so a really good business for us.

Joe Moore

Just a quick follow up on that, I mean it seems like inference is so important to Intel. If I look at Mobileye, Nervana, Altera, the way you just talked about IoT. It seems like inference is -- you are attacking it in a lot of different ways and it seems like it’s a pretty big priority, is that a fair assessment?

Murthy Renduchintala

Absolutely. I mean, I think, inference, if you look at a lot of the third-party studies and our own thinking is that, as you look at the scope of the AI opportunity, a lot of emphasis and energy has rightly been put on training up until now, particularly Data Center training.

But if you look at the future and where the majority of revenue is going to come from or at least a lot of the revenue is going to come from inference, both in the Data Center and progressively through the network towards the edge, the client environment is going to be an incredibly important area, and it’s going to require solutions of different power versus performance for the various opportunities you see. So while everybody’s focusing on training as Intel is, we shouldn’t in any way shape or form, under dimension the opportunity that exists for us in inference.

Joe Moore

Okay. I am afraid, we are out of time. Murthy, thank you very much.

Murthy Renduchintala

Great. Thank you.