QUALCOMM Incorporated (QCOM) Earnings Call Transcript & Summary

All right. Good afternoon -- well, good morning, good afternoon, and I guess, for some of you, good evening. This is Ross Seymore. I'm the head semiconductor analyst here in the U.S. for Deutsche Bank. We're very pleased to have you dialing in for the 2020 Deutsche Bank Automotive Tech Conference. Our next company to present is Qualcomm, and we're very pleased to have Nakul Duggal, the Senior VP and GM of Qualcomm's Automotive business, joining us for this next slot. This will be all fireside chat. There aren't any slides. So if you don't see slides, you're not missing anything. [Operator Instructions]

So with that being said, let's level set a little bit. Nakul, to kick things off, I wanted to level set Qualcomm's automotive business today and the trends that you're strategically focused on going forward. So conveniently, on last week's call, your last earnings call, the company highlighted that the automotive revenue for the company in its fiscal year 2020 was just about $650 million, with your cumulative design win pipeline up to about $8 billion from about $6.5 billion a year earlier. And then the final stat you guys provided was you believe the automotive SAM that you can address is growing from about $3 billion last year to about $4 billion, a couple, 2, 3 years from now, about a 12% CAGR. So with those numbers being given, not a lot of people really know what you guys do in the automotive market. So why don't we just start off, in that $644 million you've had in the last fiscal year, what constitutes that? What sort of applications chips are you selling, and so we can just get a baseline to build off of?

Yes. So thank you very much, Ross, for having me. Nice to talk to you all. Maybe let me just baseline a little bit in terms of giving you some history in terms of what our automotive business has looked like over the last many years to give you a sense as to what we shared at the earnings call and how we are projecting the next 4 to 5 years or so. We've been in automotive, focused on the wireless connectivity piece for almost 2 decades. We are in the process of supplying our ninth generation of modem products. And what that shows to you is that while it has been a market that has existed for a long time, the market is now starting to mature, and every car is shipping with the modem. The automotive market is quite different from typical consumer markets. The life cycles involved, the requirements are quite different. And what we've been able to do over the last many years is to be able to build a portfolio really with every automaker globally on our wireless connectivity piece that allows us to make sure that we are their trusted choice for putting wireless connectivity into their cars. So we've been in 4G over the last 4 to 5 years. The 5G transition is starting to happen. And that has been a large part of our overall portfolio. And that continues to be a market that keeps growing because cars are attaching more and more modems. The second part of the disclosure was our infotainment business or our digital cockpit business, which is a relatively newer business. And that is something that we got into starting in 2016. So really, only a few years ago, speaking in automotive time lines. But that business has also grown very rapidly because there is a significant transformation going on in the industry in terms of how automakers differentiate themselves, how they reset the experience, how they define their own brand identity, especially inside the cabin. And that has -- that's created a massive opportunity for us and a number of our partners to be able to help automakers and Tier 1s reinvent the in-cabin experience. So that is the second part of the portfolio that contributes to the disclosures that we made. And then the third piece is our WiFi, Bluetooth and GNSS positioning business, which is the third piece, which is also really going to full attach. So if you look at the major trends in the market, as cars become more digital, as they become more connected, really every single piece of the business that we have is getting the full attach, which is what has driven the growth, and we continue to see that going forward.

Great. Thanks for that color, Nakul. I think why don't we dig a little bit deeper into those? Talk a little bit about the modem attach to cars. Connectivity has a lot of different attributes. And maybe a way you've talked about it before is kind of a connectivity, a compute and an autonomy, the way of thinking of the world as far as 3 themes. If we start with the connectivity side of things, let's think about from a WAN perspective, talk about the 5G connectivity, the role that plays and maybe even dovetailing into your kind of the V2X side and what role 5G versus more kind of idiosyncratic standards could play?

Yes, absolutely. So I think the way to think about the car, Ross, and I learned this, I've been in the auto business here at Qualcomm for about 9 years. And the way to think about it is that the car actually has multiple personalities depending upon from what vantage point you're thinking about it. So the automaker has to think about the vehicle as something that they can manage from the cloud. So it requires a lot of outside-in capability in terms of how you manage the vehicle, how you diagnose it. Then there is the car itself in terms of the types of features that the car needs, over-the-air updates, diagnostics that are within the vehicle, sending data back to the -- back to HQ. And then finally, there is the aspect of the consumer, the driver, the passenger, whatever the use case is. And what that ends up doing is it makes the connectivity problem statement in the car quite different from a traditional. If I were to compare it to a smartphone, for example, a very different problem statement. So it is, at the same time, a high-speed connection, 4G now moving to 5G. It is a wireless hotspot because there is WiFi almost always colocated at this location. You have to be able to have a lot of security infrastructure around connectivity because that is the one ingress point into the vehicle. You have features like positioning, precise positioning, lane-level navigation that are all centralized within the modem framework. So that is another major feature, very important feature for vehicles, both in terms of navigation and the future for autonomy. And then you have technologies like V2X, which I'll talk about. So if I were to unpack that, if I were to break that down a little bit, what we started to see many years ago was really the reuse of a traditional consumer modem that goes into the vehicle, with little change. Yes, maybe there were some applications that were different, but the features look fairly straightforward. It was either for emergency calling or it was used for some concierge services, some navigation services, et cetera, fairly basic capability. Today what we are starting to see is really the modem, think of it no different really than how you might think of a customer premise equipment or a networking endpoint where it has to carry a variety of different type of traffic. There is traffic that belongs to the automaker that has to be secure. There is traffic that belongs to the consumer that is carried on its own channels. And then there's various other types of traffic depending upon the types of services that we're running. So as we start to deploy 5G here over the next 12 to 18 months, we are seeing really a complete overhaul in terms of how automakers are thinking about the connectivity portion of electronics in their vehicle. The antennas have to be designed differently because 5G requires what is known in technology is the 4x4 MIMO. You basically have to have 4 different antenna elements for transmit and receive that requires the car design to change to be able to accommodate that type of capability. The location of where you put the modem becomes important because of the distance from the antenna to, say, the unit inside the vehicle is too far away, you have losses there. So you have to be able to figure out how to put the electronics closer to the roof line. And this is really a major architectural shift for a lot of different automakers. So our role as Qualcomm has not really been to be the provider of the semiconductor, but it's been to become kind of the adviser for the system platform for the automaker. How would you design the next-generation TC with 5G? Over the last couple of years, one thing that we have started to add to our portfolio is the RF front end part of the equation. I think you heard about that in earnings last week. And since we have added the RF front end portfolio over to our larger mobile road map, we have also started to provide automotive-qualified RF components. So we now have in telematics, which is kind of how we call that category. We have everything from 4G, 5G baseband, the chips that go alongside it, the RF that accompanies it and then a lot of software that is really riding on top of these platforms that allow you to be able to do everything from positioning, the WiFi support, the V2X requirements and everything that really makes that secure. What we are then also seeing on top of this is if the requirements for connectivity in the car are increasing, you need dedicated application processing at that end point. So you have to be able to have a dedicated application processor that is capturing the capacity needed to be able to run applications, whether they be telematics applications, security applications, V2X, et cetera. So it's becoming a much more complex end point than where it started out 10, 15 years ago where it was a simple modem.

That's great. Perfect color on that. Why don't we double-click a little bit about the OEM relationship. You said that it's more of a consulting relationship than an adviser relationship. How does that transition from an advisory role to a go-to-market role? And is that done via the OEMs, the Tier 1s or some combination of the 2?

Yes. So I think there are a couple of aspects here. I think historically, given our role in just the overall wireless ecosystem, we were seen as the experts that a lot of automakers started to work very closely with to get advice, to get some direction as to what decisions they take in terms of timing, in terms of road map, in terms of choice of technology partners, et cetera. And that is really how we got into automotive. If I go back 15, 18 years, when we started off at General Motors here in the U.S., that was really kind of the inception of the business. What has happened today, and I think what we see happening over the next 4 to 5 years, is that the very -- the very solid lines that used to exist in terms of the traditional supply chain, what a Tier 1 is, what a Tier 2 is, et cetera, those probably still exist in terms of the supply of the hardware, in terms of the manufacturing requirements, in terms of how a product is installed in the vehicle, managed, spares, et cetera. But in terms of the road map, in terms of the partnerships that will define how a vehicle is going to transform, I think that has all changed. So let me give you some examples. Traditionally, automakers relied heavily on Tier 1s to be able to do a lot of, I would say, most of their development. I think that has really changed across the board to different degrees, depending upon the automaker. Some automakers do pretty much everything in-house. Others are moving in that direction. But what that does is, from a sourcing perspective, from the actual decision-making on silicon or on systems, the automakers have now become very active in terms of directly engaging with their partners, such as us in terms of being able to make choices around what silicon they want to buy, what they would pay for silicon like that, where all would they use their silicon, what requirements do they need to see there? This is what is known in automotive as directed by or a directed sourcing. At the same time, there is a very close partnership with Tier 1s that exists because you have to be able to really have the 3 elements work very closely. The OEM, the Tier 1, and the chip supplier have to work very closely in terms of being able to actually deliver the solution. Another big shift that we are seeing is that automakers are becoming very active in in-sourcing software development. So aspects of software where they do a lot of their own differentiation or a lot of their own unique identity is an area that automakers are in-sourcing, and they would like their partners to really be able to provide them with the chassis or the platform that is more complete building blocks, modular building blocks of hardware and software. So we end up having direct relationships with automakers, where we are helping them in software development. Now all of this would have been unheard of 5 years ago. The interest was obviously there. But for this to happen at scale, I think the industry was far away from it. But we are seeing this transition happen more and more across really every automaker across the globe. So my expectation is in the -- we are now targeting -- for new design wins, we are targeting the start of production '24, '25 time frame. I expect that a large majority of automakers in many parts of their business will actually have a hybrid model, where they will work very directly with system suppliers like ourselves, and they will continue to work with the Tier 1 ecosystem, whose role will also change.

Why don't we dig a little bit into the 5G as the connectivity platform? Latency is an issue that the cellular network has potential pitfalls with. 5G less so than 4G, but nonetheless, latency can be an issue, especially when you need real-time or near real-time V2X communication. Talk a little bit about 5G as a solution for that versus more application-specific wireless protocols and how Qualcomm is approaching those differing strategies.

Yes. I think it's a very good question. I think we recognized early on, I would say, back at the beginning of this last decade, that V2X was actually a very important application for 3GPP-based wireless technology. So the work in terms of what you're seeing today in cellular V2X started back in 2012. I think the way to think about this is -- maybe I think the first piece is you have to be able to have an ecosystem and you have to be able to have standards bodies. You have to be able to have an ecosystem investment that allows for technologies like V2X to continue to advance. Because what is going to happen is the use case is for V2X are going to keep evolving, and other technologies that have been focused in this space for a number of years, for multiple decades, actually, haven't really kept up with how the market was going to evolve. So let me give you some examples. When we introduced cellular V2X back in 2016, really, the main changes that we were making was that we wanted to make sure that the underlying radio technology was well suited for the needs of the vehicle-to-vehicle application for the auto industry, so a safety application. And that was our primary purpose. We wanted to make sure that from a safety perspective, the criteria that needed to be met were exceeding the expectations of the auto industry. And then on top of that, we wanted to be able to deliver features such as let -- better congestion characteristics, more capacity, et cetera. We made sure that we did not disturb any of the upper layers because the upper layers where the actual safety messages are carried, where there has been a lot of work done in the industry, those were really done extremely well. We just wanted a better radio. And we've actually been able to accomplish that, I would say, in a fairly short period of time. In China, we are now starting to see deployments of C-V2X take off. So in a fairly short period of time, this is starting to move. And I think it requires a few key ingredients to be able to get moving. I think one is automakers, obviously making a decision as to what technology they will put into their vehicles. And we've had -- so we've been working with a variety of automakers globally there. Infrastructure providers which, in this case, will vary depending upon which jurisdiction, which part of the world you are in, but really, the ultimate provider of V2X technology is the road operator. So road operators deploying the infrastructure, what are known as roadside units, that have the same capability and being able to do that at scale, once you have that, I think the potential for V2X is actually immense because it is a complementary sensor. It is a non-line-of-sight sensor. And it is able to provide a lot of information in real-time in what we call is a very hyper-local context. So if there is a roadside unit within, I would say, half a mile or even less, you were able to get real-time information on what is going on around that intersection or around that specific part of the highway, which is very important in terms of being able to reduce accidents, to be able to reduce congestion. And this isn't something that is futuristic. We are actually seeing this getting deployed in China. And we think that we are actually in the U.S. quite close to be able to seeing that same type of technology being available to deploy.

So when you say some -- whoever owns the road, is this done by municipalities, states, counties, federal mandates? I can see it being a great technology, but quite a patchwork and quote work of it that I'm not sure how reliable it is if you don't have a more ubiquitous coverage. How do you guys attack that?

So I think it varies, obviously, depending upon the part of the world you're talking about. In the U.S., it is like you described. It is something that has to be mandated federally and then has to get deployed on a state-by-state basis. And within the states, it actually gets down to the specific jurisdiction, specific cities. Yes, there is complexity. But I think in terms of the standardization of the technology, that actually makes a big difference. One of the things that was very interesting to automakers was when we decided to make the technology available as part of our standard road map. So we basically said, hey, when you put telematics in the car and you are putting a 5G-based modem in the car, we will actually make sure that it is C-V2X-enabled. So there is dedicated -- there is a dedicated application within the modem for C-V2X That reduces the overall cost of ownership for the automaker. So you can now get ubiquitous deployment of the technology across every vehicle very quickly at very little additional cost for the automaker. And then the other thing that is needed about the technology is that the same modem that goes into the vehicle, the identical modem is actually needed in the infrastructure, in the roadside unit. So it's not like this is some highly complex piece of silicon. It's actually the same type of silicon. This is broadcast-type capability. So it makes the implementation fairly straightforward. In the U.S., I think the reason for the delays have been that there has been DSRC that has been debated for a number of years. We are now seeing a lot of interest in C-V2X as the technology that gets deployed. The FCC has made a recommendation that C-V2X is what is deployed as the technology for V2X in the future. They've made a recommendation on setting aside 20 megawatts of spectrum in the 5.9 band, and we are awaiting a final decision on how that can actually get rolling going forward.

Got it. Well, thanks for the color on the connectivity side. How has the leadership role that Qualcomm has had clearly in connectivity in a cellular basis for decades, how is that leading to engagements that would pull in the next side of the equation, where you said you have application processor. So the computing end of the equation. Talk a little bit about how the connectivity relationship leads to the computing side and what Qualcomm's aspirations are on that computing plus processing side of this grand equation?

[indiscernible] for us. As you know, Snapdragon has been a leading application processor focused on smartphones for a number of years, and we've seen a lot of success there in that category. What we started to see, this was about 6, 7 years ago, as we started to engage with automakers, some leading automakers, Audi was one of them, who had the vision that the car was going to move from what is a traditional infotainment system to what will be a digital cockpit. And what I mean by digital cockpit is that there are many different audiovisual themes that are going on in the vehicle that are highly complex. There isn't really any analog that I can think of to the problem statement because there are a number of functions that as you start to make the car much more complex require you to be able to interact with the driver and the passengers in a safe way. These have to all be concurrent. They all have to be real-time in nature. Some are infotainment and entertainment-related. Others are safety-related. Some are local to the vehicle. Others are connected to the cloud. So it is a fairly complicated problem statement. And if you think about the cockpit from a software perspective, what it is really running is multiple software implementations of various applications concurrently and safely. So for us, the opportunity was how do we pivot our Snapdragon platform? We develop a tremendous amount of technology, a tremendous amount of IP that goes into Snapdragon. How do we pivot that for it to be relevant for the auto industry while making sure that we continue to provide the same time to market advantage, the same process node advantage, the same low power advantage? And this required us to make 2 very key decisions. One was how much of the IP that auto needs do we integrate into our baseline platforms? And how do we get to the quality requirements that automotive requires, but yet be able to meet the same time lines that we are used to doing, that we're used to delivering in the smartphone space? The other part, which was major learning for us, and I think we've invested heavily there, and I think we've done fairly well is to really become software-first in our thinking. So the software in the infotainment system on the cockpit is nothing like what we see in the smartphone. It is quite different. There is a lot of complexity in terms of the virtualization that is needed, the real-time nature, the fast boot capabilities, coexistence of consumer ecosystems with real-time operating systems. And the third element that makes this highly complex is how you make it scalable. So automakers sell cars of different varieties, different tiers. It is still a very cost-sensitive space. So the road map definitions that we had to make were we need to be able to build premium tier chips. We have to be able to have something in the mid-tier and then continue to go up from the entry tier because automakers and Tier 1s still need a lot of scalability and a lot of reuse. So we adopted a strategy where we wanted to make sure that we address the IP needs, we address the scalability requirements. And then we invested heavily in supporting a variety of the different software requirements of the auto industry. And that has allowed us to become, I would say, an easy choice for a number of automakers who are looking to really compete in the cockpit space, which has become a major point of differentiation. We have partnered with a number of ecosystems on the consumer side. We partnered Google on the Google Automotive Services suite that they have. We've partnered with companies like QNX and Green Hills for their real-time operating systems. So this is really an area where you have to partner heavily with the industry because no one company has it all. But to be able to integrate the solution and be able to bring a very complete offering to an automaker so that they can start to do their own differentiation on top is kind of where the magic is. And that's been our strategy. And we are now designing or are supplying to over 20 automakers. So I would say it's served us quite well.

So is it fair to say those -- that infotainment side of the equation, I would -- is that the majority of the $644 million that you did last year? Or if you split that up, would it be more on the modem side? Just rough hierarchy.