QUALCOMM Incorporated (QCOM) Earnings Call Transcript & Summary

September 22, 2022. And -- but the winning rate has not changed since that time.

Okay. What drove the increase in September?

ADAS. I think it's very significant. We talked about in the long term, probably [ digital cockpit ] and ADAS will dominate the revenue, and some big accounts with the whole chassis. But there's -- another component of it is some design wins of software and defined vehicles because that's architecture, the car is also changing. .

Got it. So let's talk about the software side. So you just bought the Arriver software from Veoneer. So the rest of Veoneer is gone now, right? So that you've carved out the PC one.

I think if I'm not mistaken, we did announce, I think they sealed the active closing.

Okay. Got it. Okay. That's right. What does that bring to the table? Why did you need that? Like you needed it enough to actually go over the top on that deal with [indiscernible], right? I guess you sold the rest of it back to -- was it [indiscernible]? Yes. Okay. But why did you feel -- why did you need do that?

Excellent question. So -- and I think that's what is really setting us apart I think, and I think it was the right target to complement our Ultra business. What we got from Veoneer is Arriver, which is a computer vision stack for assisted driving and autonomy. So -- and what we did, you should think of this -- of a combination of the Qualcomm assets, you should think about this as there are multiple pieces that fit in their puzzle. We used to have our own stack. We have been doing a lot of cars, driving in San Diego for -- and other cities for a number of years, so we developed our stack on to drive policy. Veoneer brings a proven stack for computer vision. Think about the computer vision SoC, front-facing camera, rear-facing camera, 360. And then on top of that, we have this partnership with BMW, which is very unique which is we're jointly developing the driver policy software, and we have rights to the stack that we can offer to our OEMs. So what happened when I put all those 3 things together, we came out in the ADAS as a viable alternative of a proven stack. And a stack that we can scale from entry-level ADAS, which I think is going to go to every car, all the way up to 3 plus and 4. So I think that's -- it's not about creating services, it's exactly about providing a full solution for ADAS, including a proven stack.

You mentioned BMW. So I was wondering if I was reading too much into that, but like BMW, they were the original launch partner for Intel Mobileye, first one [indiscernible]. Even before Intel bought Mobileye. Intel bought Mobileye about 6 months after that was announced, and we didn't really see very much. And then all of a sudden, I see they're going to you for vision and so symbolically, that seemed important to me. Like am I reading too much into that?

You're not. It was by design. They were the first company to design Mobileye, the first company to design Qualcomm in ADAS and autonomy.

Got it. Okay. I guess most of the auto makeup today that was not the stuff, most of it say it was. Is it connectivity and infotainment and...

Yes. Well, we're starting to see digital cockpit started to become material in the financials. Over time, it's going to dominate since it's a longer silicon content. And then ADAS will come in late.

Until now and like in 5 years or 10 years, what do you think your content opportunity?

Look, I don't think we -- I don't remember if we gave a range of the number of dollars, but I'm going to answer it qualitatively. The content today, and it varies. Like one of the things that has really positioned us well with our auto platform is actually, we can scale. We took that mindset from phones. We can scale from a premium car all the way to an entry-level car with all those technologies. But we have -- we sell a full module for connectivity on the C-box, not only the chip but with an entire module. You actually have the ability to write applications on that. And we have -- so that's module. It's one revenue stream. The other one is digital cockpit. It could be 2 SoCs, it can be 1 SoC. It has multiple tiers. It powers the dashboard, the infotainment, rear seat entertainment, smart mirrors, head-up display. Then the other 1 is the ADAS, it's the computer vision SoC. For a number of different cameras you put it in, the drive policy SoC, that is very high performance inference. That's where we use our AI 100, and it has server level performance of AI. For example, GM Super Cruise, Ultra Cruise is all running on our platform, which is based on the same technology that goes into AI 100, and that drives significantly higher silicon and revenue content. I think the last 1 that we're going to add to it is the central computing end unit for the software-defined vehicle.

Got it. So like how do you differentiate against like video or mobile? Like, is there any other sort of like public companies that have full stack solutions? I mean you mentioned this like sort of scaling from high-end [indiscernible]. Is that what it is? Or is it...

Yes, so it's a great question. We take a different approach. We take a systems approach to this. And how we started thinking about this was like, okay, so investors are asking car companies 2 questions. Are you electrical, but are you digital? That's where we come in, the digital part. So we think about, can we create a full platform with a number of competencies across multiple tiers? So unlike our competitors, we cover everything. We cover all the connectivity of the car, all the digital cockpit options from premium to low, and ADAS from premium to low in autonomy. So I think we compete with different people at their respective spaces. One thing that is very unique, for example, which has resonated very well, and you can see how our competitors don't have that capability, is what we announced Flex. In the same SoC, I can do ADAS and IVI, which now allow some of ADAS to propagate down and I can leverage the stack. I can leverage same SoC. The other thing that was differentiated, how we think about this thing.

[indiscernible] what, in vehicle infotainment or...

Yes, or digital cockpit.

Okay. Got it.

The other things that we did, we developed a lot of -- our software business -- our auto business as much as software as it is silicon. We develop a platform that we can, for example, we're running 1 type of OS and it's likely [indiscernible] or fall tolerant with a lot of redundancy for the dashboard. We run a different OS for the infotainment, all using a hypervisor in the same chip. And now with Flex, we even bring ADAS with different levels of security and redundancy. So I think it's very mature. Look, I'm very proud. I think I've been -- in a short period of time, I think we're working with virtually every company, and this year, we got the GM Supplier of the Year award. So for me, it tells that we're now inside the tent of the automotive industry. I think we graduated as an automotive supplier. We feel good about position there.

Got it. Do you want to talk a little bit about the recent acquisition you did in the space, the [indiscernible]?

Yes. Look, we have been very passionate about this. It's just -- and this is because of how we think about this, maybe it's our mobile heritage, right? When you have a -- like we talk about the scar and when we talk about the scar, with now the car is connected to the cloud. The car has services, you have all of the screens, all the computer screens. But what if you zoom out and you see a network of connected cars? So that's the concept of intelligent transportation. So we're also making some strides as part of our growth strategy in Auto with 2-wheels, scooters, e-bikes, motorcycles. So this -- with Autotalks, Autotalks allow is a technology that we can leverage on CV2X, both the SRC and CV2X, for car-to-car, car-to-bicycle, car-to-pedestrian car-to-traffic light, and basically to drive that and try to create a traffic intelligent transportation at the top. So I think there's going to be another wave of growth, and it's about building on those connected platforms and having a network effect.

Got it. Got it. So we switch over to IoT [indiscernible]. So I know the near term has been a little challenged. Is this just a sort of a typical consumer inventory correction right now?

That's correct. Then we saw some of that in the enterprise, as some of the enterprises have been careful with their IT spending in the current environment. And so that's kind of -- I think Akash talked about on the last earnings call. The IoT is important. I think the term IoT, I know we use them but [indiscernible]

By the way, this can mean anything you want it to mean.

You're not alone. But I feel our definition of IoT is different, right? So don't -- when we say IoT, I think it's a mistake to think, this is going to be a microcontroller or connectivity into a smart appliance. No. It's a number of different things. Tell me where you want me to talk about.

Well, I mean just like what do you sell in this business?

That's a very good question.

Because I've got a model, and it's got a line in it that says IoT, and I throw a number in there. And I love that something like...

No, don't do that. There's a lot of things in that IoT segment. So I am going to talk what's -- I think I'm going to take an 80-20 rule. I'm going to talk what's really important. So our IoT, we divided in -- we talked about it, in what we call consumer, networking and industrial. But consumer, it's more -- you understand when I explain it. It's consumer and commercial there. What's in each one of those buckets? In consumer, we have primarily 2 big bets. One bet is the transition of the Windows PC to modern mobile SoC. We are the sole partner of Microsoft in the transition to an arm instruction set and mobile SoC. If you look of the Apple M Series, it's very clear that there is an opportunity there, and we have been on this journey with them. We have developed our SOC. It has our own custom CPU that came after the revision, it's the Oryon CPU. It has a lot of AI. I wait to you to ask the AI question to elaborate. But that is one big piece of the IoT consumer bet. The other one, I know investors used to like that a lot. Now, it's not that much.

I'm not talking about the metaverse any more.

Yes. But it is just a matter of time. I think the merger of physical digital spaces with the reality, augmented reality, mixed reality. And here's -- so that's another bet. I think we're working with all the large ecosystems that bet on Qualcomm silicon with the sole partner of Meta for their Quest devices. We announced a partnership with Google and Samsung. We announced a partnership which we've been -- already deliver on with Microsoft. We're working with [indiscernible]. We're working with all of the other China ecosystems. So here's how I'd like to think about that. Like I told you, the PC is, if you saw what Apple is doing, you can see that there is a play for us in Windows. How I think about this Metaverse is let's just wait. There's a rumor that Apple is going to launch something. I'm sure it's going to be incredible. And when they launch something, developers are going to think, well, they may sell multiple tens of millions. They're going to develop a lot of content. All that we need to do is to just carry over on the Android ecosystem or the other ecosystem. That's an easy way to think about it. I think it's more complex than that, but that's another growth opportunity we have in the IoT consumer space. In the networking space, there are 2 big ones. It's a 5G fixed wireless access or high power 5G for broadband. The way you should think about this is in broadband in the world, there's only 2 options. Copper is no muss, right? And then with cable, there's so much you can do about the uplink. So there's fiber, and if you cannot get fiber, it's 5G. We've seen opportunities in both develop markets, we see what T-Mobile, Verizon is doing, but also developing market...

Let's talk about India in particular for this, right?

India has an ambition to connect 100 million households with 5G, so it's a big opportunity. That's into the IoT network. And the other thing in the IoT networking is WiFi access point. I think most people don't know, they think of Qualcomm still as a comms company for cellular. But we're the #1 in WiFi, both client as well as the access point. There's a lot of opportunity of Wi-Fi 6, Wi-Fi 7 enterprise. The last piece of IoT is IoT industrial. That's a very long tail, very divided by vertical front. I'll tell the big ones that is driving most of our revenue. Retail, then you have smart building, you have energy, you have oil and gas, main effects [indiscernible].

Is this Wi-Fi? is it compute? Is it cellular, like what...

All of the above. So what it is think about companies, they're building digital twins of everything. So we provide connectivity, connectivity plus processing, connectivity plus edge intelligence. So it's highly leveraged of our road map. It's -- it basically leverage the R&D and chips we do for many other segments, and that's the largest SAM. It's just take a while to build it.

Yes. Got it. So let's talk about AIs then. So where does AI sit? I mean, is it a handset play to like keep content alive? Or is it an edge compute and IoT? Or like where are the big opportunities for Qualcomm in NII? And how do you see that playing out?

Okay. Look, I'm obviously very excited about this, and it's not just because of everybody is talking about AI. We've been investing for this for 10 years, right? And here's what I would like to do. I would like to paint my view of the landscape. And then after I paint a view of the landscape, then I talk about what it means to Qualcomm. Is that okay?

All right.

So let's just talk -- I'm going to remove Qualcomm for the equation for now. A couple of things happening. First of all, those large models -- and you saw with ChatGPT, they made a lot of progress. It's in a very short period of time. And since it became open, it's remarkable to see the number of different applications, start-up companies trying to do things. Whether they are new foundational models coming from other large companies or the open source community, but also how they build on plugins or foundational models and do different applications. It's fascinating to watch. We have seen people from retail saying, look, I would like to train a model on everything that I have in that store that day. Will you come into the store, be my customer, you say, I have $50. What should -- I have 4 people for dinner, what should I buy? I'll give you a shopping list and how to navigate through the store. Whether those things will pan out or not, but there's ideas left and right. So think about what you see today and how that's going to be 12, 24 months from now? I think there's a lot of opportunities. Second layer of conversation. It's clear now with those models that you can rely solely on the data center. Like there's all those different calculations. But if you -- let's just say something as simple as you're searching on Google. And today, I think I'm sure many of you do that, and there's billions of people that were doing it. All of a sudden, you change your behavior as a user. You start chatting about the [indiscernible] -- you don't go click on all the links and see what you want, you start asking questions about it. Well, think about a ChatGPT tree with over 200 billion parameters. For every word that you start seeing, those are tokens, right? You run the entire model, you run. So the cost in the data center will grow exponentially, and I don't think the answer is you're just going to do this in the data center. That's why we believe that there's going to be this hybrid model. That model is going to get smaller. There's going to be models that you're going to run locally on devices and -- or the device will give the head start, based on context available for you, to the cloud. That is already happening as part of the conversation about what the future of those thing is. Okay. The next -- I have just 2 more context and I'll answer your question. The next way to look into this. If you look at the history of computing, and I look at AI is accelerating computing. Just use NVIDIA term, how they describe themselves. So CPU started with IBM on the mainframe, then eventually, it became a personal computer. Some people didn't believe in that time that it will become a personal computer, and now, it's in the palm of your hands over a smartphone. It's going to be no different for accelerated computing. It's going to propagate, and for different reasons. Maybe there are different applications that are going to be built on the edge, on the devices, or the edge is going to complement the cloud. Those are 2 valid answers. And when you think about that particular case on the edge, like for example, one of the things that happened with us. As soon as ChatGPT was open, I talked to our GC and we put a policy out. Our engineers can't source code to ChatGPT and ask, like, check my code because once sent, it is gone. But if you see the conversation, a lot of the enterprises want to do those things locally on the device, depending on what the application you're using. And the last point is in the history of computing, there's always the competition between the vertical player and the open horizontal computing platform for everybody else. That was the story of Macs versus Windows, that's the story of iOS versus Android. That's going to be the story of AI. So there is an incredible, I think, platform and moat built by NVIDIA with their software stack and CUDA.. But every other company that I talked to, they're looking in what is going to be the open programming interface for AI for both training and inference. That creates opportunities. So with that context, now let me just translate into why are we excited about this at Qualcomm, why are we talking about this, and some of the things we're already demonstrating. First, from a technical standpoint, we've been preparing for this. So there was no surprise. We're the only company that could demonstrate stable diffusion running on a phone. Even if you put it in airplane mode...

Was it [indiscernible]?

Well, that particular model was 1 billion parameter. A chip that we're going to demonstrate and launch this year is over 10 billion parameter. Next year, in the phone, 20 billion parameter. In the PC, 20 billion parameter. In the car, we're talking about 60 billion parameter. So that -- so the first thing on the technical side is we have a very efficient -- we have GPU assets and we have CPU assets, but also we have an NPU asset. It's a very efficient way of doing it. Just to prove if you want to believe, it's the Qualcomm way of doing inference unique, and I'll give you 2 data points. One data point is the reason we won big in auto for autonomy is because if you put a server on the trunk of a car, you're taking away from the range. So you have to be very efficient. Your power envelope's very small. On the phone in particular, so that's where our mindset comes in when we think about developing those things because it has to scale from phone up. In the phone, if I add a lot of computation for everything you're doing, you have a copilot of every app, you're going to say I still want a full day of battery life, so you have to do things a little different. So that's why we have the performance per watt advantage. And what are the use cases? In phone, we're starting to see now since we demonstrated stable diffusion, a lot of our customers thinking, I'm going to build a different set of applications, user experience. And by the way, those are the ones that we know now. We -- who knows what's going to happen? But for example, photograph. You take a photo, let's say, I take a photo of this room. And I just -- with text to image, say, change the background to Tibet. And we think comes over the new picture, unique picture you share. How you do text-to-image, image-to-image, image-to-video, all of those things. Easy to see how consumers -- but also some of the things that you see happening with enterprise application copilots. Let's shift to PC. We made an announcement with build. All this conversation about hybrid AI, Microsoft is talking about it. And we can run ChatGPT locally. It's a smaller version, locally. In our SoC, they were doing to enter the next-generation PC with Microsoft with an ARM CPU. So when you think about that, that's a confidence builder on our PC plan, right? The Microsoft talking about the ability to have all those AI experience. The car, it's just the beginning of the conversations but they're very exciting, because natural language is a very good user interface for behind the wheel. All of those things that you have to go in and I want to search or navigate, all of those things, it completely changes. I like to think as a teenager, I watch Night Rider kits, and I said, wow, we can actually, finally get this thing done. But I think the message here is there's a lot of opportunities for AI for us.

Got it. Should we go to the lightning round? We got about 5 minutes left. See what the audience has to say. Let's see if we have questions. We have a few, yes. You talk about some of the technical hurdles or work you've done to make Windows PC and ARM a reality. And maybe you could talk a little maybe more broadly about what NUVIA is bringing to the table?

Very good. Look, we knew this is going to be a journey. And we're playing -- that's a typical Qualcomm, I think we're playing for success here, and we knew that it's going to take multiple steps to climb that hill. So that's why we started early. You saw a number of different devices. We knew it will be a small volume, but consider them as going through overcoming the barriers to entry. So one of the big one was would the emulator that has provided us part of Windows run all the applications? We come a long way. Windows 11 now can run both. Used to run just 32-bit apps, 64-bit apps. We did something very unique which we're doing at Qualcomm. We're actually putting a lot of the hardwares in the cloud. The developers can log in and they be able to upload their application, see how it works. They do monkey tests. They can see how their native application works. Microsoft, we now have the A team of Microsoft working with a lot of the enterprise applications. For the past 12 months I've been on calls with CEOs of enterprise applications, working them, is it ready, when is going to be ready. So we'll be building this. And the proof of that is that we have a lot of commercial designs now across Dell, across HP, across Lenovo, Acer, Asus. Because the enterprise is really did real test, whether you can you can have the ability to do that. So it has been the journey. I think we're getting the end of it as we launch in '24.

Got it.

You want to -- You asked a question about NUVIA. .

Yes.

I think NUVIA it's -- we're aspiring to build really the Apple compete platform, I think, for the Windows ecosystem. So we know what the target is, and I think we're doing a good job. We're going to announce the specs of the chip very soon in our Tech Summit, and hopefully, we'll impress everyone.

Anything you're willing to say on the ARM lawsuit, or?

Look, it's unfortunately that [indiscernible] we also don't. But look, at the end of the day, we're their largest customer, and they're suing their largest customer. We have broad rights to the IP. We have an architecture license, which also predates NUVIA, but it's okay. The trial date is set for 2024 after we launch our SoC and we're working through it.

Got it. You might not like this question, I'm going to ask it anyway, so.

That's okay. I'll like it.

What gives you confidence that Apple won't find a way to challenge the licensing fees that they pay once [indiscernible] in-house?

Okay. So I will -- we get this question all the time. Licensing and chips are not the same thing. So here's what I'm going to tell you. If you look at the structure of the market right now, especially in markets that IP is very protected and relevant, it's a 2-horse race. There's an iPhone and there's a Galaxy. We just renew Samsung at the same terms, 2030. And basically, I cannot think of a better benchmark. That's just one data point. Second point is when we had our skirmishes, there's a lot of patent litigation. I think both companies decided to let those things continue, and some went all the way to Supreme Court and they've been validated as infringing patents. So if you think about Qualcomm is -- has a number of very relevant IP. I think as I said before, I think we have a great relationship with Apple. I think we expect things to continue as they are on the licensing front.

Got it. And we've got 30 seconds left. I will give you your self box. You've got a room full of investors here. Why should they buy Qualcomm stock today?

Look. If you're looking at Qualcomm for a quarter after quarter, we're going to a cycle on handsets. Handsets, we're always going to be focused on handsets, but there -- we've done a great job on diversifying. Now the known handset, which has much higher growth rates, it's 30% of the revenue. I was joking with you, if we lose the Apple mode, then you will be able to see how material I think the diversification is up, right? Yes. The second thing is when you think about the 5 to 10 years horizon, our IP is very relevant. If you look at our IP today, we have leading IP in connectivity. Everything is connected, both seller and Wi-Fi. We have leading CPU, leading GPU, leading AI. And with just the beginning of those things, we had unlocked the ability to take our technology to other end markets. I think the company is executing. And what I like about it, everything that I told you, whether it's the PC, whether it's the XR, industrial, you -- this is not a PowerPoint presentation. You can measure the progress. It's very easy to see. Probably those things take a little bit of time, but we're just on our way executing on our strategy.

Got it. I think we'll leave it there. Thank you so much.

Thank you so much. Really appreciate it. Thank you.