QUALCOMM Incorporated (QCOM) Earnings Call Transcript & Summary

[Audio Gap] the provider of the phone, it was the same. And at that time, they had -- there was a Symbian app story, you have a BlackBerry app store, but most of the experience was used. Then the smartphone arrives. The smartphone arrive. The big deal was there that was a computer. But when it arrive, the existing company said, oh, I can do an OS, too and there was a BlackBerry OS and Nokia didn't go [indiscernible] and they said, look, those in apps, they're there. I can recreate those apps. Pretty quickly. the 10 apps became 100 apps, 1,000 apps, hundreds of thousands of apps, and that was the experience. The experience was the app. It wasn't the actual UI of the phone, that's what changed. So when we think about AI on phones, and it's important to understand that AI is on phones, different AIs on PC, different AIs is on [ glasses ], different AI on cars, different experience, but especially on phones. Back in 2018, that feels like 6 years ago, we put the slide when it was celebrated 5G, and we said every generation of wireless has one thing that happened with it. When we said 4G was 4G and the computer in the palm of your hand, it was mobile computing. We said 5G is going to be AI. And it was very clear when you -- we didn't have a very good understanding how AI would evolve, especially a Gen AI model. But it's now clear to us that Gen AI will change the equipment computer [ interaction ]. So let me give you an example. Every single app could theoretically apply an AI front end, whether they get the AI from an open source, whether they're [ Shlama ], whether they get to OpenAI with -- from Google on the front end. Think about this, like something as simple of you go to your banking app that is based on APIs that go into the bank cloud to get information. You could ask AI. The AI get the same information. It could eventually, even if you want to render to you like your app renders. So what happens is those platforms are changing very fast. That's the reason we did AI Hub. The reason we did AI Hub, which is applicable not only to phones, to cars, to everything...

This is your collection of models, by the way?

We have now over 2,000 developers -- actually, for you to understand how it works, you could have a model that gets quantize and get optimized to run an NPU. You could get that model, I'll patch it to your app, push that to the App Store and you can start your app. You, can have them auto interface. And from that, the model can do many things. So the point that I'm going to make to you is, we're starting to see now the sink better. We see a few use cases. You saw like translate. Is this difficult to search. You see phone editing, you remove somebody from a picture, add somebody from a picture. But I will say, we are just at the beginning. And I'd like to provide the following example just to finish that answer. We have seen historically, over generations, there are more you can do with your phone. Not less. How much more you can do with your phone. And that's why we're increasing so much content and computing content into the phone. So there are things today that you can argue, I have to answer this very long e-mail right now. It's very complicated. I'm going to wait until I get to my laptop. But with the Copilot, you can probably start doing certain productivities on your phone that you could not do at older ones with this different type of interface that it can help you get things done, create an image or something you'd be able to do on a phone. So I think AI on the phone is going to be probably as natural as [indiscernible] or how apps have evolved, and we're just at the beginning of the transition right now.

How long do you think it takes that we see -- that's the unknown at this point?

That's the unknown. But I would say, here's the things that I like. The trend is going to the way we want it to be. We started to see all the flagships now with AI capabilities. We started to see some of the use cases being developed by the OEM. They're in development by the ecosystem players, to develop by the OS players. And we're starting to see a lot of developers get interested in bringing an AI front end to their apps. So it's reasonable to say within the next 5 years, this could be how are we going to interact with our phones and it's going to feel natural to us.

Is it fair to say, you guys aren't trying to monetize the AI Hub?

Look, right now, no. The AI Hub, it's a unique thing that we have developed. I think there's nobody that has anything like that or the edge. And it basically what it is, is you can bring your own model, it gets quantiles, optimized, available. We actually build even a device cloud. You can -- as a developer, you can test and hone the target device in the cloud is actually hosted in AWS and you can deploy that. And we just launched. We have over 2,000 developers. So that's exciting. It shows how much activity is for people thinking about a front-end AI for the app.

At least it must give you some visibility, too, in terms of what's going on with...

Absolutely. Absolutely.

Make sense. Let's talk about PCs as long as we're on the AI device topic. And so, this is interesting. I was making the point earlier like, first, I want to try to disaggregate, we talk about ARM PCs and AI PCs. And I think the AI PC is what's being used to sell the ARM PC narrative to the end customer. That's my -- maybe I'm wrong, but would you that's true or...

Partially. You're not entirely incorrect, but there is something fundamental happening, which exactly it's the combination of transition of a PC to a different architecture, happening at the same time that there's a transition of Windows. So I think those are 2 distinct things that happen to be at the same time. And I think at the end of the day, to be very direct and honest, I think the true AI PC, and Microsoft put together is with the Copilot+ and it happens on the Vulcan architecture. But I have a lot of things to tell you about this. It's exactly important to break this down. So I'll start in reverse order. I'll start talking about Windows before I talk about what we do and what we don't. That's my personal opinion. I'm sure Microsoft is not saying that, but I'm saying that I saw some verticals in the press about this. I think this is a big transition of Windows. It may be as significant as Windows 95 type transition. And the Copilot+ is this AI that is very pervasive. It runs all the time. And you started to see what happens when you run all the time. There's a tendency right now to think about AI in the cloud. And what do you do on devices just to get the stuff on the cloud you're doing on a device is different. They actually -- they're developing different -- the use case, the application is different. So I'm going to give you just some examples of the use cases announced by Microsoft when I talk about the Copilot+. One is recall. And think about this, you're indexing -- you're running. You're running this computing engine and you're indexing everything that happened on that screen, every pixel. And then it just tells the machine, I saw a YouTube video like this for this content a week ago. I say you saw. I went to a web page, and this is what you saw. I saw a document that has this from Stacy Raskin, they have this particular short saying Qualcomm is awesome. This is what you saw. So you -- so this thing is a different use case. Now let's shift it to another use case of PCs, which is a communication device. Teams. The ability to actually add some titles at any language to somebody that's participating and talking in a language you don't understand. This is like -- it's a new use case. This didn't exist before on something that you do on PCs. What I like most is the whole concept of co-creation, for example what they did with photo and paint. And you have a lot of company. If you do an IDH, so think about this, you're creating an image, you're doing photo editing, you're doing video editing, creating something new. You create something. You said, I don't like it. I'm going to do it again. I don't like it, I'm going to do it again. I don't like it. Every time we run this 100 -- multiple, hundred million perimeter in the cloud, who is paying that bill? But once you do it on the device, it's a whole different story. So I think there is this big change in how the operations system is going to work. And now I'm going to go from that to what we're doing. I think when we went on a mission, and our mission had 3 different vectors. I'm glad that we got to this point. I think it's showing that we executed on. One mission was the desire of Microsoft to change to -- Windows into an ARM instruction set. And we went on a journey with them. We knew we had to launch multiple products with them to get to this point, to get the ecosystem ready, applications to be noted, the emulator. It's...

You've been making chips for Microsoft laptops for -- since 2016?

We had a journey. And remember, it was not on Windows 11 that you could run Win 32 apps on the emulation. We've got to get the emulation going, you. Have to get all of the ecosystem. That's one vector. The other vector is we needed to bring the performance leadership to the Windows ecosystem. And I think we did that. And it's good to see a company like Qualcomm that we're telling investors, you look at us as a comms company, and we're telling you that we're actually a connected computing company. I think when we have the performance leadership in PC, is like graduation day. We -- that's kind of where we are. That's the second vector. The number 3 vector is the transition of windows to AI. And there is a reason, for example, when you saw the Microsoft Copilot+ launch in 22 devices from Lenovo, Dell, HP, Acer, ASUS, everybody, and. The entire service line all on X Elite. There was a reason for it. The reason is very simple. You won't run on anything else. But when you think about this AI running, and I think there's a lot of confusion in this space, the CPU is busy, has a job to do. The GPU is busy, but now you have this other computing that is running all the time. When the Copilot+ run on Qualcomm, it use the entire battery life of the device, it consumes about 20, 30 minutes. If you run on the best that you can buy from the incumbent, the fastest you can buy from the incumbent today, 5 to 6 hours. So that's kind of the different conversation is that's the reason I think our initial launch, you only see Qualcomm because it simply won't run unless they have a completely different equation. So we like that. We think it's -- AI becomes a tailwind for diversification into the PC space. And because we're new to this, we don't have any share. It's just a great expansion of the model.

Let me continue that. I mean, I understand the current products from the incumbents don't really support the study here. You've got new products that are coming that will, in theory, support it. I'm starting to get a little bit of maybe why consumers would want an AI PC. Is it a stretch to have to go and convince them to buy the ARM-powered PC versus the x86 because it's a new and weird and different maybe -- like what is it about the ARM-powered -- the ARM architecture in particular that you think is superior to x86 like -- assuming like similar AI?

So that's the reason I think when you first used that argument, I said, you're partially correct. And here's the part that I agree with you. If you have to go to the point of sale, and you tell the consumer, this PC does everything the other PC does, everything, but the instruction set is actually running on an ARM, okay? then I say what's the real value proposition. However, when you say this PC changed. This PC is the Copilot+ PC. It's a different experience altogether. On top of it, it has 22 hours of battery life, is a very thin form factor. Then it changes the conversation exactly to the point that we are right now, which is the incumbents are saying, I will have that, too. Wait for a little bit. I will have that. And I think that is what's really changing the conversation. We see that with Best Buy, we see that with [ Courage ], we see all the retail, this is something new, is this new AI-powered PC, this new experience provided by Microsoft. And some new experience from different developers, we don't know where they are yet.

Do you think other players in the ARM space see this is good or bad for you?

Fantastic. I'll tell you the reason why. First of all, when other players coming in with an ARM-based CPU for the PC space, the first thing they validate the SAM, that's an ARM SAM. Done, right? That's -- it changes the conversation. The second thing for us is, look, if you look at where we live, the neighborhood that we live, right? Phone is probably incredibly competitive in the concentrated environment. When we enter into auto, everybody said, look at who's there, everybody is there, Mobileye is there, NVIDIA is there. You're going to go into auto. How are you going to be able to compete. And I think for Qualcomm, this is good. I think we feel good about our DNA, which is about competing in very -- markets that move very fast. And we like our differentiation. It's just performance for watts. That's what we do best.

Got it. I guess, is the first like big use case coming out for smartphones is that -- are we all waiting for Apple now? Is that -- are we waiting to see if they do something with that?

Look, I think one thing that will be great for the industry, if they come in with great use cases or a platform that a lot of -- in an engine that a lot of developers are going to come up with a new experience. This is going to have a positive [indiscernible] and actually will accelerate the AI smartphone integration.

And I guess on the PC side, they've taken share since they started introducing their own ARM-powered PCs. They're creating an experience.

The question that you have to ask is, what is the difference between the Snapdragon X Elite for the Windows ecosystem is to restore the performance leadership from the Windows ecosystem. And it's just -- and so AI is a new thing, which is very unique and differentiated from Microsoft. And then what Snapdragon X Elite is the fastest CPU and is the best battery life. And I think that is what the competitiveness that now the Windows ecosystem has and got back to the front of the line.

What would you consider success in the PC? Like the PC markets, 250 million units, give or take. What would be success? Is it 20%? Is it -- is there a 50 million ARM PC, you get half? And like I don't know how to size it exactly.

Okay. We're not making share assumptions, but I can give you like a proxy answer, if I may. First of all, the reason this is great for the Qualcomm model is because we don't have any share, right? We just starting, right? So everything is gain. Second thing, when you think about our model and phones, most of the contribution in terms of gross margin and operating margin comes from the high and premium tier. A PC product for -- and we build a portfolio, they go from $600 and above, laptop, that's like a premium SoC, is equivalent to a premium smartphone.

No matter what you're selling there.

So it has some of the attributes from a financial standpoint that you see on our model from selling a premium smartphone. So think about this as like this is like better than expansion of the premium theory and enhance it.

So the margins on the PC chip should be higher on average than the margins in the handset business?

Look, we don't making a margin statement. I think you should assume that the content is much richer. ASPs are higher given the performance. And that's kind of our sweet spot in the [indiscernible] portfolio. The other part of the answer to the question. The more the -- just put yourself in Microsoft position, right? So what we like about it, if you look at historically in the PC side, the -- whether you have the strong hold of x86 with the enterprise and the enterprise will be very slow to move. Let's say, you launch Windows 10. Enterprise said, I'm going to stay on Windows 7 a little longer. You launch Windows 11, the enterprise is going to stay with Windows 10 a little longer. There's something that is different now, especially when the Microsoft Copilot, the Microsoft revenue model for the Copilot, Microsoft is driving a transition to Windows 11 to AI PCs, that is helping us having a bigger addressable market first. And even if you model, if we have 10% share, it's a significant, I think, upside for Qualcomm. And we feel like we're well positioned, not only because we have the best performance chip, but we're the only 1 of the [ 8 ] that can run Copilot+. We like our chance.

Got it. Got it. I don't really want to ask about Apple, but I'd be remiss unfortunately if I didn't ask about Apple.

Go ahead.

So what is the current baseline assumption now? It's the iPhone that launches at '26 is when you start to go away?

Whatever we said. We have been very transparent with Apple in our model. So when we extended and renew the agreements we highlighted, we expect that by '26 we start to go away. And that's the assumption right now.

Got it. Any risk to the royalty business if and when that were to happen? Or does that -- we assume that stays?

Not really. I think there's a tendency, I think, of folks to connect the chip and the royalty business, even though I think the court decided that even if you were together, it was okay for Qualcomm even though we don't do that. And I think it's basically loyalty is independent of whether Qualcomm chip is utilized or not. I think we feel really good about a situation that we have to litigate our patents, and we had good outcomes. So it was not like you even know what are the IP is already been determined. We feel the licensing business probably at a stable point in its history, right?

And is it fair to say that like in the near term, your content on Apple seems to be going up not down. I know -- again, I don't know how much you want to talk about what they're doing. But clearly, there, you have got an RF competitor of yours that seems a lot [indiscernible] you guys. Maybe more generally, how do I think about what sort of the integration between the RF modem is actually doing? Because I think that's a trend that that's helping like beyond whatever is going on at Apple.

Look, without speaking on anything on Apple, I think we had -- I think we have developed a competitive RF portfolio. I think from a performance standpoint, I think we have been differentiated. And we have seen, I think, a good interesting across the industry for premium RF components.

Let's talk about auto, and the broader diversification theme here. So you just recently took your pipeline numbers up, right, I think from $30 billion to $45 billion.

$30 billion to $45 billion.

What does pipeline mean?

It means -- the beauty of auto is when you get -- you get a design win and you have -- in that design...

These are confirmed designs. I mean, this is not like we're in engaged. We're trying to...

No, no, no. It's a little bit more than that. Those are confirmed to something wins with an associated agreement that has -- basically, you have very well-defined for the particular model, the content over the years, over the volume of their model. And I think one thing that is good is because we're designed with virtually every car company right now, we have the ability to see the overall market if you're all calibrated. So -- and I think the reason we have been disclosed in the pipeline is because it's very stable. And that's why we -- how the pipeline converts to revenue, we have been delivering on that on a quarterly basis, regardless of how the market changes up and down, for us is a story of new models with our content is a share story. And we're tracking to the statements we provide, $4 billion and $9 billion. And it was just according plan. Is that actually a very predictable and reliable business.

Did those targets already imply what the pipeline was going to extend because you didn't raise the targets when you raised the pipeline.

[Indiscernible] this.

How far does the pipeline go out, by the way? How many years?

Look, right now, we have -- we're participating in RFPs for cars in the 2027 time frame. So a lot of the design wins for cars launch in '25, '26 and '27.

Got it. What kind of things actually drove the $15 billion increase?

So one, there's a new disclosure we provide, and some of it was already in the 30, but there was a big percentage going forward. One new disclosure we provided up to $45 billion, we said 1/3 of that happens to be the increase, but even the component of that in the $30 billion before, 1/3 of that is ADAS and autonomy. And I think that drove a lot more content, especially for AI processing. It's all our high-performance NPU, is like server class performant NPU. And I think we -- our solution has been resonating. We had a very flexible architecture. Sometimes our own stacks, sometimes they're part of a stack. And we have displaced some incumbents that had a very strong position into that space, especially for autonomy. And he speaks to one of the things that have been resonating well with the automotive industry. One is the fact that we have capabilities in all domains. The fact that we actually have the ability to scale the technology up and down from a premium car to an entry-level car, especially because safety and electronic comes now table stakes will be in those cars. And then it's the power [indiscernible]. When you put a server type solution in the trunk of a car plus cooling, how it consumes energy versus what we do, it's really, I think...