QUALCOMM Incorporated (QCOM) Earnings Call Transcript & Summary

Okay. Good afternoon, everyone. Thanks for joining us this afternoon for the Rosenblatt Age of AI Scaling Conference. We're fortunate to have Ziad Asghar, Qualcomm's Vice President of Product Management; and also Mauricio Lopez-Hodoyan, Qualcomm's Vice President of Investor Relations are with us today. Ziad will be presenting today. He leads the Snapdragon road map planning and application processor technologies, and this is covering all smartphone platform products. Ziad drives the definition of products, ensuring that Qualcomm products lead in technology and best-in-class user experiences while making the trade-offs between features, power, performance and cost. He leads application processor technologies, including artificial intelligence, camera, graphics, CPU, audio, video and security, quite a few features he works on. He has more than 20 years of experience in the wireless semiconductor industry and has held a broad set of leadership positions from R&D to product management. Everyone is well aware of Qualcomm's 5G leadership. And we've asked Ziad to discuss how Qualcomm's AI development, along with the 5G, is changing the industry and expanding Qualcomm's markets well beyond the smartphones. So with that, I'll hand it over to Ziad. Thank you.

Thank you. Thank you for having me. Yes, I'd like to kind of give you guys an overview. There's an amazing amount of excitement from Qualcomm perspective and, really, from an industry's perspective with AI. The cool thing about AI, or the amazing thing, at least from a technology perspective, continues to be that every use case that we have of our end products, we feel becomes better. It becomes a lot more capable, it becomes enhanced in many ways, and then it opens up completely new opportunities for our product lines that were not possible in the past. So if you look at it and take any example, if you take the example of smartphone, for example, we think with artificial intelligence, we really gave it the ability to be able to really understand what is happening around it, to be able to comprehend and absorb the information, the stimuli around it and actually be able to perceive and then reason based on that data and actually be able to act on it too. So just some examples that come to mind are the ability for the smartphone now to be able to completely take what we have in front of it and now be able to make sense of it. So in the past, our smartphones, our connected cameras or other devices that we work on, would absorb this light but they would never know what that light means. Well, now we have this amazing ability that the smartphone camera can actually comprehend and understand what it's looking at. And that completely changes and opens up so many of the revenues, right? With this, you have the ability, for example to -- take the example of security camera. Well, you can now know what is in front of the security camera. I mean, I'll take an example. I have a security camera at home. It's not as smart and it basically sends me an alert every time somebody passes by. Well, with the application of AI, not only are you able to determine if it is -- whether it's a smartphone or if it's a safe person or if it's a car, it's also able to actually determine the face and recognize the face and say, well, this is a person that lives in this household or not, and be able to make those kinds of decisions which, possibly, even able to give you access to the home, depending on whether you determine that somebody is a safe person or not. So that's just one example of how AI is really opening up a lot of opportunities for us. We, of course, look at AI as something that not just touches the end case or end use cases, but it also touches and improves all the different technologies within the product. What I mean by that is it improves, for example, the camera. It improves the security aspect. It improves the gaming aspect. It improves audio, speech. So anything that you can sense and touch, we can markedly improve those use cases in a very, very big way by the application of AI. We've been working on AI technology for more than a decade. So we started our research in the -- our corporate R&D division more than 10 years ago. Since then, we have done various acquisitions. We've continued to invest in it because we've always looked at this as this horizontal enabling technology that makes our products and technologies much, much better. So we are applying it to those technologies. We are also applying it to the way we actually develop our silicon products. So the actual process of actually improving the way we put together our ICs, integrated circuits, is something that we're actually applying AI to also. So lots of different levels at which we actually see AI being applied to our products. We have -- what we have been really focused on at Qualcomm is the amazing ability to be able to do more processing at the lowest power possible, which means that we are actually able to process technology, we are actually able to do it at much lower power than anybody else. That advantage, actually, which we have learned, which has been our pedigree, which is in our DNA for the longest time, from a smartphone product perspective, is actually now taking us into many new markets that we are starting to enter. For example, automotive, for example, XR, for example, cloud, AI, all of those markets, when you apply this unique advantage that Qualcomm brings to them, it really sets us apart compared to many of the other vendors that are out there. In addition to that, we're, of course, working on a full stack software, hardware and all the different parts, such that we are able to develop and create these differentiated products that I'm talking about. I can spend some time talking about the applications within smartphone. And just to be clear, Kevin, how much time do I have? So I'm not taking longer or lesser than with the time that I have.

Well, the session goes until 4:45 -- 1:45 your time, and we're going to let you speak to get your point across, and then we'll open it up for questions. But feel free to make sure you get your point across.

Sounds good. Sounds good. All right. So basically, I'll go by kind of the overarching picture that we have in our mind from an AI perspective. So as you know, 5G is something that we just launched. We're very well differentiated from a 5G perspective. The way we're looking at it is that 5G and AI are kind of symbiotic technologies at this point in time, which means 5G makes AI better, and AI makes 5G better. What do I mean by that? What I mean by that is that imagine that you have multiple different devices: your smartphone, your car, your smart watch, all of them that have a certain degree of AI in it. With the advent of 5G, you have a technology that's very, very high-throughput and at the same time, very low latency. And what that allows you to do is to be able to now access the intelligence that's spread out all over the network to be able to get access to that because of the low latency of 5G. So now you are getting to a paradigm, which is a paradigm of just completely AI that's permeating through the whole network, which is a very big differentiation for us because of 5G. The other point I'd like to make is the device continues to be the right place to be able to do a lot of AI process. What is the reason for that? Well, we have heard a lot about privacy issues and all. Well, as we are putting more and more AI on the device, you're actually able to do that processing on the device. And for those aspects that don't concern privacy and all, we can leverage the distributed intelligence concept to be able to go beyond the device to be able to tap into more intelligence than that. So that's kind of the grand vision. We think we are making very good progress on it. We have products that -- attached to all of those different parts of the distributed intelligence network that I talked about. But I think the part that I really want to convey to you guys is because of the scale of Qualcomm, because of the fact that we have these millions and hundreds of millions of devices that we ship, the AI that Qualcomm can enable is much more available to everyone. And really, the smartphone is absolutely the best platform for doing AI, well, because it has all these sensors that are able to see, that are able to hear, that are able to perceive the presence of a person, and then we can work on AI on top of that. So let me maybe spend a little bit of time talking about what have we done on the [ FI ] already. If you're using a smartphone from Qualcomm, you are already using AI technology. So one great example where we started applying it was on audio and speech. So what you can do is essentially apply AI to be able to get the signal or get the voice out of the most noisiest of rooms with the application of AI. You can look into the darkest of rooms with your camera and be able to take a picture, again, with the application of AI. But as we're going forward, what we're able to do is that within a given frame, as you're looking at me right now, you can actually distinguish whether there is a certain part that's the skin or hair or cloth. And then based on that, we can process each of those segments within that frame in a way that makes the picture and image quality far better than what we were able to do in the past. And now what you can see is that we then extend this idea into video. So the processing that you would have done at one frame level, now you do it 30 or 60 frames per second, and that allows you to create these new experiences that I'm talking about. At the same time, what you can do is, as we go further out in time and we talk about virtual reality and augmented reality devices, AI actually enables a lot of those use cases because as you're looking at the world through, let's say, a pair of glasses and you're augmenting certain things that you're looking at, well, with AI, you would have to be able to figure out what exactly you're looking at. You have to do things like plane or to be able to clearly and accurately do what we call tracking the positioning of your hands. Because if you are in a world where you're interacting with the world around you by touching things, well, you need to know what the position of your hand is, another part that AI technology enables you to do. So not only is it enhancing the product lines and businesses that we have, it's actually enabling the newer and upcoming businesses that we are focused on developing. So really a very, very powerful technology, or example of technology that allows us to do all of those things. And then to extend that further, as you go from IVI or infotainment on the automotive markets and you move into ADAS, where you're doing driver assist systems and then into autonomy, these applications, this core R&D, this core investment that we have done from the perspective of mobile, we now have this unique ability to be able to take a lot of that learning, apply it to those markets, and that allows us and gives us a very good running start in a lot of those markets to really be able to make progress very, very quickly. Because many of the problems, they are analogous or similar to what we have already solved on the mobile side. So again, a very unique advantage that Qualcomm is able to bring. From the actual architecture or from the actual technology enablement perspective, we focus on developing. As you know, in the smartphone space, we do accelerators. We do an accelerator for audio, we do accelerator for videos because that is how you get the power to the right point. That's how you make a device that people use so extensively last the whole day. So that's exactly what we are trying to do and taking that pedigree, that advantage into these new spaces, which means we can do, number one, a lot of AI processing. But the unique advantage of Qualcomm, we can do that processing at the lowest power possible. So what we have done is we have created this engine that we call our artificial intelligence engine, and that constitutes an accelerator that basically have what you can call scalar, vector and tensor processors. And if you move into the detail of that, it basically maps to an exact neural network, which is exactly how you should design that. Not only have we designed it, it's already into our products. Our last smartphone products, to give you an idea, had about 26 trillion operations per second capability. We just enhanced it, and now it can basically process 32 trillion operations per second. That's a massive amount of AI processing capability at an extremely low power. And that's really the advantage that we are able to take into robotics, into all other areas that we are talking about. At the same time, we develop a full software stack that goes with AI. As you know, one of the big challenges of AI is to be able to do software that's able to scale, and that's exactly what we have done. We have a software that's basically gives you access at a very high level. We support all the frameworks like TensorFlow, PyTorch. But at the same time, we have a Snapdragon Neural Processing SDK at the highest level, which abstracts a lot of the details, allows our partners to be able to very quickly start to use AI on our devices. But for those partners who are more capable, we actually give them access at a lower level as well, much closer to the hardware so they can get a lot more out of their products. So again, software that's very well differentiated and is able to access multiple different levels of performance. So that's kind of the software stack. And then, overall, the way I continue to look at the space as we go into, they are assuming new areas that this is opening up for us. So one of the areas that we have talked about is things like smart retail or things like smart cities even. Because if you just look around, many of you who might have traveled around in China, the sheer number of cameras that are out there at any given intersection here or in many countries of the world, it is quite impossible for a person to be sitting there and processing all of that information, right? But what you can do with products like what Qualcomm is offering, you can actually have an ability to be able to look at the video stream, make sense and determine if an event has happened. For example, you are able to tell if 2 cars have come very close to each other, signifying that an accident might have happened. And then an alert can be raised. So just the unique ability of AI, coupled with Qualcomm's strengths of superb connectivity and low power, is opening up avenues and use cases that are -- that were otherwise not possible at all [indiscernible]. And maybe one of the key other things to add is, I talked earlier about, as you know, our leading 5G modem technology. But the key part is as you apply AI even to modem technology, what you're able to do is to access and to be able to get the signal information in the most complex of channel conditions where others would not be able to. So we're actually applying AI to all these technologies, giving a very unique advantage to Qualcomm across the board. And I'll take maybe a little bit of time also on the PC market. So PC market is really something that's right for the application of AI. I mean I'm sure many of you see this on a daily basis. How many times do we type something and say, "Well, let's meet up to discuss this"? With AI processing, those meetings can be created automatically. With AI processing, things like being able to summarize a long e-mail, to be able to do natural language processing, all of those things can be done uniquely and can be done at very, very low power in the case of Qualcomm technology. And maybe the other aspect which gives me the good feelings, which is basically AI for Good. I think there are applications of AI, which we've already worked on as Qualcomm. For example, we had an engagement in India, where you can actually take your smartphone and put a small lens on top of it. And what you can do with that is to be able to actually look inside your eye and be able to diagnose conditions like diabetic retinopathy, for example. And this might be in a location where there isn't as much health care available. But by some of these technologies, you're actually able to diagnose or do a course diagnosis. And then with time, of course, you can have that person see an expert and then be able to make sense of it. So really quite a lot that we've been working on. There's a long runway coming up ahead where we combine a lot of these technologies and going to a lot of new products. And I always make this example that I shared with you guys already on augmented reality, but you can just envision what we can do with this, right? You walk into a country that you don't speak the language. Augmented reality, you can pretty much take all the signs where there's any text and we're able to translate that into English. So as you walk around, you're able to just navigate your way as you'd be able to do over here. With augmented reality, the gaming experience completely can be changed as you're doing it through glasses, for example. You can actually change the plot of the game. You can change the textures in the game. You can really do some pretty amazing things that I think make the experiences for consumers and for industry significantly better than what is available today. The metric that I think we should be really focused on from an AI perspective is something that we cover across the board. It's basically looking at performance for a given amount of power. We call it performance per watt. And I think in that metric, Qualcomm continues to really shine, and that's again because of all the work that we have done in the past. We recently launched also our robotics platform, again 5G and AI capable. And really, one of the cool things that we are quite happy about, the Ingenuity chopper that's there on Mars actually uses Qualcomm technology. So something quite exciting, something that we worked on is actually sitting on another planet, but that's, again, because of the very low power and amazing amount of processing that we're able to pack into our small products. I'm hoping this has given some high-level overview, and we can probably move to some questions.

Thanks, Ziad. Yes, it's a good overview. And just a few questions I have. Maybe just first talk about the Block Diagram of the Snapdragon. I went to a presentation a few years ago, probably quite a few years ago, but where someone from Qualcomm, maybe a product engineer, said that you see Snapdragon as an apartment building, and all the different features, audio, speech, camera, like you say, are all different apartments in the chip. Where does the AI sit? Is it in each one of those apartments, or is it one chip in another apartment that can work on all the other features.

That's a great question. I think it's a good way to look at it, too. So number one, within itself, it's a big apartment inside that building. But because of its unique advantage, some of the things that I covered, there are small pieces of AI or small pieces of AI in other apartments also now. Because what you can do, like I explained, you can make security better. So you would have some help with security subsystem. At the same time, you can leverage the very large AI apartment with the large AI block to be able to access it. But for certain unique cases, where you may need to do processing at very, very even lower power, you would have a very small block in a particular engine. So for example, we are able to do a lot of all these on use cases, Kevin. And what you can do with those use cases is that even when the phone is not active, it's able to actually sense and detect and improve certain things. And what we're able to do now is to bring in all the data like video data, audio data, speech data, location data to create a complete contextual picture of what is happening around us. And by doing that, what you can do is now you can enable the use cases. For example, the phone can assert and if a person is driving, for example, and perhaps you should not be doing certain things while you're driving, or if you hear a baby crying at a certain point and late in the night, all those kinds of things are what we're enabling. So it's really both a rather large apartment inside the building and then smaller pieces in the other apartments too.

Okay. Maybe if I can probably ask a question that you can't say exactly, but as an idea, what percentage increase does -- I guess, how much bigger does [indiscernible] get when you add in the AI functions?

Yes. I think the way you should think about it is that you need AI to be able to do certain functions that other blocks might have been doing, right? So what it allows you to do now is that, actually, as you put in this AI capability, you can actually reduce some of the other blocks area as well, selectively, right? Because there are functions that now you can run it into this AI block. So I think if you look at it in its entirety, if we had to actually implement the same capabilities that we're doing in the AI block, which our consumers and our customers are asking us for, it would be a much larger increase. Actually, AI is able to do a lot of these things in a very efficient manner. So I would think of it this way as it is actually doing it more efficiently than if we use traditional approaches just like traditional signal processing in some cases.

Okay. [Operator Instructions] So with that, you had mentioned about a PC and some of the uses for AI and PCs. And we see Qualcomm having quite a bit of an opportunity with Windows 11, having the operating system being able to run natively on ARM. Can you talk about how Windows 11 could change and allow you to bring in some of these features?

Yes, absolutely, right. You can already see what Snapdragon -- Windows in Snapdragon is able to do. You don't really need to carry your charger anymore. You can be connected at any point in time wherever you are, and that's a very unique advantage. I think many of us have fiddled around, we're trying to find a WiFi connection at an airport many a times. Well, you can actually take care of a lot of that. But one of the key use cases that was shown even at the launch, like this example that we have going on right now. If I look at you, it seems like I'm not looking at the camera anymore, whereas -- so what was shown at the launch of the product was what we call gaze correction. You can actually take your face and actually adjust the eyes such that it would seem like I'm looking at the camera even though I might be looking at you. So that's just another example of, especially with video conferencing, the fact that you can basically do on device some of the capabilities that in the past would have had to be sent to the cloud and be done on the device. You can dictate a complete e-mail. You can basically translate something natively on the device completely, so you don't have to send any of your information off from the device. You can do e-mail generation. You can do e-mail summarization. There's a lot of new use cases that will be coming from a productivity perspective from the perspective of audio and speech, from the perspective of video conferencing, all that we think, with the unique advantage that Qualcomm has, we can bring very quickly to the market, and we're already working with our partners to do that.

Okay. Great. And maybe along those lines, too, the XR, and I know Cristiano, in his Mobile World Congress keynote, he said watch for XR developments. And maybe can you give us a little more hint on what you're doing with that? The augmented reality that you discussed is very interesting, that you can translate signs. But it would be also nice to have translate voice. But just some of the features you can bring out with XR?

Yes. I think, today, if you are actually traveling as many a times, you have to [ load ] your phone. Excuse me, I have a plane side over here. But basically, you have to hold your phone to be able to produce some of the augmented reality use cases. It's kind of cumbersome. But you can imagine that now if you are just viewing the world through those augmented reality glasses, well, you can do actually a lot more. I mean just imagine milestone recognition, for example. You're working anywhere, and it's able to tell you information about something. It's able to show you the path for navigation. It's able to bring in the data from your smart watch, for example, to show you. And the cool part over there would be that you don't really input your data into the device by using a keyboard anymore. What you do is you talk to the device. So now you need to be able to understand what the person is saying. So that's a natural language processing. That's what we call it from an AI perspective, to not just understand a word or a letter or understand a sound, but be able to comprehend what that sentence means. And you can imagine that with AR, you would basically talk to your device, and it's able to do whatever you're looking for, or to be able to point at something using your hand, all of which are AI problems, or to be able to take out the noise -- take out all the noise in a very noisy environment, again, something that's very unique AI problem. So things like image recognition, things like speech, things like audio, all that are really very well aligned with augmented reality are things that would be using AI. But specifically, if you look at the devices today, a lot of which are virtual reality devices. Qualcomm products are actually one of the most used products in those products -- in those end products, if you notice. So it's really the unique advantages that we've been building for VR, which we think many of them will extend into AR. Of course, there are challenges that we will be working on to solve, which means if you have a fully contained device, it would require a lot of power consumption. And that, again, like I pointed out, is one of the unique key advantages that Qualcomm always brings. So in the beginning, you should think about it that probably you would have augmented reality less when they are tethered to a phone. Because some of the heavy lifting is happening on the phone as a device. But as time goes by, in the longer-term context, we think it will become a unique, fully contained device, of course, that means connectivity is absolutely important in that form factor as well, along with AI, along with processing at very low power.

Yes. And along those lines with the virtual reality, it's been a couple of years now since you get to a conference and try out some of the latest technologies. But one thing I always thought was lacking was the sound, that you might have to view and be able to look at, but the sound doesn't match what your eyes are seeing, you get busy. Is there a -- does Qualcomm have that? You clearly have audio.

I think if you -- Kevin, if you look at the latest devices, actually, you will find that the experience is just pristine, especially some of the newer ones that have come out from Oculus. I think it's public information, those use our product. And what you see over there is that you will see, very early on, like you're pointing out, maybe a couple of years ago, I think when -- what your brain expects, if it doesn't get that, you feel a little bit unwell. But the current product lines that you have, and that was a problem with some early products that people did, but at least on all our products, you will see pristine quality now. And I think Quest, Quest 2 are some products that are public, very amazing experience and a very good response from the consumers in terms of what they are able to do.

Okay, great. Can't wait to be able to travel freely again.