Good afternoon, everyone. I'm Nicholas Johnson. I'm the publisher of Axios. Thank you for joining us again this year on the promenade at Davos. We've been here a couple of years. It's packed, it's standing room only in here. There's a line down the street. The hottest club in Davos is here at Qualcomm House. So very grateful for our partners. Nothing that we do at Davos would be possible without the support of Qualcomm. They've been with us from the very beginning. They're hosting us in their awesome house today. Super grateful for all the support they've given us to make these incredible conversations possible. And my favorite part about this is that means I get to hang out with, I think, the coolest person to hang out with in Davos. He's the CMO and SVP of Qualcomm, Don McGuire, who knows about all the coolest stuff coming down the pipeline. Okay, Don, let's talk about the future. I want up here. Great to see you. We've been having this is like following Matt and then that intro. Wow, I have a lot to live up to, man. Well, let's see. Let's try. Been a rolling conversation, I feel, about AI for the last couple of years. So let's like let's catch up. How's it been going? So like how has it been changing?
First of all, I want to return the compliment to the Axios crew. What an amazing partner you've been. We're so lucky to have you as part of the Qualcomm House experience. So thank you for being the sessions have been amazing. Don't you guys agree that Axios programming is top-notch, right? So congratulations to Axios. Thank you, Don. But yeah, I mean, wow, there's a lot going on. I mean, I was just listening to Matt and all the things that he was talking about from an OpenAI perspective. So and then we kind of have a sort of not a different perspective, but we have a we play in a different part of the value chain. So you know, kind of I think we bring a unique point of view in our own way to things. But and then as a CMO, right, where AI is, you know, marketing and sales, one of the areas where Gen AI, right, is is quickly impacting how we work, our efficiency, our PR productivity, how we create. So I have a lot of perspective just from an enterprise CMO, you know, like so so so where to begin? Like maybe go a little bit deeper on the the specific theory of the case, the angle that I know Qualcomm has on like how AI is evolving.
Well, I mean, look, I I think one of the areas because where Qualcomm's sweet spot is is at the edge, right? So historically, the the hardware and software that we have developed across our Snapdragon portfolio of chips, as well as outside of the Snapdragon product family, is is really about creating intelligent computing everywhere. And in cases where people can, you know, use it on device and have all that computing power in the palm of your hand and or on-prem, right? Whether it's in your home or on your campus or in your institution of learning, whatever that might be, without necessarily having to go to cloud. Because as as Matt, you know, talked about how tech, you know, from a technology perspective, AI is developing so fast and it's changing and computers can do more and more than they couldn't do two months ago. That same thing is happening at the chip level. Like our chips can do more and more. We can we can handle 100 billion right models on device that just a year ago was like 30. So it's it's it's happening everywhere. This this fast space innovation. So we play at the edge, and we bring, I think, a unique value proposition to how AI will scale ultimately at the individual level and at the business level, because not everything needs to go to the cloud and not everything can go to the cloud. When and when you bring that technology out of the cloud and put it in your pocket, then tell me how that will change the way you interact with the device if the AI is on it and you don't need to go ping something else, right?
Well, I think there's two things. One thing I think we should talk about is, you know, there's there is a scenario where AI is the new UI, right? So we can talk about that. But I think, look, there's so many benefits to on-device or on-prem AI. One is privacy, right? And safety. So you you control your inputs, how things are trained, and and as you develop and use the tool, whether it's in the ideation process, the creation process, whatever, you don't have to worry about whatever that IP is going somewhere where you don't want it to go, being used to train someone else's model, and then all of a sudden, you know, your voice is somewhere you don't want it to be, your art is somewhere you don't want it to be, your knowledge, your code is somewhere where you you don't want it to be, right? If if a 100 billion parameters can be handled on device, right? Unless you're trying to solve, you know, Einstein's theory of relativity, you can do a lot on device. It doesn't have to go to the cloud. So safety and security and privacy. One, latency, right? For quick decisions. And I'll use automotive as a case, right? As AI infiltrates the car more and more, you can't wait for it to go to the cloud and 20 minutes later receive a response because you've run over the pedestrian, right? It has to be milliseconds. Decision-making has to be super fast. So we we can enable that to happen on your car or in your car without having to go to the cloud. So and that's important for how AI could be ultimately because we think in your car could be a really amazing use case for how you interact with AI because there's a lot of downtime when you're driving, right? There's a lot of you're thinking, you're driving, and you could be having a conversation. I I 100% believe in what Matt said about both agentic as well as multimodal. Again, in order for AI to be truly useful.
Well, I mean, we currently live in a world of apps, right? It's a linear experience that you're in, you're out, you're in or out, and they don't sort of cross-pollinate, right? So I think we believe that, you know, there's going to be this layer, this agent or several agents, quite frankly, that, you know, going back to what Matt said, where we are today versus where we're going to be tomorrow, that you can cross-pollinate tasks and things that you want to do, whether you want to have your schedule interact with your bank account, interact with an e-commerce site where you want to buy something, you want to pay a bill. It you shouldn't have to go in and out of your banking app and then go in and out of the merch site and then go in and out of Shopify and then go in and out of your, you know, oh, do I have enough money in my and back and forth and back and forth and back and forth. You should just be able to have your agent tap into that layer and just complete whatever you wanted to complete. This is an element, I think, of what we started to talk about last year, which was big, the agent AI type thing. And so it's almost even a more comprehensive thing now where I'm not downloading a bunch of apps. I'm downloading I have one thing that will interact with all that, and when you bring the computing power onto the device, that's where that's possible, right? And how it ultimately manifests itself is not really up to us, right? We're going to be the enabler layer, right? But but it could be one or multiple. He mentioned taking a walk at the end of the day or that example where you're having a conversation and then it's sort of taking notes and it's organizing your things and everything. Well, today you're probably looking at your phone and maybe making sure you don't run into something, and you're going into this app because you remembered this, and then you're going out of it, then you're going into the next app and coming out of it. We believe that that won't have to be the case anymore.
I mean, AI is obviously going to be really horizontal. We, you know, we power the Meta Ray-Bans obviously with our Snapdragon XR and AR technology. The fact that we they've already added kind of voice-based AI to the Meta Ray-Ban experience. I think just in one generation, the Meta Ray-Bans have just come so far, super useful, less creepy. And and I also think that once the visual part of that comes into play, as you're walking down the street and you're talking to either multimodal again, you're talking to the agent and then you're looking at something, it sees what you're seeing to Matt's point, and then it's populating information visually. Like that's going to be an amazing and that's a processing power to go from text to audio to full frame of vision. Yeah. Well, you need the processing power to power the camera. You need the processing power to power your audio. Everything needs to be on that little chip that goes into those devices, right?
