Today we're thrilled to be joined by Don McGuire, the SVP and CMO for Qualcomm. Don was recently named one of the Forbes most 50 influential CMOs and we're so happy to have you today. Great to see you, Don.
Great to be here. Thank you so much for having me.
Absolutely. So, we're at CES. I'm sure this is a huge event for Qualcomm for a variety of reasons. Talk to us about what you're seeing here at CES and what you hope to get accomplished.
Sure. I think for us as well as for CES in general, the show has sort of evolved. You know, we have that little COVID thing happen in the middle. But I think over time it has evolved and new things have taken center stage. I think right just prior to COVID and carrying over even to today, CES seems to be a lot about cars. And there seems to be a lot of automotive news and a lot of automotive presence at CES. And then add on top of that this sort of constant narrative and hype cycle around AI. I think those are two of the things. And then traditionally you kind of see the big TVs and the new displays and the new screens and that's always been a mainstay of CES. So, we see a little bit of that here today as well with the transparent displays. Pretty cool. So for us, the last couple years has really been about automotive as a centerpiece with a sprinkling of other things around. And so I would say, 80% automotive, 20% Gen AI is kind of our recipe this year. And we've leaned in on Gen AI across different product categories based on kind of what's being talked about here at the show. So there's like AI PC, which has been a topic here at the show. There's other things going on and so we've kind of leaned in across those different categories. So that's kind of CES 2024 for us.
Yeah. Well, let's unpack there. When you talk about the focus on auto, we can get at in a second, but also kind of the focus on Gen AI. You know, there was a time when CES was all about this is 4K TV, it's a bigger TV, this is a smaller recording device. And I feel like that hardware has kind of become somewhat commoditized because everybody has great screens, everyone has fast devices, and now it is much more about the software. And you kind of see that happen with the iPhone development cycle and the demand kind of waning. What are your thoughts on that overall?
Yeah, I think it's either been about other industries undergoing transformation. And that's either being driven by software, and like the case of auto, right? I don't think there's any other industry that's undergone in the last 5 years more transformation than automotive where cars are less about mechanics and MPG, and MPH and torque and thrust and all that stuff. It's more about digital capabilities, right? A docking station for my life. Exactly. Or a new living space, right? And so that has been amazing to watch and participate in. And we're a huge participant in that as well as a huge enabler of it. Yeah. So with our Snapdragon digital chassis solution. So that's been super cool to see. But you're right. I mean the software-defined vehicle is now what's driving the automotive industry. So software seems to be driving a lot of the innovation and transformation versus form factor. Right. Or even performance. In some ways performance has become commoditized unless it's in the context of something new. Right. So instead of talking about, for example, I'll give you a PC example or even another device where in the past what you looked for was CPU performance, for example, like how fast is this processor, how fast is my PC, and the ecosystem built a whole narrative for years around hey CPU CPU CPU. But then gamers came about and the conversation shifted in the gaming context to no no it's all about GPU, right? GPU GPU. Well now in advent of the AI PC which has been kicked off sort of last year. We announced our Snapdragon X Elite platform at our Snapdragon Summit in October and the capabilities of that platform. We're going to see that show up in PCs in June of this year and going forward. And then here at the show, the PC ecosystem is talking about the advent of AI PC. Microsoft's launching the Copilot key right on your keyboard etc. But now when a consumer walks in to a Best Buy to buy a new PC going forward 2024, 2025 and beyond, or an enterprise CIO is saying, 'How do I upgrade my hardware, my devices for my workforce,' it's less about CPU and probably even less about GPU and it's more about NPU, right? What is the AI performance going to be on this PC so I can take advantage of all these great Copilot experiences that Microsoft has built on top of Windows? So this shift is happening where the innovation, yes it's software driven but it's also what's happening within the device that's creating this new level of performance. And so that's interesting to watch.
Is the power of AI beholden in the actual local device or in the actual servers delivering the AI? Because like, you know, there's a tool like Pika, you know, which is text to video, or Runway where it takes forever but it won't forever, just like it used to take forever to get on dialup, right? But probably two years from now you'll just be able to type 'show me a video of a cat running around the roller coaster' and it'll just get it instantly. But is the hold up there the local device that you have or is it the servers that are delivering the AI functionality?
This is a great question because in order for AI to scale both from a technology perspective and a usability perspective, it's going to have to be hybrid, meaning not everything can happen in the cloud, right? In some cases, it takes too long. The workloads are too much, right? The taxing on the data center is too much. The economic model doesn't work, right? You can't buy enough cards to put into your data center. You can't overutilize water resources to cool your data centers. So, you've got a sustainability problem. You've got an economic model problem. And you have a data privacy and personalization problem. So a certain amount of computational work that drives AI and the ability to deliver an AI solution based on algorithms or large language, small language, medium language models is where is it going to happen contextually based on what is it that you want to do. So on-device AI, which has been on device for 10 years, AI has been running in the background, whether you know it or not on devices like smartphones for a long time, right? It's been making your photos better. It's making your music sound better. It's been doing things that you didn't probably know or realize that AI was actually doing that. And AI has been part of our Snapdragon platform for years. It was really Gen AI that thrust AI into the spotlight. It was like the press release for AI, so to speak. Packaged it for consumers. Made it novel and created the hype cycle. But unlike other hype cycles, it's not going to stop, right? It's not going to go up and then go down. It's here to stay and it's moving very fast. But in order for it to scale, the technology that enables it is going to have to normalize across the spectrum of the device itself, the edge, which is really on-prem. Right. So whether that prem is your home network or is your campus network or is your office network. 5G plays a role in that. Wi-Fi 6E, 7, 8 whatever plays a role in that. And then the cloud and the data center in the big cloud. So if you're talking about something like 'hey AI assistant help me plan my day,' most of that can be handled on device and that's instantaneous. It's information, it's interaction that you want at low latency. It's text. It might be some image things like that. But we're talking about trillions of parameters like 'help me solve world hunger AI,' right? That is data center, big computing. It's going to take some time. The latency is not an issue. It's really about getting to the right answer. So it's everything in between there. So it's going to be contextual to what is the task. And if you normalize the technology stack across device, edge and cloud, then it becomes more useful for everybody. You can actually load more tools and more experiences on top of that and you can put them in the bucket that they need to go into to deliver the best experience to the user for whatever that might be. And then that's going to enable the second part which is the usability and the relevance of AI really has to move from prompted and reactive to anticipatory and suggestive, right? If it's going to actually be useful and not just novel. But that means that the user has to be able to let the AI know more about them. Exactly. Right. But it's up to the user, right? Right. Especially in the case of AI at the edge, right? You're in control. The human element is always in control. So you allow it to learn as much as you want it to learn about you and then you allow it to suggest or recommend in the context of your private network or data set. Whether it's on device and it never leaves or whether it's in your home and never leaves or whether in the cloud protected or 'hey I don't care where this goes, train your algorithms on it, I'm cool,' right? So it's everywhere in between. So I think that's really the beauty, the excitement of it all but it doesn't just stop with ChatGPT going to the data center. Sometimes you get a response, sometimes you don't, sometimes it says come back later because it's taxing a resource because everything else has to catch up. And that's where Qualcomm really plays a big role is at that edge, right? So from the on-prem to on-device, it's our sweet spot, right? We've been doing this on device for years. We make kick-ass processors and semiconductors and chips. And we have this long history and DNA in wireless connectivity. Yeah. And Wi-Fi.
So that's how people know you, right, for the what?
Historically, yeah. I mean, you know, since I've been at Qualcomm, what seven and a half years now, we were already underway of this transformation and diversification of less about connectivity, still about connectivity and first class and world leading connectivity, but more about on top of that connectivity computing. Yeah. Right. And which is why, you know, now what we stand for is really intelligent computing everywhere. So the everywhere is the nod to the connectivity. The intelligent computing is really the nod to the AI plus the processing power together in a really energy efficient package. Right. Because that's another thing, right? You can't perform really large tasks, AI tasks especially on a device that has a battery life of a day and a half unless you want a battery life of two seconds, right? You know, so that again, that's why this thing has to be hybrid. One big balancing act. Exactly. And so it's an ecosystem play. Everyone's involved. And it's exciting. We have our role. Everybody has their role. No one's right, no one's wrong, right? It's not a us versus them. It's not a Qualcomm versus Nvidia. It's not a Microsoft versus somebody else. It's really about everybody has a role to play. Who is involved in developing technology? And then where do all the puzzle pieces fit? So, even the content creators like you see New York Times and their lawsuit, OpenAI. Yeah. Well, and technology is just technology until you put it to use. Right. Right. When it becomes commercialized and it gets a little murky. Well, and you know, anything new, there's always trial and error. Yeah. Sometimes bad things happen. Happen with YouTube. It happened with Spotify, Napster. I mean, goes back. It goes back. And you know, there are bad actors and you know, and good actors in every category. In every category. And I, in my speakership I did earlier today, I talked about, look, if we could just, because I'm not a huge fan of regulatory oversight in general, because it's too slow. Yeah. And by the time they get something regulated, it's already moved beyond the regulation. Totally. Right. And so it's irrelevant. If I can't keep up, how's the government gonna keep up, right? Especially with AI, it's moving so fast. And so, it's really about standards and governance and practices and best practices that can be adopted and principles. And if we can learn the lessons from social media where there was none of that, right? It kind of went untethered so to speak for years. And you got a lot of good out of it. Absolutely. It opened the world up to so many people, connected so many people, but it also created mental health issues or enabled the environment, geopolitical issues, bullying, you know, all these types of things. So if we could learn some lessons from maybe a little earlier in the social media technology evolution, we could have anticipated some of those things and said, 'Hey, you know what? We're going to have some principles.' I think everyone involved, right? It's not just, yeah, like, yeah. It's not just Meta. It's, you know, because you can't just blame one party or can't point a finger because guess what? You maybe nobody's out. You know, everyone likes to think they're altruistic. But at the end of the day, if you are advertising on that platform that's then responsible for enabling bad behavior, you're culpable. And Meta has shareholders and those shareholders could be endowments of colleges. It's like so they need to deliver return. Right. Venture funds. Exactly. Right. Right. And so it's like it's this vast web, right? And so it's like everybody has to kind of join together and just say, 'Okay, how are we going to operate this thing, this beast?' And I know that's not easy. It's easy to say, not easy to do. But you know we've joined a bunch of companies in co-developing some basic principles for deploying AI and along with Microsoft and others and ethical principles about how we're going to create the technology. I think those types of things, you know, you look at platforms like Reddit, which are kind of community-managed, self-governed, so to speak, Wikipedia, like Wikipedia, and it works in that context. Obviously, I'm not saying it'll work everywhere, but if you take some concept of self-governance and then you add some other things in to insurance policy, so to speak. I think we can get there and I think we can prevent some of the badness, right, that came along with social media, right? And every technological innovation have badness. There's car accidents, right? Still today, right? Right. It's funny like technology is funny. There's a lot of fear, right? And I know a lot of people are fearful of AI and it's been used as a narrative, right? Was used in the writer strike and the actor strike. You know, people fear job loss and all these other things. A lot of that's just posturing. Because it's like I use the analogy of the elevator, right? The elevator first was invented, nobody would ride it because they were scared. They were scared, right? They didn't want to plummet to their death, right? But then Otis came along, who didn't invent the elevator first, by the way. They invented the latch that caught the elevator, right? Once that was invented, people are like, 'Oh, it's safe, right? This thing's going to catch it, right?' And so they start riding the elevator, right? So it's like that is a natural type of part of process that we go through as human beings. And with any technology, so I think we're going to go through those same types of things with AI as well as other things. Autonomous driving that's another thing right that you know whether we get to full autonomy and how and what context. It's going to be very interesting to watch. And you're going to have people that point to 'oh see that accident' but it's like okay well how many accidents does humans create? Exactly right. 80% of accidents are human error, right? In some way, shape or form, right? So, and so will autonomy fix that? Will it cause more? I don't know, you know, but it's going to be an interesting, it's a wild ride. It's so fun to be in tech, you know, I've been a tech marketer most of my career. So, it's always interesting.
