Clicky

The Six Five Insider Edition with Qualcomm CEO Cristiano Amon on the Future of AI
by Daniel Newman | March 3, 2023

On this episode of The Six Five – Insider Edition, hosts Daniel Newman and Patrick Moorhead sit down with Qualcomm CEO Cristiano Amon.

Their discussion covers:

  • An overview of Qualcomm’s perspective on AI and utilization
  • The benefit of running AI at the edge
  • Qualcomm’s use case examples of the power of AI
  • The potential power of bringing AI to the PC
  • AI at the device, edge, and datacenter

Be sure to subscribe to The Six Five Webcast so you never miss an episode.

You can watch the full video here:

You can listen to the conversation here:

Disclaimer: The Six Five Insider Edition is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

Transcript:

Patrick Moorhead: Hi, this is Pat Moorhead, and the Six Five is live in the Qualcomm booth in Barcelona, Spain. We’re at Mobile World Congress 2023, and I have to say, if you can’t hear the excitement around us in this booth, maybe we’ve got the sound turned in the wrong way, but I got to tell you, Dan, 2023 Mobile World Congress is on fire.

Daniel Newman: Yeah, we’re back, baby. It’s been great. After a couple of those years off, and then we came back and we were wearing masks, and it was really frustrating, the energy’s good, but also, you can just tell with what’s going on in the economy, people are focused.

Patrick Moorhead: Yes.

Daniel Newman: They are here to do business. We’ve had some great conversations. We got another one right now.

Patrick Moorhead: That’s right. One of the biggest trends out there, biggest discussions in the industry is AI, and I want to introduce Cristiano Amon, CEO of Qualcomm, who has been in AI for 15 years. Cristiano, welcome back to the show.

Cristiano Amon: Very happy to be talking to both of you. Thank you for being here in the Qualcomm booth in MWC.

Patrick Moorhead: No, absolutely, and we know that you’re going rail to rail, meeting to meeting, talking to customers and the press. We really do appreciate that.

Daniel Newman: Absolutely. So, Qualcomm has been in AI a long time, and over the last few weeks, there is a ton of talks with the onset of ChatGPT. There’s been news and headlines, but you’ve been in the business for more than a decade, maybe even 15 years that Qualcomm’s been in AI. Give us just a little bit of a rundown of Qualcomm’s overall perspective on AI and utilization.

Cristiano Amon: Absolutely. Happy to, and finally, I think the opportunity for Qualcomm as an AI company is here right in front of us. You know what is interesting, this MWC is the one in our history that we have the largest number of announcements, so we have two new modems, one for 5G Advanced, one for X35, we have modems for FWA, we have infrastructure products, we have a suite of announcements, but the interesting thing, what’s getting more attention is what we’re doing in AI. So, last week, we showed stable diffusion, large language model with over 1 billion parameters running on a smartphone.

Patrick Moorhead: That’s connected, of course.

Cristiano Amon: Yeah, but if you look at the potential now, and I think ChatGPT event, it was a key milestone. As we think about those large language models, very complex models, you are going to take the compute from the data center into the devices, and the future of AI is hybrid and that’s where Qualcomm’s going to play a major role. So, most people look at Qualcomm as a communications company, I’ve been telling them that we’re not. We’re a connected processor company, but the reality is we’re going to become as much an AI company as a communication company. And what I like about this opportunity, just think about the ability to run ChatGPT locally in your phone, locally in your PC, and you can do that from as many queries as you want in conjunction with the data center. And for you to do that with the computational power, you need a different approach to AI, and that’s what we do. Performance per watt really matters, and especially if you need to bring the AI to the device, and I think that’s the opportunity for Qualcomm right now.

Patrick Moorhead: Yeah. So, Cristiano, I think many people might be unfamiliar with the benefit of running AI at the edge. Is there classic privacy, security, latency improvements?

Cristiano Amon: It’s all of the above, but there’s a lot of other application. Let me just walk you through some examples. So, let’s talk about PCs. As we build Windows on Snapdragon and we’re working with Microsoft for the transition of PCs to next generation PCs, if you look what Microsoft demonstrated on their Studio apps-

Patrick Moorhead: Yes.

Cristiano Amon: …you are having a Teams call and your dog’s barking, if somebody has a bag of chips, you have noise, AI will do noise cancellation, have just your voice. That is done in realtime. You don’t have time to go to the data center to go do that.

Patrick Moorhead: Right.

Cristiano Amon: How to track your eyes, how to pan or zoom your camera, those are some of the ones they demonstrated. There are many security applications. For example, we now use AI in each one of our technology building blocks. This X75 modem that we announced is the very first modem that AI is integrated. Your modem actually get better over time as you start using your cellular connectivity, it gets better over time as you learn about the networks and it learns about you.

And then I go all the way to, I think, this conversation of large language models in ChatGPT. If you’re going to have a conversation with AI and you want do that in real time, just look at everybody trying to experiment with ChatGPT in the data center, how it actually got slow. The future is hybrid. You’re going to have to run that in real time on the devices.

Daniel Newman: It actually makes me think about language translation. Here we are at a very global event, you just walk the halls, you hear hundreds of languages. Could your smart device be doing all that translation in real time? You could be wearing your wireless earbuds and you could be hearing these conversations, being able to engage with anybody in any part of the world, zero latency. And then I also think about, Cristiano, sustainability because the amount of power that it takes to run these workloads at the edge has to be significantly less. Is that a couple good examples as well?

Cristiano Amon: Look, absolutely. Let me just speak on translation.

Daniel Newman: Yeah.

Cristiano Amon: And I remember we demonstrated this for the first time using AI a couple years ago on a tech summit.

Daniel Newman: Yeah.

Cristiano Amon: But it’s here now. You can see a video and AI will automatically put subtitles for you. You can do real time translation. And I want to talk about the power thing because the power, it’s an interesting conversation. In one hand, as you think about those large language models, billion parameters, 10 to 20 billion parameters and beyond, you need a lot of computation of power. That happening at the same time the companies are making Net Zero commitments-

Patrick Moorhead: Right.

Cristiano Amon: …and they need to reduce the total power consumption. That actually creates an opportunity for Qualcomm to be on both sides, not only on the device but in the data center for AI, and I’ll give a practical example of that’s already happening right now. It’s another thing about Qualcomm that nobody knows about our position in AI. We have been successful in the automotive segment right now. It’s been one of the new growth areas for Qualcomm, and one of the applications of the Snapdragon digital chassis is assisted driving ADAS and autonomy.

Patrick Moorhead: Yeah.

Cristiano Amon: All car companies for ADAS want to do reprocessing of that data in the data center. Maybe your car made a better decision than his car, and he wanted to improve the algos over time, so they’re asking to replicate this solution from the car in the data center, and we’re going into the data center with our AI processes as well. It’s a whole new world. It’s a great opportunity for AI with Qualcomm, and maybe that’s where the company’s going next.

Patrick Moorhead: Yes, Cristiano, we talked all the way from AI in modems, to AI in smartphones, even on imagery that we’ve seen for years that you’ve shared in the camera. I’m super excited about the PC because, quite frankly, the PC really needs a wake-up call, especially on the Windows side, and bringing generative AI capabilities to the PC, I think, is going to transform the experience even in little ways, even finding something you want on your files, being able to do that real time, video magic that you talked about. But one thing I’d really love to see is Qualcomm on the edge, the data center edge, being able to crank out workloads at peak efficiency. There’s one thing, after covering your company for almost 15 years, is when it comes to efficiency of doing things, whether it’s raw compute, AI, whatever workload, once your engineers get a target, you go for it, and more times than not, you’re successful at it.

Cristiano Amon: Look, I like to talk about this because our engineer, when we design some of our technology blocks, we have a mindset that we expect that’s going to be for a battery-powered device, so therefore power, it’s not separated from performance, and it’s becoming quite interesting how you see that comes into results for the company. That’s why when we got into the automotive industry, it was a no-brainer because we have such a disruptive performance per watt in an electrical car. How much electricity you’re going to consume, it matters. You won’t put a server in a trunk of a car, that’s not going to work.

But the interesting thing is this conversation about running large AI models, when we show our demo stable diffusion, one thing that, or my favorite thing to read, there was one tech analyst, I won’t say who it is, but there’s one tech analyst that basically said, Well, Qualcomm did this remarkable, and then here’s the reason why. You can run the same model on a GPU card that goes into desktops, you can imagine who that company is. And then he said, Yes, Qualcomm did it in 15 seconds. That particular GPU that I pointed to, we’re doing half the time, but Qualcomm did with 200 times less power. So now, when you think about-

Daniel Newman: There you go.

Patrick Moorhead: Yeah.

Cristiano Amon: … the potential of bringing AI into PCs and you are going to do it for many things, those things matter, and that’s the reason we’re so excited about this transition in PC. It’s not about building a new hardware.

Patrick Moorhead: Right.

Cristiano Amon: It’s actually about building a new user experience.

Patrick Moorhead: Yeah.

Daniel Newman: So, I’m going to ask you probably the question that’s on the minds of the market. We’re hearing a lot of names. You hear about Microsoft, Google, there’s an arms race. Nvidia kind of gets a lot of credit, there’s an answer, I can say their name. So, as saying they’re going to be the foundation of generative AI, but Qualcomm has a part to play in this. I’d love to hear how you’re telling that story to the market about Qualcomm’s opportunity in generative AI.

Cristiano Amon: Look, I will basically describe this in a very simple way. You have one question to answer, right? The one question to answer. Do you believe the future of AI is everything runs in the data center, or you believe in a hybrid compute model or a hybrid AI model that you have to run in a data center, the edge cloud and the device. If the answer is the hybrid compute model, there’s no way you’re not going to see Qualcomm playing in that space, and we know what to do in devices and performance per watt for high performance computing in those models that runs not only a billion parameters but more. And you are probably going to see us next year showing models running locally with 10 to 20 billion parameters, and I think that’s what we can do at the edge.

Patrick Moorhead: It’s exciting stuff here. Cristiano, we’re going to give you the last word here. Anything we haven’t asked you that we should have asked you that you want to get out to this investor and technology crowd here?

Cristiano Amon: Look, I think we have been very busy diversifying the company. One of the things that is really happening with us, the technologies that we develop for phones, and it’s a mistake to think that it’s a phone chip going into a car, it’s a phone chip going to the data center-

Patrick Moorhead: Oh, you’ve proven that was not the case.

Cristiano Amon: We’ve had the ability to scale our technology to so many industries, and all of a sudden, we found ourselves in a situation that-

Daniel Newman: Yeah.

Cristiano Amon: … the single roadmap that we developed for mobile, it’s finding its way in so many industries, and I think there’s a lot of interesting disruptions, large language models, hybrid AI, merging of physical digital spaces, convergence of phones and PCs, and the car becoming a computer on wheels. All of those things are creating great opportunities for Qualcomm.

Patrick Moorhead: Cristiano, want to thank you for coming on the Six Five again. I think this is his third time.

Daniel Newman: Third or fourth, yeah.

Patrick Moorhead: Third or fourth time.

Daniel Newman: He’s always a very generous guest with his time.

Patrick Moorhead: I know. We really appreciate that.

Cristiano Amon: And I’m looking forward to the next one.

Daniel Newman: And I’m glad you answered the question though because, like I said, I think the world really is fascinated by what’s going on in generative AI, but the whole picture, to his point, is that it’s not all going to happen in the cloud or in the data center.

Patrick Moorhead: And listen-

Daniel Newman: That’s big.

Patrick Moorhead: I’m not aware of anybody who doesn’t believe that hybrid multi-cloud is the part of the future, and I think sometimes the industry gets in its camps of what can happen and what it doesn’t, but I think it’s ready. I think it’s time for the next big disruption out there. AI on the edge, the deep edge, in devices is going to be a thing, whether that’s smartphones, cars, PCs. You have elements of it in the carrier edge, you have elements of it in the hyperscaler data center, you have elements of it in the on-prem data center, so, Cristiano, can’t wait to see what you have up your sleeve.

Cristiano Amon: Very good.

Patrick Moorhead: Thank you.

Cristiano Amon: Good talking to both of you and looking forward to the next time.

Patrick Moorhead: Definitely.

Daniel Newman: All right everybody. There you have it. Thanks so much for coming here. We are in the Qualcomm booth. This is the Six Five at Mobile World Congress, 2023, in Barcelona. Hit that subscribe button, check out all the episodes that we did here at the event, and then of course all of our shows all the time. From Patrick and myself, though, we got to go. See you later.

Patrick Moorhead: Take care.

About the Author

Daniel Newman is the Chief Analyst of Futurum Research and the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise. Read Full Bio