Welcome to Fort Knox. I'm here with Cristiano Amon, the CEO of Qualcomm. I am John Fort, as I always am. And you've got some announcements here leading up to CES, a consumer electronics show, quite a few of them, but the headline seems to be this full stack robotics platform. Even though we don't think necessarily of self-driving as being robotics, it really is. So, I guess this isn't a brand new thing for you.
No, look, we're incredibly excited in general, John. I think it's very good to see what's happening with the company. I think the strategy that we put in place to diversify, to grow, we can actually see the excitement. There are so many different things we have been doing. And I think what you saw earlier in this year, there was a lot of conversations about what we're doing in data center and now robotics. This is a massive opportunity. As you said it correctly, you mentioned about cars being physical AI, but the robot is very clear. Nobody will dispute that is an edge AI platform, and I think that's right into our wheelhouse. We have the right to win. And at CES, we're announcing not only an expansion of a roadmap for robotics for some new customer engagement and the opportunity to go after this massive market opportunity with a full stack solution.
Why is the opportunity so massive? My sense is that robotics has been relatively splintered up to this point. You've had some consumer companies going back to Ibo and before, and then the attempts in robotic vacuums, iRobot and whatnot, that have kind of flamed out to some degree, but nothing, no platform that's been as organized and robust as we've seen in technologies like the PC, the smartphone, and arguably the car.
I think the fundamental technology difference is physical AI. I think the same thing that we see, you know, AI fundamentally change, for example, I think I love the example you provided, will happen with self-driving car and assisted driving as we went from traditional machine learning and you saw how the combination of that with AI and some generative AI systems, how you change the concept of training task, how you have a self-learning device. That technology is fundamental, combined with now having a high performance processing system for those opportunities. And to answer your question, the opportunity is massive. We're really talking about, you know, there are a number of analysis showing by 2040 it's a trillion dollar opportunity. But if you unpack that, it has consumer applications, you have a lot of industrial applications, and then you have humanoid form factors. And the other way to look at this, just think about healthcare, retail, warehousing, manufacturing. Today, if you look at a warehousing, if you go to an Amazon distribution center, you see a lot of robots moving boxes and packages. And think about that doing at scale in retail, think about that in healthcare. So the opportunity is significant, and now you have both the processing technology and the physical AI to make that happen. And I think for a company like Qualcomm that is sitting on the edge, it's an incredible opportunity. The same exact reason that we were successful in automotive is what we believe is going to be success in robotics.
So break down for me the elements of the stack and your strategic approach.
Yes. So let's go back to some of the conversations we used to have years ago when we talk about what we're going to do with the Snapdragon digital chassis and automotive. You could not put a server in the trunk of a car, especially an electric car. You could not put a server. You needed a high density of computing. You needed to integrate a number of different sensors. You needed to have different type of computing, from computer vision to CPUs to GPUs to neuroprocessing units, and you had the ability to have something that meets the safety and industrial standards. And I think that's why there is no dispute at this point that our approach was successful in the automotive industry. You get the automotive problem statement, which was a physical AI, and now let's just go take that to a humanoid robot. Everything that was true from the car is even more true for the robot. The robot needs a lot of battery life, you need to be incredibly efficient in power, you need to integrate a lot of sensors. The robot has cameras, accelerometers, a number of different sensors that need to be integrated, and you need to have the scale of consumer electronics, which I think our DNA comes from the phone space. You know, one thing is for you to do a $20,000 robot. The other thing is, can you make the robot a $5,000 robot or even cheaper? Can you make the battery instead of lasting an hour or two hours, can it last for about eight hours? So that is the silicon expertise of Qualcomm. We actually build a dedicated platform for robotics, and at CES we announced the IQ10, which is probably the most advanced robotics platform that can go all the way to supporting humanoid robots. On top of it, we develop the software stack that you can apply to different elements, including the integration of the sensors, how you have the central computing and the brain of the robot, and the training mechanism and the ability to have the flywheel of AI data, which is also part of our solution. It is very different than training a large language model for example in the cloud. When you think, for example, an industrial use case, you have to train the robot on the particular task that you want the robot to do, and you perfect on that task. So it's a very well-defined training problem, and it's the same approach, for example, that we use when we develop our own stack, the training mechanism that you have on driving cars on the roads.
So what's the ecosystem approach here? I mean, with cars you already have a built-in legacy of OEMs, of drivers, their habits by country, by region, etc. In robots, you're looking to enter spaces where there are tasks that either humans have been doing or tasks that haven't been getting done at all. To what degree do you need to seed the innovation space? Do you need to even make some of these robots yourself in order to accelerate the market to the point where it's available to you at the level you want it to be?
Look, this is a great question. And I think that's what makes this incredibly exciting because first of all, you have established robotics companies. We are now in evaluation with some of the largest industrial robotics companies that actually you see a lot of robots in manufacturing. And I think when you apply AI to some robots, you make them mobile like AMRs. Think about AMR with arms. Those companies' roadmap, they see a lot of opportunity for expansion within that space. And you already have an ecosystem at scale that you see in large industrial robotics companies. The second part of it is what you're starting to see happening with some of those new companies. They're building humanoids and they're building their own training environment. One of the incredible things we're excited about this CES announcement of the IQ10 is Figure AI, a company that has done a lot of advancements in humanoid robots, just decided they are going to design IQ10 into their humanoid robots. And you have companies like Vin Group in Vietnam. Their Motion 2 robot is now also being designed with Qualcomm Dragon Wing. And so those are new ecosystems that are forming. And then you have a multitude of consumer robot opportunities that just get by enabling AI in other devices. So we also seen an interesting leverage from the ecosystem of industrial PCs. Companies like Advantech, for example, working with us and building platforms, and even some companies interested in developing development kits like arms that you can do a development kit on top of the platform to facilitate training and deployment. It's a new opportunity that is starting to expand. I think you see there's a lot of activity into that, and I think we're very happy to be entering this space at this stage, right before the big growth curve.
