Hello everyone and welcome to AR Insider Innovators. I'm your host Mike Boland, chief analyst of Artillery Intelligence and editor of AR Insider. And we're here today at Snapdragon Summit 2026. Special thanks to Qualcomm for inviting us and hosting us here at the event. We're having a great time, learning a lot. I'm here with Ziad Asghar, who is... let me see if I can get this right. SVP and GM of XR, wearables and personal AI. Did I have that right? Thanks for joining us.
You got that right and thank you for having me.
Yeah. Absolutely. So, let's get into it. We're going to talk a lot about what you talked about on stage, a lot of the themes here at the show, personal AI being a big one. But before we do that, let's sort of step back for context. Our audience is very familiar with Qualcomm as being very much a leader in XR, in the processing layer in XR devices. But to sort of start us off, tell us about Qualcomm's sort of positioning, some of its driving philosophy these days. Any figures you can give us to give us a sense of operational scale in terms of your installed base. Of course the foundation being a lot in smartphones, but really moving a lot into lots of other devices. You know, wearables, hearables, touchables, seeables, vehicles and of course XR. So, give us the sort of installed base and then we'll dive one level deeper.
Yeah, I think Qualcomm provides the technologies for all these areas and you know, we are very, you know, we have a great position on the smartphone side. The great part is the technologies that we are building, we are able to scale them. We have this notion of what we call one technology roadmap. It allows us to be able to create a technology and then really optimize, you know, and really make it more customized for each of these different areas. So, to give you an idea on the smartphone side, I mean, all the Samsung Galaxy products that you see out there, premium Samsung Galaxy out there, they are actually based on our product lines, like the Snapdragon 8 Elite. Of course, we launched two of those products over here for the first time. If you look at automotive, we are pretty much in all the cockpit designs that are coming out at this point in time. And of course, doing a lot more on the ADAS side as well. On my side, on the products like XR wearables and personal AI, I think there is amazing momentum at this point in time. Just to give you an idea, year-on-year we see almost three times the growth happening on smart glasses. And the way I think about the space is that basically on the very highest end, you have VST and OST based devices. And you know, we just announced the OST based product with XR Orion with Google. So, that's coming up very quickly over here. Valve just launched their product. Really a very different work stream frame, very different swim lane over there for our product, I'd say. And then we really expect a lot more to be happening in this short time frame over here. Then if you really look at it, the smart glasses have become a completely new form factor. And there is so much excitement around it because it's this device that's basically sees what you see, it hears what you hear, and it's able to bring in all this amazing context for AI like we talked about today. But, what it can really do is, as we continue to build the technologies, you know, we have multiple product lines, we actually can do glasses with displays.
And as you do glasses with displays and you bring in perception, and now you're actually able to do devices that are fully augmented reality-like devices. And you can see recently, just last week, Snap, of course, announced specs, also based on Snapdragon technology, very unique architecture, very unique device, but amazing capability in there. So, we are really covering the full gamut. And then if you go lower, you know, we have wearables like, you know, smart watches and pens and pendants and of course the hearable product lines like what we announced today, the Snapdragon Sound Elite Gen 2. So, we cover the full gamut of these product lines.
You certainly do. And on the XR side, one of the trends that we are observing, and I think Qualcomm very much is driving this with sort of purpose-built chips for all of those different sort of variants and variants as in device classes you just named because of course they have different compute needs when you're talking about something like a Snap specs versus Steam Frame. So, I think you guys are very much leading that in terms of, you know, powering those devices. But stepping back once again for context, one of the things that I look at as an analyst is sort of these historical trends. And leading up to this sort of discussion of personal AI, if we look at the past maybe decade and more so in the past half decade, Qualcomm has very much diversified and broadened its base. Like we talked about smartphones have always been really the stronghold. But you look at some of these macro trends where global smartphone penetration is reaching saturation points. The replacement cycles are getting longer. You know, you have Apple started making its own chips. And I think that from my observation, Qualcomm saw all these macro factors, saw the writing on the wall, and really was compelled to lean into these changing times, diversify, and really spread itself across all of those things that we talked about, all of those device classes and now just this constellation of devices. And you know, either unintentionally or intentionally, I think that now primes Qualcomm so well for this sort of new North Star, which is powering the age of agentic AI and personal AI because of course these are things that require ingesting and processing data and context from all of these different sort of touch points. And not only the inputs and the processing, but these are also the user touch points where the agentic AI is being executed. So, I guess that's a lead in for personal AI. Tell us all about personal AI and feel free to correct any of my sort of the history that got us to this point.
I think the way to think about it is Qualcomm is unbelievably unique in the sense that we make solutions for products that are in the milliwatt range to kilowatt range. As you know, we recently announced our move into data center like products as well. So, this spread and the way that our technologies can scale gives this breadth is you cannot find that anywhere else in any other semiconductor company at this point in time. So, really very unique in terms of our investment in terms of the technology leadership that we have. But I think that diversification story is also because we can do it. Right? That's the other way to think about it. We have the technologies to be able to do that. We have the technology to go into data center, into auto, into IoT, into personal AI devices. And I think that was the theme today of what I covered at the summit. But I think that the compute platform continues to morph. Right? The compute platform it was the PC for the longest time, moved to the smartphone, and now with the agentic era it's a device that's basically light, it's on your body, it's capturing that context and is actually able to send back the information to you. Devices like glasses, see what you see, hear what you hear. I think that is what is going to be the new compute platform. And that is why we see so much traction and so much interest in these devices. I mean, we have companies that are not our traditional customers in many ways, but they're coming to these devices. But I think the message that also want to give to, you know, all your listeners is the guys who have worked on XR are amazingly well positioned for this transition, right? And they're very well positioned for this transition because along with these new devices, you need new human-machine interfaces. And what are those human-machine interfaces? You need to understand hand. Well, guess who's been doing hand tracking for the longest time? The VR crew, right? And then you need to understand head position, you need to understand 3D reconstruction. Those are going to be the ingredients for these new human-machine interfaces for these devices. So, I feel that this movement really in a way advantages the people who have been working on the side. But I just feel that we need to also think about what do people like to wear? They like to wear devices that are relatively light, that makes them look good. It is also a fashion accessory at the end of the day. And that's why I think there will be devices that are glasses, there are devices that are pendants and rings and camera ear buds like the ones that we showed today as the concept. There is a lot of potential in this space, but the expertise comes from that XR side. I mean, I used to lead the XR technologies at Qualcomm in 2018, 2019. So, I've been in this space and but I think the time is right now due to the advent of agent AI.
And you're right, all the folks in the XR camp have a lot to contribute to this movement. I think the sort of hot take in the world of XR is that, you know, AI sort of breathes new life into XR and it certainly does. But I think we don't give ourselves enough credit in the XR world that AI also needs XR for all the reasons you pointed out, but also I think that the best vantage point for all of these intelligent inputs and context we need, audio is one of them, but also visual input. You talked about that today and there's no better visual input as a training set for human knowledge than the vision input that is the human's perspective eye level. So, capturing that actual perspective is important, right? Yeah. Now, you said something interesting, too. And this has been a theme at the show, where we're entering the sort of age of AI. Or sorry, excuse me, the age of agentic AI. And, you know, Qualcomm in many ways is sort of skating to where the puck is going in that respect. Because as a society, we're not quite there yet, but it's very clearly moving in that direction. So, I think the way it was put yesterday by Cristiano in his keynote was that, you know, we're moving from the age of apps, right? Our phones are filled with apps to the age of assistants. So, we have AI all around us, but it's very sort of prompted. It's still sort of somewhat manual to agentic, where things are happening in the background. So, tell us how you are designing for that, skating to where the puck is going. You talked about a lot of the key functions that you're innovating today. Everything from sensing to memory, to NPU functions, to connectivity. Like, how is this playing out at the chip level?
See, at the chip level, these technologies, to be able to bring in the context, to be able to have these devices that are tiny with the smallest of batteries, but they need to last a whole day. It requires a complete rethink. You cannot take a product that might have been done, for example, for smartphones and try to make a device like a smart glass with it. And that's what we have learned over many years because we have been at the forefront of this. If you look at all the Meta Ray-Ban's like products out there today, they use Snapdragon AR, our product. If you look at, you know, watches, for example, the Galaxy Watch, it's based on our Snapdragon Wear Elite platform. If you look at the Samsung watch that was just announced and Shahram talked about it today also from Google, it's the Warby Parker and the Gentle Monster collaboration as well. Also based on Snapdragon AR. It requires a rethink at multiple levels. So, number one, you need to have different architecture, unique architecture inside the chip to be able to get to that very low power and that long battery life. Number two, it requires that each and every one of the IPs or technologies like camera and sensing capability and all of that is redesigned for being able to sip on power, to really do ambient sensing, not this active sort of a mode to be able to conserve power. And then the third thing is basically really having the latest and greatest transistor because that's another contributor in terms of being able to reduce power. That's at a chip level. But in terms of the technologies that are enabling, you know, forces behind these products, it's basically for example, the sensing hub. The sensing hub has the ability to be able to bring in all the cues, create a full contextual picture around the user and that context goes into AI just like you said, right? XR and AI are very symbiotic. And they each need each other, but AI needs that context that technologies that come from XR in many cases provide.
That's one. Secondly, then we have the camera capability. If you need to see, you need to give eyes to AI. Well, those eyes better be very good. You need to be able to see that menu that you might be holding in a restaurant in a very dark restaurant, but be able to do OCR on that menu. Well, that requires best-in-class camera technology and that's what we've been focused on building for the longest time, but we've optimized it for these very small devices to be able to conserve power. Then there's technology like on-device AI. So, today we talked about one-bit models. So, as you take a large model and you now change to a one-bit model, the model becomes much smaller. It takes much lesser space in memory. And as a result, we can now actually fit in more capability in lesser capable devices or smaller devices, right? That's kind of the promise of that. And we have two levels of here. Micro NPUs, we have our Hexagon processor which is for more processing. We can do a small language model. If you get a chance to see the demos, we're showing a multi-model model and that's why we talked about the visually impaired like example. You can bring in any object in front of the glasses and it will tell you exactly what it is. There are also the ability to do pristine audio and speech because that's how you interact with this device, right? You need to give, I guess, ears to AI and that's exactly what we are able to provide with this thing. And then similarly rendering capability as we move to devices with, like for example, glasses with displays, you need to be able to do that extremely low power consumption. So, all of these technologies need to come together and we've been building these for more than a decade.
Yeah. So, Qualcomm's always been about efficiency with compute and one of the opportunities it seems is the more constellation devices we're talking about and that's a function of the user adoption. But as this evolves, it seems that there's an opportunity to sort of achieve economies of scale so to speak in terms of compute where you're using and you're already doing this, shared memory. When you have connectivity between these devices and the sort of ability to do a lot of load balancing and it's almost like this sort of hive mind that happens amongst all of these devices so that the whole system performs optimally with the least amount of power consumption, heat, and all that other stuff. Is that fair to say?
Absolutely. Yeah, I think you're absolutely correct. So, I've talked about in the past what I call constellation of devices, right? Think of these ambient devices that are around the person. Number one what I expect is you will have more than one of these devices on the body of the person. So, the opportunity is actually much larger to your first question, right? Because typically a person has one smartphone. But I expect people to have smart glasses, smart watch, earbuds, possibly AR pin or pendant that they use in the enterprise context. So, we launched the reference design today for an AI pin as well. But the idea is that it might be a device that you use for example in the hospital system. But since you don't want to take that device with you at home, use it only when you're there. So, you see you will have multiple of these devices available at any given point in time on the body of the person, and that's the magnitude of the opportunity. Now, to come back to your kind of your question there, if you're able to use the best place to do AI, you will optimize this whole solution, right? Which means that let's say if a query needs very quick response, you do it on the glass.
