Qualcomm Keynote: The Era of Personal AI and Endless Realities - AWE USA 2026

Good morning, everyone. Wasn't that a great device specs that Evan just shared? Really, really exciting times, by the way. Look, I think Ori and now Evan have set this up amazingly well. I have a lot to share with you today. I think there are a couple of trends that are unbelievably amazing at this point in time. The devices that you all have been working on for the longest time, being able to understand what the environment around a device is about, to be able to reason based on that, and then to be able to act. That's the vision of a personal AI device. That's the vision of what people call a spatial AI device. And that's what we all have been working on for the longest time. But there's an amazing ingredient that has come in recently, which is AI. And you guys can see with the advent of AI agents, it's really completely changing how we use these devices. But this group of people are the best positioned to be able to take advantage of this revolution. So we all lived in the world where mobile phone has been at the center of it. It is the center of all the information. It's the center of all and everything that we do from productivity to the activities that we do at home. But guess what? We're moving towards an agent-centric world, where agent becomes at the center of all the activities that we do. I really want to say that 2026 is really the year of the agents. We have multiple different types of agents. We have agents that are great for working for what we do at home. There are agents that are great for what we do at work. There are agents that are great for medical, for health care, for education, for all those different verticals. And I believe that as we build these unique devices, we actually bring these agents to these devices and we give them eyes and ears. We give them the ability to be able to see. We give them the ability to be able to understand, to reason, and then to act. And that is the opportunity that is in front of us today. Now, there's a change that's occurring in the computer architecture with the amazing sensors and the capabilities that we have built on. We can bring in the environment in. The context that we're able to bring is what makes AI and it is what makes agents more and more curated and personalized for each of us. It's not just something that's in the cloud somewhere without having the insights. It sees what you see. It hears what you hear. And that makes these devices and these capabilities truly unique. And then by combining it with these reasoning models, with world models, we're actually able to create unique end experiences that basically are configured for these different verticals that we're all talking about today. So, just to really encapsulate this challenge that we have in front of us. We have all seen these transitions. The personal computer, it was enabled by the keyboard and the mouse. The phone, it was enabled by touch. That was the human machine interface. If today the human machine interface is AI and agents, what is the device that is best suited to that world? And it's those devices. It is the devices like smart glasses, an AI pendants, an AI necklaces, and it is basically earbuds with cameras on them. And it is really critical for us to be able to bring in the understanding of the environment, and that's where all the perception technologies that this room has been working on can now form a part of the human machine interface for these new unique devices. And that's what makes me very, very excited. In terms of the amount of AI capabilities that we have into these devices. Today, I'm very excited to be talking about the world's first multimodal AI model running on a smart glass. It's not being offloaded to the phone. It's not being offloaded to a host device. It's not being offloaded to the cloud, either. It's running primarily on the glass. And actually, you can go and see this in our booth. Four different models running concurrently, taking your speech, converting it to text, taking that text and bringing in the image, tokenizing it, and then putting that into a language model, and then producing the output back as speech. That's kind of what we're able to do.

Because of these capabilities of AR1 and AR1 Plus, this is the traction that we see today. You also just saw what Evan shared. I couldn't put that on, would have loved to do it. But there is a plethora of smart glasses out there, and there's a whole pipeline of a lot more that's coming as well. But more interestingly, look at the AI hearables. Look at the wearable AI products. These are not products that even we were thinking about until recently. But it is our consumers and our partners that are coming in, and they're saying they want an AI native device that allows them to be able to do these amazing experiences. So, we are super excited about all the traction that we're seeing in this space. And thank you for all the partnerships across this room, as well. Similarly, very excited about the work we're doing with Google and Samsung to really be able to drive Android XR forward, as well. Amazing amount of traction there, as well. So, thank you for the great work there. But to put it all together, this is not one device. I think it's not even just smart glass, and it might actually be more than what I'm showing over there, too. So, it might actually even be more than that. But really, the vision is exactly what I think Ori and what Evan talked about, right? The vision really is to be able to bring in this myriad of agents to bring in these unique ideas that many of us haven't thought as we haven't thought about as Qualcomm as we design these new experiences and new platforms. To be able to make that reality real. I'm very excited today to be sharing with you the Snapdragon start program.

This is the scalable turnkey AI-ready toolkit. So, bring your amazing ideas, bring your amazing agents and use cases and we will be able to offer the complete hardware, the complete software and allow you to be able to scale to any different kind of device. This starts with first of all having the tiniest tiniest module to be able to integrate everything that you need inside to be able to make a smart glass work. Then it has a software that runs on the host and it's a software that runs on the glass and then of course scaling partners that will allow us to scale this great product. So, it's again one of my great great pleasures to be sharing with you. This is the first module, a complete system in package that allows you to be able to make a smart glass.

It is good that it's tiny by the way. If it were bigger that'd be a problem. But, this is really the most unique, the most optimized Snapdragon AR1 platform out there. It's 8.9 by 42 mm which is way smaller than anything else that you see in a smart glass today. And you can see this is a complete system in package. It has all the capabilities needed and that's why we want you guys, the builders and the tinkerers to be able to bring in their ideas to be able to use this platform to create new and amazing experiences and products out there. So now if you have a module like this, you can make it fit into any device, a glass of any different type or shape or any other form factor for that matter. And then to truly be able to scale it, we're working with our partners like Applied Materials and Thunder Comm, and of course working with scaling partners like Georgian and Pegatron to not just do AI glasses, but also do monocular glasses and binocular glasses. So we'll have experiences these amazing experiences really coming from the group that we have in this room today. And to really drive this forward, actually we have gone ahead and work with one of our partners. This is again a really an amazing interaction and a great partnership that we have created. So gives me a lot of pleasure to actually announce the partnership we have with Inspecs Group PLC. They have multiple different glasses. These guys know how to build glasses, multiple different brands. And we want to be able to work with them to take the start program and they'll be the leaders in basically driving start program with us. And I think it opens up amazing amount of opportunities. So to talk more about it, I actually wanted to share a message from Robin Totterman who is the founder and executive chairman of Inspecs Group. Please have a listen.

Thank you, Robin. Great to see. And by the way, this is the brand portfolio that InWith Specs has. We're very, very excited about working with them to be able to create these new experiences, new devices, and a whole new slew of new smart glasses and capabilities as well. Now, not to just stop over there. I have more to share. So, you know, we talked about the ring to be able to control the smart glass last year. I'm excited to be sharing the second-generation smart ring controller. And this one has amazing amount of capability. It has the Snapdragon S7 Plus inside it, which means that it can do Bluetooth and Wi-Fi connectivity back to the smart glass. It also has full AI interactions capability with a microphone built into it, with full health sensing, and hand gestures and control for devices with full three and six degrees of flight control. So, really an amazing product as well to be able to drive forward more and more adoption of these kinds of form factors. Thank you. And you can actually try it out in our booth as well. We have a great demo showing exactly how you can use the product like this. All right. So, talked a lot about smart glasses in the smaller form factors for spatial intelligence devices. But truly, the best experience for the most immersive type of experiences in virtual worlds, there is no replacement to an amazing mixed reality device as well. Now, if you look at all the traction that we're seeing today, it's amazing across our partners like Meta and Pico and Play for Dream, Samsung, Valve, and Vivo. Amazing devices, amazing experiences. And for you guys specially at AWE, we've actually worked with our partner Abbott to introduce and to create the world's first purpose-built video game for blood donation. So, if you have any anxiety around blood donation, this is AWE, right? So, you have the ability, they are on site to be able to experience this and also do a good thing, which is to donate blood as well. Additionally, our XR Sports Alliance program continues to move at a very good speed. On the right, you can see the kind of experiences that we're bringing in with using MR and the amazing immersive technologies that we have built in with our partner Fees. Again, continuing to work with our key partners Exceedo and HBS. But the really amazing part I wanted to share with you is we have a new cohort of partners for XR Sports Alliance, including partners like Manchester United, Specs, Pico, Antigravity, and many more. And with this, we now have about 33 members across four different cohorts. This is really amazing. It's really amazing critical mass at this point in time for these kinds of experiences. So, yeah, I'm still setting it up for something pretty impressive here. So, here's what I wanted to really talk about today. I'm super stoked to be sharing with you the output of the work of thousands of engineers at Qualcomm to be able to now take the mixed reality experiences to a whole different level. Please take a look.

There it is, guys. The most integrated beast of a platform for mixed reality processing. Very very excited to see what our partners and everybody is going to be able to do with this, but this is the most advanced platform out there today for mixed reality products. Thank you. All right, this is truly designed for premium mixed reality experiences from the ground up. It is longer-lasting, sleeker, and cooler devices that we can enable with it, and it's designed for GenAI. Jumping into a few of the details, this platform is the first one to bring Adreno sliced architecture onto the XR side. And if you look at the data, we can do 60% higher performance at the same level of power as our last platform, or we can do 64% lower power at the same performance level. You can do all those complex use cases and the complex rendering experiences that you want to be able to build on mixed reality products. Same time, we've enhanced ray tracing capabilities to be able to bring in those realistic lighting and images and reflections and shadows to be able to create those amazing experiences as well. On the CPU side, we have actually enhanced the capability a lot as well. We can do 30% higher performance at the same power level or 45% lower power at the same performance level compared to our previous generation XR2 Gen 2 platform. And what's very important in these products is video see-through quality. So, we're really focused very heavily on video see-through quality. We have improved the pixel-to-pixel latency by 10% and reduced the power by 33% and with the inline spatial denoising and many other techniques, this is what we are able to do comparing it to another platform that's there available on the market at this time. Much, much better image quality for video see-through. And this is possible by using platforms or blocks like IVA, which is our engine for visual analytics. We've enhanced the depth from stereo, and at the same time, we've enabled much lower power 3D reconstruction. We've added two new accelerators also for 3D point cloud and triangulation and 3D reconstruction and meshing. And with these upgrades, here is what you get. You get amazingly superb 3D reconstruction that you can see on one side, but at the same time, much, much more granularity on the depth from stereo that we're able to get on the mixed reality platform. Now, this also is designed for AI. So, this has the Hexagon NPU engine where we increase the performance by 160 percent at the same power level as the previous generation. And you can see the comparison with the other platform that's there on the market at this time, we are about 25% higher on trillions of operations per second that you can do for AI processing. So, really a beast in terms of being able to create these new experiences like NPCs in games, to be able to do 3D content creation, to be able to do actual Gaussian splatting. So, actually, if you walk over to our booth, you will see a full 3D avatar created using Gaussian splats, all of it running on the Snapdragon Reality Elite. Please check it out. It's actually connected to a large language model. It will talk back at you. So, please do try it out. With this engine, we can do almost 45 tokens per second for a language model or 1.7 seconds to create a completely new image. Here is what we can do with it. Please take a look at this demo. Here is what we are able to do. We can take an image, drag in another image to it. It uses basically a spatial diffusion model to be able to recreate that image on the fly. And you can see how quickly it's able to do that, literally in quarter of a second. We can recreate an image in quarter of a second with the shape that you're dragging on over there. That's how you create those amazing virtual worlds, folks. And then at the same time with the much better power consumption, we are able to reduce the temperature. So, this will allow 12° lower temperature on devices for the same load, 20% longer battery life as well. And we work with any type of form factor, whether tethered or all-in-one. And this in a nutshell is Snapdragon Reality Elite, the best platform out there for mixed reality processing at this time. Thank you. Now, what I'd like to do is to actually call up some of our partners who have really been working very hard on making Snapdragon taking the Snapdragon Reality Elite and creating products for it. So, this is Play for Dream and I'd like to invite their founder, Huang Feng, to share their vision with us. Huang?