In what it said will be an evolution of its principle of creating technology for “the ecosystem of you”, Qualcomm has set out its agenda for the near future, assuring that the company would transition from its current phone-centric model to an agentic-centric model that would see key original equipment manufacturers (OEMs) able to offer personalised experiences across a number of key application areas.
At the heart of the products Qualcomm says will power the agentic artificial intelligence (AI) age and the next generation of premium mobile experiences are flagship smartphones from leading global OEMs and smartphone brands including Honor, iQOO, Motorola, Nubia, OnePlus, Oppo, Redmi, RedMagic, Vivo, Xiaomi and ZTE. In addition, Qualcomm has announced a collaboration with Google to bring the Snapdragon X Elite platform to the first Googlebook laptop, as well as with Microsoft on Surface devices.
Introducing the technologies at an opening speech at its Snapdragon Summit 26 conference, Qualcomm CEO Cristiano Amon stated that for the event there was a key topic: the evolution of AI towards the transition to the AI smartphone.
“When we started with AI [we added] features in cameras so that you can take pictures with no lights, and you have a lot of AI-enhanced images,” he said. “But now we see how that evolution went and we’re going to change how we use the smartphone. Last year, we talked about the ‘ecosystem of you’, and how with AI, your experience is going to be personalised. You’re [now] going to have a lot of intelligence that is going to be with you in your own context and can be integrated not only in your phone, but in the devices you know around you.
“At this point, we’re going into this transition from what is a very phone-centric model to an agentic-centric model for the new experiences,” said Amon. “The agents become the centre of that experience – not the phone – and that changes how we think about the role of the smartphone. We can clearly see what the AI smartphone is. We have seen this transition, we have this vision and now we see this thing coming to life. We’re moving from apps and assistants into agents, and we’re just now at the beginning of the agentic age. The best AI is not going to ask more of you; it is one that does more for you. And it’s your own AI.”
However, Amon assured that despite the transition very much being underway, the smartphone was not be going anywhere soon. The difference was that for the phones of the future, the key was how it was going to interact with agentic experiences. And that while for the past two decades the smartphone experience has been organised around the apps, now, in the agentic age, the experience will begin with intent – AI agents that can understand a user’s intent and that will require a device to do much more than just execute commands.
“It’s going to be less about navigating apps and it’s going to be about completing tasks for you,” he said. “It’s going to handle a lot of the details to get things done for you. Now you have two dimensions that your device has to support.
“Today, the smartphone experience is built around your interaction with the device – you pick up your phone and you’re going to go to different apps and you’re going to do different things. You initiate the interaction; the device responds. But now there’s another dimension of the experience: that’s what you do indirectly. What your agent is going to do on your behalf. The agent now must receive what’s happening, understand the intent, understand the context, reason, plan, remember and act on your behalf. And you’re going to start to see new functionalities besides the ones expected on the phone.”
At the bedrock to this concept will be secure permission, since the agent will be gaining access to a wealth of personal knowledge. The agent will need to be granted permission to have the ability to take action on users’ behalf as a proxy of them. It will have a persistent context in interactions, real-time reasoning and be multimodal. The agent will deal not just with language, but also with vision, audio and even physical AI inputs. Another key technological aspect will be whether the agent runs on a device or is built into an operating system as it performs coordination across models, apps, services and data.
This leads to the extension of the concept of the ecosystem of you into other device categories, as experiences have to be personal, have understanding of a context, be proactive, not intrusive, and be responsive in the moment people want, with the ability to continue across different device and across multiple applications, cloud services and AI models.
Amon stressed that for a user’s experience to be useful, the agents needed to be intuitive, instant, secure and always available. Trust, he noted, was the very first and most important thing to ensure a “great” agentic experience.
“It will begin to have trust,” said Amon. “The agent is going to act on your behalf, so it needs to be secure. It needs to be private. It needs to put the user in control. The second thing is something we know in technology. It has to be effortless – no friction. Using an agent, it must be easier than you doing [things] yourself. Otherwise, you pick up your phone, you open the phone, and you’re going to go to the different apps.
“The reason so many workloads moved from the laptop to the mobile was because with the mobile, it was easier, with lower friction. And that is going to be the same recipe when you’re going to choose in those two dimensions whether you’re going to do things yourself or when the agent is going to do things for you. All the complexity will stay behind the scenes, and the experience will feel very simple. And that is how an AI smartphone is going to feel. A new generation of devices are being designed for those agentic experiences. This future is here now.”
Qualcomm’s future agentic age and the next generation of premium mobile experiences – across the varied OEM devices – will be driven by the Snapdragon 8 Elite Extreme Gen 6 and Snapdragon 8 Elite Gen 6 mobile platforms. Both platforms are attributed with combining performance and on-device intelligence with advancements across AI, gaming, camera and connectivity to deliver delivering agentic AI experiences that are immediate, personal and adaptive.
Said to be the most powerful mobile platform in the Qualcomm Technologies portfolio, Snapdragon 8 Elite Extreme Gen 6 is claimed to deliver “breakthrough” AI capabilities and advanced performance for Android smartphones. Features include enhanced mobile gaming with the new Adreno Neural Fusion, using AI to deliver intelligent graphics for more immersive gameplay, richer visuals, and longer playtime; and AI powered camera capabilities with Advanced Professional Video (APV) support and Intelligent Pixel Control to offer deeper scene understanding, enhanced image quality and more natural-looking photos and videos.
The platform is said to offer 13% improved CPU performance compared with previous generations, as well as 37% improved CPU power efficiency; 44% improved GPU performance; 40% improved GPU power efficiency; 35% improved NPU performance; and 33% improved AI performance.
Snapdragon 8 Elite Gen 6 is built on the same advanced 2nm process node, powered by the custom-built Qualcomm Oryon CPU, rearchitected Qualcomm Adreno GPU and the advanced Qualcomm Hexagon NPU. This is said to enable the platform to power many of the same AI, gaming, imaging and connectivity features as Snapdragon 8 Elite Extreme Gen 6 to a broader range of top-tier smartphones.
Compared with previous generations, it is said to offer 10% improved CPU performance; 37% improved CPU power efficiency; 35% improved GPU performance; 40% improved GPU power efficiency; 14% improved NPU performance; and 20% improved AI performance per watt.
Uplink and downlink speeds
Both platforms feature AI-powered Qualcomm X105 5G Modem-RF connectivity, architected for greater uplink and downlink speeds, coverage and better power efficiency. Indeed, it is said to be the first AI-powered connectivity system to support 3GPP Release 19 5G Advanced while providing a pathway to 6G.
The multi-antenna (6Rx) system contains six antennas to capture multiple data streams at once, while upping download speeds by 10%, improving cell-edge reception and reducing dropped connections. It is also built to accelerate satellite connectivity for 5G devices, with an integrated 5G NR over satellite offering, NR-NTN for voice, video and data, NB-NTN for messaging and RFFE compatibility.
The platforms also feature the FastConnect 8800 Mobile Connectivity System, which supports Wi-Fi 8, Bluetooth, Ultra Wideband and Thread. These are loaded with proximity AI capabilities, from locating lost items to seamless key card access. A redesigned 4×4 Wi-Fi system is said to offer double the peak speeds and up to three times longer gigabit range compared with previous generations.
Yet just as the capability of the platforms was being unveiled, Amon remarked that AI represented a system-level design challenge. That is to say, recognising that agentic AI experience in device workflow. It also requires orchestration of tasks and computing to work on an all-day battery life mode with the entire system needing to work together to deliver performance, efficiency and responsiveness.
One of Qualcomm’s key recent moves has been a multi-generation collaboration with Amazon to enable customised silicon at scale for large-scale AI datacentres, working together on AI inference and optical connectivity solutions extending up to 1.6Tbps. Amon drew on the dynamics of this business line to look at some challenges regarding mobile devices.
“Some of the problems that now you see in the datacentre – that compute needs to outpace energy availability – have been true in the smartphone since day one,” he said. “You have a lot of computing requirement, but you don’t have unlimited energy. It’s a better-powered device, so you have to design specialised computing for each one of those tasks. We have a very interesting innovation that people did not expect. When we entered the datacentre space, we talked about a brand new engine, which is the high bandwidth compute. It’s a new engine that connects memory and the accelerator for significant efficiency improvement. That technology that we designed for the datacentre is now scaled all the way into the smartphone.”
The chief noted that OEMs were now designing real products around these capabilities and experiences, with the key being that agentic AI was not just another feature to enter the device. It was creating an entirely new kind of experience in device workflow. For the likes of Qualcomm, said Amon, this presents not only a system-level design challenge, but also a “significant” innovation opportunity. One example would be running a high-level operating system all day with clock speeds of one gigahertz and more, meaning needing an all-day battery life.”
Another facet that Amon highlighted was connectivity, which he said was even more important than ever. That was mainly because connectivity was no longer about connecting with the models in the cloud, but was crucial for the increasing importance of voice once. And more than just looking at the future and issues such as 6G, the connectivity piece was crucial for, say, smart glasses where there was a need to support high-definition video upstream transport of data.
Amon outlined what this meant for technology design in the world of agentic AI. “Some of those future agents and AI devices may process up to one million tokens per day on device while remaining available throughout the day,” he said. “That’s what we see today. It’s only going to increase.
“We have developed HBC for Snapdragon as a co-processor to [the main Snapdragon chip] that is going to bring memory and AI processing closer together, delivering the speed, efficiency and performance that the next generation AI smartphones will require. And the technology has a very important role across not only Snapdragon, but everything, such as in the edge. We’re going to take that [performance] to your laptop … your smart glasses … your car and more. The agent is at the centre, but your phone, your PC, your glasses, wearables and the car are all going to be endpoints for the agents.”
Upcoming collaborations
The devices that will run the new technology are coming onstream. At the summit, Goggle unveiled a range of laptops taking advantage of Qualcomm technology, while Microsoft did the same with versions of the Surface laptop, including one specially designed with branding reflecting the company’s sponsorship partnership with Manchester United Football Club. Motorola revealed that its Signature 27 range would be incorporating Snapdragon 8 Elite technology.
In addition to the US tech giants, Chinese phone suppliers Honor and Xiaomi expressed their commitment to the platform. Honor CEO James Li noted that in working with Qualcomm technology, his company had “achieved incredible milestones in on-device AI”.
One product highlighted was the company’s robot phone, which Li said was “the perfect proof of bringing physical AI interaction to life”. Noting the imminent launch of the Magicline series of devices, he added: “As we enter the era of agentic AI, we are co-creating an AI device ecosystem … transforming the endless perspective of technology into the unlimited potential of every individual.”
Adam Zeng, senior vice-president and president of the international business department at Xiaomi, told delegates how the company had built a full human-car-home ecosystem, and has been working with Qualcomm since 2011. As Zeng was speaking, his company was launching the Xiaomi Pro 18 Pro and Pro Max, featuring the Snapdragon 8 Elite Gen 6 and Snapdragon 8 Elite Extreme Gen 6 processors, respectively, built for what was called peak performance in a range of areas.
Looking at other Snapdragon products in non-phone areas, the event also demonstrated how Qualcomm’s in-car technology had been installed in a GMC Hummer EV, Mercedes-Benz CLE and Harley Davison. The 2024 Hummer used the Snapdragon Cockpit platform to power in-vehicle experience, supporting Wi-Fi and Bluetooth in the car and providing advanced audio and sound. There is also support for Google application services, such as Google Maps, as well as Gemini AI and the Google Play Store.
It also has the Snapdragon auto connectivity platform, supporting 4G, 5G, Wi-Fi and Bluetooth, and there is support for connected services such as navigation, with over-the-air updates. The built-in Snapdragon Ride platform supports Super Cruise, and advanced driver assistance and automated driving functions on approved roads. This includes automatic lane changes, automatic braking and cruise control.
The recently launched 2026 Mercedes-Benz CLE uses the Snapdragon cockpit platforms and Snapdragon auto connectivity platforms – again, powering the in-vehicle experience, the screens and all of the connectivity.
A Harley Davison is not just a classic motorbike marque, its owners are among the most dedicated to their machines, which tradition dictates are very old-school. That said, Qualcomm has supplied very new digital technology to provide an updated user experience. This is delivered through a 12.3-inch screen offering services such as digital navigation. It also syncs with phones to make calls over Bluetooth.
