As AI becomes integrated into our everyday devices, we see tools, technologies, and capabilities changing how we work and live. Generative AI and agentic AI are integrated into our devices—from smartphones to intelligent wearables to smart glasses to AI PCs—and are more connected than ever.
Our devices will also become increasingly automated and personal, adapting to each of us and becoming more human-centric. The devices begin to understand context, nuance, and intention, using inputs of text, images, and our voices.
For example, to reschedule a meeting with a colleague, simply ask your device’s AI assistant for help. On-device agents review availability in both calendars during defined working hours and even suggest meeting locations.
By combining agents with fast connectivity and powerful processing capabilities, devices will enable an agent-centric experience rather than the app-centric paradigm we have become accustomed to today.
The cloud-edge convergence
The smartphone has been the center of our digital lives—always connected to the internet, loaded with apps, and relied on by our wearable devices. Today, AI agents on our smartphones run in the cloud. However, soon more of their inference and reasoning will occur on the device.
Personalized AI cannot solely rely on the cloud, as it requires minimal delay between a user’s action and system response, data privacy, and contextual understanding to predict what the user will do next.
Qualcomm’s platforms use different inputs, such as text, images, or video, so AI can assist the user. These work directly on the device, enabling fast, secure, and deeply personal AI. Your ecosystem responds in the moment because the AI intelligence lives where you are. When the device approaches its maximum processing capabilities, the cloud is brought in for larger compute support.
This agent-centric shift requires many innovations, and Qualcomm has been involved in two that stand out: neural processing units (NPUs) and 6G.
The new NPU
Personal computing systems traditionally utilize two types of processors: a central processing unit (CPU) and a graphics processing unit (GPU). Today, there’s a third type of processor—the NPU—optimized for AI tasks.
The NPU functions as a specialized AI processor for low-power, immediate on-device inferencing, allowing the CPU and GPU to focus on generalized workloads and graphics. For example, when you engage with your device’s AI assistant with natural language, the NPU employs advanced speech recognition, intent understanding, and text-to-speech to provide responses.
The need for a third processor type is perhaps obvious to anyone paying attention to the rapid growth of applications like Copilot+ and Gemini, but some people may be surprised to learn that AI is also affecting wireless technology.
Introducing 6G for the AI era
AI and wireless are converging to apply more intelligence to connectivity. We’re moving away from networks that simply move data, and toward adaptive systems that sense context, learn in real time, and respond dynamically.
Pushing this shift forward, 5G‑Advanced improves efficiency and device‑to‑network performance while enabling smarter, more scalable systems—from the internet of things (IoT) to edge computing. These advances accelerate the move toward AI‑native architectures where intelligence is distributed across devices, the network, and data centers.
That momentum sets the stage for 6G, which will operate as an end‑to‑end system that fuses connectivity, compute, sensing, and AI, and will cover segments such as personal AI devices, automotive, robotics, industrial IoT, and more.
AI on every device
This AI-powered agentic-centric ecosystem will include all the devices you currently use:
- Smartphones will grow more powerful, with NPUs to handle AI on-device and shift toward tools for your personalized agents.
- Wearables (or personal AI devices, including smartwatches, headphones, pendants, pins, earbuds with cameras, rings) will connect with your custom agents and work with your phone and apps to maximize usefulness.
- AI-assisted smart glasses will deliver interactive experiences that seamlessly connect your digital and physical worlds.
- AI PCs with robust on-device reasoning models perform tasks such as workflow automation for more efficient experiences, daily summarization and task preparation, and personalized assistance.
Our interactions with these devices will increasingly make them more human-centric than ever.
Alex Katouzian is the EVP and group general manager for mobile, PC, XR, and wearable AI businesses at Qualcomm.