Qualcomm at Snapdragon Summit 2026 Day 1: pushing AI onto devices

On the opening day of Snapdragon Summit 2026, Qualcomm presented a comprehensive strategy for the next era of computing, emphasizing the shift towards AI that operates directly on devices rather than relying heavily on cloud processing. The semiconductor leader, valued at $210 billion, reported a 55% surge in its stock over the past six months due to investor enthusiasm for its on-device AI strategy. However, Qualcomm challenged the industry to make AI more reliable, private, and practical for everyday use.

The summit showcased new platforms, software tools, and partner products, highlighting the ongoing transition within the market. Qualcomm and its associates emphasized the importance of developing superior AI models, enhancing developer tools, and fostering stronger ecosystem collaborations for the successful transition to agentic AI. They positioned personal AI as a distributed system, seamlessly following users across multiple devices, rather than confining it to a single app or cloud service.

Qualcomm unveiled several new platforms and programs during the event, including Liquid Context on Snapdragon, Snapdragon START, and Snapdragon Sound Elite Gen 2. Major partners such as Google, Microsoft, Canonical, HP, Lenovo, Meta, Snap, and others demonstrated products built on Snapdragon chips. The company also placed a strong focus on privacy, extended battery life, and local processing as key advantages compared to cloud-first AI competitors.

Beyond consumer devices, Qualcomm expanded its focus into the healthcare sector with the establishment of the Personal AI Health Alliance, aimed at developing on-device applications. The company's strategic emphasis was clear: AI is evolving from cloud-centric to personal, local, and context-aware experiences across a range of devices, from smartphones and wearables to XR glasses, PCs, automotive systems, and IoT devices. Users should no longer need to unlock screens, open multiple applications, or repeat instructions.

Developers will require open tools and software stacks to support various hardware types. Qualcomm also noted the industry's shift from model-centric to token-centric and inference-centric thinking, underscoring that AI's value will increasingly stem from real-time functionality rather than simply the size of the model.

Liquid AI, co-founded by Ramin Hasani, demonstrated its Liquid Foundation Models, which have been downloaded approximately 1.4 million times weekly from Hugging Face. These models range in size from 200 million to 24 billion parameters, enabling them to function efficiently across different device categories. Hasani introduced Liquid Context on Snapdragon, a smart memory layer designed to facilitate AI context retention across devices.

The company is collaborating with Mercedes-Benz on the Snapdragon Digital Cockpit to deliver in-car AI intelligence. Hasani outlined five criteria necessary for agents to become the next major computing platform: quality, context retention, better tools supporting foundation models, respect for physical constraints like battery life and heat, and open developer access.

He aimed to deploy Liquid Foundation Models on 50% of the world's annually produced devices by 2029, emphasizing the importance of making local AI accessible.