Liquid AI and Qualcomm Technologies have optimised the Liquid Context on-device context layer for AI for Snapdragon processors using the Qualcomm Hexagon neural processing unit (NPU).
Announced at Snapdragon Summit 2026 in Maui, Hawaii, the collaboration gives AI agents a persistent, privacy-conscious understanding of the user by learning, with the user’s permission, from device signals, such as messages, notes, photos, and app usage.
Liquid Context builds a running profile of routines, preferences and needs, and keeps it updated in the background. That context can then be used by AI agents running on the device, in the cloud or a hybrid setup, so they can offer more proactive and personalised assistance without sending every update to the cloud.
For original equipment manufacturers (OEMs), the integration is a ready-made foundation to add personal context as a built-in device capability. Compatible embedded, cloud and hybrid agents could tap into that context to deliver assistance that feels more aware of what the user is doing and what they might need next.
“This is a smart memory layer that allows you to read, compress, and write every sort of experience that you have throughout your, let us say, active engagement with a device and across devices. This context layer is going to enable AI systems to go from one device to another device,” said Ramin Hasani (top right with Qualcomm CEO Cristiano Amon), Co-founder and CEO of Liquid AI.
Liquid Context comprises a memory component that gathers and abstracts user preferences from multiple sources; and an agent layer that lets AI agents running locally on Snapdragon chips access that memory to perform tasks.
“By combining our industry-leading Snapdragon platforms with innovative AI technologies from companies like Liquid AI, we’re helping accelerate the next generation of Agentic AI experiences,” said Durga Malladi, Executive Vice President and General Manager of Technology Planning, Edge Solutions, and Data Centre at Qualcomm Technologies.
Liquid AI has also optimised its LFM2.5-2.6B agentic model, which powers its Liquid Agent, to take advantage of Snapdragon processors’ performance, memory access and low latency.
It is working with Mercedes-Benz on local in-car intelligence using Qualcomm’s Snapdragon Digital Cockpit technology, and aims to have its models deployed on half of annually produced devices across categories including wearables, handsets, vehicles, PCs, robots, and industrial products by 2029.
