What changed? Major mobile platforms are increasingly exposing AI capabilities that can run directly on compatible devices. In September 2025, Apple said that its Foundation Models framework lets developers use an on-device language model for intelligent app features that can be available offline. Qualcomm describes its current mobile platforms as supporting generative AI workloads running entirely on-device.
Why it matters
For the user, local processing can reduce the need for a round trip to a remote server for some tasks. This can support responsiveness and help keep certain workflows available when connectivity is limited. It does not mean every AI task runs locally: capability, device compatibility, app design and the nature of the request still matter.
The design question
The interesting question is not simply whether AI is on-device or in the cloud. It is which part of an experience should stay close to the user, and which part benefits from networked computing. For mobile products, that is increasingly a product-design decision as much as a technical one.