Tech leaders advance hybrid AI and cross-device memory technologies

Major technology players are advancing the integration of artificial intelligence across multiple devices through cloud-based memory and hybrid computing strategies.

Google DeepMind has unveiled a new design for ‘Private AI Compute’ aimed at allowing AI to maintain conversational context across different devices, such as moving from a smartphone to a laptop. This technology stores encrypted AI memory in the cloud, with decryption keys held locally on the user's device to ensure privacy. The system is designed so that data is only decrypted within an isolated processing space during active requests.

Simultaneously, Qualcomm is promoting a ‘Hybrid AI’ ecosystem in partnership with Samsung Electronics. This strategy distributes AI workloads across a spectrum of devices, including smartphones, smartwatches, and smart glasses. By utilizing 1-bit AI models for immediate tasks on wearables and offloading complex computations to smartphones or the cloud, the approach aims to overcome battery and thermal limitations of smaller devices. Qualcomm plans to supply dedicated platforms, such as Snapdragon 8 Elite and Snapdragon AR1, to Samsung’s upcoming device lineups to support this multi-device AI environment.