TechPrufer
Get started
DeepMind
DeepMind logoDeepMind · Feature trail

Private AI Compute Memory

1 post · Sep 2026

What this trail covers

Private AI Compute architecture is enhanced with a persistent, server-side memory layer that enables cross-device AI continuity while maintaining on-device privacy standards. This is achieved through a secure enclave in the cloud that temporarily decrypts user data for processing and immediately re-encrypts it, with cryptographic keys held exclusively on user devices.

All DeepMind trails
  1. Advancing Private AI Compute with secure, server-side memoryIntroduces a new persistent memory layer for Private AI Compute, enabling secure, server-side storage of AI context across devices. This layer functions as a secure digital vault where data is encrypted at rest and in transit, with decryption occurring only within hardware-isolated enclaves on the cloud and cryptographic keys managed on the user's device. The post details the architecture for secure data access and context retention, addressing the challenge of cloud-scale AI continuity while upholding strict privacy.