Google DeepMind has announced a major technical update to its Private AI Compute platform, introducing server-side memory that maintains on-device privacy standards. This breakthrough enables persistent, cross-device AI memory by creating a secure digital vault in the cloud. Under the new architecture, data is sealed within dedicated, encrypted storage, while cryptographic keys remain exclusively on users' personal devices, ensuring data remains inaccessible to anyone else, including Google. When an AI model accesses information, an authenticated, end-to-end encrypted channel connects the device to a protected cloud environment called a "secure enclave," which temporarily decrypts data in isolated memory before re-encrypting it. The system combines hardware-enforced secure enclaves, encrypted channels, and per-user databases shielded by device-derived encryption keys. Previously, Private AI Compute technology was strictly "stateless," wiping context after each task. This evolution addresses the need for cloud-scale computing power while maintaining strict privacy protections. The update allows users to seamlessly resume activities across devices, such as viewing assembly instructions on a laptop previously accessed through smart glasses.

To build trust, Google is publishing a tamper-proof public record of its server software, enabling devices to verify authenticity before transmitting personal data. The company also released an updated technical whitepaper and results from an independent cybersecurity audit. These resources invite the broader privacy community to verify the platform's protections.