See all news
Google Cloud
Global tech

GPUs put on armor! Google Cloud launches Blackwell confidential computing

Google Cloud announced the general availability of Confidential G4 virtual machines with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. The platform targets companies that process sensitive data through artificial intelligence applications and need protection while their workloads run. The service combines hardware-backed isolation, encryption and environment verification.

Encryption protects stored or transmitted information, but data can also be exposed while an application processes it. This problem becomes relevant when AI models receive private documents, business instructions or parameters of high economic value.

Confidential G4 instances aim to protect that processing with isolated execution mechanisms and cryptographic validation of the environment. Google Cloud announced the general availability of the service, powered by NVIDIA Blackwell accelerators. Organizations will be able to use it for inference tasks and agents that handle sensitive information. The company highlights the ability to verify components of the environment before authorizing the use of certain data.

The technology expands the tools available to regulated industries or teams that need to retain control over models and data. However, no isolation mechanism completely eliminates risks: identity management, permissions, application security and monitoring are still necessary.

At a glance

WhatGoogle Cloud made confidential G4 instances with NVIDIA Blackwell GPUs generally available to run AI processing with hardware-backed protection.
WhenGoogle published the announcement on October 21, 2026, after having offered a preview of the service.
WhereThe instances are part of Google Cloud's infrastructure and are aimed at enterprise customers in supported regions.
HowLa solución utiliza entornos de ejecución aislados, cifrado en hardware y mecanismos de verificación criptográfica antes de procesar datos sensibles.
WhyCompanies need to protect sensitive information not only at rest or in transit, but also while their models and agents process it.

Build your digital presence

Start today: write to us and we’ll reply within 24 hours.