Google Cloud and NVIDIA unveil Vera Rubin AI systems and new sovereign cloud tools

New A5X computing instances are scheduled for November 2026, alongside expanded support for open AI models, confidential computing and locally controlled AI deployments

Google Cloud and NVIDIA have announced new GPU-powered computing systems, sovereign AI capabilities and confidential computing technologies.

Google Cloud and NVIDIA have announced new GPU-powered computing systems, sovereign AI capabilities and security tools designed to support increasingly complex enterprise AI workloads.

Unveiled at NVIDIA GTC Berlin the updates include Google Cloud's forthcoming A5X bare-metal instances, powered by NVIDIA's Vera Rubin NVL72 platform, with general availability expected in November.

The companies are also introducing additional options for organizations that need to operate AI models within specific geographic jurisdictions or isolated computing environments, including new capabilities for Google Distributed Cloud and the public cloud.

The changes span computing hardware, networking, storage, model deployment and data protection, as Google Cloud develops infrastructure for agentic AI systems capable of carrying out multi-step tasks.

Vera Rubin infrastructure targets larger AI workloads

Google Cloud's A5X instances will bring NVIDIA's Vera Rubin platform into its AI Hypercomputer infrastructure, combining GPU acceleration with networking and storage technologies designed for large-scale model training and inference.

The company says the new systems will be available in additional regions across Europe, the Middle East and Africa (EMEA), supporting organizations with regional data residency and sovereignty requirements.

A central component is Google's Virgo Network, which the company says will support clusters of up to 80,000 Rubin GPUs at a single site and as many as 960,000 GPUs across multiple sites. These figures describe the infrastructure's intended scaling capacity rather than an announced operational deployment.

Other updates include support for NVIDIA GPUDirect Storage through Google Cloud Managed Lustre, allowing data to move directly into GPU memory, and advanced caching designed to manage longer AI interactions while reducing repeated processing.

Google Kubernetes Engine will also support workload orchestration on A5X instances, including resource scheduling and network management for distributed AI systems.

Stephan Pretorius, Chief Technology Officer at WPP, points to the potential application of the infrastructure in creative production.

“With Google Cloud’s A5X instances accelerated by NVIDIA Vera Rubin NVL72, our WPP Creative and Production teams can direct intelligent agents that perform frame-accurate quality checks and adapt brand assets seamlessly.”

Google introduces new sovereign AI and local Gemini deployment options

Beyond computing capacity, the partnership introduces new options for organizations seeking greater control over where AI systems operate and how their data is processed.

Google Cloud is adding support for open-weight models including Gemma, Mistral and NVIDIA Nemotron to its Sovereign Cloud portfolio, allowing organizations to retain custody of model weights and runtime environments.

The company has also made NVIDIA Blackwell Ultra GPUs generally available through Google Distributed Cloud, using NVIDIA HGX B300 systems for computing deployments within an organization's own facilities.

Gemini 3.5 Flash is being added to these isolated environments, enabling multimodal AI processing involving text, audio, images and video without requiring workloads to operate outside the local infrastructure.

Another addition is enterprise search powered by Gemini and multimodal retrieval-augmented generation (RAG). Available in preview for air-gapped Google Distributed Cloud deployments, the technology is designed to let organizations search, summarize and examine internal documents, media and other information while keeping processing inside their controlled environment.

Google is also extending support for NVIDIA Cosmos 3 and Nemotron 3.5 SuperVL on its Gemini platform, including tools for model customization and simulated environments used in robotics and autonomous vehicle development.

Confidential computing protections extend across Google Cloud

The companies have also introduced security updates intended to protect sensitive information while AI models are being trained or used.

Google Cloud's Confidential G4 instances, powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, are now generally available. Confidential computing is also available on A4 instances using NVIDIA HGX B200 hardware and Intel TDX.

A further development is Confidential Bare Metal on A4X Max instances, which use NVIDIA GB300 NVL72 technology and hardware-based security measures to protect AI models, intellectual property and user prompts during processing.

These capabilities are aimed at organizations managing sensitive AI workloads, particularly where data protection, local operational control and infrastructure isolation are requirements.

The announcements cover technologies at different stages of availability. Confidential G4 instances and Blackwell Ultra GPUs on Google Distributed Cloud are generally available, while the new enterprise search capability remains in preview.

Google Cloud's Vera Rubin-powered A5X instances are scheduled to become generally available in November 2026, including in additional EMEA regions.

ETIH Innovation Awards 2027 open for entries

The ETIH Innovation Awards 2027 are now open for entries, with 26 categories recognising companies, institutions, projects and individuals making a measurable impact across education technology. The global awards span AI in education, digital learning, higher education, skills and workforce, inclusion, education operations, evidence and impact, with entries managed through ETIH’s new online awards platform.

Entries are open worldwide, with the early-bird deadline on Friday 18 December 2026 and the final entry deadline on Friday 26 February 2027 at 23:59 UK time. 

Explore the 2027 categories and judging criteria, read the How to Enter guide or start an entry through the ETIH Innovation Awards 2027 platform.