Google Cloud sets Vera Rubin rollout for next month

Google Cloud’s NVIDIA Vera Rubin-based A5X instances are scheduled for general availability next month, with expansion to Europe planned. The company aims to scale the infrastructure to as many as 960,000 Rubin GPUs across multiple sites. The platform combines NVIDIA Vera Rubin NVL72, NVIDIA ConnectX-9 SuperNICs, Google Cloud Virgo Network and advanced liquid cooling to support demanding AI workloads.

Ineffable Intelligence, the London-based company founded by AlphaGo creator David Silver, was the first customer to access the platform. It is developing reinforcement learning systems that continuously learn from experience. Integrated into Google Cloud’s AI Hypercomputer, A5X combines compute, networking and storage to support the cycle of generating experience, evaluating outcomes and updating models.

WPP has demonstrated creative agents using NVIDIA Nemotron models and OpenShell on the instances. The agents translate conversational requests into workflows including video generation, upscaling and restyling, with human approval before execution. OpenShell provides governed sandboxes to control access to brand assets and tools. In initial tests, workloads typically requiring a multi-node cluster ran on a single A5X instance.

Google Cloud also plans to offer NVIDIA Confidential Computing at rack scale on bare metal instances and is making it generally available on G4 VMs. Separately, the companies are exploring NVIDIA DSX MaxLPS capabilities to improve energy efficiency and expand capacity within available data-center power budgets.