Google Cloud to launch A5X instances on NVIDIA Vera Rubin in November
Google Cloud and NVIDIA announced general availability of A5X instances on Vera Rubin in November, with up to 960,000 Rubin GPUs planned across sites.
By GennnAi··1 min read
Google CloudA5XVera RubinIneffable IntelligenceWPPAI HypercomputerGoogleInfrastructureEntreprisesFlash
A5X combines Vera Rubin NVL72, ConnectX-9 SuperNICs, Google Cloud's Virgo network and fourth-generation liquid cooling. Availability is planned to extend to Europe. London-based Ineffable Intelligence, founded by AlphaGo creator David Silver, is the first customer, for large-scale reinforcement learning. In WPP's initial tests, workloads that typically required a multi-node cluster ran on a single A5X instance. Google Cloud also makes Confidential Computing generally available on G4 VMs.
Source: NVIDIA, published on October 21, 2026.