Google Cloud and NVIDIA Announce General Availability of NVIDIA Vera Rubin Infrastructure

Google Cloud and NVIDIA announced the general availability of NVIDIA Vera Rubin infrastructure, including A5X instances, which combine NVIDIA Vera Rubin NVL72, NVIDIA ConnectX-9 SuperNICs, Google Cloud Virgo Network and liquid-cooling technology. The platform is designed to support demanding agentic inference and reinforcement learning workloads at scale, with planned expansion to Europe following initial availability. The integrated system aims to streamline the continuous cycle of experience generation, evaluation and model updates required for advanced AI systems.

Ineffable Intelligence, founded by AlphaGo creator David Silver, became the first customer to access A5X instances, using the platform to advance large-scale reinforcement learning and develop "superlearner" systems. The company highlighted the need for substantial compute and high-bandwidth connectivity to enable rapid model improvements through experience-driven learning, with Vera Rubin’s architecture meeting these requirements.

WPP demonstrated a creative agent reference architecture on Vera Rubin-based A5X instances using NVIDIA Nemotron models and NVIDIA OpenShell, enabling multistep workflows such as video generation and upscaling with human approval. OpenShell provided governed sandboxes to control access to sensitive brand assets and creative tools, reducing the need for multi-node clusters in initial tests.

Google Cloud will offer NVIDIA Confidential Computing at rack scale on bare metal instances like A4X Max, with general availability also on G4 VMs powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. The company is exploring NVIDIA DSX MaxLPS capabilities for Vera Rubin-based infrastructure to optimize performance per watt and expand capacity within data-center power budgets.