Google Cloud will launch general availability of its NVIDIA Vera Rubin-powered cloud instances next month, NVIDIA announced on Wednesday. The rollout begins ahead of a planned expansion into Europe, bringing cloud infrastructure for model training, inference and reinforcement learning.
A5X instances integrate NVIDIA Vera Rubin NVL72 racks, ConnectX-9 SuperNICs, Google Cloud Virgo Network and Google Cloud’s fourth-generation liquid-cooling technology. Google Cloud plans to scale this Rubin-based infrastructure to as many as 960,000 GPUs across multiple data center sites, tying the hardware into its AI Hypercomputer system.
Ineffable Intelligence, a London-based technology company founded by AlphaGo creator David Silver, became the first customer to gain access to the A5X platform. The firm is using the system to develop continuous reinforcement learning models that learn directly from experience.
“We’re building superlearners that continually improve through their own experience,” Silver said in NVIDIA's announcement. “That demands both substantial compute and a tight connection between generating experience and training on it, so each improvement in the agent can quickly shape its next experience. Vera Rubin’s combination of compute performance and high-bandwidth connectivity, between the Vera CPU and Rubin GPUs and between the Rubin GPUs within the rack, allow us to push the scale and ambition of our reinforcement learning research further.”
Marketing and communications group WPP also tested creative agent workflows on the A5X instances using NVIDIA Nemotron models and NVIDIA OpenShell software. The setup generates multistep workflows such as video generation, restyling and upscaling under human approval gates. In WPP’s tests, tasks that previously required a multi-node cluster ran on a single A5X instance.
Security offerings will also expand under the deployment. Google Cloud plans to offer NVIDIA Confidential Computing at rack scale on bare-metal A4X Max instances powered by NVIDIA GB300 NVL72 systems, alongside general availability of G4 virtual machines running NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Early G4 users include Vertebrae, Polaris Cloud and Confidential Core AI.
Both companies are currently testing NVIDIA DSX MaxLPS capabilities on the Vera Rubin infrastructure. NVIDIA said the tests are designed to optimize performance per watt and keep workloads within existing data center power budgets.
