Google Cloud plans Vera Rubin instance availability next month

Google Cloud's A5X instances using Nvidia Vera Rubin infrastructure will become generally available next month, Nvidia announced, with expansion to Europe planned. Google Cloud plans to scale the infrastructure to as many as 960,000 Rubin GPUs across multiple sites. A5X combines Vera Rubin NVL72 systems, ConnectX-9 networking, Google's Virgo Network and liquid cooling to support training, inference and large-scale reinforcement learning. Ineffable Intelligence was the first customer to access the instances and is using them for reinforcement learning. In WPP's initial tests, creative agent workloads that typically required a multi-node cluster ran on a single A5X instance.