Google Cloud A5X brings Vera Rubin GA next month

What changed?

At GTC Berlin, NVIDIA and Google Cloud announced that Vera Rubin-based A5X instances will be generally available next month, with planned expansion to Europe (NVIDIA blog). A5X combines Vera Rubin NVL72, ConnectX-9 SuperNICs, Google's Virgo network and fourth-generation liquid cooling, aimed at agentic inference and large-scale reinforcement learning.

Google Cloud plans to scale its Vera Rubin infrastructure to as many as 960,000 Rubin GPUs across multiple sites.

Early users

• Ineffable Intelligence, the London company founded by AlphaGo creator David Silver, was the first customer. It is building RL systems that learn continuously from experience.
• WPP ran its creative agent reference architecture on A5X with Nemotron models and OpenShell sandboxes. In WPP's initial tests, workloads that typically needed a multi-node cluster ran on a single A5X instance.

Confidential Computing at Google

Google Cloud will be among the first to offer rack-scale NVIDIA Confidential Computing on bare-metal instances (A4X Max, GB300 NVL72). Confidential Computing is now generally available on G4 VMs with RTX PRO 6000 Blackwell GPUs for confidential inference and fine-tuning.

Why it matters

Vera Rubin has mostly been a roadmap item at the big clouds. A5X is Google's first dated GA, giving frontier labs and RL workloads new capacity. There is no date yet for European regions, and the WPP result is the company's own test.