# Google Cloud and NVIDIA make A5X instances generally available next month

Digest AI · Hardware & Compute · published 2026-10-21T10:45:17Z · updated 2026-10-21T11:24:12Z

Canonical: https://digestai.news/story/google-cloud-and-nvidia-make-a5x-instances-generally-available-next-mo

## Summary

Google Cloud and NVIDIA announced that A5X instances, built with NVIDIA Vera Rubin NVL72 GPUs, ConnectX-9 SuperNICs, the Virgo Network and fourth‑generation liquid cooling, will be generally available next month, with plans to expand to Europe. The platform is designed for demanding agentic inference and reinforcement‑learning workloads at scale.

Google aims to scale the Vera Rubin‑based infrastructure to as many as 960,000 Rubin GPUs across multiple sites. Ineffable Intelligence, founded by AlphaGo creator David Silver, is the first customer and is using the system to develop large‑scale “superlearner” reinforcement‑learning systems that continuously improve from experience. WPP showcased a creative‑agent reference architecture on A5X using NVIDIA Nemotron models and OpenShell, consolidating workloads that normally required a multi‑node cluster onto a single instance while providing governed sandboxes for brand‑safe execution. Additional customers such as Vertebrae, Polaris Cloud and Confidential Core AI are using Google’s Confidential Computing offerings on G4 VMs for privacy‑preserving inference and policy‑governed AI.

## Key points

- Google Cloud and NVIDIA will offer A5X instances with Vera Rubin NVL72 GPUs, ConnectX-9 SuperNICs, and liquid cooling, GA next month.
- Google plans to scale up to 960,000 Rubin GPUs across multiple sites; first customer Ineffable Intelligence uses them for large‑scale reinforcement learning.
- WPP demonstrated creative agents on A5X with Nemotron models and OpenShell, running workloads that previously needed multi‑node clusters on a single instance.

## Why it matters

The rollout provides cloud users massive compute and high‑bandwidth connectivity for reinforcement‑learning and agentic AI, lowering hardware barriers and simplifying large‑scale training and inference.

## Sources

1. [Google Cloud and NVIDIA Announce General Availability of NVIDIA Vera Rubin Infrastructure](https://blogs.nvidia.com/blog/google-cloud-vera-rubin-infrastructure) (NVIDIA Blog, 2026-10-21, primary source)
2. [Google Cloud’s First NVIDIA Vera Rubin A5X Instances Go Live](https://unite.ai/google-clouds-first-nvidia-vera-rubin-a5x-instances-go-live) (Unite.AI, 2026-10-21)

Part of the developing story: [Nvidia Eyes Reflection AI Amid Open Model Launch](https://digestai.news/thread/reflection-may-release-open-weight-model-to-rival-chinese-ai) (6 stories)

## Cite

Digest AI, "Google Cloud and NVIDIA make A5X instances generally available next month", 21 October 2026, https://digestai.news/story/google-cloud-and-nvidia-make-a5x-instances-generally-available-next-mo

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/google-cloud-and-nvidia-make-a5x-instances-generally-available-next-mo.json