← Back
NVIDIA Blog

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

•4 min read•
#deployment#nvidia#compute
Level:Advanced
For:AI Engineers
✦TL;DR

Google Cloud and NVIDIA announced the general availability of A5X instances built on NVIDIA Vera Rubin NVL72 GPUs, combining high‑bandwidth ConnectX‑9 SuperNICs, Virgo networking, and advanced liquid‑cooling for demanding agentic inference and reinforcement‑learning workloads. The platform will scale to as many as 960,000 Rubin GPUs across multiple sites, with initial customers like Ineffable Intelligence and WPP already leveraging the infrastructure for large‑scale RL and safe AI agents. Confidential Computing is also GA on bare‑metal A4X Max instances (NVIDIA GB300 NVL72) and G4 VMs (NVIDIA RTX PRO 6000 Blackwell Server Edition), enabling privacy‑preserving inference and fine‑tuning. Engineers can now provision these integrated compute‑network‑storage systems to accelerate continuous learning loops and secure agent deployments at cloud scale.

⚡ Key Takeaways

  • Planned scale of up to 960,000 Rubin GPUs across multiple sites.
  • A5X instances integrate NVIDIA Vera Rubin NVL72, NVIDIA ConnectX‑9 SuperNICs, Google Cloud Virgo Network, and fourth‑generation liquid‑cooling.
  • WPP’s creative agent reference architecture runs NVIDIA Nemotron models within NVIDIA OpenShell sandboxes on a single A5X instance, replacing multi‑node clusters.
  • Confidential Computing GA on A4X Max (NVIDIA GB300 NVL72) and G4 VMs (NVIDIA RTX PRO 6000 Blackwell Server Edition) for secure inference and fine‑tuning.
  • First customer Ineffable Intelligence uses A5X for large‑scale reinforcement learning and “superlearner” systems.
💡 Why It Matters

The integrated A5X platform lets engineers collapse multi‑node RL pipelines onto a single instance while maintaining high‑throughput, low‑latency data movement, and the new confidential‑computing offerings protect proprietary models and data in production deployments.

✅ Practical Steps

  1. Apply the concepts from this article to your own system design.

Want the full story? Read the original article.

Read on NVIDIA Blog ↗

More like this

Introducing workload recovery and expanded infrastructure checks on Amazon SageMaker HyperPod

AWS ML Blog•#deployment

Introducing Unity Gateway Insights

Databricks Blog•#deployment

Introducing Gemini 4 Neon

Google DeepMind•#llm

DaisyRose-27B: Pushing the Frontier of Post-Training Efficiency

Hugging Face Blog•#llm

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING