Google Cloud Rolls Out G4 VMs with NVIDIA RTX PRO 6000 Blackwell GPUs and Marketplace AI Tools

October 20, 2025

Details

  • Google Cloud has launched G4 virtual machines powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, now generally available as of October 2025, and made NVIDIA Omniverse and Isaac Sim accessible as virtual machine images on the Google Cloud Marketplace.
  • The new G4 VMs are tailored for demanding AI inference and visual computing tasks, addressing the needs of sectors such as manufacturing, automotive, and logistics. Early enterprise users include WPP, leveraging the platform for photorealistic 3D ad content, and Altair, which is using it to enhance simulation and fluid dynamics workloads within its engineering suite.
  • These VMs feature up to eight GPUs per instance and boast a total of 768 GB of GDDR7 memory. The Blackwell architecture includes fifth-generation Tensor Cores for advanced AI acceleration and fourth-generation RT Cores that more than double ray-tracing performance versus earlier models.
  • NVIDIA Omniverse on Google Cloud gives businesses the tools to build accurate digital twins of products and facilities using Universal Scene Description (OpenUSD), while Isaac Sim supports robotic AI training and validation in realistic, physics-driven simulations. Omniverse Blueprints and NVIDIA Cosmos further underpin these capabilities.
  • The ecosystem incorporates leading third-party creative and engineering applications and is tightly integrated with Google Cloud’s AI Hypercomputer, Google Kubernetes Engine, and Vertex AI for streamlined machine learning operations. Major genomics workloads, like sequence alignment, see speedups of up to 6.8 times over legacy solutions.

Impact

This expansion rounds out Google Cloud’s Blackwell-based infrastructure, positioning it as a leader in cloud GPU innovation for digital industries. With these offerings, Google Cloud remains ahead of the curve in delivering powerful GPU instances and AI tools compared to its main cloud rivals. The move is set to accelerate adoption of digital twins, robotics, and AI-powered automation in industries increasingly seeking cloud flexibility and robust simulation capabilities.

Read more