NVIDIA and Google Cloud are expanding access to accelerated computing to transform the full spectrum of enterprise workloads, from visual computing to agentic and physical AI.
Google Cloud today announced the general availability of G4 VMs, powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Plus, NVIDIA Omniverse and NVIDIA Isaac Sim are now available as virtual machine images (VMIs) on the Google Cloud Marketplace to unlock physical AI-driven applications for key industries like manufacturing, automotive and logistics.
This powerful combination creates a versatile, multi-workload platform for enterprises to accelerate their most demanding challenges on Google Cloud.
NVIDIA RTX PRO 6000 Blackwell GPUs excel at high-performance AI inference for multimodal, generative and agentic AI deployments, while also powering complex visual and simulation workloads ranging from computer-aided engineering and content creation to robotics simulation.
Customers like WPP are using G4 VMs with NVIDIA Omniverse to instantly generate photorealistic 3D advertising environments at global scale, while Altair is using the platform within Altair One to accelerate demanding simulation and fluid dynamics workloads.
The G4 VM: A Universal Platform for AI and Visual Computing
At the core of the new G4 VM is the NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, the ultimate data center GPU for AI and visual computing.
Built on the NVIDIA Blackwell architecture, it serves as a universal platform for a broad range of workloads. Its design uniquely combines two powerful engines:
- Fifth-Generation Tensor Cores that deliver a massive leap in AI performance, supporting new data formats like FP4 to enable faster performance with lower memory usage.
- Fourth-Generation RT Cores that provide over 2x the real-time ray-tracing performance over the previous generation, enabling cinematic-quality graphics and photorealistic simulations.
On Google Cloud, these G4 VMs are built for massive scale, configurable with up to eight RTX PRO 6000 GPUs — totaling 768 GB of GDDR7 memory — and supported by high-throughput local and network storage.
As part of Google Cloud’s AI Hypercomputer architecture, G4 VMs natively integrate with services like Google Kubernetes Engine and Vertex AI to simplify containerized deployments and streamline machine learning operations for AI workloads. This flexibility extends to accelerating large-scale data analytics on Apache Spark and Hadoop with Dataproc.
For design and simulation, the VMs also support a broad ecosystem of popular third-party engineering and graphics applications like Autodesk AutoCAD, Blender and Dassault SolidWorks.
Industrial Digitalization at Scale With NVIDIA Omniverse
Google Cloud customers can now tap into NVIDIA Omniverse, a collection of integration-ready libraries and frameworks for developing industrial digitalization applications built on Universal Scene Description (OpenUSD).
With the availability of the Omniverse VMI and G4 VMs on Google Cloud, enterprises can:
- Build and Operate Digital Twins: Easily create physically accurate, real-time virtual replicas of factories and products to simulate and optimize operations. These workflows are powered by the NVIDIA Cosmos world foundation model platform and NVIDIA Omniverse Blueprints for creating digital twins.
- Accelerate Robotics Development: Use NVIDIA Isaac Sim, a reference application built on Omniverse, to train, simulate and validate AI-driven robots in physics-based virtual environments before deployment.
NVIDIA AI Accelerates Every Enterprise Workload
In addition to Omniverse, Google Cloud customers can use the full NVIDIA software stack to accelerate a range of high-demand workloads, including:
- Agentic AI: Developers can use the NVIDIA Nemotron family of open reasoning models and NVIDIA Blueprints to get started quickly with building sophisticated AI agents that can reason and act. For deployment, NVIDIA NIM — a set of easy-to-use microservices — provides optimized, high-performance inference for AI models with enterprise-grade security and support.
- Scientific and High-Performance Computing: Solve complex problems in fields like drug discovery and genomics using NVIDIA CUDA-X libraries and microservices. On the RTX PRO 6000 Blackwell GPU, core genomics algorithms used in sequence alignment can see up to 6.8x faster throughput compared with the previous generation.
- Design and Visual Computing: Power remote creative and design pipelines with NVIDIA RTX Virtual Workstation software, which delivers high-performance virtual workstation instances from G4 VMs to any device, anywhere.
These latest announcements establish a complete, end-to-end platform built on the NVIDIA Blackwell platform — from NVIDIA GB200 NVL72 (A4X VMs) and NVIDIA HGX B200 (A4 VMs) for massive-scale AI training and inference, to the RTX PRO 6000 Blackwell for AI inference and visual computing on G4 VMs.
This unified architecture provides a seamless experience for accelerating every workload, empowering enterprises to tackle complex, multistage pipelines from data analytics to physical AI within a single, consistent cloud ecosystem.
Learn more about G4 VMs on Google Cloud, and deploy the NVIDIA Omniverse and Isaac Sim VMIs from the Google Cloud Marketplace. Explore NVIDIA Nemotron models and NVIDIA Blueprints