Google Cloud has significantly advanced its artificial intelligence offerings by rolling out new virtual machines, the G4 series, which integrate Nvidia's powerful RTX Pro 6000 Blackwell GPUs. This strategic enhancement aims to supercharge a diverse range of enterprise AI applications, including multimodal AI inference, detailed photorealistic design, and complex robotics simulations. Concurrently, Google Cloud has made Nvidia Omniverse broadly accessible as a virtual machine image on its Marketplace, along with Nvidia Isaac Sim, providing robust tools for sophisticated digital twin and robotics architectures. This move underscores a deepening collaboration between the two tech giants, fostering an integrated platform for accelerating AI workloads within a unified cloud environment.
Google Cloud Unveils High-Performance AI Virtual Machines and Expands Nvidia Collaboration
On October 22, 2025, Google Cloud announced the general availability of its new G4 Virtual Machines (VMs), marking a significant milestone in its ongoing partnership with Nvidia. These cutting-edge VMs are equipped with Nvidia's RTX Pro 6000 Blackwell Server Edition GPUs, promising a substantial boost in performance for advanced artificial intelligence applications across various industries. Simultaneously, Google Cloud has made Nvidia Omniverse, a platform for building and operating metaverse applications, and Nvidia Isaac Sim, a robotics simulation tool, widely available as virtual machine images on the Google Cloud Marketplace.
According to Roy Kim and Dai Vu from Google Cloud, the G4 VMs deliver up to nine times the throughput compared to previous G2 instances. This remarkable performance leap is set to transform operations for multimodal AI inference, hyper-realistic visualization, and robotic simulations that leverage applications developed on Nvidia Omniverse. Initially, these G4 VMs are available with configurations of one, two, four, or eight RTX Pro Blackwell GPUs, collectively offering an impressive 768 GB of GDDR7 memory. Plans for fractional GPU options are also in the pipeline.
Nvidia highlights that the Blackwell GPUs' design features fifth-generation tensor cores for unparalleled AI performance and fourth-generation RT cores that enable cinematic-grade graphics and highly accurate photorealistic simulations. This technological synergy allows Google Cloud customers to achieve enhanced efficiency and flexibility. For instance, the G4's ability to divide a single GPU into four isolated instances, each with dedicated high-bandwidth memory and compute cores, facilitates the concurrent execution of multiple smaller workloads on a single VM. This capability supports the scaling of AI models ranging from under 30 billion to over 100 billion parameters, providing immense adaptability.
The integration of Nvidia Omniverse into the Google Cloud Marketplace further strengthens the platform for enterprise users. As Kim and Vu elaborated, G4 VMs offer the essential infrastructure for demanding real-time rendering and physically accurate simulations required for industrial digital twins. Together, Google Cloud and Nvidia provide a scalable cloud ecosystem for developing, deploying, and interacting with sophisticated applications for digital twins and robotics simulations, establishing a comprehensive, end-to-end platform built on Blackwell's unified architecture.
The Future of AI: Unlocking Unprecedented Capabilities with Cloud-GPU Synergy
The introduction of Google Cloud's G4 VMs with Nvidia Blackwell GPUs is more than just a hardware upgrade; it represents a pivotal moment in the evolution of enterprise AI. This development provides a powerful, accessible infrastructure that democratizes advanced AI capabilities, allowing businesses of all sizes to tackle previously intractable problems. The emphasis on multimodal AI, photorealistic design, and robotics simulation signals a shift towards more immersive and intelligent applications. This synergistic relationship between cloud providers and GPU manufacturers is crucial for driving innovation, enabling faster development cycles, and fostering a new generation of AI-driven solutions that will redefine industries and human-computer interaction.
