TLDR: Google Cloud has announced the general availability of its new G4 Virtual Machines, equipped with Nvidia’s RTX PRO 6000 Blackwell Server Edition GPUs and AMD EPYC Turin CPUs. These powerful VMs are designed to significantly accelerate enterprise and industrial AI workloads, offering substantial performance improvements for tasks ranging from generative AI and robotics simulation to high-fidelity design and digital twin development. The rollout also includes the availability of Nvidia Omniverse and Isaac Sim on the Google Cloud Marketplace, further enhancing capabilities for AI-driven industrial operations.
Google Cloud has officially rolled out its highly anticipated G4 Virtual Machines (VMs), marking a significant advancement in cloud-based AI infrastructure. These new VMs are now generally available, providing enterprises with enhanced capabilities for a wide spectrum of artificial intelligence and industrial digitalization workloads.
The G4 VMs are powered by Nvidia’s cutting-edge RTX PRO 6000 Blackwell Server Edition GPUs, combined with AMD’s EPYC Turin CPU platform. This powerful hardware configuration is engineered to deliver a profound leap in performance, with Google stating that the G4 VM offers up to 9x the throughput of previous G2 instances. Each G4 VM integrates eight RTX PRO 6000 GPUs, providing 3,753 teraFLOPS of sparse FP4 compute power and 768 GB of GDDR7 memory. This represents a 4x increase in compute and memory capacity and a 6x boost in memory bandwidth compared to the prior G2 generation.
These machines are optimized for a diverse range of compute-intensive tasks, including multi-modal AI inference, photorealistic design and visualization, and robotics simulation. They are also instrumental in accelerating generative AI applications, such as multimodal and text-to-image creation models, and improving the speed of training, fine-tuning, and inference for large language models.
Further expanding their utility, Nvidia Omniverse and Nvidia Isaac Sim are now available as virtual machine images through the Google Cloud Marketplace. These tools are crucial for powering physical, AI-driven applications across manufacturing, automotive, and logistics sectors, enabling the design of virtual factories, product modeling, and safe robot training in virtual environments.
Google Cloud emphasized that the G4 VMs run natively alongside Google’s existing tools like Vertex AI and Kubernetes Engine, making it easier for developers to build smarter agents, accelerate scientific discovery, and create photorealistic 3D visuals on a unified platform. This collaboration between Nvidia and Google Cloud is seen as a strategic move to democratize AI and fuel the growth of the AI infrastructure market, which is projected to expand significantly in the coming decade.
Also Read:
- The Unprecedented Boom in AI Server Market: Key Players and Future Trajectories
- Google Unveils LLM-Evalkit: A Structured Approach to Prompt Engineering for Large Language Models
According to Google, ‘The G4 VM offers a profound leap in performance, with up to 9x the throughput of G2 instances, enabling a step-change in results for a wide spectrum of workloads, from multi-modal AI inference, photorealistic design and visualization, and robotics simulation using applications developed on NVIDIA Omniverse.’ Nvidia also noted that ‘These latest announcements establish a complete, end-to-end platform built on the Nvidia Blackwell platform — from Nvidia GB200 NVL72 and Nvidia HGX B200 for massive-scale AI training and inference, to the RTX PRO 6000 Blackwell for AI inference and visual computing on G4 VMs.’ This partnership strengthens Google Cloud’s AI infrastructure and solidifies Nvidia’s position as a leading provider of end-to-end computing solutions for enterprise transformation.
