Google Cloud and NVIDIA have announced the general availability of G4 Virtual Machines powered by NVIDIA RTX PRO 6000 Blackwell GPUs, alongside streamlined access to NVIDIA Omniverse via the Google Cloud Marketplace. This partnership between two of the most influential players in AI and cloud computing signals a major leap forward in enterprise infrastructure. This integration brings industrial-grade simulation tools into a scalable cloud environment, accelerating how enterprises build and deploy digital twins of factories, warehouses, and logistics hubs. With Omniverse now natively supported on Google Cloud, teams can model, simulate, and iterate complex physical systems faster, to reduce deployment cycles and improve operational planning across global infrastructure.

STRATEGIC BUSINESS IMPLICATIONS

Agentic AI, capable of autonomous reasoning and decision-making, demands significantly more compute than traditional Gen AI. The G4 VMs meet this demand with up to 9x the throughput of previous generations, allowing organizations to move from pilot projects to production-grade deployments. With Google Cloud enabling fractional GPU options and Multi-Instance GPU (MIG) support, businesses can optimize resource allocation and cost-efficiency across diverse workloads.

This infrastructure upgrade makes it feasible to deploy scalable, autonomous systems across industries. Enterprises can now build real-time applications that were previously constrained by hardware limitations, such as multimodal assistants, robotics simulations, and high-fidelity digital twins.

TECHNICAL IMPLICATIONS

The G4 VMs introduce FP4 precision and fifth-generation Tensor Cores, which significantly enhance throughput for LLM inference and simulation. These capabilities support real-time Generative AI and agentic reasoning, while fourth-generation Ray Tracing cores deliver over 2x the performance of prior generations for photorealistic rendering.

Additionally, Google Cloud’s custom Peripheral Component Interconnect Express (PCIe) peer-to-peer fabric accelerates multi-GPU communication, boosting performance for large-scale models exceeding 100B parameters without requiring code changes.

USE CASES ALREADY UNDERWAY

The combined strength of NVIDIA’s Blackwell-powered G4 VMs and Google Cloud’s scalable infrastructure is already delivering measurable impact across key enterprise use cases. In warehousing, digital twins now support simulations of larger, more intricate logistics environments with greater physical accuracy, thanks to expanded memory and rendering capabilities. In manufacturing, Gen AI industrial assistants benefit from faster AI-driven diagnostics and more detailed digital twins, improving operational insight and reducing downtime. This is just the beginning.

WHAT COMES NEXT

With capabilities for NVIDIA Omniverse™ enabling high-quality digital twin simulations, the next bridge is into physical AI to further elevate workflows. It creates a unified environment for simulation, robotics, and real-world operations.

Enterprises across the spectrum will increasingly favor embedded agents within platforms like NVIDIA Omniverse™ to simplify integration and reduce operational overhead. Made available through their existing preferred cloud providers, such as Google Cloud, these services offer the added benefits of enterprise-grade security and global scalability. This eliminates the need for costly data center investments, allowing organizations to test and scale these solutions efficiently while streamlining deployment and improving interoperability across systems.

FINAL THOUGHTS

The NVIDIA-Google Cloud announcement marks a turning point in enterprise AI infrastructure. Organizations that understand the operational and economic demands of agentic AI will be best positioned to lead in this new era.

SoftServe has deepened its leadership in industrial AI through two major milestones: a multi-year partnership with Google Cloud and its selection as NVIDIA’s 2025 Americas NPN Service Delivery Partner of the Year. The company earned this recognition for delivering eight Gen AI solutions built with NVIDIA Blueprints, including applications that integrate NVIDIA Omniverse™ and NIM microservices. As both a Google Cloud Premier Partner and NVIDIA Elite Service Delivery Partner, SoftServe brings a tested approach to deploying simulation-ready infrastructure.

Don't want to miss a thing?

Jack’s recent stories

NVIDIA-Google Cloud Announcement Brings Omniverse to Industrial AI Workloads