“NVIDIA and Google Cloud Collaborate on AI Factories”

NVIDIA and Google Cloud have deepened their partnership with a series of announcements at Google Cloud Next 2026, focusing on new GPU infrastructure, confidential computing, and Google’s revamped enterprise agent stack. This move is part of a broader strategy to advance physical AI factories. Google is shifting its enterprise AI focus towards autonomous agents, consolidating Vertex AI and related services under the Gemini Enterprise umbrella.

Google revealed plans to offer A5X bare-metal instances based on NVIDIA Vera Rubin NVL72 systems, making it one of the first cloud providers to do so when the platform becomes available later in 2026. This infrastructure design can scale to support up to 80,000 Rubin GPUs in a single-site cluster and 960,000 across a multisite cluster, catering to large training and inference environments rather than traditional enterprise AI projects.

More than just a numerical milestone, this move signals a shift in demand towards integrated infrastructure for agents, simulation, robotics, and digital twins. NVIDIA and Google are aligning their offerings to meet the evolving needs of industries, focusing on engineering and manufacturing workflows rather than just data-center applications.

Enhancing Security in Physical AI Factories

Security is a critical aspect of deploying physical AI factories. NVIDIA and Google are expanding the deployment options by previewing Gemini on Google Distributed Cloud, powered by NVIDIA Blackwell and Blackwell Ultra GPUs. This setup is targeted at customers requiring advanced models closer to sensitive data. Additionally, Google announced the global preview of Confidential G4 VMs with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, providing hardware-backed security for regulated or multi-tenant AI workloads.

Empowering Industrial Workflows with Agents and Models

On the software front, Google introduced the Gemini Enterprise Agent Platform at Next, offering a unified environment for developing, managing, and scaling agents. NVIDIA is integrating Nemotron 3 Super models and NeMo-based training tools into this platform, creating a seamless path from model development to production deployment. This integration is particularly beneficial for industrial users leveraging robotics simulation, digital twins, vision systems, and workflow automation.

While the recent announcements primarily focus on platform alignment rather than immediate product releases, they provide insight into the strategic positioning of both companies. Google Cloud aims to become the go-to platform for enterprise agents, while NVIDIA is positioning itself as the infrastructure and model toolkit provider for the next generation of physical AI factories. The collaboration between NVIDIA and Google Cloud is poised to drive innovation in AI infrastructure and services, catering to the evolving needs of industries embracing advanced technologies.