OpenNebula joins VAST Cosmos for AI factory infrastructure

OpenNebula Systems has joined the VAST Cosmos Community as a technology partner to pair its cloud orchestration platform with VAST Data’s AI operating system. The partnership is aimed at sovereign AI factories, gigafactories and neoclouds that need high-performance data services, multi-tenant controls and large-scale GPU infrastructure.

Why it matters: - AI factories and gigafactories need more than GPU capacity. - The joint approach is designed to combine compute, networking, virtualization, Kubernetes, bare metal and data services in one operating model. - The goal is to help organizations run large AI environments with stronger isolation, consistent performance and easier operations.

What happened: - OpenNebula Systems announced that it has joined the VAST Cosmos Community as a Technology Partner. - The collaboration links OpenNebula’s sovereign cloud and AI Factory management platform with VAST Data’s AI operating system. - The announcement was made Aug. 10, 2026, from Madrid. - The combined solution is aimed at AI Factories, Gigafactories and next-generation neoclouds.

The details: - The integration is built to support model training, fine-tuning, inference and scientific computing across GPU-heavy environments. - The architecture supports multiple deployment patterns depending on operational needs. - For performance-sensitive workloads, customers can use direct high-throughput access between compute and data. - For shared environments, storage can run through the virtualization layer so OpenNebula manages tenant-aware VM lifecycle, isolation and infrastructure policies. - VAST provides the underlying data services in that model. - The setup can also expose shared file services directly to workloads for virtual machines, containers and AI applications. - OpenNebula manages provisioning and lifecycle for compute resources. - VAST provides the scalable data foundation for training, inference, retrieval-augmented generation, checkpointing, model repositories and data-heavy scientific workloads. - The solution is designed for NVIDIA accelerated computing systems, including NVIDIA Grace Blackwell environments. - Supported deployment models include on-premises infrastructure, sovereign cloud, hybrid environments, commercial GPU clouds, research infrastructures and geographically distributed AI Factory deployments.

Between the lines: - The partnership reflects a shift from standalone GPU clusters toward policy-driven AI infrastructure that treats compute and data as shared services. - That matters for sovereign AI projects, where operators want control, tenant isolation and predictable performance without locking into one storage model. - The companies are also positioning the stack for organizations moving from HPC and virtualization into AI-native infrastructure. - OpenNebula Systems and VAST Data said they are already working on opportunities in Europe and international markets.

What's next: - The companies plan to define a joint reference architecture and deployment blueprints. - Those resources will combine accelerated computing, AI-native data services, virtualization, Kubernetes, workload orchestration and infrastructure automation. - The companies expect the materials to help customers move toward shared, policy-driven AI Factory operating models. - Target markets include national AI infrastructure, enterprise AI platforms, scientific computing environments, AI cloud services, HPC centers and neocloud providers.

The bottom line: - OpenNebula and VAST are packaging compute and data into one operating model for organizations building large, sovereign AI environments.