OpenNebula and Waldur Unite to Power Federated European AI Gigafactories

Two Open Source Platforms, One European Vision

In a significant step for European digital sovereignty, open source cloud management platforms OpenNebula and Waldur have announced a technical integration aimed at powering federated European AI gigafactories. The collaboration positions both projects at the heart of Europe's ambition to build sovereign, scalable AI infrastructure — free from dependence on US or Chinese hyperscalers — by combining their respective strengths in cloud orchestration and resource management into a unified, interoperable stack.

For IT decision makers, developers, and policy professionals watching Europe's evolving technology landscape, this integration is more than a product announcement. It represents a concrete technical response to the European Union's push for AI sovereignty, data localisation, and federated compute capacity — goals that have gained urgency as AI workloads grow exponentially and regulatory pressure around data residency intensifies under frameworks like the GDPR and the EU AI Act.

OpenNebula is a battle-tested, open source cloud and edge computing platform used by enterprises, research institutions, and telcos across Europe and beyond. Waldur, developed by the Estonian-based company Estonia's CloudLinux spinout UT/uleandur team and widely adopted in academic and research cloud environments, specialises in self-service cloud management, resource accounting, and multi-provider orchestration. Together, they address a critical gap: the ability to federate distributed AI compute resources across borders while maintaining governance, accountability, and compliance at each node.

What Exactly Is an AI Gigafactory — and Why Does Europe Need One?

The term "AI gigafactory" — borrowed from the manufacturing world, where it describes massive production facilities — refers in this context to large-scale, high-performance computing facilities specifically designed to train and run advanced AI models. The concept has gained traction across European policy circles following calls from industry leaders and EU officials for the continent to develop indigenous AI compute capacity capable of competing with the massive data centre investments being made in the United States and China.

Europe currently lags in raw AI compute capacity. According to research tracked by the European Commission's Joint Research Centre, the majority of the world's frontier AI model training takes place in hyperscaler facilities located outside the EU — raising serious concerns about data sovereignty, GDPR compliance, and strategic dependency. The EU's AI Act, which introduces risk-based regulation of AI systems, adds further urgency: organisations processing sensitive data through non-European infrastructure face increasing legal and reputational exposure.

The gigafactory model addresses this by aggregating smaller, distributed compute clusters — spread across universities, national supercomputing centres, and private data centres — into a coherent, federated resource pool. Importantly, this federation must happen without forcing all data and workloads through a single centralised point, which would recreate the very dependency problem Europe is trying to solve. This is precisely where OpenNebula and Waldur's combined capabilities become relevant.

How the OpenNebula–Waldur Integration Actually Works

The technical integration between OpenNebula and Waldur creates a layered architecture designed for federated multi-site deployments. OpenNebula handles the low-level cloud orchestration layer — managing virtual machines, containers, storage, and networking across heterogeneous hardware — while Waldur operates at the service management layer, providing self-service portals, quota management, billing, and multi-tenancy governance.

In practical terms, this means that a research institution in Helsinki and a private data centre in Barcelona can both contribute compute resources to a federated AI gigafactory, with Waldur managing user access, resource allocation, and cost accounting across the federation, and OpenNebula managing the actual workload execution at each site. The data never needs to leave its local jurisdiction unless the workload explicitly requires it — a critical feature for GDPR compliance and national data sovereignty requirements.

The integration also leverages standardised APIs, reducing the vendor lock-in that has plagued many public sector cloud projects in Europe. Because both platforms are fully open source, organisations can audit the code, customise deployments, and ensure that no proprietary dependencies introduce hidden compliance risks. This openness is increasingly valued by public procurement teams across the EU, who are under growing pressure to justify cloud spending under the European Data Governance Act and related frameworks.

OpenNebula's architecture supports both KVM-based virtualisation and LXC containers, with growing support for GPU passthrough — essential for AI training workloads. Waldur's recent development roadmap, as detailed on its official platform documentation, has emphasised research cloud infrastructure and HPC (high-performance computing) integration, making the two platforms technically complementary rather than overlapping.

Why Federated European AI Infrastructure Matters for Digital Sovereignty

The broader context for this integration is Europe's accelerating push toward what the European Commission calls "technological sovereignty" — the ability to develop, deploy, and govern critical digital infrastructure without strategic dependence on foreign providers. This agenda has been formalised through initiatives including the European Cloud Federation, the Gaia-X project, the European High Performance Computing Joint Undertaking (EuroHPC JU), and most recently through calls for dedicated AI gigafactories as part of the EU's AI strategy.

Gaia-X, despite its turbulent early years and criticism over slow progress, established important groundwork for federated data space standards and interoperability rules that projects like the OpenNebula–Waldur integration can now build upon. According to reporting by Politico Europe's tech desk, European policymakers have increasingly prioritised funding for open source cloud alternatives that align with Gaia-X principles, viewing proprietary lock-in to US hyperscalers as both a regulatory and a geopolitical risk.

The EuroHPC JU, which oversees Europe's network of supercomputing centres, has similarly been expanding its mandate toward AI workloads, investing in petascale and exascale machines that could form the backbone of AI gigafactories. A federation layer like the one OpenNebula and Waldur are building would allow these national supercomputing assets to be more effectively pooled and allocated across research and commercial use cases, maximising the return on public infrastructure investment.

What This Means for Developers, Enterprises, and Public Sector Organisations

For developers and infrastructure teams, the OpenNebula–Waldur integration means access to a credible, production-ready open source stack for building or joining federated AI compute environments — without the compliance headaches associated with hyperscaler dependencies. Both platforms have mature APIs, active communities, and documented integration patterns that make them practical choices for real deployments, not just research prototypes.

For enterprises and small businesses operating under GDPR, the federated model offers a path to using powerful AI infrastructure while keeping sensitive training data within EU borders — or even within a specific member state if national regulations require it. This is particularly relevant for sectors like healthcare, finance, and legal services, where data localisation requirements are strictest and the consequences of non-compliance are most severe.