The Eclipse Foundation launches the Sovereign AI Foundation with 17 organisations to help members escape AI vendor lock-in

Picking a second AI provider a year into a build can mean rebuilding the integrations, moving the data, re-documenting the system and re-testing every downstream feature that ever touched the first vendor's model, tool shape or prompt idioms.

The Eclipse Foundation, one of the largest open-source software foundations, announced the Sovereign AI Foundation on 30 September 2026 as a vendor-neutral group whose job is to help member organisations map those dependencies before the switching cost gets large. It launches with 17 participating organisations, and the Eclipse Foundation says more are in the process of joining.

Who is in it

The 17 initial members, per the Eclipse Foundation's own announcement, are CEA LIST, EclipseSource, Engineering Ingegneria Informatica, Ericsson, Eurotech, Infosys, Kentyou, KU Leuven, Open Elements, Red Hat, Renesas Electronics, Robert Bosch, Thales, The IO Foundation, TypeFox, the University of York and Vector Informatik. The group spans automotive, telecommunications, aerospace and defence, research institutes and open-source vendors, and most of the founding members are based in Europe.

Mike Milinkovich, executive director of the Eclipse Foundation, told The New Stack that the risk of not adopting an open approach depends on how deeply AI is embedded in an organisation. "The deeper the dependency, the greater the potential risk," Milinkovich said. "But however embedded the underlying systems are, the costs show up in predictable places: higher prices that make switching difficult, and the challenge of rebuilding integrations, moving data and workloads, documenting systems after the fact, compliance delays, and operational disruption."

What the group will actually do

The Sovereign AI Foundation says it will not build software. Its work sits across six areas: mapping how members deploy and pilot open-source AI, assessing models and evaluation frameworks, evaluating existing Eclipse projects against member needs, sharing lessons from real deployments, producing whitepapers and blueprints, and turning EU-funded research into practical guidance.

The related Eclipse projects the group can point members at include Eclipse Theia and Theia AI for AI-enabled developer tools, Eclipse Enclave for running AI coding agents in isolated policy-controlled environments, Eclipse PanEval for evaluating model capability and safety, Eclipse LMOS for building multi-agent systems, and the Open VSX Registry, which the Foundation says handles close to one billion extension downloads a month. The Foundation also says it is looking at a broader AI Registry for Model Context Protocol servers, agent skills and plugins.

Milinkovich pushed back on the reading that this is a European or China-focused initiative. "This is not about China or any single provider. The same questions apply to AI technologies from the United States, Europe, Asia, or anywhere else," he told The New Stack. "Closed models can limit visibility, while access to model weights alone does not provide the transparency, control, or the independent governance that organisations may need."

The bigger point for a team about to pick a model, an agent framework or an MCP tool for a system that will run for years: the switching cost is a real number, and it grows quietly with every integration you add. Naming it out loud, before the next stack decision, is what this group is asking members to do.

Source

• Eclipse Foundation Launches the Sovereign AI Foundation, Eclipse Foundation, 30 September 2026
• Eclipse wants companies to be free to switch AI providers, The New Stack, 30 September 2026