New Eclipse initiative tackles costly AI provider lock-in for enterprises

Organizations deeply invested in one AI provider often face prohibitive switching costs—rebuilding entire workloads, untangling integrations, and moving sensitive data. The Eclipse Foundation has just announced the Sovereign AI Foundation, a vendor-neutral initiative designed to help companies maintain flexibility and control over their AI infrastructure. The effort brings together 17 founding member organizations spanning Europe, Asia, and North America.

The lock-in problem#

Dependency on a single AI service creates compounding operational risks. As AI systems become embedded into mission-critical applications, the cost of pivoting to competitors rises sharply—requiring engineering teams to rewrite integrations, migrate data, document legacy decisions, and navigate compliance reviews. According to Mike Milinkovich, executive director of the Eclipse Foundation, "The deeper the dependency, the greater the potential risk. 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."

This dynamic mirrors patterns seen in earlier waves of vendor lock-in. The Sovereign AI Foundation aims to gather real-world evidence of these risks and develop strategies to mitigate them.

A focus on transparency and modularity#

The foundation operates from a premise that organizations need visibility into where their data flows, which models process it, and how easily they can change course if costs or vendor terms shift. Rather than advocating for a specific technology, Milinkovich emphasizes that the risk applies equally to AI systems from any region or provider—whether proprietary closed models or openly available ones. "The real question for our members is whether they can understand, adapt, and, if necessary, replace the systems they depend on."

Across existing Eclipse projects, a common pattern already emerges: modularity—keeping models, infrastructure, and applications sufficiently independent so that swapping one component doesn't trigger a full rebuild. Examples include EclipseSource's work on AI-enabled development tools, Eurotech's edge AI deployments for long-lifecycle equipment, and Deutsche Telekom's Eclipse LMOS framework for multi-agent systems.

Five workstreams#

The foundation's initial roadmap covers five areas: