The European Union has launched three generative AI projects with €36.8 million in total funding—half of it from Brussels—aimed at deploying AI tools across public administrations by 2029-2030. The initiative faces a familiar question: can the EU encourage innovation without repeating past mistakes, where EU-backed schemes promised breakthroughs but delivered little more than rebranded consultancy?
The projects span climate disaster management, cross-border AI adoption, and what one consortium calls an escape from "AI fast food" imported from non-EU providers like ChatGPT, Copilot, and Gemini. Each reflects Brussels' dual ambition: digital sovereignty and practical public-sector modernisation. Whether €18.4 million in EU funds can achieve both remains to be seen.
The Three Projects
The "Floods and droughts" consortium is led by Italy with partners from France, Spain, and Greece. It's valued at €8.8 million, with €4.4 million from the EU, and is expected to conclude by 30 June 2030—four years later. The project aims to give public administrations Gen-AI tools for anticipating climate disasters, managing crises, and communicating with citizens during floods and droughts.
"Eunomia AI" is led by Spain and includes 13 more EEA countries: Austria, Estonia, Finland, Greece, Portugal, Latvia, Germany, the Netherlands, Poland, Italy, Hungary, Norway, and Malta. Total value: €13.9 million, with €6.9 million from the EU. The project ends 30 June 2029—three years later—with first pilots expected in one and a half years. Its goals are to deploy GenAI in 25 public administrations and at least 15 public services, and to reinforce European tech sovereignty by using EU open-source models. The project description says this will be achieved by procuring and hosting "centralized, high-performance GenAI infrastructure" at the University of Granada, with an option for federated local nodes.
The third project, "EuropAI," is led by the Netherlands with Denmark, Belgium, and Luxembourg. It costs €13.9 million—€6.9 million from EU funding—and is also expected to finish 30 June 2029, three years later. Its goals are to accelerate adoption of sovereign and trustworthy GenAI solutions across European governments by creating a framework for member states to co-create, test, and share GenAI solutions in three domains: legislative and administrative simplification, urban analysis and spatial planning, and digital assistance for citizens.
The 'AI Fast Food' Problem
The EuropAI consortium describes European public administrations as relying on imported "AI fast food" from non-EU providers. The plan is to offer member states a trusted "community kitchen where Member States co-create, test, and share GenAI solutions." The system's backbone is built on Denmark's OS2ai platform, which already serves more than 28,000 employees, and the Netherlands' GovChat-NL. The consortium describes ready-made tools as "pre-cooked meals ready to serve," reusable building blocks as European-made AI models, blueprints that let other countries replicate what works, compliance checks as "nutrition labels," and an app marketplace as "an organised pantry."
The Track Record Question
The EU faces a delicate balancing act: encouraging AI adoption across public and private sectors without funding a wave of "AI-washing" projects that promise innovation but deliver little, as happened with EU-backed schemes in the past. The total EU contribution of €18.4 million is modest by Brussels standards, but the risk isn't just wasted money—it's wasted credibility. If these projects don't produce tangible results by 2029-2030, the EU's digital sovereignty agenda will look like another exercise in branding over substance.
Why This Matters:
European governments can't afford to fall further behind in AI adoption, but they also can't afford to throw money at projects that don't deliver. The €18.4 million in EU funding is a test of whether Brussels can support genuine innovation or whether it'll repeat the pattern of past schemes that produced glossy reports and little else. The focus on "tech sovereignty" is welcome—Europe shouldn't be entirely dependent on American platforms—but sovereignty means nothing if the tools don't work or if member states can't deploy them efficiently. The real question isn't whether the EU can fund AI projects. It's whether these projects will actually be used by the 25 public administrations and 15 public services they're supposed to serve. If they're not, the EU will have spent €18.4 million proving that good intentions and PowerPoint decks aren't a substitute for execution. The story was originally published on EU Tech Loop and shared on Euronews as part of a syndication agreement.