Archive — US — techcrunch
Open source LLMs hit Europe’s digital sovereignty roadmap
A new initiative, OpenEuroLLM, aims to create truly open-source large language models (LLMs) for all EU languages, backed by €37.4 million in funding and collaborations from approximately 20 organizations. Project leaders emphasize the importance of digital sovereignty while addressing challenges in collaboration and resource allocation compared to corporate AI investments.
Key points
- OpenEuroLLM aims to develop multilingual open-source LLMs for all 24 EU official languages and candidates like Albania, promoting digital sovereignty in Europe.
- The project is co-led by Jan Hajič and Peter Sarlin and has a budget of €37.4 million, with €20 million funded by the EU, although significant corporate funding is absent.
- Previous related projects, like HPLT, provide foundational datasets and expertise, allowing OpenEuroLLM to advance without starting from scratch.
- Critics question the project's viability and whether a large consortium can match the speed and focus of smaller AI firms.
- OpenEuroLLM is designed to deliver high-quality, multilingual LLMs while adhering to EU regulations on AI and data sharing.
Why it matters
- The rise in European initiatives like OpenEuroLLM aligns with the 2025 trend towards digital sovereignty, emphasizing local data control and infrastructure.
- The competition between OpenEuroLLM and similar projects, like EuroLLM, reflects the fragmented nature of European AI development amidst larger corporate players.
- As more cloud giants adapt to local regulations by establishing EU infrastructure, projects like OpenEuroLLM potentially provide competitive alternatives for European enterprises.
The OpenEuroLLM initiative reflects Europe's commitment to digital sovereignty by developing a robust, multilingual framework for open-source AI, although its success hinges on navigating collaboration challenges and resource allocation in a competitive landscape.