Last week, large language models (LLMs) made a significant entry onto Europe’s digital sovereignty agenda with the announcement of a new initiative aimed at developing a set of “truly” open-source LLMs that encompass all official languages of the European Union. This initiative not only includes the current 24 EU languages but also extends to those of countries in the process of negotiating EU membership, such as Albania, emphasizing a commitment to future-proofing.
The OpenEuroLLM project is a collaborative effort involving around 20 organizations, co-led by Jan Hajič, a computational linguist from Charles University in Prague, and Peter Sarlin, the CEO and co-founder of the Finnish AI lab Silo AI, which was acquired by AMD last year for $665 million. This initiative aligns with a broader trend of Europe prioritizing digital sovereignty, which aims to localize critical infrastructure and tools. Major cloud providers are investing in regional infrastructure to keep EU data within the EU, while OpenAI has recently launched a service enabling clients to process and store data in Europe.
Additionally, the EU has recently secured an $11 billion agreement to establish a sovereign satellite constellation to compete with Elon Musk’s Starlink, making the OpenEuroLLM project a fitting part of this narrative. However, the budget allocated specifically for developing the models is €37.4 million, with approximately €20 million sourced from the EU’s Digital Europe Programme — a relatively small amount compared to the investments made by major corporate AI players. The overall budget increases when considering funding for related work, with one of the largest costs being computational resources. The partners involved in the OpenEuroLLM project include EuroHPC supercomputer centers located in Spain, Italy, Finland, and the Netherlands, with the broader EuroHPC initiative boasting a budget of around €7 billion.
Despite this, the involvement of numerous diverse stakeholders from academia, research, and industry has raised questions about the feasibility of achieving the project’s ambitious goals. Anastasia Stasenko, co-founder of the LLM company Pleias, expressed skepticism about whether a “sprawling consortium of 20+ organizations” could maintain the same focused direction as a dedicated private AI firm. She noted, “Europe’s recent advancements in AI are evident in smaller, focused teams like Mistral AI and LightOn — companies that have true ownership over their creations.” Stasenko emphasized that these companies bear immediate responsibility for their decisions, whether related to finances, market strategy, or reputation.
