Dec. 12, 2025
Strategic access to EuroHPC resources granted to OpenEuroLLM
The EU has set out a comprehensive roadmap to secure European technological and economic sovereignty in AI by building its own world‑class AI infrastructure, data space and talent base. Aligned to EU goals, OpenEuroLLM is the first AI project to be granted strategic access across multiple EuroHPC centres, in the amount of over 10 million GPU hours.
“The compute is granted at four of the EuroHPC’s supercomputers, LUMI, Leonardo, Jupiter, and Marenostrum5. These compute resources enable the continuation of the OpenEuroLLM project towards its stated goals.” says Jan Hajič, the OpenEuroLLM project leader. Work conducted thanks to this compute allocation will help advance all of OpenEuroLLM’s objectives, with a main focus on preparation for the coming flagship models for all official EU languages. Experimentation with smaller models informs the architecture choices that will be applied to larger models.
“We appreciate EuroHPC’s recognition of both the importance of OpenEuroLLM and the unique computational needs of the project. This compute allocation will allow the project to continue and expand its efforts to build the next generation of fully open LLMs for all European languages.” comments Sampo Pyysalo, group lead at the University of Turku and technical lead in OpenEuroLLM.
The collaborative spirit of open tech ecosystems in Europe enables innovation and strengthens competitiveness. Open models are a crucial building block for developing sovereign AI capabilities. They enable nations and companies to fine-tune models for specific use cases, while remaining in control of the models. This ensures transparency and trustworthiness of the models. The OpenEuroLLM consortium is committed to contributing to the advancement of open, multilingual LLMs and supporting innovation across the European tech ecosystem.
“The compute is granted at four of the EuroHPC’s supercomputers, LUMI, Leonardo, Jupiter, and Marenostrum5. These compute resources enable the continuation of the OpenEuroLLM project towards its stated goals.” says Jan Hajič, the OpenEuroLLM project leader. Work conducted thanks to this compute allocation will help advance all of OpenEuroLLM’s objectives, with a main focus on preparation for the coming flagship models for all official EU languages. Experimentation with smaller models informs the architecture choices that will be applied to larger models.
“We appreciate EuroHPC’s recognition of both the importance of OpenEuroLLM and the unique computational needs of the project. This compute allocation will allow the project to continue and expand its efforts to build the next generation of fully open LLMs for all European languages.” comments Sampo Pyysalo, group lead at the University of Turku and technical lead in OpenEuroLLM.
The collaborative spirit of open tech ecosystems in Europe enables innovation and strengthens competitiveness. Open models are a crucial building block for developing sovereign AI capabilities. They enable nations and companies to fine-tune models for specific use cases, while remaining in control of the models. This ensures transparency and trustworthiness of the models. The OpenEuroLLM consortium is committed to contributing to the advancement of open, multilingual LLMs and supporting innovation across the European tech ecosystem.