EU Picks EUROPA Consortium to Build 400B Parameter Sovereign AI
The European Commission selected the EUROPA consortium to build a sovereign AI model with over 400 billion parameters.
The project will produce an open-source AI covering all 24 official EU languages as a strategic priority.
Per the European Commission, EUROPA is the winner of the Frontier AI Grand Challenge launched in February 2026.
What Is the EU EUROPA Sovereign AI Project?
EUROPA is a European AI consortium led by Italian enterprise Domyn, selected to build a frontier AI model.
The model will exceed 400 billion parameters, placing it at the same scale as the world’s top AI systems.
It will be fully open-source and cover all 24 official EU languages for broad public and enterprise access.
The project is backed by a 6,000-chip NVIDIA Blackwell GPU cluster for training at unprecedented scale.
EUROPA also has access to up to 2.5% of EuroHPC’s high-performance supercomputing capacity for the project.
The goal is to demonstrate that Europe can build frontier AI on its own infrastructure and its own values.
Why the EU Is Building a Sovereign AI Model Through EUROPA
Europe has lagged behind the US and China in developing home-grown frontier AI model capabilities.
The EU AI Act, adopted in 2024, requires responsible AI but depends on models built outside the EU’s control.
A sovereign AI model ensures Europe is not dependent on OpenAI, Anthropic, Google, or Chinese AI providers.
EUROPA will be able to train on European data and reflect EU values around privacy, democracy, and rule of law.
Per EU Monitoring, the EUROPA model will be multilingual from the ground up rather than retrofitted.
Open-source release means any EU company or public institution can use and adapt the model freely.
EUROPA AI Technical Specs: 400B Parameters and NVIDIA Blackwell
The EUROPA model will exceed 400 billion parameters, a scale previously only achieved by closed proprietary models.
A 6,000-chip NVIDIA Blackwell cluster provides the compute needed for pre-training at that parameter count.
EuroHPC’s supercomputers supplement the NVIDIA cluster for specialized training and inference experiments.
Supporting all 24 EU languages requires significantly more training data and compute than English-only models.
The team includes AI researchers from across Europe assembled through the Frontier AI Grand Challenge process.
The scale connects to the big tech AI spending context where 400B parameter models are now a standard ambition.
How EUROPA Fits Into Europe’s Broader AI Sovereignty Strategy
The European Commission has committed to strengthening tech sovereignty alongside strict AI regulation.
EUROPA complements the EU AI Act by giving Europe a safe, compliant AI foundation it controls entirely.
The project is part of a broader package that includes investing in quantum computing, chips, and cloud infrastructure.
The EU AI Act enforcement context is covered in our EU AI Act breakdown article.
France’s Mistral, Germany’s Aleph Alpha, and now EUROPA represent Europe’s three-pronged sovereign AI strategy.
EUROPA is the most ambitious of the three, targeting frontier capability rather than efficiency or specialization.
What the EUROPA Sovereign AI Model Means for Developers and Enterprises
European companies will gain access to a frontier-class AI model without sending data to US or Chinese servers.
Open-source availability means developers can fine-tune EUROPA for specific industry and language applications.
Public sector institutions across the EU can use EUROPA for citizen services in their national languages.
Healthcare, legal, and financial AI applications that require EU data residency will benefit most from EUROPA.
The model is expected to be released in stages, with an initial version available before the end of 2027.
EUROPA signals that open-source sovereign AI is now a serious geopolitical priority, not just a research project.
