Europe Bolsters AI Independence with National Language Models
Amsterdam & Lisbon – Several European nations are actively developing their own large language models (LLMs) to foster technological sovereignty and compete in teh rapidly evolving artificial intelligence landscape.The Netherlands and Portugal are leading the charge with projects focused on creating AI capable of understanding and generating text in their respective national languages – Dutch and Portuguese.
these initiatives represent a broader European effort to reduce reliance on dominant US-based AI models and ensure data privacy,cultural relevance,and economic benefits remain within the continent. Unlike commercially available LLMs, these “sovereign AI” projects prioritize transparency, open-source accessibility, and reciprocal value for content creators.
In the Netherlands, three non-profit organizations launched GPT‑NL in 2023, aiming for a “reliable, transparent, reciprocal and sovereign” model for the dutch language and culture. The project utilizes a combination of copyrighted data obtained through agreements, public data, and synthetically generated data. A recent accord signed with dutch publishers, united within NDP Nieuwsmedia, and the press agency ANP, will allow GPT-NL to train on their articles in exchange for a share of future profits. Researchers began training the model in June 2025, with a first version anticipated before the end of the year. Access may require a small fee for non-professional use. Academic institutions, researchers, and public authorities will have open access to experiment with the model’s applications in fields like health, education, and public services.
Meanwhile, Portugal’s Amalia, a sovereign AI developed by a consortium of Portuguese universities since 2024, is capable of answering questions, generating code, summarizing texts, and interpreting information in portuguese, with a focus on local context. The nova School of Science and Technology confirmed these capabilities. Beta testing concluded in September, and public availability is targeted for mid-2026. The Portuguese government plans to integrate Amalia into public administration services and scientific analysis. While Amalia will not be released as a public chatbot, its code will be open source, enabling portuguese companies to leverage the LLM for their own AI applications.
