Three years after OpenAI rolled out ChatGPT and brought artificial intelligence (AI) to the mainstream, several European countries are making their own sovereign systems.<\/p>\n
Sovereign AI is a country\u2019s ability to develop, host, deploy, and govern AI systems that are made in its country for its citizens instead of being dependent on foreign systems or cloud jurisdictions.<\/p>\n
The European Parliament recognised in a June report<\/a> that it is \u201ccurrently heavily dependent on foreign technologies,\u201d specifically American technologies, which prevent the bloc from having its own tech champions. Europe\u2019s reliance \u201clooks set to continue,\u201d the report said, due to the US\u2019s recent $500 billion (\u20ac432.9 billion) domestic AI investment<\/a>. <\/p>\n The EU stated that to regain its advantage, it should invest in research and develop new systems. This is where national governments can come into play.<\/p>\n A handful across Europe are building their own sovereign AI systems. Euronews Next takes a look at what has been built so far.<\/p>\n Germany<\/p>\n Germany is the latest country to announce its own AI plan, called Sovereign Open Source Foundation Models (SOOFI).<\/p>\n SOOFI is an attempt<\/a> to build a fundamental \u201cadvanced AI\u201d open-source model that can be adapted by other companies building AI products, according to the German government. <\/p>\n The technology will be used for highly complex tasks, such as AI-controlled robots, according to a statement from the German government.<\/p>\n \u201cWith SOOFI, we are laying the foundation for the next generation of European AI models \u2013 sovereign, powerful, and entirely in European hands,\u201d said Wolfgang Nejdl, a professor from Leibniz University Hannover, one of the universities involved in the project.<\/p>\n \u201cLarge AI models that respect European values are essential for building trust in AI, especially in sensitive areas such as education, medicine, administration, and production,\u201d he added. \u200b<\/p>\n Telecommunications company Deutsche Telekom and T-systems said the goal is for SOOFI to have 100 billion parameters or settings that regulate the model\u2019s behaviour.<\/p>\n Both companies are providing technical support for the large language model in one of their AI factories. To train the model, Deutsche Telekom will use around 130 NVIDIA chips and over 1,000 graphics processing units (GPUs) that will be in place by next March.<\/p>\n TU Darmstadt, one of the German universities involved with the project, said<\/a> SOOFI will also determine what is needed to build expertise in every area of developing large AI models, from collecting and preparing data, to building and training the software.<\/p>\n Switzerland<\/p>\n In September, the Swiss AI initiative launched Apertus, the country\u2019s first multilingual language model.<\/p>\n Apertus, the Latin word for \u201copen,\u201d lets researchers, professionals and the public customise the model for their specific needs.<\/p>\n The developers say everything about the model is open for use, including the training architecture, datasets, source code and model weights, the parameters that instruct the LLM on how to interpret data.<\/p>\n ETH Zurich, one of the collaborating universities, said<\/a> that Apertus was trained on 15 trillion tokens or pieces of information across more than 1,000 languages, like Swiss German and Romansh.<\/p>\n Apertus was uploaded on Public AI<\/a>, an online access point for sovereign models, so people can access the model around the world.<\/p>\n The Swiss AI Initiative said it will explore domain-specific models with expertise in law, climate, health, and education.<\/p>\n \u201cThis release isn’t a final step\u2014it’s a beginning,\u201d Antoine Bosselut, the co-lead of the Swiss AI initiative, said on Public AI. \u201cWe’re building toward a long-term commitment to sovereign, open AI foundations that serve the public good, worldwide”.<\/p>\n Poland<\/p>\n In February, Polandlaunched<\/a> its domestic large language model, Polish Large Language Model (PLLuM).<\/p>\n PLLuM is \u201ctailored to the specificities of the Polish language,\u201d so any AI speech or writing projects will be able to \u201chandle the challenges of inflection and complex syntax very well,\u201d a government release said at the time of the launch.<\/p>\n \u200bThe government suggests<\/a>that PLLuM models can be turned into AIs that can help write texts and emails, summarise documents, help students prepare for class, generate chatbot content, plan a trip, or create outlines.<\/p>\n Dariusz Standerski, Poland\u2019s deputy minister of digital affairs, said at the time that PLLuM \u201cis an investment in the digital state\u201d.<\/p>\n \u200bStanderski said at the model\u2019s launch that it will be expanded into Hive AI<\/a>: a system that will eventually be integrated into the government\u2019s public administration operations and help develop \u201cthe national AI ecosystem\u201d.<\/p>\n For example, it means the public will get access to a virtual assistant that will help them get public information, an \u201cintelligent\u201d office assistant that will automate document processing and information retrieval.<\/p>\n Down the line, PLLuM will also be used to help teachers \u201cconduct engaging lessons\u201d with the latest technologies in their classrooms.<\/p>\n Spain<\/p>\n In January, the Barcelona Supercomputing Centre (BSC) launchedAlia<\/a>, the \u201cfirst European open and multilingual infrastructure,\u201d which will develop \u201cresponsible AI for the service of the people\u201d.<\/p>\n The BSC developed Alia with the help of MareNostrum 5, a supercomputer that can run 314 quadrillion calculations per second. <\/p>\n Alia provides an open database<\/a>of resources, such as datasets, language models and integration tools in Spanish, Basque, Catalan and Galician that will help startups build their own domestic models.<\/p>\n Eventually, the Spanish Agency for the Supervision of Artificial Intelligence (AESIA) said<\/a> Alia will be developed into a tax agency chatbot and built into an application that will be able to easily diagnose heart failure.<\/p>\n The Alia project also builds on the back of Ilena, another Spanish government initiative that built over 100 AI resources in Spanish, Basque, Catalan, and Galician languages for domestic companies to use.<\/p>\n In 2020, the government of Catalonia launched<\/a> Aina, a pilot project that generates computer models in Catalan for other companies that want to build AI products like voice assistants, automatic translators or conversational AI agents.<\/p>\n The government trained that model on an initial Catalan database that has 1.7 million words combined into 95 million sentences.<\/p>\n The \u200bNetherlands<\/p>\n In 2023, three non-profit organisations started developing a Dutch-speaking open-source AI model called GPT-NL.<\/p>\n A website dedicated<\/a>to the project describes GPT-NL as a model \u201cfor the Dutch language and culture: reliable, transparent, reciprocal and sovereign\u201d.<\/p>\n The consortium is using<\/a>a mix of data obtained by copyright agreements from high-quality sources, public data and generating its own synthetic data.<\/p>\n The consortium recently signed a deal<\/a> with Dutch publishers under the trade organisation NDP Nieuwsmedia and news agency ANP to use their articles for GPT-NL\u2019s training. In return, the publishers get a share of the LLM\u2019s profits when it is eventually released.<\/p>\n The project will also be open source, meaning that academic institutions, researchers and government can try out its applications in health, education, and service fields. Users who aren’t using the LLM for professional reasons might have to pay a small fee<\/a> to access it once it\u2019s available.<\/p>\n The researchers started training the model in June 2025 and are expecting a first version to be available before<\/a> the end of the year, according to a recent update.<\/p>\n Portugal<\/p>\n Since 2024, a consortium of Portuguese universities has been working on a sovereign AI called Amalia.<\/p>\n The Nova School of Science and Technology, one of the research groups behind Amalia, said<\/a> it is able to answer questions, generate code, explain concepts, summarise texts and interpret information in Portuguese and with local context.\u200b<\/p>\n So far, researchers have tested the beta version of Amalia in September and are working towards a mid-2026 public release of the AI.<\/p>\n The government is already planning to use this large language model for public administration services through its online portal and in science to help with analysis.<\/p>\n