Great language models are high on the European digital agenda, as evidenced by a new initiative: OpenEuroLLM.
The EU’s new initiative, called OpenEuroLLM, is intended to provide greater digital sovereignty. This program serves to develop a set of open-source language models covering all European languages. With this, the EU aims to support not only the 24 official languages, but also languages of countries that will soon join, such as Albania.
Strategic advantage
An initiative of 20 organizations, OpenEuroLLM is led by Jan Hajič, a linguist at the University of Prague, and Peter Sarlin, CEO of Finnish AI lab Silo AI. AMD bought that lab last year for $665 million. Europe wants to become digitally independent, and like tech giants investing in local cloud infrastructure and OpenAI enabling European data processing, the EU is building its own AI future with OpenEuroLLM.
According to Sarlin, OpenEuroLLM has sufficient resources because many costs, such as computing power, are already covered by EuroHPC. In addition, the project focuses on developing a foundation for AI models rather than a turnkey chatbot like ChatGPT.
read also
InvestAI: EU invests €200 billion to build AI advantage
OpenEuroLLM does not have to be a direct competitor to the tech giants. The goal is a European alternative that provides AI infrastructure for companies and institutions within the EU. “What we contribute is an open source foundation model that functions as the AI infrastructure for companies in Europe to build on. We know what it takes to build models. It’s not something that requires billions,” Sarlin said in an interview with TechCrunch.
In a world where AI is increasingly dominated by big players, an independent European model is much needed. Even if OpenEuroLLM does not become the best model on the market, a reliable, European alternative is a strategic advantage. As Hajič points out, “Even if our model is not number one, we still have a fully European model. That’s a win in itself.”