AI from the alps: A blueprint for the EU’s tech ambitions?

The European Union wants large AI models made in the Europe – models that are state-of-the-art, trustworthy, and fully aligned with EU copyright and privacy standards. Is this a realistic goal, or will competitive European AI remain a political pipe dream?

Switzerland could show the way: A joint initiative by researchers at two leading technical universities – EPFL and ETH Zurich – and the Swiss National Supercomputing Centre (CSCS) has launched a multi-lingual open-source Large Language Model (LLM) with 70 billion parameters. Their LLM has been trained on publicly available compute infrastructure, with Swiss privacy and copyright law as well as the European AI Act in mind. It is freely available, with its source code, weights, and training data openly accessible.

To learn more about the initiative and discuss what policy makers can learn from it, I have invited Dr. Imanol Schlag for a 60-minute conversation. Imanol is a research scientist at the ETH AI Center and co-leads the initiative. We will discuss what it takes to turn the political ambition of AI sovereignty into reality.

Sign up for this discussion and Q&A with Imanol:

DATE: October 28th 2025
LOCATION: Online
TIME: 3.00 - 4.00 pm (CET)