Yitang and IdeCon Join Forces to Consolidate European Computing Power: Domyn Plans to Introduce Over 400B Open Source Model to Challenge China and the U.S.

According to Dynapulse Beating monitoring, Italian AI startup Domyn has announced plans to release a fully open-source, reproducible large model with over 400 billion parameters within a year. The EUROPA alliance formed by Domyn in collaboration with the Fraunhofer-Gesellschaft of Germany has been selected for the European Commission's Frontier AI Grand Challenge project and has received computational support from the European Public Supercomputing Network EuroHPC.

Reducing reliance on overseas hosting systems is a core focus of Europe's quest for local alternatives. The open-source AI ecosystem is mainly led by Chinese companies, while leading models in the United States are mostly closed-source and require remote access. Due to concerns about U.S. export controls on large models like Anthropic, as well as restrictions by Italy and the Czech Republic on DeepSeek remote services, European businesses and governments are accelerating their shift to localized systems that can be freely deployed on local infrastructure.

The required training data will be collected from institutional partners, with the first batch of government data sharing agreements expected to be reached in the weeks following meetings with several European heads of state. In terms of computational power requirements, Domyn CEO Uljan Sharka noted that the computational power consumed in training a cutting-edge large model from scratch is much lower than providing remote inference services to hundreds of millions of users, and the European public computational network is now able to meet the R&D requirements. In terms of funding, the project has received support from Abu Dhabi's G42, Eurizon Capital, Rabobank, and BNY Mellon.