Soofi Introduces First Sovereign Open Source Foundation Model for European Industry
Key Takeaways
- ▸Soofi S is a 30B Mixture-of-Experts foundation model trained on 27 trillion tokens, optimized for German and English and industrial applications
- ▸The model prioritizes transparency, adaptability, and sovereignty—enabling deployment on independent European infrastructure without vendor lock-in
- ▸Soofi is actively recruiting industrial partners across sectors to test and validate the model in real-world applications
Summary
Soofi, a German-led consortium of research institutions and startups, has unveiled Soofi S, the first foundation model in an open European AI model family designed for industrial applications. The model is a 30-billion parameter Mixture-of-Experts system trained on 27 trillion tokens, with emphasis on German and English language content, optimized for industrial use cases including technical and regulatory document processing, code generation, agentic AI systems, and enterprise-specific applications.
The initiative represents a significant step toward European AI sovereignty and independence from US-dominated AI systems. Soofi S is currently undergoing practical testing with industrial partners, with the consortium actively recruiting enterprises, mid-market companies, government agencies, research institutions, and startups to participate in the next testing phase. A general public release has not yet been scheduled.
The model is designed to enable powerful AI applications that are transparent, adaptable, and can run on sovereign infrastructure—addressing the needs of organizations where control, transparency, and compliance are critical. Soofi S serves as the foundation for a broader family of specialized models in development, including variants optimized for dialogue, reasoning, and agent applications.
- The model forms the foundation for a broader family of specialized Soofi models currently in development
Editorial Opinion
Soofi S represents a crucial milestone for European AI autonomy at a time when the sector is increasingly dominated by US tech giants. By prioritizing open-source development, sovereignty, and industrial transparency, the initiative directly addresses legitimate concerns from enterprises and governments about data control and vendor independence. However, the initiative's ultimate success will depend on whether it can match the capabilities and scale of better-resourced models like OpenAI's GPT or Meta's Llama—a gap that requires sustained investment and broad ecosystem engagement.
