At the conference “European Data Spaces: Towards Sovereign and Sectoral AI” in Paris, Prof. Dr. Jakob Rehof, Co-Director of the Lamarr Institute, emphasized how crucial trustworthy data spaces are for the development of powerful, vertical AI models. In his presentation “Data Sharing and Vertical AI – An Opportunity for European R&D,” he explained that Europe’s innovative strength arises primarily where research and industry have access to high-quality, domain-specific data. Only on this basis can AI systems be developed that have a profound impact on industrial processes, scientific workflows, and societal applications.
Rehof placed particular emphasis on Inria’s current research approaches to sovereign AI, agent-based LLMs, and novel data infrastructures. They illustrate how important European basic research is for resilient vertical AI models and how crucial the close integration of technical and scientific excellence with robust data spaces is becoming for Europe’s digital independence.
The much-cited Draghi Report on European competitiveness lends additional relevance to this perspective: it recommends that European AI development be specifically aligned with Europe’s own strengths – i.e., those sectors in which Europe has traditionally been a leader. Vertical AI models that map specific industry logic and data structures are considered a strategic priority in this regard. For this approach to work, interoperable data spaces, robust infrastructure, clear governance, and a regulatory environment that enables both innovation and security are needed.
The conference discussions addressed precisely these questions: how Europe can establish independent and resilient data and AI architectures, how traceability and cybersecurity can be ensured, and how sector-specific AI can be scaled from health to mobility to energy. Rehof made it clear that Europe’s opportunity lies less in competing with global general-purpose models and more in building trustworthy, resource-efficient, and highly specialized AI that directly addresses real social and industrial needs.