Mapping Europe’s AI Media Landscape

Abstract

This report maps Europe’s media sector’s position in the AI landscape across three layers: compute, models, and applications. It examines how media organisations are affected by AI, how they contribute to AI systems, and how they can capture value from them. The central finding is that media organisations are no longer merely users of AI. Their archives, content, editorial expertise, trusted brands, and audience relationships are increasingly valuable for AI training, retrieval, and product development. However, this does not automatically translate into control, revenue, or strategic autonomy.

At the compute layer, Europe has developed stronger public infrastructure through EuroHPC, LUMI, JUPITER, VSC, and AI Factory Antennas, but media applications will likely continue to depend on hybrid environments combining public compute and commercial cloud. At the model layer, Europe has produced sovereign and language-oriented initiatives, yet media organisations remain only selectively involved. Cases such as GPT-NL, NorLLM/Kopinor, and OpenGPT-X/Teuken show that media archives can become strategic and monetisable assets when governance, IP, licensing, and revenue-sharing are addressed from the start. At the application layer, European media organisations are most active, particularly in archive search and retrieval, newsroom productivity, and audience-facing products.

Across all layers, two themes emerge: local language and cultural specificity matter, and AI adoption alone does not guarantee value capture. The report identifies several routes for governed value creation, including bilateral licensing, collective licensing, collaborative sovereign models, internal application development, B2B commercialisation, and governed attribution or monetisation layers. The main conclusion is that future European media AI initiatives should move beyond siloed experimentation toward collective approaches that align technology with governance, data ownership, IP, and value-sharing.