The TDM equation

The TDM Equation assesses the economic importance of text and data mining (TDM) for Europe’s AI competitiveness and examines how changes to the current commercial TDM regime would affect Europe’s ability to realise the economic value of AI. 

Key findings of the study

• Economic opportunity: Adoption of generative AI offers Europe a large economic opportunity, with annual GDP gains of EUR 1.2 trillion at widespread adoption, equivalent to around 8% of GDP after a ten-year adoption period, under the current commercial TDM regime. Beyond adoption gains, GenAI could unlock further 200 billion from producing AI in the EU and EUR 250 billion from using AI to accelerate European research and innovation, making Europe’s total economic potential of generative AI EUR 1.65 trillion annually. 
• Why data access matters: Access to commercial text and data mining is vital for modern AI development because frontier models and European AI applications rely on very large and diverse datasets. Training models on smaller and less diverse datasets materially weakens model capabilities, with benchmark tests showing a 50% drop in complex reasoning performance when access to copyrighted training data is restricted. 
• What is at risk: Restricting the use of commercial TDM in the EU for AI training puts EUR 600 billion of Europe’s AI opportunity at risk through slower adoption, less powerful models, lower R&D productivity, and reduced prospects for European AI development. A one-year delay in adoption would cost EUR 175 billion in economic value. Other initiatives, such as a mandatory licensing regime, disproportionate transparency requirements and unworkable opt-out mechanisms, can lead to less powerful AI and lost R&D and development potential in the EU, putting a further EUR 425 billion at risk.
• The way forward: Maintaining a balanced commercial TDM regime is crucial to support European innovation and competitiveness while sustaining incentives to create new content. Preserving a workable commercial TDM regime requires transparency measures that are proportionate and operationally feasible and ensuring that opt‑outs are practical and machine‑readable. Europe should also leave room for voluntary, commercial partnerships and data-sharing collaborations to develop where they add value, rather than replacing TDM with prescriptive or premature licensing interventions.