Generative AI could add 440 billion euros in annual value creation for Germany by 2034. If access to training data gets restricted, an ifo model calculation shows this figure dropping by 155 billion euros to about 285 billion euros.
A model calculation by the ifo Institute shows that generative AI could add 440 billion euros in annual value creation for Germany by 2034. If access to training data for AI developers gets restricted, this figure drops by 155 billion euros to roughly 285 billion euros, a cut of more than a third.
The potential breaks down into 330 billion euros from broad adoption, 65 billion euros from more productive research and 45 billion euros from AI development inside Germany. Restrictions hit through four channels: slower rollout costs 45 billion euros, less capable models cost 75 billion euros, lost research potential costs 25 billion euros, and weaker development in Europe costs 10 billion euros. Europe largest copyright-free dataset holds only two to three trillion tokens, while leading models already train on 15 to 36 trillion.
The study was commissioned by CCIA Europe and Google, and the authors themselves note the figures are potential estimates, not forecasts. They argue for keeping the text and data mining exemption with proportionate transparency duties rather than tightening data access further. The European Commission has been reviewing the underlying copyright directive since May 2026.
What this means for decision-makers
- Track the EU review of the copyright directive running since May 2026 and its possible impact on training data.
- Assess how much your business model depends on access to large, diverse training datasets.
- Factor in that the study was funded by Google and CCIA Europe when weighing its conclusions.
This story was produced automatically from the source named above and checked by software before publication. The image is symbolic and shows neither the event nor a real person. How this paper is made