Germany could forfeit €155 billion ($173.4 billion) in additional annual economic value if access to AI training data becomes more restricted, according to a new study from the ifo Institute, one of the country's most influential economic research bodies.

The findings, published on Tuesday, arrive at a pivotal moment for European technology policy. Policymakers in Brussels and Berlin are under pressure to protect publishers and rights holders whose work is used to train AI systems, while Europe's own AI industry argues that overly restrictive rules could cripple its ability to compete with the United States and China. The study adds hard numbers to that debate, and it feeds a growing transatlantic argument over whether Europe's regulatory instinct is an economic liability. For anyone following AI policy news, the ifo analysis is set to become a central reference in the fight over the EU's text-and-data-mining rules.

The Stakes: €440 Billion on the Table

The ifo researchers model a scenario in which generative AI adds €440 billion annually to Germany's economy by 2034 — a substantial productivity windfall spread across industries from manufacturing to services. But if AI training data restrictions intensify, the institute calculates that €155 billion of that annual value would never materialize, wiping out more than a third of the projected gains.

That is not a niche effect. It is comparable to a significant share of Germany's federal budget, and the study argues the damage would compound over time as European AI development falls further behind.

Four Channels of Economic Damage

The institute's model breaks the €155 billion loss into four distinct channels:

  • Weaker AI models — €75 billion. The largest single impact. If models trained on less data perform worse, every downstream application of AI in the German economy becomes less valuable.
  • Slower AI adoption — €45 billion. Businesses hesitant to deploy AI solutions without adequate training data and capable models would delay investments, compounding productivity losses.
  • Lower research output — €25 billion. Restrictions would dampen the research ecosystem that feeds both academic progress and commercial innovation.
  • Reduced AI development in Europe — €10 billion. Development work would migrate to jurisdictions with more permissive data regimes, eroding Europe's domestic AI capacity.

The numbers underscore how data regulation cascades through the entire innovation pipeline — from the models themselves to the companies that deploy them.

"You Cannot Have Both"

ifo expert Oliver Falck distilled the study's central tension into a blunt formulation: "Anyone who wants a leading European AI model cannot at the same time make access to training data unnecessarily complicated. You cannot have both."

That message is aimed squarely at ongoing debates in the EU over the copyright framework that governs how AI companies ingest European text, images, and other creative work. Europe currently grants AI developers a text-and-data-mining exemption, but it allows rights holders to opt out — and publishers, media groups, and creators have been pushing to tighten or commercialize that opt-out mechanism.

What the Researchers Want From Policymakers

The study's authors do not argue for a regulation-free zone. Instead, they call for a calibrated middle path:

  • Preserve the EU's text-and-data-mining exemption for commercial uses, which they describe as essential infrastructure for AI development.
  • Introduce proportionate transparency requirements so rights holders can see how their work is used without imposing impossible technical burdens.
  • Adopt a practical opt-out standard that meaningfully protects creators while remaining workable at the scale of modern web-scale training.

The researchers also warn that regulatory uncertainty is already doing damage. Companies are hesitating to commit resources to European AI development, they argue, which risks accelerating a brain drain of AI talent and projects toward regions with clearer rules.

A European Debate With Global Consequences

The ifo study lands as European leaders openly fret about the continent's AI competitiveness. German industry figures, including Chancellor Friedrich Merz and Siemens leadership, have recently pushed for lighter EU regulation of industrial AI, while France and Germany court AI investments and data center projects meant to bolster digital sovereignty.

At the same time, the pressure from publishers and creators is real and politically potent. Many European media groups have watched AI chatbots summarize their content while traffic and advertising revenue drift elsewhere, and they argue that without licensing leverage, the economics of journalism and creative work will collapse.

The €155 billion figure gives the AI industry a powerful talking point in that negotiation; the opt-out debate gives publishers theirs. How the EU resolves the collision between those two positions will shape not only Germany's AI economy, but whether Europe fields frontier models of its own in the decade ahead.

For now, the ifo Institute's message to Berlin and Brussels is that data access policy is industrial policy — and that the cost of getting it wrong is measured in tens of billions of euros a year.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →