Risky Business or Safe Bets?
Recommendations for accelerated compliance with the EU AI Act based on the Risk Classification of 628 German AI Startups
How is the EU AI Act affecting German AI startups?
The EU AI Act establishes a risk-based framework for AI systems. While the legal categories are well defined, little empirical evidence exists on how they translate into real AI applications.
To address this gap, the appliedAI Institute for Europe conducted the largest study of its kind since the AI Act entered into force. As part of the Bavarian AI Act Accelerator, funded by the Bavarian State Ministry of Digital Affairs, we classified the highest-risk AI use case of 628 German AI startups using a novel methodology that combines legal expertise, LLM-assisted classification and manual validation.
The study provides a first evidence base for understanding where compliance obligations are likely to arise—and what this means for startups, policymakers and the wider AI ecosystem.
What you'll find in the report
- The first empirical assessment of how German AI startups are affected by the EU AI Act.
- Classification of 628 startups and 2,400+ AI use cases across the AI Act's risk tiers.
- Insights into where high-risk AI is concentrated across sectors, business functions and regions.
- A dedicated analysis of Generative AI and transparency obligations.
- Policy recommendations to strengthen innovation while enabling trustworthy AI.
Why this matters now
The EU AI Act is entering its implementation phase. For startups, understanding whether an AI system is classified as low-risk, subject to transparency obligations or high-risk will increasingly shape product development, governance and investment decisions.
Our findings suggest that the regulatory impact on Germany's AI ecosystem is broader than often assumed.
The study also highlights that compliance challenges will not be evenly distributed. Instead, they are concentrated in specific sectors, enterprise functions and AI application areas.
This evidence can help policymakers design more targeted support measures and enable startups to prepare for upcoming obligations under the AI Act.