The EU AI Act gives the European Commission authority to demand direct inspection of AI models, restrict market access across the bloc, and levy fines of up to 15 million euros or 3% of turnover. Anthropic and OpenAI are among the firms now facing new scrutiny under that enforcement framework.
The inspection power is the most operationally direct of the three instruments. Compelling providers to make their AI systems available for regulatory review is a different category of obligation than submitting documentation. The Commission can examine the models themselves, placing the core technology within reach of regulators rather than only the paperwork around it.
Market access restriction is the harder sanction. The ability to block or limit a provider's EU distribution operates independently of any financial penalty, with immediate effect on commercial operations.
The fine ceiling anchors the financial exposure. At up to 15 million euros or 3% of turnover, the maximum scales with provider revenue rather than sitting as a flat figure that falls equally across companies of different sizes.
Anthropic and OpenAI, identified by name under the new scrutiny, sit within reach of all three instruments.