EU AI Act Full Effect: Global Tech Confronts Brussels Effect

On 2 August 2026, the EU AI Act enters its most consequential enforcement phase, forcing technology companies around the world to confront the so-called Brussels Effect. Although the EU's Digital Omnibus package pushed back standalone high-risk AI obligations to 2 December 2027, Article 50 transparency rules, expanded AI Office powers and a fresh ban on non-consensual intimate imagery are now live. Non-compliance carries fines of up to €35 million or 7% of global turnover, and with 78% of enterprises still unprepared, the regulation is already reshaping AI development far beyond Europe.

What is the EU AI Act's risk-based framework?

The AI Act, Regulation 2024/1689, classifies AI applications into four tiers: unacceptable, high, limited and minimal risk, plus a separate category for general-purpose AI. Unacceptable uses such as social scoring and real-time public biometric surveillance have been banned since February 2025. High-risk AI systems used in healthcare, recruitment, credit scoring and critical infrastructure face strict obligations around risk management, data governance, human oversight and Fundamental Rights Impact Assessments. Limited-risk systems must meet transparency duties, while minimal-risk applications are largely unregulated. Unlike the GDPR compliance framework, which creates individual rights, the AI Act is product regulation that places duties on providers and deployers.

August 2026 deadline: what actually changes?

Despite the headline date, the Council of the EU approved the Digital Omnibus on 29 June 2026, delaying Annex III high-risk requirements for standalone systems to 2 December 2027 and product-embedded high-risk systems to 2 August 2028. Harmonised standards from CEN-CENELEC were not ready. However, Article 50 transparency obligations remain on their original 2 August 2026 schedule: chatbots must disclose AI interaction, AI-generated content must be labelled, and deepfakes marked. A four-month watermarking grace period runs to 2 December 2026. New Article 5 prohibitions on non-consensual intimate imagery take effect 2 December 2026.

Global fragmentation: EU, US and China

The EU's rights-based, prescriptive model contrasts sharply with the US sectoral, innovation-first approach, where more than 1,100 state AI bills coexist without a federal framework, and China's state-centric system with over 4,000 registered AI systems. This divergence is the core of the Brussels Effect: because the EU is a market of 450 million consumers and €16 trillion GDP, global firms often adopt EU standards worldwide rather than maintain separate versions. Microsoft, Google, OpenAI and Apple have already embedded AI Act requirements into global products, while Meta has refused and faces investigation.

Impact on enterprises

Compliance is expensive. Large enterprises face first-year costs of €8–15 million, and each high-risk system adds roughly $1 million annually. Over half of organisations still lack a basic AI inventory, and only 8 of 27 EU member states have designated competent authorities, raising concerns about uneven enforcement. The result is consolidation and a reordering of global supply chains as smaller vendors struggle with AI compliance costs and notified-body capacity remains constrained.

Expert perspectives

Columbia Law School professor Anu Bradford, who coined the Brussels Effect, argues that unilateral EU regulation can set global standards because multinationals find it more efficient to apply one strict rule everywhere. “The AI Act is following the GDPR playbook, but faster, because the compliance infrastructure already exists,” she has said. Others warn that the delay creates false confidence: the EU AI Office's enforcement powers are expanding, and Article 12 logging demands tamper-evident audit trails from a system's operational lifetime.

FAQ: EU AI Act full effect

When does the EU AI Act fully apply?

Article 50 transparency obligations and expanded AI Office powers apply from 2 August 2026. Standalone high-risk AI obligations were delayed to 2 December 2027 by the Digital Omnibus.

What are the maximum fines?

Fines reach €35 million or 7% of global annual turnover, whichever is higher, exceeding GDPR's 4% maximum.

Does the AI Act apply to non-EU companies?

Yes. It applies extraterritorially to any provider or deployer whose AI systems serve users in the EU.

What is the Brussels Effect?

The Brussels Effect describes how EU regulation becomes a de facto global standard because companies adopt EU rules worldwide and other governments model laws on them. Japan, Canada, Brazil and South Korea are doing so with AI.

What should companies do now?

Build AI inventories, map high-risk use cases, implement Article 50 transparency and begin logging for future conformity assessments—the Digital Omnibus delay is a planning reprieve, not a stand-down.

Conclusion

The 2 August 2026 enforcement date is less a finish line than a starting gun. Whether the Brussels Effect drives global convergence or deepens geopolitical divergence will depend on how quickly the US and China respond, but for now, the EU AI Act is the world's de facto AI rulebook.