EU AI Act: general application from 2 August 2026 — high-risk deadlines set to move to 2027/2028

Regulation (EU) 2024/1689 (the AI Act) entered into force on 1 August 2024 and its remaining provisions largely apply from 2 August 2026. Prohibited practices and AI-literacy duties have applied since 2 February 2025, and general-purpose AI model obligations since 2 August 2025. The high-risk timetable, however, is being amended: under the 'digital omnibus on AI' provisionally agreed by the co-legislators on 7 May 2026, requirements for stand-alone (Annex III) high-risk systems would apply from 2 December 2027 and for high-risk AI embedded in regulated products from 2 August 2028, once the amending regulation is formally adopted.

2 August 2026 remains a real milestone for everything outside the deferred high-risk track — governance, transparency, penalties and Member State enforcement structures. Treating the omnibus deferral as a blanket delay of the whole Act would be a mistake.

Providers and deployers of AI systems in the EU, providers of general-purpose AI models, and any organisation with Annex III use cases planning its compliance timeline.

Split your AI Act programme: confirm which obligations still bite on 2 August 2026, and verify the final adopted text and dates of the AI omnibus in the Official Journal before re-baselining the high-risk workstream.