AI Governance Engineer
GitHub

EU AI Act deadlines after the Digital Omnibus: what applies when

The 13 dates on which something under Regulation (EU) 2024/1689 starts to apply, as Regulation (EU) 2026/1744 left them, with the articles each date switches on, who it concerns and what the Omnibus moved. For Spain: the supervisor, the sandbox and the national AI bill, which is still in parliament.

Leer en español

Next deadline

as of 30 Sep 2026 New prohibitions and the end of the marking grace

Indicative, not legal advice

Dates as Regulation (EU) 2024/1689 states them after the Digital Omnibus, Regulation (EU) 2026/1744. The page was last checked on 30 Sep 2026 and is due for review on 2 Dec 2026.

The timeline

Every date on which something under the AI Act starts to apply, from entry into force to the last transitional deadline.

The AI Act on one axis13 AI Act milestones from 2024-08-01 to 2030-12-31 on one time axis, marked as applying or upcoming on the date of the last check of this page.StatusAppliesUpcomingBasisRegulation (EU) 2024/1689Set or moved by the Omnibus2024202620282030As of 2026-09-30The AI Act enters into force · 2024-08-01 · Applies2024-08-01Entry into forceProhibited practices and AI literacy · 2025-02-02 · Applies2025-02-02Prohibitions, AI literacyGeneral-purpose AI models, governance and penalties · 2025-08-02 · Applies2025-08-02GPAI models, governanceThe Digital Omnibus enters into force · 2026-07-27 · Applies2026-07-27Omnibus in forceGeneral application · 2026-08-02 · Applies2026-08-02General applicationNew prohibitions and the end of the marking grace · 2026-12-02 · Upcoming2026-12-02New prohibitionsEarlier general-purpose AI models and national sandboxes · 2027-08-02 · Upcoming2027-08-02Earlier GPAI, sandboxesPost-market monitoring guidance and template · 2027-09-02 · Upcoming2027-09-02Monitoring templateHigh-risk systems in Annex III · 2027-12-02 · Upcoming2027-12-02High-risk, Annex IIINotified bodies for Annex I, Section A · 2028-01-28 · Upcoming2028-01-28Notified bodiesHigh-risk systems in Annex I · 2028-08-02 · Upcoming2028-08-02High-risk, Annex IEarlier high-risk systems used by public authorities · 2030-08-02 · Upcoming2030-08-02Earlier public-sectorsystemsLarge-scale EU IT systems (Annex X) · 2030-12-31 · Upcoming2030-12-31Annex X IT systems2024-08-01 · The AI Act enters into force · Applies2025-02-02 · Prohibited practices and AI literacy · Applies2025-08-02 · General-purpose AI models, governance and penalties · Applies2026-07-27 · The Digital Omnibus enters into force · Applies2026-08-02 · General application · Applies2026-12-02 · New prohibitions and the end of the marking grace · Upcoming2027-08-02 · Earlier general-purpose AI models and national sandboxes · Upcoming2027-09-02 · Post-market monitoring guidance and template · Upcoming2027-12-02 · High-risk systems in Annex III · Upcoming2028-01-28 · Notified bodies for Annex I, Section A · Upcoming2028-08-02 · High-risk systems in Annex I · Upcoming2030-08-02 · Earlier high-risk systems used by public authorities · Upcoming2030-12-31 · Large-scale EU IT systems (Annex X) · UpcomingSource: Regulation (EU) 2024/1689[1] and Regulation (EU) 2026/1744[2]As of 2026-09-30
The AI Act on one axis13 AI Act milestones from 2024-08-01 to 2030-12-31 on one time axis, marked as applying or upcoming on the date of the last check of this page.StatusAppliesUpcomingBasisRegulation (EU) 2024/1689Set or moved by the OmnibusAs of 2026-09-302024202620282030The AI Act enters into force · 2024-08-01 · AppliesEntry into force2024-08-01General-purpose AI models, governance and penalties · 2025-08-02 · AppliesGPAI models, governance2025-08-02General application · 2026-08-02 · AppliesGeneral application2026-08-02Earlier general-purpose AI models and national sandboxes · 2027-08-02 · UpcomingEarlier GPAI, sandboxes2027-08-02High-risk systems in Annex III · 2027-12-02 · UpcomingHigh-risk, Annex III2027-12-02High-risk systems in Annex I · 2028-08-02 · UpcomingHigh-risk, Annex I2028-08-02Large-scale EU IT systems (Annex X) · 2030-12-31 · UpcomingAnnex X IT systems2030-12-31Prohibited practices and AI literacy · 2025-02-02 · AppliesProhibitions, AI literacy2025-02-02The Digital Omnibus enters into force · 2026-07-27 · AppliesOmnibus in force2026-07-27New prohibitions and the end of the marking grace · 2026-12-02 · UpcomingNew prohibitions2026-12-02Post-market monitoring guidance and template · 2027-09-02 · UpcomingMonitoring template2027-09-02Notified bodies for Annex I, Section A · 2028-01-28 · UpcomingNotified bodies2028-01-28Earlier high-risk systems used by public authorities · 2030-08-02 · UpcomingEarlier public-sector systems2030-08-022024-08-01 · The AI Act enters into force · Applies2025-08-02 · General-purpose AI models, governance and penalties · Applies2026-08-02 · General application · Applies2027-08-02 · Earlier general-purpose AI models and national sandboxes · Upcoming2027-12-02 · High-risk systems in Annex III · Upcoming2028-08-02 · High-risk systems in Annex I · Upcoming2030-12-31 · Large-scale EU IT systems (Annex X) · Upcoming2025-02-02 · Prohibited practices and AI literacy · Applies2026-07-27 · The Digital Omnibus enters into force · Applies2026-12-02 · New prohibitions and the end of the marking grace · Upcoming2027-09-02 · Post-market monitoring guidance and template · Upcoming2028-01-28 · Notified bodies for Annex I, Section A · Upcoming2030-08-02 · Earlier high-risk systems used by public authorities · UpcomingSource: Regulation (EU) 2024/1689 [1] and Regulation (EU) 2026/1744 [2]As of 2026-09-30
The AI Act on one axis Every milestone below on one time axis: solid once it applies, outlined while it is upcoming, and square where the Digital Omnibus set or moved the date; the bar runs up to the dashed line, the date of the last check of this page. The next deadline is in bold, and each label leads to its entry.
The milestones
The milestones
DateMilestoneStatusBasis
2024-08-01The AI Act enters into forceAppliesRegulation (EU) 2024/1689
2025-02-02Prohibited practices and AI literacyAppliesRegulation (EU) 2024/1689
2025-08-02General-purpose AI models, governance and penaltiesAppliesRegulation (EU) 2024/1689
2026-07-27The Digital Omnibus enters into forceAppliesSet or moved by the Omnibus
2026-08-02General applicationAppliesRegulation (EU) 2024/1689
2026-12-02New prohibitions and the end of the marking graceUpcomingSet or moved by the Omnibus
2027-08-02Earlier general-purpose AI models and national sandboxesUpcomingSet or moved by the Omnibus
2027-09-02Post-market monitoring guidance and templateUpcomingSet or moved by the Omnibus
2027-12-02High-risk systems in Annex IIIUpcomingSet or moved by the Omnibus
2028-01-28Notified bodies for Annex I, Section AUpcomingSet or moved by the Omnibus
2028-08-02High-risk systems in Annex IUpcomingSet or moved by the Omnibus
2030-08-02Earlier high-risk systems used by public authoritiesUpcomingRegulation (EU) 2024/1689
2030-12-31Large-scale EU IT systems (Annex X)UpcomingRegulation (EU) 2024/1689
  1. Applies

    The AI Act enters into force

    Regulation (EU) 2024/1689 enters into force. Nothing applies yet: every obligation has a later application date. 12

    Articles
    Art. 113
  2. Applies

    Prohibited practices and AI literacy

    Chapters I and II apply: the definitions, the AI literacy duty and the original list of prohibited practices. 1

    Articles
    Art. 113(a) · Art. 4 · Art. 5
    Who
    Every provider and deployer; the bans bind anyone who places on the market, puts into service or uses the practice.
    In the register
    Art. 3(1) , Art. 4 , Art. 5
  3. Applies

    General-purpose AI models, governance and penalties

    The rules on notified bodies, the obligations of providers of general-purpose AI models, governance, penalties (except Art. 101) and confidentiality apply; Member States publish their national contact points. 1

    Articles
    Art. 113(b) · Art. 70(2) · Arts. 53 to 55
    Who
    Providers of general-purpose AI models; Member States.
    In the register
    Art. 53 , Art. 53(1)(c) , Art. 55
  4. Applies

    The Digital Omnibus enters into force

    Regulation (EU) 2026/1744 enters into force: the reworded AI literacy article (Art. 4), the new Art. 4a on processing special categories of data to detect and correct bias, and the amendments to other acts in Arts. 102 to 110, brought forward from 2 Aug 2026, apply from this day. 23

    Articles
    Omnibus Art. 4 · Art. 4 · Art. 4a · Art. 113(d) · Arts. 102 to 110
    In the register
    Art. 4 , Art. 4a
  5. Applies

    General application

    The Act applies in general: Art. 50 transparency duties, Commission fines on providers of general-purpose AI models, the Chapter VI measures including testing in real-world conditions, and the enforcement chapter including Arts. 75a to 75d. 124

    Articles
    Art. 113 · Art. 50 · Art. 60 · Arts. 75a to 75d
    Who
    Providers and deployers of systems under Art. 50; providers of general-purpose AI models.
    In the register
    Art. 50 , Art. 60 , Art. 87
  6. Upcoming

    New prohibitions and the end of the marking grace

    The new prohibited practices in Art. 5(1)(ba) and (bb) (AI-generated non-consensual intimate imagery and child sexual abuse material) apply, and generative systems placed on the market before 2 Aug 2026 must mark their outputs under Art. 50(2). 234

    Articles
    Art. 113(a) · Art. 5(1)(ba) · Art. 5(1)(bb) · Art. 5(1a) · Art. 5(1b) · Art. 111(4) · Art. 50(2)
    Who
    Providers of generative AI systems; anyone using the new banned practices.
    In the register
    Art. 5 , Art. 50
  7. Upcoming

    Earlier general-purpose AI models and national sandboxes

    General-purpose AI models placed on the market before 2 Aug 2025 must comply; each Member State must have at least one AI regulatory sandbox operational; the delegated acts limiting duties for Annex I, Section A products are due. 123

    Moved by the Omnibus from . The original text required the national sandboxes to be operational by 2 Aug 2026; the Omnibus moved the deadline to 2 Aug 2027.

    Articles
    Art. 111(3) · Art. 57(1) · Art. 2(13)
    Who
    Providers of earlier general-purpose AI models; Member States; the Commission.
    In the register
    Art. 53 , Art. 55
  8. Upcoming

    Post-market monitoring guidance and template

    The Commission guidance and template for the post-market monitoring plan are due. 12

    Moved by the Omnibus from . The original text had the Commission adopt an implementing act with the template by 2 Feb 2026; the Omnibus replaced it with guidance and a template due by 2 Sep 2027.

    Articles
    Art. 72(3)
    Who
    The Commission.
  9. Upcoming

    High-risk systems in Annex III

    The high-risk rules apply to Annex III systems: classification, requirements, provider and deployer duties, and the fundamental rights impact assessment. 1234

    Moved by the Omnibus from . The original text applied the Annex III rules from 2 Aug 2026; the Omnibus moved them to 2 Dec 2027.

    Articles
    Art. 113(c)(i) · Annex III · Arts. 6 to 27
    Who
    Providers, deployers, importers and distributors of Annex III systems.
    In the register
    Art. 6 , Art. 9 , Art. 10 , Art. 14 , Art. 26 , Art. 27
  10. Upcoming

    Notified bodies for Annex I, Section A

    Notified bodies under the Annex I, Section A legislation apply for designation under the AI Act. 2

    Articles
    Art. 43(3)
    Who
    Notified bodies.
  11. Upcoming

    High-risk systems in Annex I

    The high-risk rules apply to systems that are safety components of, or are themselves, products under the Annex I legislation (Art. 6(1)). 1234

    Moved by the Omnibus from . The original text applied the Annex I route from 2 Aug 2027; the Omnibus moved it to 2 Aug 2028.

    Articles
    Art. 113(c)(ii) · Art. 6(1) · Annex I
    Who
    Providers and deployers of AI in regulated products (machinery, medical devices, toys and others).
    In the register
    Art. 6 , Art. 9 , Art. 10 , Art. 14 , Art. 26 , Art. 43
  12. Upcoming

    Earlier high-risk systems used by public authorities

    High-risk systems already on the market and intended for use by public authorities must comply. 12

    Articles
    Art. 111(2)
    Who
    Providers and public-authority deployers of those systems.
    In the register
    Art. 9 , Art. 26
  13. Upcoming

    Large-scale EU IT systems (Annex X)

    AI components of the large-scale IT systems listed in Annex X, placed on the market before 2 Aug 2027, must comply. 1

    Articles
    Art. 111(1) · Annex X
    Who
    The EU agencies and Member State authorities that run those systems.

What changed with the Digital Omnibus

The Omnibus moved the high-risk dates and added new ones. Earlier dates in the original text are shown next to the dates that now apply.

What the Omnibus moved4 AI Act dates the Digital Omnibus deferred, each drawn from its original date to the date that applies now, with the delay in months.Original dateDate now2026202720282029As of 2026-09-30National AI sandboxesNational AI sandboxes · Original date: 2026-08-02National AI sandboxes · Date now: 2027-08-02+12 monthsPost-market monitoring templatePost-market monitoring template · Original date: 2026-02-02Post-market monitoring template · Date now: 2027-09-02+19 monthsHigh-risk, Annex IIIHigh-risk, Annex III · Original date: 2026-08-02High-risk, Annex III · Date now: 2027-12-02+16 monthsHigh-risk, Annex IHigh-risk, Annex I · Original date: 2027-08-02High-risk, Annex I · Date now: 2028-08-02+12 monthsSource: Regulation (EU) 2026/1744[2]As of 2026-09-30
What the Omnibus moved4 AI Act dates the Digital Omnibus deferred, each drawn from its original date to the date that applies now, with the delay in months.Original dateDate now2026202720282029As of 2026-09-30National AI sandboxesNational AI sandboxes · Original date: 2026-08-02National AI sandboxes · Date now: 2027-08-02+12 monthsPost-market monitoring templatePost-market monitoring template · Original date: 2026-02-02Post-market monitoring template · Date now: 2027-09-02+19 monthsHigh-risk, Annex IIIHigh-risk, Annex III · Original date: 2026-08-02High-risk, Annex III · Date now: 2027-12-02+16 monthsHigh-risk, Annex IHigh-risk, Annex I · Original date: 2027-08-02High-risk, Annex I · Date now: 2028-08-02+12 monthsSource: Regulation (EU) 2026/1744 [2]As of 2026-09-30
What the Omnibus moved Each date the Digital Omnibus deferred, from the date in the original text to the date that applies now, with the delay in months; the dashed line marks the date of the last check of this page. The table below adds the dates the Omnibus created.
The deferred dates
The deferred dates
WhatOriginal dateDate nowDelay
National AI sandboxes2026-08-022027-08-02+12 months
Post-market monitoring template2026-02-022027-09-02+19 months
High-risk, Annex III2026-08-022027-12-02+16 months
High-risk, Annex I2027-08-022028-08-02+12 months
What Original date Date now Basis
Earlier general-purpose AI models and national sandboxes Regulation (EU) 2026/1744, Art. 111(3)
Post-market monitoring guidance and template Regulation (EU) 2026/1744, Art. 72(3)
High-risk systems in Annex III Regulation (EU) 2026/1744, Art. 113(c)(i)
High-risk systems in Annex I Regulation (EU) 2026/1744, Art. 113(c)(ii)
The Digital Omnibus enters into force New date Regulation (EU) 2026/1744, Omnibus Art. 4
New prohibitions and the end of the marking grace New date Regulation (EU) 2026/1744, Art. 113(a)
Notified bodies for Annex I, Section A New date Regulation (EU) 2026/1744, Art. 43(3)

Spain: supervisor, sandbox and a bill

Spain has a working AI supervisor and a sandbox, and a national AI bill that is still in parliament.

  • In force

    AESIA, the Spanish AI supervisor

    The Spanish Agency for the Supervision of Artificial Intelligence, seated in A Coruña, has its statute in Royal Decree 729/2023. The Commission lists it as Spain's single point of contact under the AI Act, with the designation pending final adoption: the bill would make it the single point of contact and main market surveillance authority. Its 16 sandbox guides are not binding and predate the Omnibus. 651310

  • In force

    The sandbox, Royal Decree 817/2023

    Royal Decree 817/2023 set up a controlled testing environment for AI Act compliance. Its first call, published on 20 Dec 2024, sought up to 12 high-risk systems for about 12 months; 44 applied and 12 were selected. No second call had been published by 30 Sep 2026. The bill would have AESIA run the sandbox every Member State must have operational by 2 Aug 2027, and would repeal the decree. 789132

  • Bill, not law

    The Spanish AI bill, in parliament

    The Council of Ministers approved a first draft on 11 Mar 2025 and, on 26 May 2026, the bill for an organic law on the good use and governance of AI, published by the Congreso on 12 Jun 2026 (BOCG-15-A-97-1, file 121/000096). On 30 Sep 2026 it was at the amendments stage in committee, with the deadline extended to 7 Oct 2026. As proposed, it names the market surveillance authorities (AESIA, the data-protection agency AEPD, the judiciary council CGPJ, Banco de España, CNMV and the insurance directorate DGSFP), sets fines of up to EUR 35 million or 7% of turnover, and makes failing to disclose a deepfake a serious infringement. It is a bill, not law. 11121314

How this page is updated

The page is updated by milestone: when a date passes, the milestone is checked against its sources, marked as applying and dated. The site's test suite fails if a milestone passes without that check, so the page cannot quietly go stale. Last checked 30 Sep 2026; next review 2 Dec 2026.

Changelog

  • : First version: 13 milestones from chapter 18, checked against Regulations (EU) 2024/1689 and 2026/1744 and the consolidated text; the Spanish supervisor, sandbox and bill checked against the BOE, the Council of Ministers, the Congreso and AESIA.

Downloads

Related

Sources

  1. [1] Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) (original text, OJ L 2024/1689 of 12.7.2024; Arts. 57, 70, 72, 111 and 113 set the dates). Publications Office of the EU (EUR-Lex). 2024-07-12. https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng (verified: primary)
  2. [2] Regulation (EU) 2026/1744 (Digital Omnibus on AI), of 8 July 2026 (amends Arts. 2(13), 5, 43(3), 50, 57(1), 72(3), 111 and 113 of Regulation (EU) 2024/1689; OJ L 2026/1744 of 24.7.2026; in force 2026-07-27). Publications Office of the EU (EUR-Lex). 2026-07-24. https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng (verified: primary)
  3. [3] Regulation (EU) 2024/1689, consolidated text as amended by Regulation (EU) 2026/1744 (CELEX 02024R1689-20260727; documentation only, no legal effect). Publications Office of the EU (EUR-Lex). 2026-07-27. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02024R1689-20260727 (verified: primary)
  4. [4] Timeline for the implementation of the EU AI Act (AI Act Service Desk; the post-Omnibus milestones). European Commission. 2026. https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act (verified: primary)
  5. [5] Market Surveillance Authorities under the AI Act (single points of contact by Member State; the Spanish designation is marked pending final adoption). European Commission. 2026-09-07. https://digital-strategy.ec.europa.eu/en/policies/market-surveillance-authorities-under-ai-act (verified: primary)
  6. [6] Real Decreto 729/2023, de 22 de agosto, por el que se aprueba el Estatuto de la AESIA (the agency statute; seat in A Coruña). Boletín Oficial del Estado. 2023-09-02. https://www.boe.es/eli/es/rd/2023/08/22/729 (verified: primary)
  7. [7] Real Decreto 817/2023, de 8 de noviembre (controlled testing environment (sandbox) for compliance with the AI regulation). Boletín Oficial del Estado. 2023-11-09. https://www.boe.es/eli/es/rd/2023/11/08/817 (verified: primary)
  8. [8] Sandbox IA: convocatoria (first call published 2024-12-20; up to 12 high-risk systems; about 12 months; resolution 2025-04-30). Ministerio para la Transformación Digital y de la Función Pública. 2026-06-09. https://digital.gob.es/digitalizacion/areas-interes/sandbox-ia/convocatoria (verified: primary)
  9. [9] El Gobierno activa el primer entorno de pruebas de la UE (44 applications, 12 high-risk systems selected). La Moncloa. 2025-04-03. https://www.lamoncloa.gob.es/serviciosdeprensa/notasprensa/transformacion-digital-y-funcion-publica/paginas/2025/030425-primer-entorno-pruebas-ia.aspx (verified: primary)
  10. [10] Guías (16 non-binding guides from the sandbox pilot; the page still says they will be updated once the Omnibus is approved). AESIA. 2026-09-30. https://aesia.digital.gob.es/es/guias (verified: primary)
  11. [11] Referencia del Consejo de Ministros, 11 de marzo de 2025 (Anteproyecto de Ley para el buen uso y la gobernanza de la Inteligencia Artificial, first reading, urgent processing). La Moncloa. 2025-03-11. https://www.lamoncloa.gob.es/consejodeministros/referencias/paginas/2025/20250311-referencia-rueda-de-prensa-ministros.aspx (verified: primary)
  12. [12] Referencia del Consejo de Ministros, 26 de mayo de 2026 (approves the Proyecto de Ley Orgánica on the good use and governance of AI and sends it to the Congreso). La Moncloa. 2026-05-26. https://www.lamoncloa.gob.es/consejodeministros/referencias/paginas/2026/20260526-referencia-rueda-de-prensa-ministros.aspx (verified: primary)
  13. [13] Proyecto de Ley Orgánica para el buen uso y la gobernanza de la inteligencia artificial (BOCG-15-A-97-1, the Government text: authorities (Arts. 5 and 6), sandbox (Art. 11), infringements and fines (Arts. 13 to 30)). Congreso de los Diputados. 2026-06-12. https://www.congreso.es/public_oficiales/L15/CONG/BOCG/A/BOCG-15-A-97-1.PDF (verified: primary)
  14. [14] Iniciativa 121/000096 (parliamentary stage: amendments in the Comisión de Economía, Comercio y Transformación Digital, deadline extended to 2026-10-07). Congreso de los Diputados. 2026-09-30. https://www.congreso.es/es/busqueda-de-iniciativas?p_p_id=iniciativas&p_p_lifecycle=0&p_p_state=normal&p_p_mode=view&_iniciativas_mode=mostrarDetalle&_iniciativas_legislatura=XV&_iniciativas_id=121%2F000096 (verified: primary)