The European Commission will be able to enforce the EU AI Act's obligations for providers of general-purpose AI models from 2 August 2026, when its supervisory and enforcement powers take effect. The obligations themselves, set out in Chapter V of the Act, have applied since 2 August 2025, and providers were given a one-year adjustment period before the Commission could act against them.
Responsibility for enforcement sits with the Commission's AI Office. From 2 August 2026, the AI Office will be able to require providers to produce technical documentation under Article 91 and make models available for evaluation under Article 92. Where it identifies non-compliance, Article 93 also allows it to order corrective measures, including risk mitigation, compliance actions and, where necessary, restricting, recalling or withdrawing a model from the EU market. Providers also face enforcement if they provide incorrect, incomplete or misleading information, refuse model access, or fail to comply with measures ordered by the Commission.
Under Article 101, the Commission may fine a GPAI provider up to the greater of €15 million or 3% of its total worldwide annual turnover in the preceding financial year where it determines that a provider intentionally or negligently infringed the Act, failed to comply with a request for documents or information, failed to comply with a required measure, or failed to make a model available for evaluation.
Providers whose GPAI models were placed on the EU market before 2 August 2025 have until 2 August 2027 to bring them into full compliance.
The 2 August 2026 date for GPAI enforcement was not moved by the Digital Omnibus agreed earlier in 2026, which deferred the high-risk obligations: The package pushes the Annex III obligations for stand-alone high-risk AI systems to 2 December 2027, while AI embedded in regulated products under Annex I moves to 2 August 2028. The timetable for GPAI enforcement, along with the Article 50 transparency obligations, remains unchanged.
