Summary Under the proposed Cloud and AI Development Act (CADA), while Member States are mandated to establish Experience and Acceleration Centres for AI ('Centres for AI'), the specific procedural rules governing their creation are set by the European Commission. Article 5(4) explicitly empowers the Commission to adopt implementing acts detailing the establishment procedure, participant organisation profiles, selection criteria, and the implementation of tasks. These acts are adopted under the examination procedure referred to in Article 46(2) of the proposal, ensuring Member State oversight through a committee system.

Detail

The proposed Cloud and AI Development Act (CADA), COM(2026) 502 final, establishes a comprehensive framework to strengthen the Union's cloud and AI ecosystem. A critical component of this framework is the deployment of a network of Experience and Acceleration Centres for AI (hereinafter 'Centres for AI'). These centres are designed to serve as regional and local accelerators, bridging the gap between advanced research and the practical adoption of AI and cloud technologies by small and medium-sized enterprises (SMEs), small mid-caps (SMCs), and public sector bodies.

While the legislative text places the primary obligation of establishment on Member States, the detailed regulatory architecture that ensures uniformity, quality, and strategic alignment across the Union is delegated to the Commission. This structure balances national sovereignty in implementation with the need for a harmonised EU-wide approach.

The Legal Basis: Article 5 and the Commission's Delegated Power

Article 5(1) of the proposal mandates that "Each Member State shall establish Experience and Acceleration Centres for AI ('Centres for AI')." These centres must build upon the existing network of European Digital Innovation Hubs (EDIHs) established under Article 16 of Regulation (EU) 2021/694, or any successor entities established under Union law. The objective is to leverage existing infrastructure rather than creating entirely new administrative bodies from scratch.

The proposal outlines the strategic objectives of these centres in Article 5(2), which include supporting the integration of AI use cases in strategic sectors, accelerating broad adoption at regional and local levels, and leveraging infrastructure for model development. Furthermore, Article 5(3) details specific tasks, such as helping organisations accelerate digital transformation, providing access to upskilling schemes in collaboration with the AI Skills Academy, and facilitating the transfer of expertise across regions.

However, the proposal recognises that high-level objectives alone are insufficient to guarantee a coherent network. To prevent fragmentation and ensure that all Centres for AI meet minimum standards of capability and governance, Article 5(4) grants the Commission specific rule-making powers.

The Scope of Implementing Acts under Article 5(4)

Article 5(4) states: "The Commission may adopt implementing acts detailing the procedure for establishing Centres for AI and further arrangements concerning the participant organisation profile, selection criteria and details on the implementation of the tasks and functions."

This provision is the cornerstone of the regulatory framework for the Centres. It authorises the Commission to flesh out the following critical elements:

  1. Establishment Procedure: The implementing acts will define the step-by-step administrative process that Member States or candidate organisations must follow to be formally recognised as a Centre for AI. This ensures a transparent and consistent application process across all Member States.
  2. Participant Organisation Profile: The acts will specify the characteristics, qualifications, and structural requirements for the organisations that constitute the Centres. This ensures that only entities with the requisite technical and operational capacity can participate.
  3. Selection Criteria: The Commission will define the specific metrics and standards used to select which existing hubs or new entities qualify as Centres for AI. This prevents arbitrary selection and ensures that the network is composed of the most capable actors.
  4. Implementation of Tasks and Functions: The acts will provide detailed guidance on how the centres must execute their mandated tasks. This includes specific methodologies for connecting organisations with European providers, delivering upskilling schemes, and facilitating expertise transfer.

The Adoption Procedure: Article 46(2) Examination Procedure

Crucially, the Commission does not adopt these implementing acts unilaterally. Article 5(4) explicitly stipulates that these acts "shall be adopted in accordance with the examination procedure referred to in Article 46(2)."

The examination procedure, as defined in Regulation (EU) No 182/2011, involves a committee composed of representatives of the Member States, chaired by a representative of the Commission. Under this procedure:

  • The Commission submits a draft of the implementing act to the committee.
  • The committee votes on the draft.
  • If the committee delivers a positive opinion (a qualified majority in favour), the Commission adopts the act.
  • If the committee delivers a negative opinion, the Commission cannot adopt the act and must either submit a revised draft or refer the matter to an appeal committee.
  • If no opinion is delivered, the Commission may generally adopt the act, though this is subject to specific conditions.

This mechanism ensures that Member States retain a direct and substantive role in shaping the rules that will govern the Centres for AI within their territories. It acts as a check on Commission power, ensuring that the detailed rules are politically acceptable and practically feasible for national authorities.

Autonomy vs. Regulation

It is essential to distinguish between the rules for establishment and the operational autonomy of the centres once established. While Article 5(4) empowers the Commission to set the rules for how centres are created and selected, Article 5(5) guarantees that "Centres for AI shall have substantial overall autonomy as regards their organisation, composition and working methods, in compliance with the objectives set out in this Regulation."

This means that once a centre is established in accordance with the Commission's implementing acts, it retains significant freedom in how it organises its internal governance, selects its staff, and manages its daily working methods. The Commission's rules focus on the entry criteria and outcome objectives, not on micromanaging the internal administration of each centre, provided they remain compliant with the Regulation's overarching goals.

What this means for you

For legal counsel, compliance officers, and strategic planners within the AI and cloud ecosystem, the interplay between Article 5(4) and Article 46(2) has significant implications:

  1. Monitoring Secondary Legislation: The primary rules for establishing Centres for AI are not yet final. They reside in the proposed Regulation, but the operational details will be defined in future implementing acts. Stakeholders must monitor the Commission's work under the examination procedure to anticipate the specific selection criteria and establishment procedures that will apply.
  2. Preparation for Selection: Organisations aiming to become part of a Centre for AI (either as a lead entity or a partner) should prepare to demonstrate compliance with the forthcoming selection criteria. These will likely focus on technical expertise, the ability to support SMEs and public bodies, and alignment with EU strategic priorities such as sovereignty and sustainability.
  3. National Coordination: Member States will need to align their national strategies and selection processes with the Commission's implementing acts. Legal teams advising public authorities should ensure that national procedures are flexible enough to adapt to the EU-wide rules once adopted.
  4. Autonomy in Operations: For entities already operating as EDIHs or similar hubs, the guarantee of "substantial overall autonomy" under Article 5(5) is a key selling point. While they must meet the entry criteria set by the Commission, they will retain control over their internal working methods, allowing them to adapt to local market needs.

Common misconceptions

Misconception 1: Member States have complete freedom to decide how Centres for AI are established. Correction: While Member States are responsible for the physical establishment, they do not have unfettered discretion. Article 5(4) mandates that the Commission adopt implementing acts detailing the establishment procedure and selection criteria. Member States must adhere to these EU-wide standards to ensure coherence across the Union.

Misconception 2: The Commission can set these rules without Member State input. Correction: The Commission's power is constrained by the examination procedure under Article 46(2). This requires a vote by a committee of Member State representatives. The Commission cannot adopt the implementing acts without the committee's approval (or under specific conditions if no opinion is given), ensuring Member State oversight.

Misconception 3: The Centres for AI are fully independent of EU oversight. Correction: While Article 5(5) grants "substantial overall autonomy" regarding organisation and working methods, this autonomy is conditional on compliance with the Regulation's objectives and the detailed rules set by the Commission's implementing acts. The centres are not independent of the regulatory framework.

Misconception 4: The rules are already in force. Correction: CADA is currently a proposal (COM(2026) 502 final). The implementing acts under Article 5(4) have not yet been adopted. They will only be developed and adopted after the Regulation enters into force and the Commission initiates the examination procedure.

Related

This is general information about a draft EU regulation, not legal advice.