Summary Under the proposed Cloud and AI Development Act (CADA), the European Commission selects "frontier AI priority projects" exclusively through open calls for expression of interest. As set out in Article 8, this mechanism ensures that only pioneering, cross-border initiativesβ€”undertaken by a consortium involving at least three Member States and pooling computing resourcesβ€”can be recognised by a formal Commission decision. This process is designed to align strategic AI development with the Union's "Grand Challenge 3" and ensure transparent, merit-based allocation of high-performance computing resources.

Detail

The proposed Cloud and AI Development Act (CADA), COM(2026) 502 final, establishes a comprehensive framework to strengthen Europe's cloud and AI ecosystem. A critical component of this framework is the identification and support of "frontier AI priority projects." These are not ordinary research initiatives; they are strategic assets intended to push the boundaries of AI capabilities, specifically in areas like advanced reasoning, cross-modal understanding, and agentic systems.

The Act deliberately avoids ad-hoc selection. Instead, it mandates a rigorous, transparent procedure to ensure that only projects meeting the highest strategic criteria receive the status of "priority project" and the associated support.

The Exclusive Entry Route: Open Calls for Expression of Interest

The primary and exclusive gateway for any entity wishing to have its project recognised as a frontier AI priority project is the open call for expression of interest.

Article 8 of the proposal explicitly states:

"The Commission may, by means of a decision, recognise as frontier AI priority projects, projects selected through open calls for expression of interest..."

This provision establishes two fundamental rules:

  1. Exclusivity: The Commission cannot simply designate a project on its own initiative or through a closed negotiation. The project must be selected through an open call. This ensures that the opportunity to compete for priority status is available to all eligible entities across the Union, fostering competition and innovation.
  2. Transparency: The "open call" mechanism requires the Commission to publish the criteria, the selection process, and the timeline publicly. This prevents opaque decision-making and ensures that the selection is based on merit and alignment with Union objectives rather than political discretion.

These open calls are specifically linked to Grand Challenge 3 set out in Annex I of the Regulation. Grand Challenge 3 focuses on "Developing the next generation of multimodal frontier AI models and systems and pioneering novel capabilities." Therefore, any expression of interest submitted must demonstrate how the project addresses these specific architectural and capability goals.

The Three Cumulative Criteria for Recognition

Submitting an expression of interest is merely the first step. Article 8 sets out three cumulative criteria that a project must fulfil to be recognised. Failure to meet even one of these criteria disqualifies the project.

1. Pioneering Nature and Strategic Focus

The first criterion requires that the project be "a pioneering project, focused on the support and scaling-up of frontier AI technologies." This definition excludes incremental improvements or the deployment of existing, mature AI models. The project must aim to develop next-generation multimodal models that push the boundaries of current algorithmic capabilities. As detailed in the explanatory memorandum and Annex I, this includes advancing capabilities in:

  • Advanced reasoning and cross-modal understanding.
  • Agentic capabilities (AI systems that can perceive and act autonomously).
  • Novel approaches to model efficiency and cognitive modelling.

2. Consortium Structure and Cross-Border Participation

The second criterion imposes a strict structural requirement on the legal entity undertaking the project. It must be:

  • Undertaken by a European digital infrastructure consortium (EDIC) established pursuant to Decision (EU) 2022/2481, OR
  • Undertaken by "another legal entity eligible for funding under Union law."

Crucially, regardless of the legal form, the project must involve the participation of at least three Member States. This requirement is a deliberate policy choice to:

  • Foster Collaboration: Ensure that frontier AI development is a collective European effort rather than a fragmented national one.
  • Distribute Benefits: Guarantee that the strategic advantages of frontier AI are shared across the Union.
  • Mitigate Risk: Pool resources and risks among multiple jurisdictions.

This cross-border element is non-negotiable. A project led by a single Member State or a private entity without the involvement of three Member States cannot be recognised, regardless of its technical merit.

3. Resource Pooling

The third criterion addresses the practical implementation of the project. The participating Member States must "pool computing time and other relevant resources to support the implementation of the designated project." This criterion acknowledges that frontier AI development is computationally intensive. It requires Member States to commit tangible resourcesβ€”specifically computing timeβ€”to the project. This ensures that the project has the necessary infrastructure to succeed and that the recognition is backed by concrete national commitments.

The Commission Decision: Formal Recognition

Once a project is selected through the open call and verified to meet the three cumulative criteria, the Commission formally recognises it. Article 8 specifies that this recognition is granted "by means of a decision."

This is a significant legal instrument. The decision is not merely an administrative note; it is a binding act that confers the status of "frontier AI priority project" upon the initiative. This status triggers specific rights and obligations under the Regulation:

  • Allocation of Resources: Under Article 9, the Union and Member States are obliged to ensure that sufficient AI computing resources are allocated to support these designated projects.
  • Matching Mechanism: Article 9(2) states that the Union shall "at least match the AI computing resources contributed by Member States to frontier AI priority projects" to the extent that sufficient capacity is available within the Union's share of European High-Performance Computing (EuroHPC) access time.
  • Strategic Importance: The designation signals that the project is of strategic importance to the Union, potentially unlocking additional support mechanisms and funding streams.

Transparency and Accountability in Selection

The combination of open calls, cumulative criteria, and formal decisions creates a robust framework for transparency and accountability:

  • Clear Criteria: The criteria in Article 8 are explicit and public. Applicants know exactly what is required: pioneering tech, a cross-border consortium (3+ MS), and resource pooling.
  • Open Process: The "open call" requirement ensures that the selection is competitive and accessible.
  • Documented Outcome: The Commission's decision provides a clear, documented record of which projects have been selected and why, based on the published criteria.

This structure allows stakeholders, including cloud service providers, data centre operators, and research institutions, to understand the selection logic and plan their strategies accordingly. It also provides a mechanism for accountability, as the Commission's decisions can be scrutinised against the statutory criteria.

What this means for you

For cloud service providers, data centre operators, and research consortia, understanding the role of Commission open calls is critical for strategic planning and market positioning.

1. Strategic Partnership Formation

The requirement for at least three Member States to participate means that no single entity can succeed alone. You should proactively seek partnerships with:

  • National Research Institutes: To access the necessary expertise and national backing.
  • Other Cloud Providers: To pool infrastructure and computing resources.
  • Public Sector Bodies: To align with public interest goals and secure Member State commitment.

Forming a consortium that includes entities from three different Member States is a prerequisite for even entering the selection process.

2. Monitoring Open Calls

Since selection is exclusively via open calls, you must monitor the Commission's announcements closely. These calls will be the only entry point for frontier AI priority projects. Missing an open call means missing the opportunity to be recognised as a priority project for that cycle.

3. Aligning with Grand Challenge 3

Your project proposals must explicitly demonstrate how they address Grand Challenge 3 (frontier AI). This means focusing on:

  • Next-generation multimodal models.
  • Advanced reasoning and agentic capabilities.
  • Novel algorithmic approaches. Projects focused on incremental improvements or mature AI applications will not meet the "pioneering" criterion.

4. Preparing for Resource Pooling

If your consortium is selected, you will be expected to pool computing time and resources. As a provider, you should be prepared to:

  • Offer high-performance computing capacity that can be integrated into a pooled resource.
  • Demonstrate the ability to support the specific technical requirements of frontier AI (e.g., low latency, high throughput).
  • Engage in the resource allocation mechanisms under Article 9, where Union resources will match Member State contributions.

5. Leveraging the "Priority" Status

Once recognised, a project gains significant advantages:

  • Guaranteed Access: Priority access to EuroHPC capacity.
  • Matching Funds: The Union will match Member State contributions.
  • Market Signal: Recognition serves as a strong market signal, potentially attracting further private investment and partnerships.

Common misconceptions

"Any AI project can apply for priority status."

  • Reality: No. Only projects that are pioneering, involve a consortium of at least three Member States, and pool computing resources are eligible. Small, national-scale projects or those focused on mature AI technologies do not qualify.

"The Commission can select projects directly without an open call."

  • Reality: Article 8 is explicit: recognition is possible only for projects selected through open calls for expression of interest. The Commission cannot bypass this process.

"Meeting the criteria guarantees selection."

  • Reality: While meeting the three cumulative criteria is a prerequisite, it does not guarantee selection. The Commission evaluates proposals against strategic priorities, resource availability, and the overall quality of the proposal. The open call is competitive.

"Recognition is permanent."

  • Reality: While the decision grants priority status, the ongoing support (such as the matching of compute resources under Article 9) depends on the continued fulfilment of the project's objectives and the availability of EuroHPC capacity. The Commission may also review and update the list of priority projects as technological landscapes evolve.

"Only large tech giants can win."

  • Reality: The requirement for cross-border consortiums and the focus on strategic, rather than purely commercial, goals opens the door for academic institutions, public research bodies, and mid-sized specialised providers to participate and lead projects, provided they can form the required consortium.

Official sources

Related

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