Summary Under the proposed Cloud and AI Development Act (CADA), a frontier AI priority project must involve the participation of at least three Member States to qualify for recognition and the associated allocation of strategic computing resources. This requirement, explicitly set out in Article 8(b), is designed to ensure that these capital-intensive and technically complex initiatives benefit from a "collaborative approach at Union level." By mandating cross-border participation, the proposal prevents the fragmentation of strategic assets and ensures that the development of frontier AI models serves the collective competitiveness and technological sovereignty of the entire Union, rather than isolated national interests.

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 support for "frontier AI," defined in Article 2(4) as "AI models or AI systems built upon such models that can perform a wide variety of tasks and that approach, reach or exceed the current state of the art." Recognizing that these technologies represent critical strategic assets, the proposal introduces a mechanism for the Commission to designate specific initiatives as "frontier AI priority projects."

However, the proposal imposes strict eligibility criteria to ensure that public support and strategic resources are directed toward projects that genuinely advance the Union's collective capabilities. The most significant structural constraint is found in Article 8(b), which stipulates that a project must "involve the participation of at least three Member States." This is not merely a procedural formality but a substantive condition rooted in the unique challenges of frontier AI development.

The Rationale: A Collaborative Approach at Union Level

The requirement for multi-state participation is directly linked to the nature of the technology and the scale of resources required. Recital 34 of the proposal provides the legislative context, stating that "Given the unprecedented scale of resources required for frontier AI development, it is necessary to set criteria for the designation of a project as a frontier AI priority project."

The recital explicitly identifies two key characteristics of these projects:

  1. Technical Complexity: The development of frontier AI involves "technical complexity" that often exceeds the capacity of individual national research infrastructures.
  2. Capital Intensity: These projects are "capital-intensive," requiring massive investments in computing power, data, and talent.

Because of these factors, Recital 34 concludes that such projects "require a collaborative approach at Union level." A single Member State acting alone may lack the necessary scale, diversity of expertise, or financial capacity to compete globally in frontier AI. By requiring the participation of at least three Member States, the CADA ensures that:

  • Resources are pooled: As mandated by Article 8(c), participating Member States must "pool computing time and other relevant resources to support the implementation of the designated project." This pooling is essential to achieve the scale necessary for training and deploying models that approach the state of the art.
  • Strategic Sovereignty is reinforced: Frontier AI is viewed as a strategic asset. A multi-state structure ensures that the resulting technologies are developed under a shared Union framework, reducing reliance on third-country providers and ensuring that the benefits of the innovation are distributed across the Union.
  • Fragmentation is avoided: The requirement prevents the emergence of competing, isolated national AI stacks that could undermine the single market and the Union's collective bargaining power in the global AI landscape.

Eligibility Criteria and Legal Structure

To be recognized as a frontier AI priority project, an initiative must satisfy a cumulative set of criteria outlined in Article 8. Beyond the three-Member-State requirement, the project must:

  • Be a "pioneering project, focused on the support and scaling-up of frontier AI technologies" (Article 8(a)).
  • Support "grand challenge 3" set out in Annex I, which focuses on developing next-generation multimodal frontier AI models and systems.
  • Be undertaken by a specific type of legal entity. Article 8(b) requires the project to be undertaken by "a European digital infrastructure consortium established pursuant Decision (EU) 2022/2481 or another legal entity eligible for funding under Union law."

The involvement of a European Digital Infrastructure Consortium (EDIC) is significant. EDICs are legal structures specifically designed to facilitate cross-border cooperation in digital infrastructure. Their use ensures that the project has a robust governance framework capable of managing the complex legal, financial, and operational coordination required between multiple Member States.

The Commission's Role and Resource Allocation

The Commission has the authority to recognize projects "by means of a decision" once it is satisfied that all criteria in Article 8 are fulfilled. This recognition is the gateway to significant strategic support.

Once designated, a frontier AI priority project gains access to the allocation of AI computing resources. Article 9 establishes that "The Union and the Member States shall ensure that sufficient AI computing resources from their compute capacities are allocated to support the development of frontier AI priority projects." Crucially, Article 9(2) mandates 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 AI computing capacity is available within the Union's share of European high performance computing access time."

This matching mechanism creates a powerful incentive for Member States to collaborate. By pooling their resources in a project that meets the three-Member-State threshold, they unlock additional computing capacity from the Union's EuroHPC infrastructure, effectively amplifying their national investments.

What this means for you

For legal counsel, compliance officers, and strategic planners at organizations involved in AI research and development, the three-Member-State requirement has profound implications for project structuring and consortium building.

  • Consortium Architecture is Mandatory: You cannot apply for frontier AI priority project status as a standalone national entity or a bilateral partnership. Your consortium must be structured to include active, substantive participation from at least three distinct Member States. This requires early engagement with national authorities and research entities in those jurisdictions.
  • Legal Vehicle Selection: The project must be undertaken by an EDIC or an equivalent entity eligible for Union funding. Your legal team must verify that the chosen vehicle complies with Decision (EU) 2022/2481 and is capable of managing the cross-border pooling of resources. The consortium agreement must clearly define the contributions of each Member State to ensure compliance with Article 8(c).
  • Resource Pooling Documentation: Compliance is not just about signing up; it is about demonstrating active resource contribution. You must be prepared to document how computing time, data, and other relevant resources are being pooled across the participating Member States. This documentation will be scrutinized during the application process and subsequent monitoring.
  • Strategic Alignment with Grand Challenge 3: Your project proposal must explicitly align with "grand challenge 3" in Annex I. This involves demonstrating how the project advances the development of next-generation multimodal models, particularly in areas like cybersecurity, scientific discovery, or complex data interpretation.
  • Unlocking Matched Compute: The primary tangible benefit of meeting the three-Member-State threshold is access to matched computing resources. Without this designation, your organization would not be eligible for the Union's matching of AI computing resources under Article 9, potentially placing you at a competitive disadvantage in training large-scale models.

Common misconceptions

"Any three companies from different countries count." No. The requirement is for the participation of at least three Member States, not just three private companies. The involvement must reflect the strategic interest and resource contribution of the Member States themselves, typically through their national entities, public research organizations, or designated consortia. The legal entity undertaking the project must be an EDIC or similar Union-eligible structure.

"Frontier AI projects are automatically eligible if they are advanced." Recognition is not automatic. Even if a project is technically advanced, it must be selected through open calls for expression of interest and must fulfill all criteria in Article 8. The Commission has discretion to recognize projects by decision only if the criteria are fully met.

"The requirement is only about funding eligibility." While funding is a factor, the three-Member-State requirement is a fundamental eligibility criterion for the status of a frontier AI priority project. This status is the prerequisite for accessing the matched AI computing resources under Article 9 and aligning with the Union's broader sovereignty goals.

"Participation can be nominal or passive." The requirement implies active involvement. Article 8(c) explicitly requires that participating Member States "pool computing time and other relevant resources." Passive membership without substantive resource contribution would likely fail to satisfy the "collaborative approach at Union level" mandated by Recital 34.

Official sources

Related

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