Summary Yes, as proposed in the Cloud and AI Development Act (CADA), the network of Experience and Acceleration Centres for AI is explicitly required to cooperate with existing EU networks. Article 5(7) mandates that Member States and the Commission cooperate with networks established under other Union initiatives, specifically citing those in the fields of semiconductors and data. This ensures the Centres for AI function as integrated nodes within a broader European digital infrastructure rather than isolated entities. This external cooperation complements the internal network structure established in Article 5(6), creating a cohesive ecosystem for AI adoption.

Detail

The Cloud and AI Development Act (CADA), as proposed in COM(2026) 502 final, establishes a dedicated network of Experience and Acceleration Centres for AI (hereinafter "Centres for AI") to drive the adoption of cloud and AI technologies across the Union. A critical design feature of this network is its interoperability and mandatory cooperation with pre-existing European structures. The proposal does not envision these Centres operating in a vacuum; instead, it legally binds them to collaborate with other key pillars of the EU's digital ecosystem to avoid fragmentation and maximize resource efficiency.

The Legal Mandate for External Cooperation

The primary legal basis for this external cooperation is found in Article 5(7) of the CADA proposal. The text states:

"Member States and the Commission shall cooperate with existing networks established under other Union initiatives, including Union initiatives in the field of semiconductors and data."

This provision creates a statutory obligation for both national authorities (who are responsible for establishing the Centres) and the European Commission to align the activities of the Centres for AI with other major EU initiatives. The explicit mention of "semiconductors and data" is significant. It signals that AI adoption cannot be decoupled from the underlying hardware supply chain (semiconductors) or the fuel that powers AI models (data). By mandating cooperation, the CADA ensures that Centres for AI can leverage the expertise, infrastructure, and resources of these adjacent networks, preventing the creation of siloed AI initiatives that ignore hardware constraints or data availability.

Integration with the Internal Network Structure

This external cooperation mandate complements the structural requirements set out earlier in Article 5. Article 5(6) establishes a formal "network of Centres for AI" to support collaboration and the exchange of best practices among the Centres themselves. It further specifies that this internal network should provide specialized services across regions where required skills or compute capacity are not available locally.

Article 5(7) extends this collaborative logic outward. While Article 5(6) focuses on horizontal cooperation between the Centres for AI, Article 5(7) focuses on vertical and lateral cooperation with other EU-wide initiatives. This dual approach ensures that the Centres for AI are both internally cohesive and externally connected. The internal network ensures that a Centre in a peripheral region can access expertise from a Centre in a hub; the external network ensures that all Centres can access the broader EU infrastructure for chips and data.

Specific Areas of Cooperation

The CADA proposal highlights two primary areas for this mandated cooperation, reflecting the strategic priorities of the EU's digital sovereignty agenda:

  1. Semiconductors: The proposal references initiatives such as the Chips Act 2.0 (referenced in Recital 27 and Recital 31 of the explanatory memorandum). The Centres for AI are expected to work with semiconductor networks to address hardware constraints, facilitate access to AI-optimized processors, and support the co-design of hardware and software. This is crucial for operational objective 2 (supporting cloud computing stacks for technological autonomy) and operational objective 4 (advancing capabilities in physical AI). The cooperation ensures that the software and AI models promoted by the Centres are compatible with and can leverage the latest EU-designed hardware.
  2. Data: The proposal references the Data Union Strategy and common data spaces (Recital 14, Recital 25). Centres for AI are tasked with establishing synergies with initiatives launched under the Data Union Strategy. This cooperation is vital for ensuring that AI models have access to high-quality, compliant data, which is a frequent bottleneck for SMEs and public sector bodies. By integrating with data networks, the Centres can help users navigate data sharing, interoperability, and the legal frameworks surrounding data use.

Broader Context of Integration

Beyond Article 5(7), the CADA proposal positions the Centres for AI as part of a wider, interconnected ecosystem. Article 5(1) states that these Centres shall "build on the European Digital Innovation Hubs (EDIHs)." This means the Centres for AI inherit the existing relationships and infrastructure of the EDIHs, which already have established cooperation mechanisms with other EU networks. This "building on" approach ensures continuity and avoids the duplication of efforts.

Furthermore, Article 5(3) tasks the Centres with ensuring access to upskilling schemes in close collaboration with the AI Skills Academy. This implies a necessary cooperation with education and training networks to deliver on the human capital component of AI adoption. The synergy between the Centres, the AI Skills Academy, and the semiconductor/data networks creates a comprehensive support system that addresses technology, infrastructure, and talent simultaneously.

What this means for you

For CTOs, architects, and SMEs evaluating the practical impact of the proposed CADA, this cooperation mandate has several direct implications:

  • Access to Holistic Solutions: When you engage with a Centre for AI, you should not expect it to only provide AI software advice. Due to the mandated cooperation with semiconductor and data networks, these Centres are positioned to offer guidance on hardware procurement (e.g., access to EU-designed chips) and data governance (e.g., access to common data spaces). This integrated support can simplify the complex task of building a sovereign AI stack.
  • Regional Resource Sharing: If your local Centre for AI lacks specific expertise or compute capacity, Article 5(6) ensures they can tap into the broader internal network. Combined with the external cooperation mandates in Article 5(7), this means you may have access to resources from semiconductor testbeds or data hubs that are not physically located in your immediate region but are connected through these formalized networks.
  • Standardization and Interoperability: The cooperation with existing networks promotes the use of common standards. For architects, this means that the solutions and architectures promoted by Centres for AI are more likely to be interoperable with other EU-backed infrastructure, reducing vendor lock-in and facilitating multi-cloud or hybrid-cloud strategies.
  • Strategic Alignment for Funding: For SMEs seeking funding or support, aligning your projects with the priorities of these interconnected networks (semiconductors, data, AI) will likely be advantageous. The CADA encourages a cohesive ecosystem, so projects that demonstrate integration across these domains may receive greater support from the Cloud and AI Leadership Initiatives.

Common misconceptions

  • Misconception: The Centres for AI replace existing EDIHs.
    • Reality: No. Article 5(1) explicitly states that Centres for AI shall "build on" the European Digital Innovation Hubs. They are an evolution or specialization of the EDIH network, not a replacement. The existing relationships and infrastructure of EDIHs are preserved and expanded.
  • Misconception: Cooperation with other networks is optional or informal.
    • Reality: No. Article 5(7) uses the mandatory verb "shall cooperate." This is a legal obligation for Member States and the Commission, ensuring that cooperation is structured and sustained, not left to chance.
  • Misconception: The Centres for AI only focus on software.
    • Reality: No. By mandating cooperation with semiconductor initiatives, the CADA explicitly links AI adoption to hardware. The Centres are expected to address the full stack, including the physical infrastructure and chips required for AI workloads.

Related

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