Summary As proposed in the Cloud and AI Development Act (CADA), Member States and Union entities must actively lower barriers for small and medium-sized enterprises (SMEs) in public procurement for cloud and AI. Article 33(5)(c) specifically obliges contracting authorities to promote the "development of public contract clauses that are favourable for innovative SMEs." This mandate works in tandem with requirements to divide contracts into lots (Article 33(2)) and facilitate matchmaking (Article 33(5)(b)). While CADA does not prescribe exact legal text, it would require public buyers to adjust commercial termsβ€”such as payment schedules, liability caps, and intellectual property rightsβ€”to ensure smaller European innovators can compete effectively against larger incumbents.

Detail

The proposed Cloud and AI Development Act (CADA) identifies the dominance of non-European hyperscalers as a strategic risk. To counter this, the proposal introduces demand-side measures designed to stimulate the supply of sovereign, European cloud and AI services. A central pillar of this strategy is the restructuring of public procurement to favour innovative SMEs.

The Legal Mandate: Article 33(5)(c) The core of this requirement is found in Article 33, titled "Monitoring of procurement of innovation in cloud and AI." Specifically, Article 33(5)(c) states that Union entities and contracting authorities shall promote:

"(c) development of public contract clauses that are favourable for innovative SMEs."

This is a binding obligation for public buyers under the proposal. It moves beyond general encouragement, requiring authorities to actively design contract terms that accommodate the specific constraints of smaller, innovative providers. The objective is to prevent standard "boilerplate" clauses, often drafted for large corporations, from effectively excluding SMEs from the market.

What Would "Favourable" Clauses Look Like? The text of CADA does not provide an exhaustive list of specific clauses. However, the legislative intentβ€”supported by the explanatory memorandum's focus on reducing barriers for SMEs (Recitals 66–68)β€”and standard procurement best practices suggest that "favourable" would primarily refer to commercial and legal terms rather than technical performance standards.

  1. Payment Terms and Cash Flow: SMEs often lack the working capital reserves of large hyperscalers. Favourable clauses would likely include:

    • Shorter Payment Periods: Mandating payment within 30 days (or faster) rather than the standard 60 days often seen in public contracts.
    • Advance Payments: Allowing for partial upfront payments to cover initial development or infrastructure setup costs.
    • Milestone-Based Payments: Structuring payments to align with the delivery of specific, verifiable outputs rather than waiting for final project completion, thereby reducing the financial risk for the SME.

  2. Scope and Liability: Standard public contracts often impose disproportionate liability caps or indemnity requirements that can be fatal to a smaller provider. Favourable clauses might:

    • Cap Liability: Limit the total liability of the SME to a reasonable percentage of the contract value, rather than exposing them to unlimited damages.
    • Exclude Indirect Damages: Explicitly excluding liability for indirect or consequential losses, which are often difficult for SMEs to insure against.
    • Proportionate Penalties: Ensuring that penalty clauses for delays are proportionate to the specific lot or component delivered, rather than applying to the entire project scope.

  3. Intellectual Property (IP) Rights: For innovative SMEs, their IP is their primary asset. Standard contracts often demand full assignment of all IP developed during the project. Favourable clauses would:

    • Retain Background IP: Allow the SME to retain ownership of pre-existing IP and background technology.
    • License Instead of Assign: Grant the public authority a perpetual, royalty-free license to use the deliverables, rather than demanding a full transfer of ownership.
    • Future Commercialisation: Permit the SME to continue using the developed technology for other clients, ensuring the innovation can scale beyond a single public contract.

  4. Flexibility and Milestones: Innovative AI and cloud projects often require iterative development. Rigid, waterfall-style contracts can stifle this. Favourable clauses would:

    • Allow for Iteration: Include provisions for adjusting technical specifications based on testing results without triggering breach-of-contract penalties.
    • Simplified Change Management: Streamline the process for approving necessary changes to the scope, reducing administrative burden.

The Role of Division into Lots (Article 33(2)) The requirement for favourable clauses does not operate in a vacuum. It is intrinsically linked to Article 33(2), which mandates Member States to "support the design of simplified, proportionate and SME-friendly procurement strategies, including division into lots, where appropriate."

Dividing a large cloud or AI procurement into smaller, distinct lots is a prerequisite for SME participation. A single, monolithic contract for a national cloud infrastructure is often too large for an SME to bid on. By breaking the project into lots (e.g., one lot for data migration, another for specific AI model training, another for regional edge nodes), authorities create opportunities for SMEs to compete.

Article 33(5)(c) ensures that once an SME wins a lot, the terms of that specific contract are also tailored to their size. Without favourable clauses, an SME might win a lot but still face contractual terms (e.g., massive liability caps or 90-day payment terms) that make the contract financially unviable. Thus, the combination of Article 33(2) (structural division) and Article 33(5)(c) (contractual terms) creates a holistic framework for SME access.

Monitoring and the 25% Target To ensure these measures translate into results, Article 33(3) requires Member States to monitor and report annually on SME participation trends, including the number of contracts awarded and their share of total contract value.

Furthermore, Article 33(4) sets a concrete objective: Member States must pursue a target where "at least 25% of their procurement for cloud computing services and AI systems be awarded to innovative SMEs." The development of favourable contract clauses is a critical mechanism to achieve this statutory target.

Matchmaking and Preliminary Consultations Complementing these measures, Article 33(5)(b) requires authorities to promote "matchmaking between public buyers and innovative solutions provided by European SMEs and start-ups." Additionally, Article 33(5)(a) mandates "preliminary market consultations." These steps ensure that SMEs are aware of upcoming opportunities and that authorities understand the specific constraints SMEs face, informing the design of the favourable clauses required under Article 33(5)(c).

What this means for you

For cloud service providers, data centre operators, and AI developersβ€”particularly SMEs and small mid-caps (SMCs)β€”these provisions would fundamentally alter the negotiation landscape with public authorities.

  • Leverage in Negotiations: You can now cite Article 33(5)(c) as a legal basis to request adjustments to standard contract templates. If a contracting authority proposes clauses that are disproportionately burdensome regarding liability, IP, or payment terms, you can argue that these fail to meet the CADA mandate to develop clauses "favourable for innovative SMEs."
  • Highlight Innovation: The provision specifically targets "innovative SMEs." Ensure your proposals clearly articulate the innovative nature of your cloud stack, AI models, or sustainability features. This aligns your offering with the specific category of providers the Act aims to support.
  • Prepare for Modular Bidding: With the push for division into lots under Article 33(2), you should prepare modular service offerings. Public buyers will increasingly break down large infrastructure projects. Having clear, modular contract templates ready for these specific lots will give you a competitive edge.
  • Engage in Market Consultations: Participate actively in the preliminary market consultations mandated by Article 33(5)(a). This is your opportunity to educate public buyers on what "favourable" terms look like in practice (e.g., specific payment schedules or IP retention models) before the tender is published.
  • Advocate for Standardisation: Engage with national procurement bodies and industry associations to help define what "favourable" means in your sector. The development of standardised model clauses for SMEs could reduce legal costs for both buyers and providers, accelerating the adoption of CADA's goals.

Common misconceptions

"Favourable clauses mean lower quality or security standards." Incorrect. CADA does not lower technical, cybersecurity, or "Union assurance level" requirements. The "favourable" nature of the clauses refers strictly to commercial and legal terms (payment, IP, liability, milestones). An SME must still meet the same rigorous technical standards as a large provider; the contract terms are simply adjusted to make compliance financially and administratively feasible.

"All SMEs automatically receive these favourable clauses." No. The provision requires authorities to promote the development of such clauses. It is not an automatic right that triggers for every SME bid. However, it creates a strong legal expectation that authorities will design tenders with these terms in mind. If standard clauses are used, the SME may need to demonstrate why they are disproportionately burdensome to trigger the CADA mandate.

"This only applies to EU-level procurement." Incorrect. Article 33 applies to both Union entities and the contracting authorities of Member States. National public procurement for cloud and AI services must align with these principles, meaning SMEs in France, Germany, or Italy would see similar changes in national tenders.

"CADA replaces the AI Act's procurement rules." No. The AI Act regulates the safety and fundamental rights of AI systems. CADA regulates the infrastructure and market conditions for cloud and AI procurement. They are complementary: an SME might need to comply with the AI Act for the safety of its software, while CADA ensures the public contract for that software is structured to allow the SME to win the bid.

Official sources

Related

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