How to evaluate Union added value in CADA cloud and AI tenders

Under the proposed Cloud and AI Development Act (CADA), contracting authorities must include specific non-price award criteria in public procurement for in

Summary Under the proposed Cloud and AI Development Act (CADA), contracting authorities must include specific non-price award criteria in public procurement for innovative cloud computing services and AI systems to evaluate a tenderer's contribution to the European cloud and AI ecosystem. As proposed in Article 32, these "Union added value" criteria focus on strengthening the Union's digital supply chain, integrating Union-developed technologies, and using hardware designed or manufactured in the Union. Crucially, Article 32(2)(d) mandates that these criteria must remain ancillary and not decisive in the final award, ensuring they support but do not override core technical and financial performance requirements.

Detail

The Cloud and AI Development Act (CADA), as proposed in COM(2026) 502 final, introduces a structured mechanism for public-sector buyers to leverage their purchasing power to strengthen Europe's technological sovereignty. A central component of this framework is Article 32, titled "Union added value," which mandates that contracting authorities include specific non-price award criteria when procuring innovative cloud computing services and AI systems.

This provision transforms the abstract goal of "technological sovereignty" into a concrete, measurable procurement requirement. It moves beyond voluntary preferences to a mandatory inclusion of ecosystem-strengthening criteria in relevant tenders, ensuring that public spending actively contributes to the resilience and autonomy of the Union's digital infrastructure.

The Legal Basis: Article 32

Article 32(1) of the CADA proposal explicitly states that in public procurement procedures for innovative cloud computing services and AI systems, contracting authorities "shall include, as part of the quality evaluation of the tender, non-price award criteria that allow them to evaluate the tenderer's contribution to the development of a European cloud and AI ecosystem."

This obligation applies specifically to innovative procurements. It requires buyers to look beyond the immediate functionality of the service and assess the broader strategic impact of the tenderer's solution on the European industrial base.

The Four Pillars of Evaluation: Article 32(3)

Article 32(3) provides the specific dimensions against which a tenderer's contribution must be evaluated. Contracting authorities must assess the extent to which the tenderer meets the following four criteria:

Strengthening the Digital Technology Supply Chain in the Union. The first criterion evaluates whether the tenderer contributes to strengthening the digital technology supply chain in the Union. This includes, but is not limited to, the use of "software or hardware designed or manufactured in the Union." This criterion encourages tenderers to source components from within the EU, thereby reducing reliance on third-country supply chains and fostering local industrial capacity. It rewards supply chains that are rooted in the Union's industrial ecosystem.
Integration of Union-Developed Technologies. The second criterion assesses whether the tenderer "has integrated technologies developed in the Union." This specifically highlights the integration of "research and development results stemming from Union funded research and development programmes." It also encompasses the use of "tools, such as standards, specifications, software, models or other technology developed in the Union." This rewards tenderers who build upon and commercialize European innovation, ensuring that public R&D investments translate into market-ready solutions.
Strengthening Security of Supply and Ecosystem Development. The third criterion evaluates whether "the innovation required to deliver the service contributes to strengthening the security of supply and the development of a European cloud and AI ecosystem." This criterion looks at the strategic value of the proposed innovation. It asks whether the specific technological approach proposed by the tenderer enhances the resilience and autonomy of the EU's digital infrastructure, rather than merely delivering a functional service.
Use of EU-Designed or Manufactured Hardware. The fourth criterion is hardware-specific. It requires that "the service is delivered, to the greatest extent feasible with regard to market availability and technical requirements, through critical computing, storage and networking hardware components designed and/or manufactured in the Union."

This criterion prioritizes EU-made critical infrastructure components (such as servers, storage arrays, and networking gear). However, it includes a pragmatic caveat: if EU-made components are not feasible due to market availability or technical requirements, tenderers may use hardware components from a third country, provided that those components still "contribute to strengthening the security of supply and the development of a European cloud and AI ecosystem." This ensures that the requirement does not force the use of inferior or unavailable technology but still incentivizes the use of EU components where possible.

The Principle of Ancillarity: Article 32(2)

A critical safeguard in Article 32 is the requirement that these criteria must be ancillary. Article 32(2) lists several conditions for applying non-price award criteria, including that they must be:

Linked to the subject matter of the contract.
Not conferring unrestricted freedom of choice on the contracting authority.
Expressly set out in the procurement documents or in the contract notice.
Ancillary and not decisive in the award of the contract.

The term "not decisive" is legally significant. It means that while Union added value can influence the final score and help differentiate between otherwise similar bids, it cannot be the sole or primary factor for awarding the contract. The core technical and financial criteria—those directly connected to the performance requirements of the service—must remain paramount. This ensures that procurement decisions remain driven by value for money, quality, and functionality, with sovereignty and ecosystem strengthening acting as a positive tie-breaker or differentiator rather than a barrier to competition or quality.

Recital 67 of the CADA explanatory memorandum provides further guidance on scoring, suggesting that contracting authorities could consider a maximum weighting of 15 out of 120 points to be allocated to European added value. This example illustrates how the criterion can be quantified while remaining subordinate to the core award criteria. The recital emphasizes that the criterion should be applied in a manner that "preserves the primacy of technical and financial criteria directly connected to the performance requirements."

What this means for you

For public-sector procurement officers, Article 32 changes how you structure your tender documents for cloud and AI services. You can no longer treat "European origin" or "sovereignty" as vague political aspirations; you must embed them into your evaluation matrix as mandatory, yet ancillary, criteria.

1. Redefine Your Award Criteria

When drafting your tender for innovative cloud or AI solutions, you must explicitly add a section for "Union Added Value." You should break this down into the four categories listed in Article 32(3). For example, you might create sub-criteria for:

Supply Chain: Percentage of hardware components sourced from the EU or evidence of software designed in the Union.
R&D Integration: Documentation linking the solution to specific EU-funded projects or the use of EU-developed standards.
Security of Supply: A narrative explaining how the innovation strengthens the EU ecosystem.
Hardware Feasibility: A declaration on the use of EU-made critical components, with justification for any third-country alternatives.

2. Quantify the Weighting Carefully

While the law does not prescribe a fixed percentage, the guidance in Recital 67 suggests keeping the weighting proportionate (e.g., around 12.5% or less of the total score). Ensure your evaluation methodology clearly demonstrates that these points are secondary to the core technical and financial scores. If a tenderer has a superior technical solution but lower Union added value, they should still be able to win if their technical superiority outweighs the difference in added value points. The "ancillary" nature means these criteria cannot tip the scales if the core requirements are not met.

3. Request Specific Evidence

To evaluate these criteria fairly, you must ask for verifiable evidence in your tender requirements.

For hardware: Ask for supply chain declarations, certificates of origin, or manufacturer attestations.
For R&D integration: Ask for documentation linking their solution to specific EU projects, standards, or open-source foundations.
For third-country hardware: Ask tenderers to explain how those specific components still contribute to security of supply (e.g., through robust contractual safeguards, interoperability standards, or the absence of viable EU alternatives).

4. Ensure Feasibility

When evaluating the hardware criterion (Article 32(3)(d)), remember the "feasibility" clause. Do not disqualify a tenderer solely for using non-EU hardware if they can demonstrate that no EU alternative exists that meets the technical requirements, or that the non-EU component is necessary for security of supply. Your evaluation must be nuanced, not binary. The phrase "to the greatest extent feasible" allows for flexibility while maintaining the strategic objective.

Common misconceptions

Misconception 1: Union added value criteria are decisive. Many believe that CADA allows public buyers to award contracts primarily based on how "European" a provider is. This is incorrect. Article 32(2)(d) explicitly states these criteria must be "ancillary and not decisive." You cannot award a contract to an inferior technical solution simply because it has higher Union added value. The primary focus must remain on the quality, functionality, and price of the service.

Misconception 2: All hardware must be EU-made. Article 32(3)(d) does not mandate 100% EU-made hardware. It requires delivery through EU-designed/manufactured components "to the greatest extent feasible." If a critical component is only available from a third country, you may accept it, provided the tenderer demonstrates that its use still contributes to security of supply and ecosystem development. This prevents market distortion where EU providers lack specific high-end components.

Misconception 3: This applies to all cloud procurement. Article 32(1) specifies that these criteria apply to "innovative cloud computing services and AI systems." While the spirit of the Act encourages broad adoption, the strict legal mandate for these specific non-price criteria is tied to innovation procurement. Standard, off-the-shelf cloud services may not trigger the same mandatory inclusion, though buyers are still encouraged to consider sovereignty through other means (such as the Union assurance levels under Article 30).

Misconception 4: "Union developed" only means EU-funded projects. While Article 32(3)(b) highlights R&D from Union-funded programmes, it also broadly includes "technologies developed in the Union," including standards, specifications, software, and models. A technology developed by a private EU company without direct EU funding still counts as "developed in the Union" for the purpose of this criterion.

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