What is Grand Challenge 8 (Public Sector AI) under the proposed CADA?

Summary Grand Challenge 8, titled "Public Sector AI," is a cornerstone of the proposed Cloud and AI Development Act (CADA), COM(2026) 502 final. As proposed, it aims to develop AI models and systems based on high-quality data from the public sector, specifically targeting critical domains such as healthcare, public administration, law, and crisis management. The challenge explicitly mandates the use of privacy-preserving frameworks—such as federated learning and high-fidelity synthetic data generation—to enable data sharing and model training across national borders without compromising the confidentiality of underlying datasets. This initiative serves as the practical implementation vehicle for Operational Objective 7 of the Cloud and AI Leadership Initiatives, seeking to avoid fragmentation and scale user-oriented solutions across the Union.

Detail

The Cloud and AI Development Act (CADA) is a legislative proposal designed to strengthen Europe's cloud and AI ecosystem by boosting domestic capabilities, simplifying deployment, and ensuring technological sovereignty. A core component of this framework is the establishment of the Cloud and AI Leadership Initiatives, which are supported by large-scale, cross-sectoral projects known as "grand challenges." These challenges are detailed in Annex I of the proposal and are designed to address major technological and industrial hurdles facing the Union.

The Scope of Grand Challenge 8

Grand Challenge 8 is explicitly dedicated to developing AI models and systems based on high-quality data from the public sector. According to Annex I(8) of the CADA proposal, the focus is on critical domains such as "healthcare, public administration, law and crisis management as well as public services." The objective is to create public service solutions that are expected to have a "high positive impact on the most critical public services" and are "shared across different levels of public sector organisations."

The proposal emphasizes that these solutions should be developed using data that is often siloed within national or local authorities. By unlocking this data, the EU aims to avoid fragmentation and enable the scaling-up of successful, user-oriented solutions. The text notes that one target will be to "enable data sharing and frontier model development across national public services to increase the impact on the overall Union's public sector, including also in areas handling sensitive data."

Privacy-Preserving Frameworks as a Technical Mandate

A central technical requirement of Grand Challenge 8 is the adoption of privacy-preserving frameworks. Annex I(8) explicitly highlights mechanisms such as "federated learning and high-fidelity synthetic data generation." These technologies allow AI models to be trained on sensitive data without the data itself leaving its secure environment or being exposed.

The proposal states that these frameworks "make it possible to train of models without compromising the confidentiality of underlying datasets." This is crucial for domains like healthcare and law enforcement, where data protection regulations (such as the GDPR) are strict, and public trust is paramount. By enabling training on distributed data sets, the challenge seeks to overcome the legal and technical barriers that currently prevent the pooling of public sector data for AI development.

Connection to Operational Objective 7

Grand Challenge 8 is the practical implementation vehicle for Operational Objective 7 of the Cloud and AI Leadership Initiatives, as defined in Article 4(7) of the CADA proposal. While Operational Objective 7 sets the strategic goal—to "increase the development and adoption of AI models and systems across the Union's public sectors"—Grand Challenge 8 provides the specific project-based structure to achieve it.

Article 4(7) outlines several specific aims that Grand Challenge 8 projects must support:

Accelerating the technological development and uptake of AI in critical public sector domains.
Developing AI models that "increase the effectiveness of public service delivery, improve decision-making, and simplify administrative procedures."
Promoting the "sharing and reusing of training data and AI models across the Union's public services."
Facilitating "secure, privacy-enhancing health data reuse for AI models and tools in healthcare."
Facilitating the development, testing, and deployment of AI models in the automotive sector, including for autonomous driving (noting the intersection between public infrastructure and private sector innovation).

Strategic Importance for Sovereignty

The proposal frames public sector AI as a matter of strategic autonomy. Recital 22 of the CADA explanatory memorandum states that the Union should foster the availability of highly secured computing infrastructures for the training, testing, and deployment of defence-related AI models, but also broadly for public sector AI. By developing these models within the EU, using EU data and adhering to EU values, the proposal aims to reduce dependence on third-country providers.

This aligns with the broader CADA goal of ensuring that critical public services are not subject to the jurisdictional reach or operational disruptions of non-EU entities. The proposal notes that "AI models and systems should be used to support better decision-making, simplify administrative procedures and reduce unnecessary burdens, in particular for critical public domains such as healthcare where data reuse for AI models and tools should be facilitated while ensuring security and data protection."

Funding and Implementation Mechanisms

Projects under Grand Challenge 8 would be implemented through the Cloud and AI Leadership Initiatives, as detailed in Article 6. These initiatives may receive funding from Union programmes such as Horizon Europe and the Digital Europe Programme, as well as from the European Competitiveness Fund (ECF) under the next Multiannual Financial Framework (2028-2034).

Article 6(3) states that the initiatives "may be supported by funding from Union programmes, including Horizon Europe and the Digital Europe Programme." Furthermore, Recital 29 notes that private-sector stakeholders should be encouraged to take into consideration the Cloud and AI Leadership Initiatives when developing their investment strategies. Member States are also encouraged to support these projects through national research, development, and innovation measures, ensuring coherence between national strategies and Union-level priorities.

The implementation is further supported by the network of Experience and Acceleration Centres for AI (Centres for AI) established under Article 5. These centres are tasked with supporting the integration and scaling-up of AI use cases in strategic industrial and public sectors, acting as entry points for public bodies to access the ecosystem.

What this means for you

For public-sector procurement officers, digital transformation leaders, and data stewards, Grand Challenge 8 represents a significant shift in how AI capabilities will be sourced, developed, and shared across Europe.

New Opportunities for Collaboration

If your organisation holds high-quality, structured data in critical domains like healthcare or public administration, you may be eligible to participate in or benefit from Grand Challenge 8 projects. The proposal encourages the pooling of data through trusted third parties and the use of federated learning. This means you could contribute to training a powerful, sovereign AI model without having to transfer sensitive citizen data to a central server, maintaining compliance with strict data protection laws. The focus on "high-quality data from the public sector" implies that organisations with well-governed datasets will be prioritised for these collaborative efforts.

Procurement of Innovation

CADA introduces specific procurement measures to support these initiatives. Article 33 of the proposal requires Member States to monitor their procurement of innovation in cloud and AI, with a specific objective to award "at least 25% of their procurement for cloud computing services and AI systems be awarded to innovative SMEs." As Grand Challenge 8 matures, procurement officers should look for tenders that leverage these shared, open-source, or sovereign AI models. The proposal promotes the use of open standards and components released under open-source licences (Article 41), which can reduce vendor lock-in and long-term costs.

Alignment with National Strategies

Member States are required to adopt national cloud and AI strategies under Article 7. These strategies must include measures to support the broad deployment of AI in strategic public sectors. Grand Challenge 8 provides a template for what these deployments should look like: secure, privacy-preserving, and interoperable across borders. Procurement officers should align their local AI procurement strategies with these national plans to ensure eligibility for funding and support from the Centres for AI established under Article 5.

Data Quality and Governance

Participating in Grand Challenge 8 initiatives requires high-quality data. Annex I(8) explicitly mentions the need for "high-quality data from the public sector." Procurement and IT leaders must invest in data governance and cleaning. The proposal emphasizes that AI systems must be trained on relevant, representative, and error-free data. Poor data quality will not only hinder the performance of AI models but may also disqualify organisations from participating in these high-level collaborative projects.

Common misconceptions

Misconception 1: Grand Challenge 8 is only about buying commercial AI software. Reality: Grand Challenge 8 is focused on developing and deploying sovereign AI models and systems. It emphasizes building European capabilities, often through open-source components and shared public sector data, rather than simply procuring black-box solutions from third-country vendors. The goal is technological autonomy, not just consumption.

Misconception 2: Data must be centralized to be useful for AI. Reality: The proposal explicitly promotes privacy-preserving frameworks like federated learning. This allows data to remain in its original, secure location while still contributing to the training of AI models. Centralization is not a prerequisite for participation in Grand Challenge 8; in fact, decentralized, secure approaches are preferred for sensitive public data.

Misconception 3: Only large national governments can participate. Reality: While the challenges are large-scale, the proposal actively encourages the involvement of SMEs and local authorities. Article 33 sets a target for SME participation in innovation procurement. Furthermore, the network of Centres for AI (Article 5) is designed to support regional and local adoption, ensuring that smaller public bodies can access and benefit from these advanced AI capabilities.

Misconception 4: This replaces existing national AI initiatives. Reality: Grand Challenge 8 is designed to complement and coordinate with national strategies. Article 7 requires Member States to align their national cloud and AI strategies with CADA's objectives. The challenge acts as a framework for cross-border cooperation and shared best practices, rather than a replacement for local innovation efforts.

Official sources

GDPR (Regulation (EU) 2016/679)

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