Summary As proposed in Article 7 of the Cloud and AI Development Act (CADA), Belgium would have to adopt a national cloud and AI strategy within one year of the Regulation's entry into force. Article 7(2) sets out eight mandatory elements (points (a)–(h)), from governance aligned with the "AI first" principle to concrete measures for data centre capacity and for SMEs and SMCs. Belgium would notify the Commission within three months of adoption and review the strategy at least every three years (Article 7(5)).

Detail

CADA, COM(2026) 502 final, is a proposed Regulation. A cornerstone of the proposal is the obligation on every Member State, including Belgium, to develop and maintain a national strategy. Article 7 sets the requirements, the timeline, and the monitoring and review mechanism.

The obligation to adopt a strategy

As proposed in Article 7(1), Member States "shall establish national cloud and AI strategies" by the date of entry into force plus one year. The strategy must be consistent with the Regulation's objectives (Article 7(3)) and must be consistent with, and contribute to, the digital targets under Article 4 of Decision (EU) 2022/2481, the Digital Decade Policy Programme 2030 (Article 7(4)).

The eight mandatory elements

Article 7(2) lists the minimum content of Belgium's strategy:

  1. Key objectives and governance (Article 7(2)(a)): objectives and priorities for cloud and AI adoption in line with the "AI first" principle — which, per the proposal's recitals, is the principle defined in the Apply AI Strategy, urging organisations to reflect on their business processes by considering the needs and opportunities offered by AI while taking potential risks into account — plus a governance and monitoring framework.
  2. Acceleration at all levels (Article 7(2)(b)): measures to accelerate adoption at national, regional and local level, targeting public sector bodies, SMEs and SMCs, including support for the Experience and Acceleration Centres for AI (Article 5).
  3. Strategic sectors (Article 7(2)(c)): measures for the broad deployment of AI in strategic industrial and public sectors, with the proposal naming healthcare, energy and mobility.
  4. Data centre capacity (Article 7(2)(d)): measures to support the deployment of capacity, focused on high-value data centres meeting high environmental and energy-efficiency standards.
  5. High-intensity computing infrastructure (Article 7(2)(e)): investment in AI factories, AI gigafactories and quantum computers as strategic national and cross-border assets.
  6. Procurement and innovation (Article 7(2)(f)): measures to develop capabilities and promote innovation, including public procurement measures and the procurement-of-innovation measures in Article 33.
  7. Open hardware and software (Article 7(2)(g)): support for cloud-stack technologies built on open hardware and software to strengthen technological sovereignty.
  8. Data accessibility (Article 7(2)(h)): measures to ensure access to high-quality data for AI development, preventing data bottlenecks.

How the elements fit together

The eight elements are designed to interlock rather than stand alone. Article 7(2)(a) expressly requires "a governance and monitoring framework," so the "AI first" objectives in point (a) are meant to steer the adoption measures in points (b) and (c); those depend on the infrastructure built under points (d) and (e); and points (f), (g) and (h) supply the procurement levers, the open technological base and the data that make adoption possible. For Belgium — where competences over digital policy, spatial planning and procurement are split across the federal and regional levels — the governance-and-monitoring requirement is particularly significant: the strategy has to show a coherent line of sight across those levels, not merely aggregate separate regional initiatives.

The "AI first" principle

The "AI first" principle is referenced in Article 7(2)(a) as a required orientation for the strategy's objectives. Per the proposal's recitals it is the principle defined in the Apply AI Strategy, urging organisations to reflect on their business processes by weighing the opportunities of AI against potential risks. For Belgian counsel, this means the strategy's implementation guidance should encourage public and private entities to assess AI opportunities proactively rather than reactively — and, given the "while taking into account potential risks" qualifier, to do so within a risk-aware framework rather than as an unconditional push to deploy AI everywhere.

Support for SMEs and SMCs

Article 7(2)(b) expressly requires measures to accelerate adoption among SMEs and SMCs, reinforced by the Centres for AI under Article 5, which the proposal frames as "entry points to the European AI innovation ecosystem." Belgium's strategy must therefore set out how it will give smaller companies practical access to these resources — for example access to compute, AI tools, testing environments and upskilling — rather than treating SME support as an aspiration. Because the procurement-of-innovation provisions also bear on smaller firms (Article 33 sets an objective that at least 25% of cloud and AI procurement be awarded to innovative SMEs, and requires that objective to be reflected in the national strategy), the SME dimension threads through both the adoption element in point (b) and the procurement element in point (f).

Consistency with the Digital Decade targets

Article 7(4) requires Belgium's strategy to be consistent with, and contribute to, the digital targets established under Article 4 of Decision (EU) 2022/2481 (the Digital Decade Policy Programme 2030). Those targets include ambitions such as a high share of Union enterprises taking up cloud, big data and AI, and the deployment of secure, climate-neutral edge nodes. The strategy therefore cannot be drafted in isolation from Belgium's existing Digital Decade reporting; the two are meant to reinforce each other.

Notification and review

Under Article 7(5), Belgium must notify the Commission within three months of adoption, must assess the strategy at least every three years on the basis of key performance indicators, and update it where necessary; the Commission monitors adoption and revision. The three-month window runs from adoption, not from the Regulation's entry into force, so the sequence is: adopt within one year of entry into force, notify within three months of adoption, then review at least triennially. Because the review is anchored to key performance indicators, the indicators chosen at the outset effectively determine how the strategy is judged later — making indicator design an early priority rather than an afterthought.

Coordination with the AI Board

Article 7(6) provides that the European Artificial Intelligence Board (the "AI Board"), established under Regulation (EU) 2024/1689, advises and assists Member States on coordinating national strategies and facilitates the exchange of best practices.

The strategy as a delivery vehicle for other obligations

The strategy is not a stand-alone document; several other provisions of the proposal route through it. Article 33(4) requires Member States to include in their national strategy their plans for achieving the objective that at least 25% of cloud and AI procurement be awarded to innovative SMEs. The Article 5 Centres for AI are expressly named in Article 7(2)(b) as something the strategy must support. And the strategy's infrastructure elements (points (d) and (e)) sit alongside the data centre acceleration-zone and strategic-project regimes in the proposal's data centre Title. For Belgium, the drafting task is therefore partly one of cross-referencing: the strategy must show how the country's procurement, support-structure and infrastructure commitments cohere, rather than addressing each in isolation. Counsel reviewing a draft strategy should check that each Article 7(2) element is matched by a concrete, monitorable measure, since the three-yearly review under Article 7(5) is conducted against key performance indicators.

What this means for you

For in-house counsel and compliance officers in Belgium:

  1. Strategic alignment. Track the national strategy once published; it will shape procurement and data-governance priorities.
  2. Governance frameworks. The "AI first" principle suggests formalised processes for assessing AI opportunities and risks.
  3. Data and infrastructure. If you rely on data centres or high-intensity compute, watch the strategy's sustainability and capacity measures.
  4. Procurement compliance. Suppliers to the public sector should expect emphasis on open source, European supply chains and sovereign cloud.
  5. Continuous monitoring. Given the three-year review cycle, build internal tracking so you do not miss updates.

Common misconceptions

  • "The strategy is optional or advisory." Article 7(1) uses "shall" — it is a binding obligation as proposed.
  • "The strategy only applies to the public sector." Its measures expressly target SMEs, SMCs and strategic industrial sectors (Article 7(2)(b) and (c)); private entities are affected through procurement and funding.
  • "The 'AI first' principle is just a recommendation." It is embedded in the mandatory element at Article 7(2)(a).
  • "Once adopted, the strategy is static." Article 7(5) requires review at least every three years.

Official sources

Related

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