What has been established at this stage must be framed precisely. The Commission has adopted a proposal for a regulation, but the text is not yet applicable law. It must now be negotiated by Member States and the European Parliament. Therefore, nothing has yet been voted on in the full legislative sense. However, the political signal is strong: Brussels intends to create a European framework for cloud and AI sovereignty, steer public demand, accelerate data center deployment, and foster the emergence of a digital value chain that is less dependent on non-European providers.

Brussels has delivered a stark diagnosis. The European Union remains heavily dependent on foreign technologies in areas that have become essential to the functioning of modern economies. The cloud supports messaging, databases, public services, hospitals, energy grids, banks, defense infrastructure and AI applications. Yet, a very large share of the digital products, services, and infrastructure used in Europe originates from third countries, while major American providers largely dominate the European cloud market. AWS, Microsoft Azure and Google Cloud represent about 70% of the European market according to the data provided, placing the Union in a position of operational, legal and geopolitical dependency.

The sensitive point is not merely commercial; it is legal and strategic. European authorities are concerned about exposure to extraterritorial legislation, notably the US Cloud Act, which can compel US-based providers to hand over certain data to American authorities, even when that data is stored outside US territory. Added to this is a more political risk: that of an interruption, restriction, or technical dependency in areas deemed critical. This is the meaning behind the reference to the "kill switch" risk mentioned in the provided materials: Europe wants to prevent an external actor from being able, directly or indirectly, to disrupt essential services.

CADA is built on three pillars: innovation, capacity, and autonomy. Regarding innovation, the Commission aims to support next-generation cloud and AI technologies, particularly frontier AI, industrial AI, and physical AI. The text also provides for "grand challenge" type initiatives to direct research and development efforts toward technologies where Europe can regain a strategic position. This section aligns with previously stated ambitions regarding AI Factories, AI Gigafactories and the EuroHPC program.

The second pillar is likely the most visible for financial markets: capacity. Brussels aims to at least triple European data center capacity over the next five to seven years. To achieve this, the Commission intends to reduce administrative delays, facilitate permitting, and improve access to land, energy, water, and financing. The message sent to the markets is crystal clear: European digital sovereignty will require a wave of infrastructure spending.

The third pillar is autonomy. CADA introduces a single European framework for cloud and AI sovereignty with four levels. Level 1 is based on data storage and processing within EU-based infrastructure. Level 2 adds requirements for independence from third countries and transparency regarding the software chain. Level 3 mandates European ownership and control of the provider, although the Commission may recognize certain third-country providers under specific conditions. Level 4 represents the maximum degree of sovereignty, with full control of the software chain and the absence of foreign interference.

This framework could become the true instrument of market transformation. Public administrations and bodies will have to assess their needs based on their risk level and select cloud or AI services compatible with these requirements. Sensitive sectors such as defense, healthcare, energy, banking, or public infrastructure will be the most affected. The Commission also plans a common public procurement framework at the European level, which could pool demand from administrations and increase the public sector's bargaining power against suppliers.

At this stage, the text does not represent a brutal closure of the European market to American hyperscalers. The Commission insists on maintaining a market largely open to the Union's partners. However, openness is becoming conditional. The more sensitive the use cases, the higher the sovereignty requirements are expected to be. This means that Microsoft, Amazon, and Google (Alphabet) could continue to operate in Europe, but with increased pressure on localization, governance, auditability, software transparency and independence from foreign jurisdictions.

The implicit signal is more powerful than the text itself. Brussels is not just saying it wants more data centers - it is indicating that public demand, funding, permits, and sovereignty criteria could be directed towards actors capable of demonstrating a real European contribution. In other words, public procurement could become an industrial lever. This is precisely what can alter stock valuations: not because revenues change immediately, but because the market begins to revise the addressable market size and the probability of public contracts for certain European players.

The first segment to watch is that of the European sovereign cloud. OVH and Ionos are the most obvious names. These two players already have a European base, a recognized brand, and direct exposure to the theme of digital sovereignty. CADA can offer them dual support: increased demand for offerings qualified as sovereign and better visibility in public tenders. The potential is particularly significant if Levels 3 and 4 become benchmarks in critical public markets. In this scenario, American hyperscalers would retain their dominance over standard use cases, but European providers could gain share in sensitive workloads.

OVH presents a pure-player profile that is more exposed to the theme. The group could benefit from a resurgence of interest in European cloud solutions, particularly in hosting, Infrastructure as a Service (IaaS), and offerings aimed at administrations or companies subject to sovereignty constraints. The risk lies in its execution capacity: CADA may create a commercial opportunity, but it does not automatically resolve competitiveness issues against hyperscalers regarding service depth, software ecosystems, capital intensity and operating margins.

Ionos appears as another natural beneficiary. Its European positioning, exposure to SMEs, cloud services, and hosting give it positive sensitivity to rising sovereignty concerns. Its profile may be perceived as less speculative than a player exclusively focused on the sovereign cloud, depending on its revenue structure and financial discipline. CADA can strengthen Ionos's strategic relevance, especially if European administrations and companies seek to diversify their providers beyond the American trio.

Nebius belongs to a different category. Its exposure is more closely linked to AI infrastructure, high-performance computing, and the growing need for GPU capacity. The goal of tripling European data centers and supporting AI Factories or AI Gigafactories can bolster interest in players capable of providing dedicated AI computing power. The theme is promising, although riskier: it depends on access to chips, energy costs, infrastructure financing and the ability to sign long-term contracts with solvent clients. The market might assign a growth premium to this type of actor, but volatility will remain high.

European semiconductors constitute the other major market segment. CADA cannot be analyzed in isolation from the Chips Act 2.0. The two texts complement each other. Tripling data center capacity and developing AI capabilities requires more chips, power components, equipment, sensors, cooling solutions, network infrastructure, and computing architectures. Brussels wants to bring Europe's share of semiconductors to 20% of the global market by 2030, up from a significantly lower level today. The ambition is very high, especially since the 2025 audit highlighted the unlikely nature of this target without a major acceleration.

The beneficiaries are players positioned in equipment, materials, power chips, data center components, industry, and critical electronics. ASML remains the most obvious European strategic asset in the global semiconductor chain, even if its exposure to CADA is indirect. Infineon and STMicroelectronics could also benefit from a cycle of reshoring, electrification, and industrial investment, although their trajectory will depend more on the automotive and industrial cycles and final demand than on the European package alone. Soitec can be included in the discussion for its role in advanced materials, but with a stronger sensitivity to technological cycles and volumes.

The energy component is equally decisive. Tripling data center capacity in five to seven years implies a significant increase in electricity demand (beneficiaries: Schneider Electric, Legrand, Nexans, Prysmian, Siemens Energy, etc.). The Commission explicitly links data center development to energy efficiency, cooling, grid integration, and heat recovery. Indirect winners could therefore be grid operators, utilities, electrical equipment suppliers, energy management specialists, and companies exposed to data center cooling. CADA is a digital text, but its implementation will be constrained by energy. Without connections, grid capacity, and competitive electricity, cloud sovereignty will remain theoretical.

The established fact is that the Commission proposes to strengthen European capacities, triple data centers, accelerate permitting, structure a sovereignty grid, and steer certain public procurement. The implicit signal is that European providers could benefit from a functional preference in critical use cases, even if the market officially remains open. A reasonable interpretation is that valuation premiums could shift toward actors capable of ticking three boxes: European footprint, industrial execution capacity, and compatibility with future sovereignty criteria.

Europe has previously announced ambitious industrial goals, notably in chips, without always managing to rapidly shift global power dynamics. Funding remains fragmented between the European budget, Member States, and the private sector. Procedures can be simplified, but data centers will still need to secure energy, land, water, and local acceptance. Financing will also be decisive, as AI infrastructure is extremely capital-intensive.

Even if the Commission claims not to target American groups, the logic of the text can be perceived as a progressive challenge to their access to sensitive markets. In an already tense trade environment, the United States could contest certain criteria if they deem them discriminatory. This is why the final version of the text will likely be the result of a compromise: firm enough to address European concerns, but open enough to avoid a frontal confrontation with Washington. Most stocks highly exposed to the theme have already seen strong performances since the beginning of the year (one only needs to look at a European YTD performance ranking, available below; most are from the semiconductor sector). However, the fundamental effect will depend on three variables: the final version of the regulation, the sovereignty levels required by sector, and the actual volume of public procurement or funding mobilized.

European YTD performance ranking through June 4, 2026 (market capitalization over €5bn):