Policy
Brussels Unveils the Cloud and AI Development Act — Europe's $1 Trillion Infrastructure Bet
The European Commission has proposed the Cloud and AI Development Act (CADA), targeting a tripling of EU data centre capacity within five to seven years, a four-tier cloud sovereignty framework, and accelerated AI deployment in critical sectors. US tech industry groups warn of market fragmentation.
By Michael G ·
The European Commission has proposed the Cloud and AI Development Act — known as CADA — as the centrepiece of its Tech Sovereignty Package announced on June 4, 2026. The legislation sets out an ambition to triple EU data centre capacity within five to seven years, establishes a four-tier cloud sovereignty assurance framework for providers operating in the European market, and creates a common procurement framework for public administrations across member states. The proposal represents the most significant European intervention in AI infrastructure policy since the EU AI Act entered into force.
The four sovereignty assurance levels — L1 through L4 — create a tiered framework that determines which cloud providers can serve which categories of public and critical infrastructure workloads. L4, the highest level, requires data to be stored and processed exclusively within EU jurisdiction on infrastructure owned by EU-domiciled entities. L1 permits standard commercial cloud services for non-sensitive workloads. The framework is designed to give public sector organisations a clear procurement guide while creating commercial incentives for cloud providers to invest in EU-sovereign infrastructure.
The capacity tripling target is the most ambitious element of the proposal. Current EU data centre capacity is estimated at approximately 10 gigawatts. Tripling that figure within seven years would require sustained investment of roughly $100 to $150 billion annually — a figure that the Commission acknowledges cannot come from public sources alone. CADA includes provisions to simplify permitting for data centre construction, improve access to energy and water resources, and create financing mechanisms through the European Investment Bank to catalyse private investment.
The proposal also establishes what the Commission calls 'grand challenges' — structured research and deployment programmes for frontier AI, industrial AI, and physical AI. These are modelled loosely on the US DARPA challenge framework and are intended to accelerate AI adoption in healthcare, energy, manufacturing, and defence — sectors where European companies have competitive strengths but have been slower to deploy AI at scale than their US and Chinese counterparts.
The industry response has been divided along predictable lines. European cloud providers and telecommunications companies have broadly welcomed the proposal, seeing the sovereignty framework as a commercial opportunity that favours domestic players. US technology industry groups have been more critical. The Computer and Communications Industry Association warned that certain provisions of CADA are 'discriminatory' and risk fragmenting the European cloud market in ways that increase costs for European businesses without delivering meaningful security benefits. The Business Software Alliance raised concerns about provisions that it argued go beyond legitimate security policy into industrial protectionism.
The legislative timeline for CADA is uncertain. The proposal must pass through the European Parliament and the Council of the EU, a process that typically takes 18 to 24 months for complex technology legislation. The EU AI Act took three years from proposal to final text. CADA's proponents argue that the urgency of the AI infrastructure gap justifies an accelerated timeline, but the complexity of the sovereignty framework and the commercial interests at stake make rapid passage unlikely.
The deeper question CADA raises is whether European industrial policy can meaningfully close the AI infrastructure gap with the United States and China within the timeframe that matters for competitive positioning. The US is currently investing $180 to $190 billion annually in AI infrastructure through Alphabet alone, with Microsoft, Amazon, and Meta adding comparable sums. China's state-directed investment in AI infrastructure is estimated at $100 billion or more annually. Against that backdrop, CADA's ambition to triple EU capacity over seven years — starting from a much smaller base — represents a significant acceleration but not necessarily a convergence.