Leggi in italianoCloud, Artificial Intelligence and European Sovereignty: A New Arena of Transatlantic Tensions
The remarks made by the US ambassador to the European Union, Andrew Puzder, about Europe’s plans to regulate cloud and artificial intelligence services strike me as particularly revealing, especially against the background of growing transatlantic tensions over technological infrastructure. On 3 June 2026, the European Commission presented its proposed Cloud and AI Development Act (CADA), intended to strengthen European cloud and AI capabilities, stimulate investment and establish common criteria for assessing digital sovereignty and resilience. Puzder acknowledges Europe’s right to protect sensitive government information, but warns against using legislation to favour European suppliers at the expense of American competitors.
I think the issue is more complicated than that distinction suggests. Protecting technological sovereignty does not necessarily amount to protectionism. When cloud services support public administration, critical infrastructure or strategic functions, it matters who ultimately controls the systems, where the data are stored and which legal jurisdictions can exercise authority over the providers. Locating infrastructure in Europe is not, by itself, a guarantee that decisions about its operation, access to information and continuity of service are insulated from external pressures. Nor does this mean that any restrictive requirement is justified: measures must be proportionate, transparent and consistent with applicable EU and international obligations. But the security of digital infrastructure cannot be assessed solely through competition and price.
I also find the American position revealing of a broader tension in transatlantic relations. The United States considers it legitimate to protect strategic technologies and restrict access where national security interests are involved. When Europe seeks to reduce its technological dependencies, concerns arise that European policy may disadvantage US companies. Of course, the measures adopted on either side are not necessarily equivalent, but the underlying tension between national security, open markets and strategic autonomy is much the same.
For me, the decisive issue lies elsewhere. Europe cannot establish technological autonomy merely by adding new regulatory requirements. It also needs industrial capacity, infrastructure, investment and credible alternatives. At the same time, it cannot abandon security conditions for services underpinning essential public functions simply for fear of commercial retaliation. Technological sovereignty does not necessarily mean producing everything in Europe. It means retaining the ability to choose, supervise and, where necessary, replace the suppliers on which we depend.This, more than declarations about strategic autonomy, will be the real test of European policy.
Sources and references
- European Commission, 3 June 2026, Proposal for the Cloud and AI Development Act (CADA). Official presentation of the proposal and its three main objectives: innovation, infrastructure capacity and technological autonomy.
- European Commission, COM(2026) 502 final, 3 June 2026, Proposal for a Regulation establishing a framework of measures for strengthening Europe’s cloud and AI ecosystem. Legislative proposal based on Articles 114 and 173(3) TFEU.
- Financial Times, 19 October 2026, US threatens European companies over Starlink restrictions. Reporting on Ambassador Andrew Puzder’s remarks and broader transatlantic tensions over technological infrastructure.
- European Commission, 1 June 2026, Sovereign Cloud Framework explained. Explanation of the framework used to assess sovereignty, jurisdictional control, technological autonomy and resilience in Commission cloud procurement. This framework is separate from the CADA proposal.
- EUR-Lex, Legislative procedure 2026/0138(COD). Documentation on the legislative procedure concerning the proposed CADA Regulation.
