To achieve tech sovereignty, Europe must not mimic its rivals

The European Union has a goal of ‘technological sovereignty’. The aim is to develop and control the chips, cloud capacity and artificial-intelligence capabilities needed for Europe’s economy and security, rather than depending on potentially unreliable imports. This does not mean a switch to self-sufficiency, which is unrealistic: Europe will continue to rely on foreign technology. Nevertheless, the current dependencies are real and the end goal – more domestic capacity – is sound.

The question is how to get there. A so-called ‘tech sovereignty package’ published by the European Commission on 3 June, correctly diagnoses the need for more research funding and identifies Europe’s shortcomings in energy and capital. But the package also embeds a subtle protectionist turn. The proposed measures suggest that, in a bid to reinforce sovereignty, Europe is imitating its competitors – primarily the United States and China.

The package would update the 2023 EU Chips Act (Regulation (EU) 2023/1781), intended to reduce EU dependencies on foreign semiconductor suppliers, and proposes a Cloud and AI Development Act (CADA), which would seek to boost EU data-centre rollout. To build such capacity, the Commission plans to mimic the US and Chinese approaches: state subsidies, buy-local procurement and domestic-preference rules, echoing the US CHIPS and Science Act and Buy American Act, and China’s Big Fund and procurement law.

In other words, the Commission proposes to implement the standard protectionist toolkit the EU has spent decades opposing in the name of open markets, while labelling it “openness-grounded” sovereignty and “fair competition”. The proposals would make it easier for EU governments to justify and secure approval for national subsidies for strategic first-of-a-kind projects, semiconductors and data centres, potentially leading market distortions to multiply. Smaller, fiscally constrained EU countries would struggle to keep up in the intra-EU subsidy race, which would be paid for by taxpayers in larger countries.

Countries would be allowed to favour domestic producers in public procurement. A bidder using EU-made software or hardware might receive favourable weightings. For cloud services, CADA would establish four ‘assurance levels’. Each level would add stricter requirements on bidders, such as being established in the EU and employing EU citizens. Non-EU providers cannot attain the highest levels, which EU countries may nevertheless require to mitigate security risks in critical infrastructure.

Thus, in contract awards, rather than relying on legal constraints, such as localisation rules that keep data stored and processed in the EU, or technological solutions, such as encryption or auditing, the Commission would make a company’s nationality a proxy for security, closing the market to foreign competition.

The consequence of favouring local suppliers could be higher chip prices and lower quality in critical infrastructure. The evidence that buy-local rules raise procurement costs and dampen incentives to innovate is extensive. The Commission assumes only a 5% price premium for sovereign services, but this is unrealistic. It also clashes with the core premise of the tech sovereignty plan: that EU companies are currently not close substitutes for their foreign competitors and thus need a leg-up. A higher premium means higher recurring costs. By the Commission’s own estimate, migrating the public sector to sovereign services could cost up to €86 billion.

Protectionism also invites retaliation, potentially slowing, rather than accelerating, tech development in Europe. Being heavily dependent on global value chains for exports and imports, the EU has everything to lose from an international protectionist spiral. The Commission does call for cooperation with “trusted partners […] for supply diversification purposes”, but it is hard to see how like-minded third countries could be open to digital partnerships when the EU is simultaneously designing origin tests to exclude them.

The US and China are not necessarily successful in tech because they implement protectionist policies. Their competitive advantage has deep roots: cheap energy and capital, and a far less-fragmented market than the EU. In these areas, the Commission’s proposal is tepid, offering a road map for industrial electricity prices and a promise to reflect on financing approaches, rather than the structural measures and capital-market integration actually needed.

A bolder sovereignty strategy would steer clear of protectionism and leverage single-market competition to reduce costs in upstream input markets, while increasing incentives to innovate downstream. Domestic producers may need anchor demand to reach scale, but the efficient way to provide this is open-innovation procurement and a market deep enough to supply that scale, not origin tests. More structural measures may be less appealing to EU countries that want quick solutions and fewer concessions to market integration, but it would be far more impactful in the long run than joining the global race towards isolationism.