The missing ingredient for European tech sovereignty

In early June, the European Commission unveiled its “European Technical Sovereignty Package,” a new policy agenda designed to strengthen Europe’s strategic autonomy through investments in artificial intelligence, domestic semiconductor manufacturing, cloud infrastructure and open-source software. While ambitious, the package is not enough. To ensure that AI and digital technologies advance the EU’s economic and security interests requires confronting the root causes of Big Tech’s dominance over Europe’s digital economy.
The plan’s objective — to reduce the EU’s dependence on foreign technologies and, in turn, its vulnerability to external coercion — is laudable. Today, about 80 percent of the bloc’s digital technologies are imported, primarily from the US. This has created a power imbalance that US President Donald Trump’s administration has not hesitated to exploit in pursuit of its political interests.
Yet the European Commission’s plan is unlikely to shift that balance. Consider one of the package’s main goals: tripling the EU’s data center capacity over the next five to seven years. While expanding computing capacity is important, it will do little to foster a competitive domestic ecosystem unless Europeans capture a greater share of the economic value generated by AI.
Three American companies — Amazon, Google and Microsoft — currently control 70 percent of Europe’s cloud market, while AI development is dominated by a handful of trillion-dollar US firms, including OpenAI and Anthropic, whose partnerships with Big Tech give them unparalleled access to capital and computing power. Without a concerted effort to tackle industry consolidation, most of the value created by Europe’s data center buildout will ultimately flow back to the US.
Even if the new package succeeds beyond expectations and helps create a new generation of fast-growing European startups, Big Tech could still use its market dominance to acquire or otherwise eliminate these challengers. Between 2010 and 2023, Big Tech companies acquired nearly 100 AI startups. Among them was DeepMind, whose 2014 purchase by Google has been widely viewed as a strategic loss for the UK, where the company was founded.
Acquisitions are only one mechanism through which dominant firms protect their position. Self-preferencing, tying and restricting access to critical digital infrastructure can be just as effective in suppressing potential rivals. Meta’s recent attempt to block competing AI providers from accessing WhatsApp — a move that the European Commission swiftly reversed — illustrates how easy it is to undermine promising European alternatives. 

Europe should have the confidence to pursue a different innovation path from Silicon Valley’s model.

Giorgos Verdi

The good news is that Europe already has many of the tools it needs to prevent these outcomes. The EU’s Digital Markets Act, for example, prohibits gatekeeper platforms from engaging in unfair practices that stifle competition from European innovators.
But the act has not kept pace with developments in cloud computing and AI, and political pressure from the Trump administration has discouraged robust enforcement against US firms. As long as the law is enforced only half-heartedly, European innovators will continue to compete on an uneven playing field.
Merger control is another underused instrument. Existing competition rules and national security screening regimes already allow EU governments to prevent acquisitions that undermine strategic autonomy. In May, for example, the Netherlands blocked IBM spinoff Kyndryl’s takeover of Dutch cloud provider Solvinity, citing risks to the public interest.
Other EU member states — and the European Commission — should follow this example and scrutinize acquisitions that risk absorbing Europe’s most promising tech companies. Last year, Apple reportedly considered acquiring the French AI startup Mistral. Although the deal never materialized, it served as a reminder that the EU must be prepared to intervene when takeovers threaten not only competition but also its technological sovereignty.
At the same time, the tech sovereignty package itself should be more ambitious. Its centerpiece, the Cloud and AI Development Act, seeks to ensure that sensitive public sector functions — particularly those related to national security and defense — rely on European providers. But its strongest requirements apply only to a narrow range of activities, while member states retain considerable discretion over implementation. Given many governments’ reluctance to provoke the Trump administration, this framework is unlikely to generate sufficient demand to foster a globally competitive European cloud industry.
Critically, Europe should also have the confidence to pursue a different innovation path from Silicon Valley’s highly centralized and resource-intensive model. A growing number of experts argue that today’s dominant AI paradigm may be approaching its limits, making it unwise for Europe to devote all its resources to replicating it.
Instead, the EU should invest more heavily in areas such as physical, spatial and so-called frugal AI, which aims to design systems that maximize impact while minimizing computational and energy requirements. Rather than simply imitating the American model, Europe must develop technologies that generate economic value, improve citizens’ lives and align with its sustainability goals.
By recognizing that dependence on US tech giants is both an economic challenge and a strategic vulnerability, the tech sovereignty package marks a welcome shift in the EU’s approach. But without equally ambitious efforts to curb market concentration and break Big Tech’s hold on Europe’s digital economy, it is bound to fall short.

• Giorgos Verdi is Policy and Advocacy Lead at Open Markets Institute Europe.
©Project Syndicate