Quote from Alex bobby on July 8, 2025, 3:02 AM

Despite efforts to build European champions, EU investments may strengthen US tech giants, a new report finds.

As the European Union pours billions of euros into the development of artificial intelligence (AI), a new warning has surfaced: rather than empowering homegrown innovation, the EU may be inadvertently deepening the grip of American tech giants. A report published by Dutch non-profit SOMO (Centre for Research on Multinational Corporations) on Monday reveals that most top AI start-ups—even those emerging in Europe—are fundamentally reliant on infrastructure and services controlled by U.S. companies like Microsoft, Amazon, Google, and Nvidia.

While the European Commission has made AI a cornerstone of its digital strategy, promoting the rise of companies such as the French start-up Mistral, SOMO’s findings suggest that these start-ups are building their future on foundations laid by a handful of dominant American players.

The Illusion of Independence

“These start-ups may look like fresh challengers on the surface,” said Margarida Silva, the report’s author, “but scratch a little deeper and you’ll see they’re built on infrastructure owned by Nvidia, Amazon, Google and Microsoft.”

The AI sector, especially when it comes to developing and running generative AI models, requires staggering amounts of computing power—something only a few companies can currently offer. Nvidia, for instance, supplies the specialised chips that power most advanced AI systems. According to the SOMO report, 11 out of the 12 leading generative AI start-ups rely on Nvidia’s hardware.

Beyond the chips, these companies also depend heavily on massive cloud infrastructures, with Microsoft Azure, Amazon Web Services (AWS), and Google Cloud emerging as the only realistic providers. SOMO found that 10 of the 12 start-ups analysed host their training and deployment operations on these platforms.

Exclusive Deals and Market Access

Access to infrastructure doesn’t come without strings. In many cases, start-ups must enter into exclusive agreements with Big Tech providers to use their resources. A notable example is Microsoft’s investment in Mistral, which grants the U.S. tech giant exclusive early access to Mistral’s AI models through its Azure cloud platform.

This arrangement has sparked criticism within the European Parliament, especially as it appears to conflict with the recently adopted EU AI Act—legislation designed in part to reduce the continent’s dependency on foreign technology providers. Instead, critics argue, such partnerships reinforce the very monopolies the EU seeks to avoid.

Moreover, Big Tech also controls the distribution channels for AI applications. Nine out of the 12 start-ups surveyed by SOMO sell their AI models through marketplaces operated by Amazon, Microsoft, or Google. These platforms are not just conduits for sales—they shape visibility, pricing, and ultimately, access to customers, including public institutions.

One revealing example is the European Parliament’s decision to adopt Anthropic’s Claude model for managing digital archives. Because AWS holds an EU contract, only AI models available on its platform were considered—a process that effectively excluded many other options.

AI Sovereignty at Risk

The report casts a long shadow over Europe’s hopes of achieving digital sovereignty. While initiatives to fund AI development are accelerating, the foundational infrastructure—chips, cloud services, and even distribution—remains firmly in U.S. hands.

“Europe wants to build its own AI capacity,” Silva noted, “but the foundations are still controlled by a handful of American companies.”

This dependence raises strategic concerns: without controlling the full AI value chain, Europe risks becoming a perpetual junior partner in the global AI ecosystem. The SOMO report compares the situation to previous tech booms—such as in smartphones and social media—where a few companies rapidly became gatekeepers, stifling competition and innovation.

Recommendations and Next Steps

SOMO is calling on EU and national competition regulators to urgently investigate the terms of cloud service contracts and scrutinise exclusive partnerships that may hinder market access for other players. The watchdog also recommends regulatory measures to limit market concentration and ensure interoperability between cloud platforms—so that AI developers can switch providers without prohibitive costs or delays.

Such measures would align with the goals of the EU’s Digital Markets Act, but according to SOMO, they need to be tailored specifically to the AI ecosystem, where the stakes are even higher due to the scale and strategic value of the technologies involved.

Final Thought:

Europe’s drive to foster AI innovation is commendable, but without structural independence, it risks becoming a well-funded extension of U.S. tech dominance. True digital sovereignty means more then building AI models—it requires controlling the infrastructure, platforms, and policies that support them. If Brussels truly wants to lead in ethical, competitive, and independent AI development, it must urgently address the power imbalance embedded in today's AI ecosystem. Otherwise, Europe may find itself investing in its own dependency.

Conclusion

The EU’s ambition to lead in ethical and responsible AI development is clear—but ambition alone is not enough. Without a more robust, self-sufficient AI infrastructure, Europe risks reinforcing the global dominance of U.S. tech giants, even as it invests in local innovation. To truly build an independent AI future, Brussels must tackle the systemic dependencies now—before it’s too late.

Meta Description:
A new report warns that EU efforts to support AI start-ups could deepen reliance on U.S. tech giants like Microsoft and Amazon, raising questions about Europe’s digital sovereignty.