Is the European Union its own technological enemy? While the United States launches projects worth hundreds of billions of dollars and China rapidly automates its factories, Brussels is refining regulatory texts and risk classifications. The gap is widening, the figures are undeniable, yet the debate is more nuanced than it appears.
A quantifiable lag that doesn't lie
The numbers speak for themselves. The United States invests four to ten times more in AI four to ten times more than the European Union: between $60 and $70 billion per year on the American side, compared to $7 to $8 billion on the European side. Over the past decade, private investments in AI in the United States have exceeded $400 billion, while all EU countries combined have attracted only about $50 billion. In 2024, Europe captures only 5% of global AI investments, or $5.9 billion, even though this figure shows a 43% increase.
Robotics is even more symptomatic. According to a study published by Allianz Research in June 2025, over 50% of global industrial robots are currently deployed in China. Europe, once at the forefront of the sector with players like Germany, Sweden, or Switzerland, now risks definitive technological decline. The parallel with the automotive industry and electric vehicles is painful: a market that Europe let slip away.
The AI Act: scapegoat or real obstacle?
TheAI Actentered into force on August 1, 2024, with progressive application until August 2026, is at the center of all criticism. The text classifies AI systems according to four risk levels, from the most trivial to the unacceptable, and imposes strict obligations on so-called "high-risk" systems used in critical sectors such as energy, transport, or health. The concrete problem: between 33% and 50% of European startups integrating AI could be affected by this high-risk classification, far exceeding the 15% initially estimated by the European Commission.
For these companies, the stakes are existential. Their model relies on agility, speed of development, and a limited capacity to absorb high compliance costs. Yet, Europe had over 8,000 active startups in the AI field in 2023, with strong concentrations in Germany, France, the United Kingdom, and the Netherlands. These companies form the continent's innovation pool, but the regulatory pressure could push them to relocate elsewhere or slow down their development.
However, some experts temper the diagnosis. Boston Consulting Group states, in an analysis published in January 2026, that the AI Act is not a brake on innovation but a necessary governance framework for scaling AI. The argument: safeguards create trust, and trust accelerates adoption, especially in industrial sectors where European decision-makers remain more hesitant than in the United States regarding unregulated systems.
The accumulated weight of digital regulation
Europe's real problem may not be the AI Act itself, but the accumulation of texts. After GDPR, the Digital Services Act, and the Digital Markets Act, European companies must now integrate the AI Act into an already saturated legal environment. Each new regulation adds a layer of compliance, additional delays, and legal teams to expand. For a startup of 15 people, this is a disproportionate burden.
The Draghi report, published in September 2024 and with an anniversary in 2025, formalized this observation with rare brutality for a European institutional document. It identifies the Regulatory burden as a structural obstacle to competitiveness, estimates the investment need at 750 to 800 billion euros per year to close the gaps with the United States, and calls for urgent administrative simplification. One year after its publication, some initiatives have started, but many remain at the proposal stage.
Europe reforming, too slowly?
Brussels is not insensitive to criticism. In November 2025, the European Commission presented the Digital Omnibus Package, an initiative aimed at simultaneously easing five major digital regulations: the GDPR, the AI Act, the e-Privacy directive, the Data Act, and NIS2. The official objective is clear: to simplify, harmonize, and accelerate, with expected savings of 10 to 20 billion euros per year for businesses. Trilogue negotiations between the Commission, Parliament, and Council are scheduled for the first half of 2026.
In parallel, the European Parliament is studying targeted adjustments to the AI Act in February 2026, particularly to lighten certain obligations deemed disproportionate and to strengthen SMEs' access to 'regulatory sandboxes', these supervised experimentation spaces. On the financial front, the action plan for an AI continent mobilized 20 billion euros for scaling up AI in April 2025, supplemented by an additional billion in October 2025 for the deployment strategy.
A sovereignty to build, not to endure
The 'regulation versus innovation' debate is actually a false dilemma. The real European challenge is structural: the absence of a mature venture capital market hinders access to funding for innovators, the fragmentation of national policies prevents economies of scale, and Europe excels in academic research without managing to transform these innovations into prosperous companies. Regulations only exacerbate pre-existing weaknesses.
The European technological sovereignty will not be built by copying the American model or by ignoring the real risks of AI. It requires serious regulatory simplification, massive and coordinated investments, and political will to take calculated risks. Europe has the engineers, the labs, and the markets. What it still lacks is speed.
No comments! Be the first one.