Eurostat published data in December 2024 showing that 20% of European Union enterprises with at least ten employees now use artificial intelligence in some part of their business. This marks a significant jump from 13.5% the previous year. The aggregate figure suggests progress, but a closer look reveals deep disparities across member states.

The spread of AI adoption across Europe

Denmark leads with 42% of enterprises using AI, followed by Finland and Sweden. At the other end, Romania reports only 5.2%. This spread underscores that the EU is not a single market when it comes to AI adoption. The average of 20% flatters high-performing Nordic countries and obscures the struggles of Eastern and Southern economies.

The standard explanation for Europe lagging behind the United States is regulatory burden, specifically the AI Act. However, the deeper issues are structural: capital does not flow freely, skills are scarce, and the single market remains incomplete on paper.

Capital concentration in the US

According to OECD figures released in February 2025 and cited by Christine Lagarde, approximately three-quarters of global AI venture capital in 2025 went to US firms, totaling around $194 billion. The European Union collectively attracted only $15.8 billion. This is not a gap but two different orders of magnitude. Mario Draghi’s earlier report noted that about 70% of the per-capita GDP gap between the EU and US is a productivity gap, with technology explaining two-thirds of that shortfall since the turn of the century.

These numbers have concrete consequences. A French SME considering an AI pilot often finds a non-existent budget and then a service that is almost always American. This leads to a second structural problem.

Cloud dependency on US providers

Three US providers held roughly 70% of the European cloud infrastructure market in 2025, while European providers held about 15%. Every enterprise AI rollout that does not deliberately design around this fact ends up training on US compute, billed in dollars, and governed by a foreign court’s interpretation of data protection. Mistral’s CEO Arthur Mensch has argued that Europe must own and operate its own AI infrastructure, and the company has put $830 million of debt behind a Paris data centre. However, this is a long way from being delivered.

Skills shortages and SME challenges

Inside firms, the limiting factor is people. The OECD’s December 2025 report on AI adoption by SMEs, prepared for the G7 presidency, found that half of all surveyed SMEs cite a skills shortage as their primary barrier. Forty percent point to maintenance costs, 32% flag hardware, and 26% say they cannot understand the digital regulations they must comply with. These are not the answers of executives frightened by Brussels but of executives who would adopt AI if they could find someone to install, run, and explain it in their own language.

Large enterprises in the EU adopt AI at around 55%, while small ones sit at 17%. The gap is not philosophical but practical: having a data engineer in-house makes all the difference.

The role of the AI Act

The AI Act’s most invasive provisions for high-risk systems do not apply until August 2026. The European Commission has moved to soften the edges, proposing a Digital Omnibus in November 2025 that targets a 25% reduction in compliance burden overall and 35% for SMEs by 2029. Yet industry analyses suggest EU and UK developers report launch delays in nearly six in ten cases due to the Act, and about two-thirds of European companies still cannot articulate their obligations. Regulation is not the main hindrance but it is not negligible either.

Bright spots and uneven progress

Denmark’s enterprise AI adoption is now higher than the US enterprise average reported by Stanford. Finland and Sweden are close behind. McKinsey’s State of AI 2025 survey, with nearly 2,000 respondents across 105 countries, found that 88% of organizations globally regularly use AI in at least one function. However, only 6% see material enterprise-wide impact (5% or greater contribution to EBIT). On that measure, Europe lags less severely—Americans are also running pilots, just more of them.

What separates high performers everywhere is not country but commitment: senior leadership ownership, end-to-end workflow redesign, and willingness to spend on infrastructure before measuring returns. These are habits, not regulations, and Europe can choose them at any time.

European industry is not absent from the productive end of the curve. Siemens has pushed its Industrial Copilot into factory-floor workflows with new agentic capabilities announced at Automate 2025. SAP has woven Joule into its core ERP. Mistral has signed multi-year deployment deals with Accenture and a major European bank. The picture is not paralysis but unevenness—and the unevenness has a shape.

The firms doing AI well in Europe are large, well-capitalized, internationally minded, and concentrated in a handful of countries. The firms not doing AI are small, regionally bound, and disproportionately in the East and South. The single market, on this technology, is two markets. The real bottleneck is the absence of a European capital and skills base that would allow a Slovenian logistics firm or a Portuguese clinic to adopt AI as easily as a Danish bank already does.

The AI Act will get its share of blame, and some of it will be earned. But the more durable failure is older and has nothing to do with AI. It is the failure to finish the single market for capital, skills, and cloud infrastructure—the very issues Mario Draghi described in his report. If the gap between Denmark and Romania narrows, it will be because Europe finally decided that adopting AI was about industrial policy and human capital rather than ethics frameworks. If the gap widens, the explanation will sit in the same survey data that has been available for a decade.

Source: TNW | Eu News