Eurostat released data last December showing that one in five enterprises in the European Union with ten or more employees deployed artificial intelligence in some capacity during 2023, up from 13.5% the previous year. The six‑point jump signals momentum, yet the figure conceals a pronounced north‑south divide. Denmark tops the chart at 42%, while Romania lags behind at just over five percent.
Industry observers say the average masks a market that no longer moves as a single entity. Capital flows heavily toward the United States, where roughly three‑quarters of global AI venture funding was directed in 2025, amounting to $194 billion. The EU attracted a modest $15.8 billion, an order of magnitude smaller. The disparity translates into budget shortfalls for French SMEs and a reliance on American cloud services for most AI projects.
Three U.S. cloud providers commanded about 70% of the European cloud‑infrastructure market in 2025, leaving European players with roughly 15%. Consequently, many European AI deployments run on U.S. compute, billed in dollars and subject to foreign legal interpretations of data protection. French AI firm Mistral has tried to counter this trend by financing a Paris data centre with $830 million of debt, but the facility remains under construction.
Human capital emerges as the next hurdle. The OECD’s December 2025 report on SME AI adoption found that half of surveyed small businesses cite a shortage of skilled personnel as the primary barrier. Maintenance costs, hardware availability, and regulatory uncertainty also rank high. Large firms report a 55% adoption rate, while small firms sit at 17%, reflecting the gap between having an in‑house data engineer and not.
Regulation, specifically the AI Act, is often blamed for the lag. Its most stringent provisions on high‑risk systems will not take effect until August 2026, and the European Commission has already proposed a Digital Omnibus package aimed at cutting compliance burdens for SMEs by up to 35% by 2029. Nevertheless, surveys indicate that two‑thirds of European companies still cannot articulate their obligations under the Act.
Despite the challenges, pockets of progress stand out. Denmark’s AI adoption now exceeds the United States’ enterprise average reported by Stanford, and Finland and Sweden are close behind. Global surveys show that 88% of organizations worldwide use AI in at least one function, though only 6% see a material impact on earnings before interest and taxes. The difference lies not in geography but in commitment: senior leadership buy‑in, end‑to‑end workflow redesign, and upfront investment in infrastructure.
European giants such as Siemens, SAP, and Mistral demonstrate that the continent is not inert. Siemens has rolled out its Industrial Copilot across factory floors, SAP integrated its Joule AI engine into core ERP, and Mistral secured multi‑year deals with Accenture and a major European bank. These successes are concentrated among large, well‑capitalized firms in a handful of countries, while smaller, regionally bound companies—especially in the East and South—struggle to keep pace.
The underlying bottleneck appears to be the absence of a unified European market for capital, talent, and cloud infrastructure, a gap that predates the AI Act. Bridging that divide could narrow the adoption gap between Denmark and Romania, turning the current uneven landscape into a more cohesive AI ecosystem across the EU.
This article was written with the assistance of AI.
News Factory SEO helps you automate news content for your site.