Reasons for the increasing adoption of AI in EU businesses and the challenges preventing full integration are being examined.

      Last December, Eurostat released data that, if it had been published in another continent, would have likely made front-page headlines. They reported that 20% of European Union businesses with a minimum of ten employees are now utilizing artificial intelligence in some capacity, an increase from 13.5% the previous year. This marks a rise of six and a half percentage points over twelve months. The announcement was met with quiet relief in Brussels. An economist from a Berlin think tank forwarded the information to a colleague with a single word: "finally." Meanwhile, an SME owner in a Bucharest co-working space calculated her home country's figure, finding Romania at just 5.2%.

      This wide disparity—from 42% in Copenhagen to 5% in Bucharest—serves as a starting point for any candid discussion about AI adoption in Europe. The continent is not static; it has seen rapid advancements in certain areas while lagging in others. The aggregate figure of twenty percent can be misleading, presenting an average that doesn't truly reflect the diverse economy where, on this issue, countries do not seem to operate as a unified market.

      The common explanation for Europe’s lag behind the United States in enterprise AI adoption is regulatory pressures. It’s often said that the AI Act has unsettled boards and burdened legal departments. While there is some truth to this, it is less significant than advocates might suggest. The core issue is that European AI adoption remains low for the same reasons European technology has been relatively small for the past two decades: a lack of investment and a scarcity of skills.

      The concept of a single market exists only on paper; companies investing in AI continue to primarily purchase from American cloud providers. Regarding investment, data from the OECD released in February and referenced by Christine Lagarde in a November speech revealed that around 75% of all AI venture capital in 2025 was directed toward U.S. firms, totaling approximately $194 billion. In contrast, the European Union gathered only $15.8 billion. This is not just a gap; it represents two vastly different scales. Lagarde's speech also referenced Mario Draghi’s research indicating that about 70% of the per-capita GDP divide between the EU and the U.S. is attributable to productivity disparities, with the technology sector accounting for about two-thirds of that gap since the early 2000s.

      These figures carry real implications. They explain why a French SME contemplating an AI pilot often first seeks a budget that does not exist, followed by a service that does—typically American.

      This leads to a second significant issue: three U.S. companies controlled roughly 70% of the European cloud infrastructure market by 2025, while European providers held about 15%. Any enterprise AI implementation in Europe that fails to account for this reality typically relies on U.S. computing power, billed in dollars and governed by a foreign interpretation of data protection regulations. This is not merely a theoretical concern. Mistral’s CEO, Arthur Mensch, has spent the last year advocating for Europe to “own and operate” its AI infrastructure, backing a Paris data center with $830 million in debt. Nevertheless, achieving this is still a long-term goal.

      Internally, companies face another barrier: personnel. The OECD’s December 2025 report on AI adoption among small and medium-sized enterprises, prepared for the G7 presidency, indicated that half of surveyed SMEs identified a skills shortage as their main obstacle to adoption, while 40% cited maintenance costs, 32% mentioned hardware limitations, and 26% expressed difficulties understanding the digital regulations they must comply with. These responses do not stem from executives who have been scared off by Brussels regarding AI; rather, they reflect a desire to adopt AI if they could simply find someone capable of installing, operating, and explaining it in their own language. The Eurostat statistics highlight this gap: the adoption rate of AI among large enterprises in the EU hovers around 55%, while small businesses sit at just 17%. This gap is not theoretical; it stems from the availability of in-house data engineers.

      At this juncture, it may be tempting, especially for an American audience, to view the AI Act as evidence that Europe prioritizes process over progress. However, that interpretation is more complex. The Act's most stringent provisions concerning high-risk systems won't take effect until August 2026. The European Commission has already sought to ease pressures with a Digital Omnibus proposal introduced on November 19, 2025, targeting a 25% reduction in the overall compliance burden and a 35% reduction for SMEs by 2029, while also expanding the simplified SME framework to include firms with up to 750 employees and €150 million in revenue.

      The Commission appears to be aware of these survey results, but whether it has taken action in time is uncertain. Industry analyses indicate that developers in the EU and the UK report launch delays in nearly 60% of cases due to the Act