The AI adoption gap between small and large firms in Europe (2026 data study)
Eurostat data for 35 European geographies: how far firms with 10 to 49 staff trail firms with 250 or more in using AI, country by country, 2021 to 2025.
Published · Updated · By Edgars Lapins
On this page
How much less likely is a small firm to use AI than a large one, and does it depend on the country? We took Eurostat’s official survey figures for 35 European geographies (the EU, the euro area, the 27 member states and 6 other countries) and compared enterprises with 10 to 49 employees against those with 250 or more, for 2021 to 2025. The headline EU shares are published by Eurostat. The gap, ratio and rankings below are our own calculations from those figures.
In short:
- EU-wide in 2025, 55.0% of large firms used AI against 17.0% of small firms (10 to 49 staff). Large firms were 3.2 times as likely to use AI, a gap of 38.0 percentage points.
- AI use among EU firms with 10 or more staff rose from 8.1% in 2023 to 20.0% in 2025, up 11.9 points in two years. Denmark rose most, by 26.9 points to 42.0%.
- The widest gap is in Slovenia: 53.4 points (18.3% of small firms against 71.7% of large firms). In Poland, large firms were 7.9 times as likely as small ones to use AI (45.8% against 5.8%).
- A small Danish firm (37.5%) is more likely to use AI than a large firm in Romania (20.8%), Bulgaria (26.2%) or Turkey (23.5%).
- Lowest overall use among firms with 10 or more staff: Romania 5.2%, Turkey 7.4% and Poland 8.4%.
Looking for UK figures? Eurostat does not cover the UK, so see our UK AI and automation statistics, which use ONS data.
The gap, country by country
The chart shows how many percentage points separate large firms from small firms in each country in 2025. A longer bar means small firms are further behind.
Gap in AI use, large minus small firms, 2025 (percentage points)
- Slovenia53.4
- Belgium47.6
- Finland45.9
- France43.0
- Austria42.1
- Sweden41.1
- Norway40.8
- Czechia40.7
- Ireland40.5
- Spain40.3
- Poland40.0
- Portugal39.7
- Euro area39.3
- Netherlands38.9
- Italy38.9
- EU (27 countries)38.0
- Latvia37.3
- Denmark37.1
- Germany33.9
- Malta33.7
- Lithuania33.7
- Estonia32.2
- Hungary31.3
- Greece30.2
- Slovakia28.2
- Luxembourg27.5
- Cyprus27.4
- Croatia26.8
- Bulgaria19.0
- Albania16.9
- Turkey16.9
- Romania16.6
- Serbia8.9
- Bosnia and Herzegovina8.2
Check your country
Pick a country to see its 2025 figures.
Full results table
All figures are the share of enterprises using at least one AI technology. “Gap” is large minus small in 2025. “Ratio” is how many times as likely a large firm is to use AI as a small one. “Change” is the movement for all firms with 10 or more staff from 2023 to 2025. “n/a” means Eurostat publishes no value.
| Slovenia | 18.3% | 71.7% | 21.6% | 53.4 pts | 3.9x | +10.2 pts |
| Belgium | 28.8% | 76.4% | 34.5% | 47.6 pts | 2.7x | +20.7 pts |
| Finland | 33.5% | 79.4% | 37.8% | 45.9 pts | 2.4x | +22.7 pts |
| France | 15.0% | 58.0% | 18.2% | 43.0 pts | 3.9x | +12.3 pts |
| Austria | 26.2% | 68.3% | 29.9% | 42.1 pts | 2.6x | +19.2 pts |
| Sweden | 30.7% | 71.9% | 35.0% | 41.1 pts | 2.3x | +24.7 pts |
| Norway | 25.2% | 66.0% | 28.9% | 40.8 pts | 2.6x | +19.7 pts |
| Czechia | 13.4% | 54.1% | 17.6% | 40.7 pts | 4.0x | +11.7 pts |
| Ireland | 16.8% | 57.3% | 19.6% | 40.5 pts | 3.4x | +11.6 pts |
| Spain | 17.2% | 57.5% | 20.3% | 40.3 pts | 3.3x | +11.1 pts |
| Poland | 5.8% | 45.8% | 8.4% | 40.0 pts | 7.9x | +4.7 pts |
| Portugal | 9.4% | 49.1% | 11.5% | 39.7 pts | 5.2x | +3.7 pts |
| Euro area | 18.4% | 57.6% | 21.4% | 39.3 pts | 3.1x | +12.5 pts |
| Netherlands | 28.8% | 67.6% | 33.2% | 38.9 pts | 2.4x | +19.1 pts |
| Italy | 14.2% | 53.1% | 16.4% | 38.9 pts | 3.7x | +11.3 pts |
| EU (27 countries) | 17.0% | 55.0% | 19.9% | 38.0 pts | 3.2x | +11.9 pts |
| Latvia | 9.8% | 47.1% | 12.2% | 37.3 pts | 4.8x | +7.7 pts |
| Denmark | 37.5% | 74.5% | 42.0% | 37.1 pts | 2.0x | +26.9 pts |
| Germany | 23.1% | 57.0% | 26.0% | 33.9 pts | 2.5x | +14.4 pts |
| Malta | 18.0% | 51.8% | 21.5% | 33.7 pts | 2.9x | +8.3 pts |
| Lithuania | 18.1% | 51.8% | 21.3% | 33.7 pts | 2.9x | +16.4 pts |
| Estonia | 20.7% | 52.9% | 23.4% | 32.2 pts | 2.6x | +18.2 pts |
| Hungary | 8.7% | 40.0% | 10.4% | 31.3 pts | 4.6x | +6.7 pts |
| Greece | 7.3% | 37.6% | 8.9% | 30.2 pts | 5.1x | +4.9 pts |
| Slovakia | 15.6% | 43.7% | 18.0% | 28.2 pts | 2.8x | +11.0 pts |
| Luxembourg | 30.5% | 58.0% | 33.6% | 27.5 pts | 1.9x | +19.2 pts |
| Cyprus | 7.7% | 35.1% | 9.3% | 27.4 pts | 4.5x | +4.6 pts |
| Croatia | 12.8% | 39.6% | 15.2% | 26.8 pts | 3.1x | +7.3 pts |
| Bulgaria | 7.2% | 26.2% | 8.6% | 19.0 pts | 3.7x | +4.9 pts |
| Albania | 7.9% | 24.8% | 9.0% | 16.9 pts | 3.1x | n/a |
| Turkey | 6.5% | 23.5% | 7.4% | 16.9 pts | 3.6x | +1.9 pts |
| Romania | 4.1% | 20.8% | 5.2% | 16.6 pts | 5.0x | +3.7 pts |
| Serbia | 9.4% | 18.3% | 10.1% | 8.9 pts | 1.9x | +8.3 pts |
| Bosnia and Herzegovina | 10.4% | 18.6% | 10.8% | 8.2 pts | 1.8x | +5.4 pts |
| Montenegro | n/a | n/a | 10.1% | n/a | n/a | +4.4 pts |
Download the results as CSV (35 rows, Eurostat values to two decimals, plus our change and gap columns). Eurostat does not publish a figure for medium-sized firms (50 to 249 staff) in the extract we used, so the table shows small, large and all firms with 10 or more staff.
What this means for a small business
The pattern holds almost everywhere: large firms adopt AI first, and small firms follow more slowly. In 2025 the gap was at least 26.8 points in every EU member state except Bulgaria and Romania, where use was low for firms of every size. Bosnia and Herzegovina and Serbia had the smallest gaps of all, also because large firms there used AI far less than elsewhere.
Size is not the whole story. Small firms in Denmark, Finland and Sweden used AI more than large firms in several other countries. If you run a small business and want to close the gap, start with one repeatable task. Our guide to where AI helps a small business first lists common starting points, and if you would rather have help, you can get matched with an AI automation agency for free.
Method and caveats
- Source. Eurostat dataset isoc_eb_ai (“Artificial intelligence by size class of enterprise”), indicator E_AI_TANY: enterprises using at least one AI technology, as a percentage of enterprises. Survey years 2021, 2023, 2024 and 2025. We pulled the data from the Eurostat API on 29 October 2026 with a short script that needs no key.
- Our calculations. Gap = share of large firms (250+) minus share of small firms (10 to 49) in 2025. Ratio = large divided by small. Change = 2025 minus 2023 for all firms with 10 or more staff, in percentage points. The table rounds to one decimal; the CSV keeps Eurostat’s two decimals.
- No UK data. The UK left the Eurostat ICT survey. The ONS measures UK business AI use in a different survey with a different question, so UK and EU figures are not directly comparable.
- Firms with fewer than 10 staff are not covered. The survey only includes enterprises with 10 or more employees, so microbusinesses and sole traders are not in these figures.
- Definitions change. “Using AI” means using at least one of the AI technologies listed in Eurostat’s questionnaire. The list and wording have changed between survey rounds, and 2021 is the most likely to differ from 2023 onwards, so treat long-run comparisons with care.
- Financial sector excluded. The figures cover enterprises in the sectors Eurostat surveys excluding the financial sector (Eurostat code C10-S951_X_K).
- Gaps in the data. Eurostat publishes no size breakdown for Montenegro and no 2023 figure for Albania, so those cells show n/a.
How to cite
Suggested citation
TrustedAIAgencies (2026). The AI adoption gap between small and large firms in Europe (2026 data study). Analysis of Eurostat isoc_eb_ai data. https://trustedaiagencies.com/guides/ai-adoption-gap-small-business-europe-2026/
Link for your article:
<a href="https://trustedaiagencies.com/guides/ai-adoption-gap-small-business-europe-2026/">AI adoption gap between small and large firms in Europe, TrustedAIAgencies</a>Please also credit Eurostat as the source of the underlying data. You are free to reuse our CSV with a link back.
