Last updated: . Refreshed annually after the December Eurostat release and the June EC country report. Figures are reproduced with the decimals printed at source.
How to cite: Murarum (2026). AI adoption in Romania: 2026 statistics. https://murarum.ro/en/insights/ai-adoption-romania-2026. Compiled from Eurostat (isoc_eb_ai, December 2025) and the European Commission Romania 2025 Digital Decade Country Report (June 2025). CC BY 4.0.
What the numbers say
Romania has the lowest enterprise AI adoption rate in the EU, and the gap widened in 2025. Across the EU, the share of enterprises using AI rose from 13.5% in 2024 to 19.95% in 2025. Romania moved from 3.1% in 2024 to 5.21% in 2025. Both series measure enterprises with 10 or more employees, excluding the financial sector, using the same Eurostat survey.
The gap is not an AI-only story. Romanian SMEs have almost closed the gap on basic digital intensity (69.1% against an EU average of 72.9% in 2024), but they trail on every advanced indicator: cloud, data analytics, AI and the digital skills of the workforce. For a scaling company, that ordering matters. Basic systems are in place; the next layer, connected data and automated decisions, is not.
Adoption also splits by company size. In Romania, large enterprises used AI at four times the rate of SMEs in 2024 (11.26% against 2.78%). In the EU as a whole, 55.03% of large enterprises used AI in 2025, against 30.36% of medium and 17% of small enterprises.
AI use in enterprises, 2025
| Indicator | Romania | EU | Year | Source |
|---|---|---|---|---|
| Enterprises (10+ employees) using at least one AI technology | 5.21% | 19.95% | 2025 | Eurostat, isoc_eb_ai (Dec 2025) |
| Enterprises using AI, previous edition | 3.1% | 13.48% | 2024 | EC Romania 2025 Country Report |
| SMEs using AI | 2.78% | not stated in the source | 2024 | EC Romania 2025 Country Report |
| Large enterprises using AI | 11.26% | 55.03% (2025) | 2024 / 2025 | EC report; Eurostat isoc_eb_ai |
| Highest EU rates | Denmark 42.03%, Finland 37.82%, Sweden 35.04% | 2025 | Eurostat, isoc_eb_ai | |
| Lowest EU rates | Romania 5.2%, Poland 8.4%, Bulgaria 8.5% | 2025 | Eurostat news, 11 Dec 2025 |
The most used AI technologies in EU enterprises in 2025 were text mining and analysis of written language (11.8%), generation of images, video or audio (9.5%), generation of written or spoken language (8.8%) and speech recognition (7.2%). Eurostat does not publish this breakdown for Romania in the release above, so we do not report it.
Digital intensity of SMEs, 2024
| Indicator | Romania | EU | Year | Source |
|---|---|---|---|---|
| SMEs with at least a basic level of digital intensity | 69.1% | 72.9% | 2024 | EC Romania 2025 Country Report |
| SMEs with high or very high digital intensity | 21.69% | 32.66% | 2024 | EC Romania 2025 Country Report |
| SMEs with basic digital intensity, earlier reading | 52.49% | 2022 | EC Romania 2025 Country Report |
The Digital Intensity Index counts how many of twelve digital technologies an enterprise uses. "Basic" means at least four. The 2024 value uses version 4 of the index, which the Commission states is comparable with the 2022 reading used for the annual growth figure (+14.7% per year) but not with the national trajectory.
Cloud, data analytics and skills
| Indicator | Romania | EU | Year | Source |
|---|---|---|---|---|
| Enterprises buying cloud computing services | 15.52% | 38.97% | 2023 | EC Romania 2025 Country Report |
| Enterprises performing data analytics | 21.87% | 33.25% | 2023 | EC Romania 2025 Country Report |
| Enterprises using AI, cloud or data analytics | 28.74% | 54.7% | 2023 | EC Romania 2025 Country Report |
| People (16 to 74) with at least basic digital skills | 27.7% | 55.56% | 2023 | EC Romania 2025 Country Report |
| ICT specialists as a share of employment | 2.8% | 5.0% | 2024 | EC Romania 2025 Country Report |
Cloud, data analytics and skills were last measured in 2023; the Commission notes that these indicators were not surveyed in 2024. They are the prerequisites for AI use in production: an AI workflow needs data it can reach and people who can supervise it.
Method and sources
Indicators: share of enterprises using at least one AI technology (Eurostat survey on ICT usage in enterprises, dataset isoc_eb_ai); share of SMEs with at least a basic level of digital intensity (Digital Intensity Index, version 4); share of enterprises buying cloud services and performing data analytics; share of individuals aged 16 to 74 with at least basic digital skills (Digital Skills Indicator 2.0). Enterprise indicators cover enterprises with 10 or more employees, all sectors except finance. SMEs are enterprises with 10 to 249 employees; large enterprises have 250 or more.
Years: the AI figures are the 2025 survey (published December 2025), the digital intensity figures are 2024, and cloud, data analytics and digital skills are 2023, the latest year available at the time of publication. We report the decimals as printed by each source; where a source rounds differently in two places (5.21% and 5.2%), we show the more precise value and note the rounded one.
- Eurostat, Use of artificial intelligence in enterprises, data extracted December 2025
- Eurostat news release, 20% of EU enterprises used AI technologies in 2025, 11 December 2025
- European Commission, Romania 2025 Digital Decade Country Report, factpage, 16 June 2025
- European Commission, Romania 2025 Country Report, full PDF, SWD(2025) 294 final
Why these indicators: they are the only annual, official, EU-comparable series on enterprise AI use and its prerequisites. We do not use vendor surveys, because their samples are self-selected and their definitions of "using AI" vary from one year to the next.
Change since the last edition
This is the first edition. From the 2027 edition on, this section will list every indicator that moved, with the previous and current values side by side.
What this means for a scaling company
Two readings are consistent with the data. First, the gap is a prerequisites gap before it is a technology gap: cloud, data and skills trail the EU by 20 to 28 percentage points, and AI use follows them. Second, the low base is an advantage for a company that moves now, because the competitive bar in most Romanian sectors is set by the 94.79% of enterprises that do not yet use AI at all. Our AI integration service starts from that reading: process and data first, model second, production always.
