This box presents new information about the adoption of artificial intelligence (AI) technologies by euro area firms, and about their plans for AI investment up to the end of 2026. The Survey on the Access to Finance of Enterprises (SAFE) for the fourth quarter of 2025 (ECB, 2026) included a set of ad hoc questions about the adoption of AI and the reasons for using, or not using, these technologies. Firms were asked about the extent of their adoption of specific technologies, including predictive tools (such as text mining, voice and image recognition, and machine learning), generative tools (such as chatbots and text/image generation) and robotic process automation. They were then asked to indicate their investment plans in AI over the next 12 months and to assess the diffusion of AI investment among their competitors in their own country up until June 2025.[1]
Large firms, listed or venture capital-backed companies and young firms are adopting AI more frequently than small, unlisted and established firms (Chart A). The survey results show that 38% of euro area firms are at an advanced stage of AI adoption, indicating significant or moderate use of AI. 33% of firms are still at an early stage, with very infrequent or experimental use of AI. Around 45% of large firms and listed or venture capital-backed companies are at an advanced stage of use of AI – this rises to 56% for young firms. However, the share of firms reporting significant use of AI is similar across size classes and ownership types, suggesting that AI adoption is spreading evenly among a core group of firms. Interestingly, this share is twice as high among young firms, likely reflecting the driving role of start-ups.[2]
Chart A
Use of AI by firm characteristics
(percentages of respondents)
Sources: SAFE (ECB, 2026) and ECB staff calculations.
Notes: SMEs stands for small and medium-sized enterprises (firms with up to 250 employees). Young firms are less than five years old. Private ownership refers to single individuals, families or other enterprises holding majority ownership in the firms. Listed/venture capital refers to majority ownership through public shareholders or venture capital firms. The figures refer to the October-December 2025 survey round.
The evidence indicates the most common reasons for not using AI are a perceived lack of usefulness and challenges linked to implementation (Chart B). Specifically, 30% of firms that are currently not using AI identify a lack of usefulness as the main barrier, while around 20% report incompatibility with their existing systems and a shortage of relevant skills. Large firms are less likely than small and medium-sized enterprises (SMEs) to view AI as lacking usefulness, but they more frequently report implementation challenges such as insufficient AI expertise and system incompatibilities. Listed or venture capital-backed firms, where AI adoption is generally more widespread, appear to recognise the potential value of AI even when they have not yet adopted it, while privately owned firms more commonly report barriers like perceived lack of utility. Compared with established firms, younger firms more frequently express ethical concerns and cite distrust in AI output as a barrier to adoption.[3]
Chart B
Reasons for not using AI by firms’ characteristics
(percentages of respondents)
Sources: SAFE (ECB, 2026) and ECB staff calculations.
Notes: “All firms” includes only firms that do not currently use AI. SMEs stands for small and medium-sized enterprises (firms with up to 250 employees). Young firms are less than five years old. Private ownership refers to single individuals, families or other enterprises holding majority ownership in the firms. Listed/venture capital refers to majority ownership through public shareholders or venture capital firms. The figures refer to the October-December 2025 survey round.
Firms making significant use of AI are more likely to expect increases in turnover and investment in fixed assets compared with firms not using AI (Chart C). A regression analysis highlights the relationship between the degree of AI use and firms’ expectations about real outcomes. Conditional on size, sector and location, firms making significant use of AI are more likely to expect higher turnover and increased investment in fixed assets within the next three months (by 21% and 13%, respectively) compared with firms that do not use AI (Chart C, panel a). Additionally, the expected proportion of future investment allocated to AI increases with the frequency of AI usage. Firms using AI infrequently anticipate allocating 3.2 percentage points more investment into AI than those not using AI (Chart C, panel b). For firms making moderate or significant use of AI, there are corresponding percentage point differences of 5.5 and 11.5 compared with firms not using AI. Moreover, over the next 12 months, firms making significant use of AI expect an additional 0.6 percentage point increase in wages and a 1.3 percentage point spike in employment compared with firms that do not use AI.[4]
Chart C
AI use and firms’ expectations about real outcomes
a) Investment and turnover
(percentages)
b) Wages, employment and AI investment
(percentage points)
Sources: SAFE (ECB, 2026) and ECB staff calculations.
Notes: Panel a) shows the coefficients of firm level regressions of expected turnover/investment (which are dummy variables that take value 1 if the firm expects an increase in the next three months) on AI usage with omitted category “No use”. Panel b) shows the coefficients of firm level regressions of three continuous variables – expected increases, over the next 12 months, of wages, number of employees and share of investment on AI usage. Survey-weighted regressions with industry, country and firm-size fixed effects. The whiskers represent 90% confidence intervals. The chart is based on the October-December 2025 survey round.
Firms currently using AI expect to invest more in AI this year compared with non-users, indicating a reinforcing cycle of adoption and innovation (Chart D). On average, firms expect to allocate 9% of their total investment to AI, though significant variation exists depending on firm characteristics and the degree of AI adoption. Firms that do not currently use AI anticipate allocating a relatively small share of their investment to AI (4% on average), with large firms among these non-users forecasting slightly higher rates (6%) compared with SMEs (4%). In contrast, firms at a more advanced stage of AI adoption plan much higher investment rates. Firms making moderate use of AI expect to allocate 11% of their investment to it, while those making significant use report the highest planned investment rates (20%). Among these significant users, SMEs lead with an expected allocation of 21%, compared with 17% for large firms. The overall AI investment pattern suggests the existence of a reinforcing cycle, where firms already using AI invest more to further develop and integrate these technologies.
Chart D
Expected investment in AI over the next 12 months by current usage intensity and firm characteristics
(percentages of overall investment)
Sources: SAFE (ECB, 2026) and ECB staff calculations.
Notes: SMEs stands for small and medium-sized enterprises (firms with up to 250 employees). Young firms are less than five years old. Private ownership refers to single individuals, families or other enterprises holding majority ownership in the firms. Listed/venture capital refers to majority ownership through public shareholders or venture capital firms. The figures refer to the October-December 2025 survey round.
Ownership structure also correlates with AI investment patterns. While firms making moderate use of AI show similar investment tendencies regardless of their ownership structure, differences emerge in the early and advanced stages of AI adoption (Chart D). Listed or venture capital-backed firms have the highest expected AI investment rate if they are early-stage AI adopters, allocating 17% of their total investment to AI, compared with 8% for privately owned firms. This leadership is likely driven by their funding advantages and focus on high-growth opportunities. In contrast, privately owned firms dominate advanced-stage investments, expecting to allocate 21% of their total investment compared with 12% for listed or venture capital-backed companies. This likely reflects their concentrated ownership structures and long-term strategic focus, which give them greater flexibility to capitalise on the proven benefits of AI.
Young firms invest little in AI at first but scale up as adoption deepens. They report the lowest expected AI investment rates in the initial stages of adoption (3% of their total investment – Chart D). As AI adoption intensifies, however, this share increases to 17%, getting close to depth of investment seen at established firms. Several factors may explain this pattern: learning-by-doing reduces uncertainty about which applications create value and successful early pilots justify subsequent commitments.
References
Aldasoro, I., Gambacorta, L., Pal, R., Revoltella, D., Weiss, C. and Wolski, M. (2026), “AI adoption, productivity and employment: evidence from European firms”, BIS Working Papers, No 1325, Bank for International Settlements, January.
Bencivelli, L., De Masi, L., Falck, E., Fernández Cerezo, A., Formai, S., Hidalgo Bricio, I., Mattevi, E. and Nagengast, A. (2026), “Embracing AI in Europe: New evidence from harmonised central bank business surveys”, VOXEU Column.
ECB (2026), Survey on the Access to Finance of Enterprises in the euro area - Fourth quarter of 2025.
Yotzov, I., Barrero, J.M., Bloom, N., Bunn, P., Davis, S. J., Foster, K.M., Jalca, A., Meyer, B. H., Mizen, P., Navarrete, M. A., Smietanka, P., Thwaites, G. and Wang, B. Z. (2026), “Firm Data on AI,” Working Papers, No 34836, National Bureau of Economic Research, February.
For additional information on other ad hoc questions on AI included in the SAFE, see ECB (2026).
Similar percentages of AI adoption across firm characteristics are also found for firms across countries – see Yotzov et al. (2026) for the United Kingdom, the United States, Germany and Australia and Bencivelli et al. (2026) for Germany, Italy and Spain.
The survey also asked firms about their reasons for using AI. Responses show that AI is primarily adopted to enhance core and non-core business processes, while significantly fewer firms mention reducing personnel costs, supporting R&D and innovation, or expanding their product and service offerings. Regardless of the extent of AI adoption, firms tend to cite similar reasons for implementation, with no significant differences across size classes (ECB, 2026).
For non-AI adopters the unconditional weighted mean of expected wages increases is 3.1% compared with 3.7% for firms with significant use of AI. In the case of employment, the respective figures are 0.6% and 2.8%. Similar results are reported by Aldasoro et al. (2026).