JRC Generative AI Outlook 2025: key figures from the EU Joint Research Centre
The European Commission’s Joint Research Centre brought together the technology, economy, labor market and regulation of generative AI in one review. The EU holds 7% of the world’s generative AI players, and China 60%. Below are the key JRC generative AI figures on the market, patents, companies, occupations and energy.
Key figures
7%
of the world’s generative AI players are in the EU, third place after China (60%) and the United States; the US share is 12% in the text and 11% in the figure
120 000
generative AI patents filed worldwide by 2024; the EU has 2% of priority filings, China 80%
2nd place
the EU’s world rank by number of scientific publications on generative AI
41%
of large EU companies used AI in 2024, about 21% of medium and 11% of small ones, according to Eurostat
82%
of active users and downloads of conversational AI in the EU belong to the five largest services, mostly from the United States and China; ChatGPT has the most
90%
of workers are less exposed to generative AI than teachers; the top group also includes electrical engineers, software developers, office clerks and secretaries
45–65 TWh
of electricity was used by EU data centers in 2022, 1.8–2.6% of total EU consumption, by the JRC estimate
The report in brief
- Published by
- Joint Research Centre of the European Commission (JRC)
- Published on
- 13 June 2025
- Original title
- Generative AI Outlook Report: Exploring the Intersection of Technology, Society and Policy (JRC142598, EUR 40337)
- Length and language
- 167 pages, in English
- Checked
- 16 October 2026
- Next edition
- frequency not announced; no new edition as of 16 October 2026
“Digital maturity is crucial for GenAI adoption, especially for SMEs”Executive summary, p. 7
How the report was made
This is an analytical review by the JRC for people who shape EU policy: technology, the economy, society, regulation and examples by industry. The figures on the market and its players come from the JRC’s own DGTES datasetdataset. A dataset is a set of data collected and formatted to train or test a model: a table, a folder of pictures, or a body of texts in which every example is described in the same way. on organizations and their generative AI activity for 2009–2024 (section 1.2, p. 16).
The authors take the shares of companies using AI from Eurostat (section 3.2, p. 47). The rest relies on Sensor Tower app data and the scientific literature. The labor market section compares occupations by exposure to generative AI (section 3.4, p. 54), and the energy section gives a JRC estimate for EU data centersAI data center. AI data centers are buildings with thousands of servers on graphics cards and other accelerators, with strong power supply and cooling. Neural networks are trained and run there. (section 4.5, p. 70).
What the report does not measure
This is not a new statistical survey. The figures on company adoption are for 2023–2024; fresh Eurostat data for 2025 is on the Eurostat page in this library. The authors themselves write that the review does not claim to be a final research analysis.
The B2B market and company spending on generative AI are not measured. The figures on patents and players show activity in research and filings, not revenue or market share. In section 1.2 the text and figures sometimes differ: 12% for the United States in the text and 11% in Fig. 1.
What this means for a company in Russia
Russia in the report: there is no data on Russia; in the distribution of players it is not shown separately and falls into the group of other countries.
- In the EU, the five largest conversational AI services, mostly from the United States and China, hold 82% of users and downloads (p. 49). The report gives no shares for Russia, and the data cannot be carried over directly. The lesson for a company: if contracts and personal data go into promptsprompt. A prompt in AI is the text with which you give a task to an AI model: a question, an instruction, examples and the data the model uses to write its answer., consider your own AI on company servers.
- The top group by exposure to generative AI includes software developers, office clerks and secretaries, and teachers are more exposed than 90% of workers (p. 54). In the report, exposure means that the mix of tasks changes. Start with employees in these roles: give them free AI courses before you change processes.
- The gap by size is large: 41% of large EU companies against 11% of small ones (Fig. 9). The authors call digital maturity a key condition, especially for small businesses (p. 7). It is easier for a small company to start with ready-made solutions, which an AI solution map for small business brings together.
What to compare it with
- Eurostat: AI in EU companies 2025: fresh data on EU companies using AI in 2025
- ILO: generative AI and work 2025: occupational exposure to generative AI by the ILO method
- IEA: key questions on energy and AI: energy use of data centers and AI worldwide