AI adoption statistics 2026

This roundup collects the AI adoption figures StatOrigin holds from primary sources, each linked to its record, its publisher's file and the sentence or table row it came from. Business adoption comes from official surveys on both sides of the Atlantic; personal use of generative AI comes from Eurostat, Pew Research Center, China's CNNIC and the Reuters Institute.

• United States, businesses: 23.8% ¹ of nonfarm employer businesses used AI in the two weeks to 20 September 2026 (Census Business Trends and Outlook Survey).
• European Union, enterprises: 19.95% ² of enterprises with 10 or more persons employed used at least one AI technology in 2025 (Eurostat).
• Highest European rate: Denmark, 42.03% ³ of enterprises in 2025.
• People using generative AI: Norway leads Europe at 56.32% ⁴ of people aged 16–74 in 2025.
• Workers: 38% ⁵ of employed US adults say they use chatbots for tasks at work (Pew Research Center, February 2026).
• China: generative AI reached 42.8% ⁶ of the population at the end of 2025 (CNNIC).

The US and EU business figures come from different surveys with different populations, questions and reference periods. Their closeness does not make them a like-for-like comparison, and this roundup keeps them in separate sections.

The Census Bureau's Business Trends and Outlook Survey asks US nonfarm employer businesses every two weeks whether they used AI in any business function. Current use was 17.3% ⁷ on 30 November 2025 and 23.8% ¹ on 20 September 2026, across 22 consecutive estimates. The US business AI adoption tracker shows every release as it is published.

Expected use over the next six months was 21.1% ⁸ at the first estimate and 27.6% ⁹ at the latest. Expectations include businesses already using AI, so they are not added to current use.

Census broadened the question in November 2025 from AI used to produce goods or services to AI used in any business function. The series here starts at that revision; joining it to the earlier question would overstate continuity. Census explains the revision.

In the period ending 6 September 2026, reported use was 43.5% ³⁰ in Information, 43.3% ³¹ in Professional, Scientific, and Technical Services and 36.5% ³² in Finance and Insurance. Manufacturing stood at 23.4% ³³ and Retail Trade at 15.5% ³⁴.

By size, 38.4% ³⁵ of businesses with 250 or more employees used AI, against 34.5% ³⁶ of those with 100–249. Businesses with 1–4 employees reported 23.6% ³⁷, so adoption does not rise smoothly with every added employee. A firm-level share counts each business once; it does not measure the share of workers, tasks or revenue touched by AI.

Eurostat's survey on ICT usage in enterprises puts the 2025 EU rate at 19.95% ². Denmark led at 42.03% ³, ahead of Finland (37.82% ³⁸) and Sweden (35.04% ³⁹). The lowest rates were in Romania (5.21% ⁴⁰), Poland (8.36% ⁴¹) and Bulgaria (8.55% ⁴²).

The large economies sit around the middle: Germany at 25.97% ⁴³, Spain at 20.27% ⁴⁴, France at 18.16% ⁴⁵ and Italy at 16.4% ⁴⁶. Outside the EU, Norway reported 28.89% ⁴⁷. Every reporting country, with small, medium and large firm rates side by side, is on the enterprise AI adoption ranking.

Each country has its own mix of industries and firm sizes, so a higher rate is not proof of a more advanced economy or of demand for any particular product.

Size is the strongest divide in the EU data: 17% ⁵¹ of small enterprises used AI in 2025, against 30.36% ⁵² of medium and 55.03% ⁵³ of large enterprises.

The EU rate was 7.65% ⁵⁴ in 2021, 8.06% ⁵⁵ in 2023, 13.48% ⁵⁶ in 2024 and 19.95% ² in 2025. The 2025 questionnaire added image, video and audio generation to the list of AI technologies, so the latest step is not an unchanged basket.

Eurostat's household survey measured generative AI use for the first time in 2025. Norway reported the highest share of people aged 16–74 using a generative AI tool in the last three months, 56.32% ⁴, followed by Denmark (48.44% ⁵⁷) and Switzerland (47.02% ⁵⁸). Germany reported 32.25% ⁵⁹ and France 37.46% ⁶⁰; Italy (19.86% ⁶¹), Romania (17.76% ⁶²) and Turkiye (17.19% ⁶³) were lowest.

Age matters more than country. In Denmark, 77.64% ⁶⁴ of people aged 16–24 used generative AI against 24.15% ⁶⁵ of those aged 55–74; in Italy the figures were 47.17% ⁶⁶ and 7.69% ⁶⁷. Use for work ranged from 35.44% ⁶⁸ of people aged 16–74 in Norway to 8% ⁶⁹ in Italy. Every country with each age group is on the generative AI use ranking.

Pew Research Center found that 44% ⁷² of US adults had used ChatGPT in February 2026, up from 34% ⁷³ a year earlier. Adults under 50 were about twice as likely as those 50 and older to have used it (57% ⁷⁴ against 28% ⁷⁵), and 38% ⁵ of employed adults used chatbots for tasks at work.

In China, CNNIC counted 602 million ⁷⁶ generative AI users at the end of 2025, up from 249 million ⁷⁷ a year earlier, or 42.8% ⁶ of the population. The Reuters Institute found weekly generative AI use across six markets rose from 18% ⁷⁸ in 2024 to 34% ⁷⁹ in 2025, while 10% ⁸⁰ of people use AI chatbots for news each week and 1% ⁸¹ call AI their main source of news.

Company figures answer a different question. OpenAI reported 700 million ⁸² weekly active users of ChatGPT in September 2025; that is the company's own count of its product, not a survey of a population.

The most common AI technology among EU enterprises in 2025 was analysis of written language, used by 11.8% ⁸³ of all enterprises, followed by generation of written or spoken language at 8.8% ⁸⁴. Enterprises can use several technologies, so these shares overlap; they are technology categories, not supplier market shares.

Among enterprises that had considered AI but did not use it, 70.89% ⁸⁵ cited missing expertise, 52.52% ⁸⁶ unclear legal consequences and 48.83% ⁸⁷ privacy or data-protection concerns. The denominator is that group of interested non-adopters, not all enterprises.

Every number on this page is a published StatOrigin record. US business figures come from the Census Bureau's Business Trends and Outlook Survey dashboard files, which cover nonfarm employer businesses; businesses without employees are outside it. BTOS programme description. Current use refers to the two weeks before each survey; dates are the end of that window.

EU enterprise figures come from Eurostat's survey on ICT usage in enterprises (dataset isoc_eb_ai and the December 2025 release), covering enterprises with 10 or more persons employed in industry and market services other than finance. Generative AI use by people comes from Eurostat's survey on ICT usage in households and by individuals (isoc_ai_iaiu), people aged 16–74. Country values are read from the publisher's data files; the snapshot of each response is kept with the record.

Pew Research Center, CNNIC and the Reuters Institute figures are quoted from their published reports; OpenAI's figure is the company's own count. These are survey estimates, subject to sampling and nonresponse error, and the sources differ in population, question and period. No global total, causal effect, spending estimate or forecast is derived here. Analysis is StatOrigin editorial interpretation of the linked evidence.