AI statistics 2026: AI adoption in Germany in numbers
These AI statistics bring together the key figures on AI use in German companies: official data from Destatis and Eurostat, surveys by Bitkom and KfW, plus figures on data protection, cloud dependence, shadow AI, the market and the EU AI Act. Every figure comes with its source, year and base, so you can read it correctly and cite it.
As of 3 November 2026 · next update: December 2026 (new Eurostat data)
¹ Among enterprises in Germany that considered AI but do not use it (Eurostat 2025).
Key points
26% of enterprises with 10+ employees used AI in 2025 (Destatis). The EU average was19.95% (Eurostat).- Bitkom reports
57% for 2026 among companies with 20+ employees. This is not directly comparable with the official figure: different year, size and a different question (wording not published). - Size matters:
57% of large enterprises use AI, but only23% of small ones (Destatis 2025). - Among enterprises that considered AI but do not use it,
62.22% in Germany name unclear legal consequences and60.01% data protection, well above the EU average (Eurostat 2025). 71% of companies use US cloud providers, but only8% prefer them;91% prefer providers from Germany (Bitkom 2026).12% of employed people who use AI at work do so without their employer knowing (Bitkom 2026).
How many companies in Germany use AI?
The official answer:
Eurostat reports 25.97% for Germany and 19.95% for the EU. Germany is therefore about six percentage points above the EU average. The share has more than doubled since 2021 (10.56%).
Higher figures from surveys do not contradict this. They measure differently; see the table below and the methodology section.
Three figures, three measurements: 26%, 57% and 20%
| Source | AI use | Year | Who was counted? | Method |
|---|---|---|---|---|
| Destatis (Federal Statistical Office), table on enterprises using AI technologies, reference year 2025, as of 24 Nov 2025 | 26% | 2025 | Enterprises with 10+ employees, excluding finance | Official survey, EU-harmonised; asks about specific AI technologies |
| Bitkom press release “Erstmals nutzt die Mehrheit der Unternehmen KI”, 14 Sep 2026 (603 companies with 20+ employees) | 57% | 2026 | Companies with 20+ employees | Telephone survey, 603 companies; different question (wording not published) |
| KfW Research press release on AI in German SMEs, 11 Feb 2026 (KfW SME Panel) | 20% | 2022–2024 | SMEs including micro firms, almost 780,000 companies | KfW SME Panel |
The figures cannot be converted into one another. Always cite source, year and base.
Large companies use AI more than twice as often as small ones.
In 2025,
Compared with the EU, Germany’s lead is largest among small enterprises: 23.06% vs 17.0%. Among large enterprises, Germany and the EU are almost level.
57%
of large enterprises (250+ employees) used AI in 2025.
36%
of medium-sized enterprises (50–249) used AI in 2025.
23%
of small enterprises (10–49) used AI in 2025.
| Figure | Value | Source |
|---|---|---|
| SMEs with 50+ employees using AI (2022–2024) | 36% | KfW Research press release on AI in German SMEs, 11 Feb 2026 (KfW SME Panel) |
| SMEs with their own R&D using AI | 53% | KfW Research press release on AI in German SMEs, 11 Feb 2026 (KfW SME Panel) |
| SMEs in construction using AI | 8% | KfW Research press release on AI in German SMEs, 11 Feb 2026 (KfW SME Panel) |
| Staff access to generative AI: 20–99 staff 23%, 100–499 staff 36%, 500+ staff | 43% | Bitkom press release “Beschäftigte nutzen vermehrt Schatten-KI”, 21 Oct 2025 (604 companies with 20+ employees) |
Data protection and legal questions hold AI back more in Germany than in the EU.
Eurostat asks enterprises that
Important when citing: these percentages refer only to the group that considered AI, not to all enterprises. So write “60% of enterprises that considered AI but do not use it”, not “60% of enterprises”.
Companies that already use AI also put data protection first: 66% of AI-using companies see data protection requirements as an obstacle, and 45% fear company data could fall into the wrong hands (Bitkom 2026).
62.22%
of enterprises in Germany that considered AI but do not use it name unclear legal consequences (EU: 53.6%, 2025).
66%
of AI-using companies see data protection requirements as an obstacle (2026).
45%
of AI-using companies fear company data could fall into the wrong hands (2026).
¹ Base: enterprises with 10+ employees that considered using AI but do not use it. Multiple answers possible.
Companies use the US cloud but want a provider from Germany.
The Bitkom Cloud Report 2026 shows a clear gap:
85% think Germany is too dependent on US cloud providers. In a separate Bitkom survey from October 2026, 73% are worried about dependence on the US (2024: 38%). At the same time demand is growing: 42% use AI from the cloud today, and 69% expect to in five years.
Running AI on your own servers or in a data centre in Germany closes exactly this gap; RAGSuite is built for both.
85%
think Germany is too dependent on US cloud providers (2025: 78%).
Source: Bitkom Cloud Report 2026, press release of 17 Jun 2026 (603 companies with 20+ employees)
87%
require protection against access by the provider as a must-have; just as many see a lack of transparency as a weakness of international providers.
Source: Bitkom Cloud Report 2026, press release of 17 Jun 2026 (603 companies with 20+ employees)
16 months
is how long companies could survive on average without digital imports (2025: 14 months).
Digital sovereignty for AI →The US CLOUD Act explained →Data sovereignty check →
Employees often use AI before their company provides it.
58% of people aged 16+ in Germany use AI, 34% at least once a week. Among employed people, 48% use AI at work, 8% daily and 18% at least weekly (Bitkom 2026).
The three most used AI assistants come from US providers; 71%, 50% and 43% of AI users use them. One European offering reaches 4% (Bitkom 2026).
¹ Base: employed people who use AI at work.
Training also matters legally: Article 4 of the EU AI Act requires measures to ensure staff AI literacy.
The German AI market grows by 48% in 2026.
The market figures are
€28.7bn
AI market Germany 2026, +48% (forecast; 2025: €19.4bn).
Source: Bitkom press release “28,7 Milliarden Euro für KI”, 22 Sep 2026 (market data: IDC, forecast)
€11.5bn
generative AI spending 2026, double 2025 (€5.7bn) and 40% of all AI spending (forecast).
Source: Bitkom press release “28,7 Milliarden Euro für KI”, 22 Sep 2026 (market data: IDC, forecast)
€40.3bn
expected AI market 2027, +40% (forecast).
Source: Bitkom press release “28,7 Milliarden Euro für KI”, 22 Sep 2026 (market data: IDC, forecast)
| Figure | Value | Source |
|---|---|---|
| AI software 2026, +65% (forecast; services €6.8bn, hardware €5.8bn) | €16.0bn | Bitkom press release “28,7 Milliarden Euro für KI”, 22 Sep 2026 (market data: IDC, forecast) |
| Companies spending more on AI in 2026 than the year before (significantly more 13%, somewhat more 32%) | 45% | Bitkom press release “Erstmals nutzt die Mehrheit der Unternehmen KI”, 14 Sep 2026 (603 companies with 20+ employees) |
| AI users for whom infrastructure such as cloud and compute is a major cost (data preparation: 50%) | 51% | Bitkom press release “Erstmals nutzt die Mehrheit der Unternehmen KI”, 14 Sep 2026 (603 companies with 20+ employees) |
| AI users who see their competitive position improved by AI | 66% | Bitkom press release “Erstmals nutzt die Mehrheit der Unternehmen KI”, 14 Sep 2026 (603 companies with 20+ employees) |
The key deadlines and figures of the EU AI Act.
The EU AI Act applies in stages. The Digital Omnibus on AI, Regulation (EU) 2026/1744, moved the deadlines for high-risk AI. Older sources often still give August 2026 for this; that is out of date.
| Date | What applies | Source |
|---|---|---|
| 27 July 2026 | Digital Omnibus on AI, Regulation (EU) 2026/1744, enters into force | European Commission, “Regulatory framework for AI”, updated 19 Oct 2026 |
| 2 August 2026 | Transparency duties under Art. 50 (e.g. notices for chatbots and AI content) and enforcement powers for general-purpose AI models | European Commission, “Regulatory framework for AI”, updated 19 Oct 2026 |
| 2 December 2027 | Duties for stand-alone high-risk AI systems (Annex III) | European Commission, “Regulatory framework for AI”, updated 19 Oct 2026 |
| 2 August 2028 | Duties for high-risk AI embedded in regulated products (Annex I) | European Commission, “Regulatory framework for AI”, updated 19 Oct 2026 |
€15M / 3%
fine ceiling for, among others, breaches of the Art. 50 transparency duties: up to €15M or 3% of worldwide annual turnover.
Source: European Commission AI Act Service Desk, Article 99 (penalties)
37%
of companies expect to be affected by the AI Act as deployers (2025: 23%); 89% of companies that are affected or still checking expect high implementation effort.
66%
of companies think the AI Act brings more disadvantages than advantages (2025: 56%).
How to read the figures.
Statistics on AI use differ mainly in four respects: who was counted, how the question was asked, which year, and whether it is a measurement or a forecast. Check these four points before comparing two figures.
Official statistics
Destatis and Eurostat use the annual EU survey on ICT use in enterprises. It covers enterprises with 10+ employees, excluding finance. From reference year 2025 the EU enterprise concept applies, so earlier years are only partly comparable. We give Eurostat values with up to two decimals, some rounded.
Association surveys
Bitkom surveys around 600 companies with 20+ employees, or around 1,000 people aged 16+, by telephone. The questions are broader and the results more recent, but they are not the same as official statistics. KfW figures come from its SME panel and include micro firms.
Watch the base
Many values refer to a subgroup: Eurostat’s reasons against AI apply only to enterprises that considered AI but do not use it; some Bitkom obstacles apply only to AI users. The base is given with every figure.
Forecasts are not measurements
Market figures are IDC estimates published by Bitkom. The EU target of 75% for 2030 is a policy goal. We label both.
We only included figures we checked in the primary source. Values we only knew from secondary sources were left out. Eurostat values were retrieved from the database on 3 November 2026.
You are welcome to use these figures.
Journalists, bloggers and data protection officers may use the figures and charts on this page with attribution. For individual values, please also name the primary source, i.e. Destatis, Eurostat, Bitkom or KfW, with year and base.
We update the page as soon as new official data appears. The next update is planned for December 2026.
RAGSuite (2026): AI statistics 2026 – AI adoption in Germany in numbers. As of 3 November 2026. https://ragsuite.de/en/guides/ai-statistics/
Questions about AI statistics.
How many companies in Germany use AI?
According to official statistics, 26% of enterprises with 10 or more employees used at least one AI technology in 2025 (Destatis). A Bitkom survey of companies with 20 or more employees finds 57% for 2026. Both figures are correct but measure different things: a different year, company size and question.
Why do the Destatis and Bitkom figures differ so much?
Destatis and Eurostat use the EU-wide harmonised method, ask whether an enterprise uses specific AI technologies and cover all enterprises with 10 or more employees. Bitkom surveys around 600 companies with 20 or more employees by telephone and asks more broadly whether AI is in use. Larger companies use AI more often, and the Bitkom figure is from a later year. This explains much of the gap.
Is Germany above the EU average in AI use?
Yes. In 2025, 25.97% of enterprises with 10 or more employees in Germany used AI, against an EU average of 19.95% (Eurostat). The gap is largest among small enterprises (23.06% vs 17.0%) and smallest among large ones (56.99% vs 55.03%).
What stops companies from using AI?
Among German enterprises that considered AI but do not use it, 72.07% named a lack of expertise in 2025, 62.22% unclear legal consequences and 60.01% data protection and privacy concerns (Eurostat). Legal and privacy concerns are named far more often in Germany than on EU average. Companies that already use AI also see data protection as the biggest obstacle (66%, Bitkom 2026).
How widespread is shadow AI?
12% of employed people who use AI at work do so without their employer knowing (Bitkom, 2026). From the company side, private AI tools are used at work, at least presumably, in 4 out of 10 companies, but only 26% give their staff access to generative AI themselves (Bitkom, 2025).
How big is the AI market in Germany?
According to an IDC forecast published by Bitkom, the German AI market grows by 48% to €28.7bn in 2026. Generative AI accounts for €11.5bn. €40.3bn is expected for 2027. These are estimates, not official statistics.
When will new figures be available?
Eurostat and Destatis usually publish the results of the annual survey on ICT use in enterprises towards the end of the year. We will update this page as soon as the 2026 data is out, expected in December 2026.
Use AI without handing over your data.
Data protection and dependence are the most cited barriers. Jürgen shows you how RAGSuite runs on your own servers or in a data centre in Germany.