Two years ago, AI was something most German companies were "looking into". Today, it is part of daily business in a majority of them. The latest Bitkom study, published in September 2026, shows how fast this has happened — and how much work remains.
- 57% of German companies use AI (2025: 36%, 2024: 20%)
- 11% already use AI agents; another 29% are planning to
- 0% say they fully use AI's potential
- 66% of AI users say AI improved their competitive position
- 91% consider AI the most important technology of the future
From experiment to majority in two years
The growth curve is steep: from 20% in 2024 to 36% in 2025 to 57% in 2026. Another 38% are planning or discussing AI use. Only 4% of companies now say AI is not relevant to them — two years ago, that figure was 41%.
Attitudes have shifted just as quickly. Only 8% still see AI as overhyped (2024: 26%). And 61% now believe that companies without AI have no future.
The value gap
The most telling number in the study is a zero: not a single company says it fully exploits the potential of AI. Only 3% say they use it to a large extent, 28% to a small extent, and 59% not at all in a meaningful way.
In other words: most companies have started with AI — through chat assistants, writing tools or individual pilots — but very few have built it into their core processes. Yet where AI is used, it pays off: two thirds of users say it has improved their competitive position.
Where the money really goes
When companies invest in AI, the largest costs are not the AI models themselves:
| Cost item (AI users) | Share naming it among the largest costs |
|---|---|
| Infrastructure | 51% |
| Data preparation | 50% |
| System integration | 41% |
| Tokens / model usage | 8% |
This is the key lesson for decision-makers: AI success is mostly a data and process problem, not a model problem. Companies with clean data, documented processes and integrated systems get value faster.
45% of companies plan to spend more on AI in 2026 than in 2025; half keep budgets flat.
What holds companies back
Among companies not yet using AI, the main barriers are:
- lack of technical know-how (85%)
- legal uncertainty (66%)
- hard-to-calculate costs (63%)
- lack of staff (51%)
Companies already using AI name data protection requirements (66%), legal uncertainty (56%) and costs (54%) as their biggest challenges. Regulation weighs heavily: 66% believe the EU AI Act creates more disadvantages than advantages for German companies.
AI and the workforce
70% of companies now offer AI training to employees (in 2025, 43% offered none). At the same time, 57% expect their headcount to decrease over the next two years because of AI — 45% slightly, 12% significantly. The message is clear: AI is changing jobs, and qualification is becoming a core management task.
From pilots to real impact: five steps
- Start with processes, not tools. Which recurring tasks cost the most time — invoice checking, customer requests, reporting, document handling?
- Fix your data first. Structured, accessible data is the foundation for every AI use case.
- Pick two or three use cases with measurable impact and take them into production, rather than running ten pilots.
- Integrate, don't isolate. AI creates value when it is connected to ERP, CRM and document systems.
- Build skills and governance in parallel: training, clear rules for data use, and an overview of AI Act obligations.
Conclusion: Germany has crossed the threshold — AI is no longer the exception. The next competitive divide will not be between companies that use AI and those that don't, but between those that use it superficially and those that build it into how they work.