Breakthrough in artificial intelligence

One in three companies in Germany is now using AI – almost twice as many as a year ago
8 out of 10 companies see AI as the most important future technology
93 per cent would prefer an AI provider from Germany

Berlin, 15 September 2025 – Artificial intelligence has arrived across the breadth of the German economy in recent months. Around one in three companies (36 per cent) is now using AI. This means the share has almost doubled compared to just one year ago, when it stood at 20 per cent. In addition, almost one in two companies (47 per cent) is currently planning or discussing the use of AI, also significantly more than in the previous year (37 per cent). By contrast, only 17 per cent now say that AI is not an issue for them, compared with 41 per cent in the previous year. These are the results of a representative survey of 604 companies in Germany with at least 20 employees, conducted on behalf of the digital association Bitkom. “Artificial intelligence has achieved a breakthrough in the German economy,” says Bitkom President Dr Ralf Wintergerst. “Companies have not only recognised the potential of AI, they are using it and investing in it. This is good news for the competitiveness and future viability of the German economy.”

A positive view of AI dominates

Eight out of ten companies (81 per cent) are now convinced that AI is the most important future technology (2024: 73 per cent), while only 17 per cent still see it as a passing hype (2024: 26 per cent). For the first time, a slim majority of 51 per cent believe that companies which do not use AI have no future (2024: 48 per cent). Conversely, only 31 per cent still think that AI may look spectacular but brings no concrete benefits to companies. A year ago, this figure was significantly higher at 46 per cent.

Looking at their own company, 83 per cent see AI as an opportunity, compared with 78 per cent a year ago and just 68 per cent in 2023. Only 14 per cent see AI more as a risk, and just 1 per cent believe that AI has no impact on their own company. Around a quarter (24 per cent) expect AI to change their business model. Almost as many (23 per cent) are concerned that AI could threaten the company’s existence. “AI offers enormous opportunities for companies, regardless of size or sector. The greatest risk is simply ignoring AI and missing the AI train,” says Wintergerst.

29 per cent of companies want to increase their AI investment

In the current year, 8 per cent of companies that use, plan or discuss AI want to invest significantly more in AI than in 2024. A further 21 per cent are planning moderately higher investments. By contrast, only 5 per cent want to reduce their investments somewhat, and no company wants to cut them significantly. Four per cent have never invested and have no plans to do so, apparently using only free offerings. The clear majority of 60 per cent are keeping AI investments at the previous year’s level. “Entering the world of AI is cheaper for companies than ever before. There is a wide range of free offerings available to gain an initial overview of the possibilities,” says Wintergerst. “However, particularly powerful and legally secure AI that is tailored to the specific needs of a company and deeply integrated into its IT systems does not come free of charge.”

AI use: customer contact and communication dominate

AI is used primarily in customer contact (88 per cent) and in marketing and communication (57 per cent). At a considerable distance behind follow applications in research and development (21 per cent), production processes (20 per cent), controlling and accounting (17 per cent), human resources (14 per cent) and internal knowledge management (11 per cent). AI is hardly used in management, legal or tax departments and sales (each 5 per cent), as well as in IT departments (2 per cent). Around one in eight companies (12 per cent) that use AI integrate artificial intelligence into their own products and services. “Many companies are still using AI only in a selective way. With the initial experience and know-how gained in the process, further fields of application can be opened up. This is the only way to fully realise the potential of AI,” says Wintergerst.

The often still selective use of AI is reflected in the number of AI applications used by companies. A quarter of companies (24 per cent) that use AI rely on just one application, a further 27 per cent use two applications and 24 per cent use three. Only 6 per cent use four applications, and just 2 per cent use five or more. Seventeen per cent cannot or do not wish to provide information on this.

20 per cent expect fewer jobs due to AI, 7 per cent more

It is currently difficult to assess the impact AI will have on the labour market. The clear majority of companies in Germany (67 per cent) expect AI to have no impact on employment levels. One fifth (20 per cent) believe that AI will reduce the number of employees – by an average of 7 per cent. At the same time, 7 per cent expect AI to increase employment, by an average of 8 per cent. Among companies already using AI, 28 per cent expect employment to fall by an average of 7 per cent, while 9 per cent expect an increase, averaging 9 per cent. Fifty-seven per cent expect AI to have no impact on the number of jobs. Around one third of all companies (31 per cent) believe that AI will help to alleviate the skills shortage in Germany. “AI will change job profiles, and some will even disappear. At the same time, other new jobs will emerge,” says Wintergerst. “For the German labour market, with its challenging demographic structure and already significant skills shortages, AI is a major opportunity.”

At present, only 5 per cent of companies specifically recruit specialists with AI skills. A further 27 per cent plan to do so, 24 per cent are discussing it, but for 43 per cent it is not an issue. And only 8 per cent offer AI training to all employees, a further 21 per cent to the majority of staff and 25 per cent to selected employees. However, 43 per cent offer no such training. “Companies themselves are responsible for building up AI expertise. Every company should therefore train its employees in AI. Companies that use AI are even obliged to do so under the AI Act,” says Wintergerst.

AI barriers: legal uncertainty, lack of know-how and lack of staff

The biggest obstacles to using AI in the German economy are uncertainty caused by legal hurdles and ambiguities (53 per cent), lack of technical know-how (53 per cent) and insufficient personnel resources (51 per cent). Forty-eight per cent complain about strict data protection requirements, 39 per cent fear that data could fall into the wrong hands, 38 per cent cite a lack of transparency of results, 36 per cent poor quality of results, another 36 per cent a lack of financial resources, and 35 per cent fear future legal restrictions on the technology. For 31 per cent, a lack of acceptance among employees is one of the biggest obstacles, 24 per cent lack data for AI use, 23 per cent see no use cases and 17 per cent have ethical concerns.

For 9 out of 10 companies, the country of origin of the AI provider matters

For a clear majority of companies (88 per cent), the country of origin of the AI provider is important. Ninety-three per cent would prefer AI from Germany, followed at a considerable distance by the USA (51 per cent), Japan (43 per cent), EU countries other than Germany and France (40 per cent), France (38 per cent), the United Kingdom (37 per cent), South Korea (29 per cent), Israel (27 per cent), India (22 per cent), Ukraine (19 per cent) and China (18 per cent). AI from Russia is not an option for any company. “Companies want AI providers from Germany. However, these will only be successful if they are just as powerful as foreign offerings and competitively priced,” says Wintergerst. “We must not accept a role for Germany as merely a user of AI – we must become a provider of AI.”

A majority see more disadvantages than advantages in the AI Act

Companies are critical of current AI regulation. Fifty-six per cent believe that the European AI Act creates more disadvantages than advantages for German companies. Twenty-three per cent expect to be affected by the AI Act as users, 1 per cent as providers. Thirty-two per cent do not see themselves as affected, 30 per cent are still reviewing the issue and 11 per cent have not yet engaged with the AI Act.

Among companies that expect to fall under EU regulation, 93 per cent anticipate a high level of effort: 49 per cent expect very high effort and 44 per cent fairly high effort. Around one third of affected companies (37 per cent) expect to operate at least one so-called high-risk AI system, while 29 per cent expect two. Only 4 per cent expect three or more high-risk systems, and 2 per cent expect none. However, 29 per cent cannot or do not yet wish to comment. “With regard to the AI Act, companies urgently need clarity on how it will be implemented in detail. They also need a federal-level point of contact to support them in using AI in compliance with the AI Act,” says Wintergerst. The implementation deadlines for high-risk requirements should be postponed by at least two years. Otherwise, due to a lack of standards and resulting legal uncertainty, there is a risk of “a full-scale innovation brake in the high-risk AI sector”. In addition, high-risk requirements in already heavily regulated sectors such as medical devices and machinery should be made more flexible to avoid inconsistencies and redundancies with existing regulation.

What policymakers should do for AI

Above all, companies want policymakers to promote German AI providers (51 per cent), reform the AI Act (46 per cent) and improve access to data (45 per cent). Around one third would prioritise promoting AI research (36 per cent), investing in AI computing infrastructure (34 per cent) and promoting the use of AI in public administration (31 per cent). Twenty-eight per cent call for support for the use of AI in the economy, 17 per cent for support for AI start-ups and 6 per cent for promoting AI talent. Only 8 per cent believe the government should place no focus on AI. More than one third of companies (37 per cent) would even like AI to go unregulated for ten years – an idea that has been discussed in the USA. “When it comes to AI support, we need a holistic strategy,” says Wintergerst. “This includes becoming much more ambitious both in the breadth of measures and in the level of investment.”