AI study: investments are increasing, but scaling remains a challenge
According to the Lenovo study, managers in the EMEA region are extremely satisfied with AI projects. Despite increasing budgets, scaling remains a hurdle.
According to Lenovo's "AInomics CIO Playbook 2025", 94 % of AI projects in the EMEA region have been successful. Companies are increasingly investing in generative AI, while interpretive and predictive AI are becoming less important. The share of IT budgets for AI will increase to 20 % by 2025, up from 13 % in the previous year. The willingness to invest is particularly high in Germany, France and the UK.
Despite this positive development, challenges remain. Scaling problems (30 %) and insufficient data quality (29 %) are slowing down implementation. In addition, there is often a lack of clear guidelines: 22 % of companies have no plans for AI governance, which makes integration more difficult. At the same time, the use of professional AI services is becoming increasingly important. While 27 % of companies currently use such services, 53 % are planning to implement them and a further 19 % are reviewing their use.
The demand for hybrid IT infrastructures is also increasing. 65 % of companies in the EMEA region primarily rely on on-premises or hybrid architectures to meet data protection and compliance requirements. This is confirmed by Giovanni Di Filippo, President of the EMEA Infrastructure Solutions Group at Lenovo: "Most companies have left the hype phase behind them and are focusing on the sustainable implementation of AI."
The type of applications is also changing. The focus is shifting from AI-supported applications to the development and management of AI models. Companies need to ensure that their AI strategies are not only innovative, but also sustainable and scalable. According to Lenovo manager Greg Smith, the success of AI projects depends not only on technology, but also on expertise and professional support: "Executives are quickly gaining confidence and are increasingly turning to diversified approaches such as AI-as-a-Service."
According to the study, four key challenges need to be overcome in order to successfully implement AI: maintaining data sovereignty and compliance (32 %), seamless integration into existing systems and processes (32 %), employee training and development (31 %) and access to high-quality data (31 %).