A Serverelite nationwide survey of nearly 500 Hungarian company leaders, IT specialists and system administrators found that 92% of respondent firms already use artificial intelligence (AI) in some form. Behind that high figure, however, there is wide variation: nearly two-thirds of companies report AI is still experimental or confined to a few narrow areas, while only 27% say AI is used broadly across multiple business processes.
Which tools are used and in what share?
Among generative AI tools, ChatGPT is the most widely used, cited by 69% of respondents. Google Gemini follows at 56%, Anthropic Claude at 37%, and Microsoft Copilot at 29%. The survey also found that 22% of companies already deploy internally developed AI solutions, indicating rising demand for bespoke AI tailored to specific business needs.
Andrea Marosán, CEO of Serverelite, warned of risks connected to unmanaged personal accounts: if employees use non-company-managed ChatGPT or Gemini accounts, firms can lose control over the data handled through those services.
Main application areas and motivations
Companies primarily apply AI for information processing, data analysis and support of digital workflows. The most common use cases are programming (64%), data analysis (50%) and document processing (48%). In contrast, AI currently plays a much smaller role in HR (5%), manufacturing (4%) and logistics (3%).
Key motivations for adopting AI include saving time and costs, reducing human resource needs, and decreasing errors originating from manual work.
Infrastructure, spending and technological constraints
For running AI workloads, 62% of respondent firms use public cloud services, mostly for quick scalability and to avoid large upfront investments. On-premises server solutions are used by 33%, and private cloud by 17%.
Serverelite highlights significant preparedness gaps in infrastructure: only 16% of companies consider their existing server capacity fully adequate for AI tasks, while 46% are dissatisfied or skeptical about current hardware capabilities. GPU acceleration is still uncommon: 58% of respondents do not use GPUs to speed up computations.
Top technological barriers to development were high costs (36%), limited memory (29%) and constrained compute capacity (27%). On the spending side, companies also act cautiously: 42% spend less than HUF 50,000 per month on AI services, while only 2% spend more than HUF 1,000,000 per month.
Conclusions: progress, but large-scale transformation remains ahead
Serverelite's assessment is that Hungarian companies have entered the AI era, yet systemic, enterprise-wide integration of the technology remains in early stages. According to Andrea Marosán, the next major challenges are scaling successful pilots to company-wide solutions and embedding AI in strategic processes — steps that current infrastructure weaknesses may impede.
Her recommendation is that companies first define the business goals, processes and ways they intend to use AI, then assess the necessary infrastructure and make development decisions based on that analysis.
Methodology
Nearly 500 Hungarian company leaders, IT professionals and system administrators participated in the survey.
