Companies' costs for AI agents are rising sharply, yet most plan to invest more
A McKinsey survey shows a year-on-year increase in companies' AI costs of 5 to 600 percent due to the operation of AI agents. A fifth of companies have limited consumption as a result, yet more than 60 percent still plan to increase investment.
According to an extensive survey by the consulting firm McKinsey, companies' costs for artificial intelligence are rising year on year, mainly as a result of the expansion of autonomous AI agents, whose operation is more costly. Czech and Slovak companies contacted by the E15.cz server cite specific figures: the development studio Ackee recorded an approximately fivefold year-on-year increase in AI spending, while the Slovak company Sudolabs reports an increase of nearly 600 percent in costs for AI agents in software development. According to the IT company Trask, the development and operation of internal AI infrastructure is today among its largest investment items.
The McKinsey survey further shows that 20 percent of respondents have limited the use of AI within their company for financial reasons, for example by setting limits on token consumption. According to the survey, this proportion is similar across organizations of different sizes and industries. According to the cited representatives, companies generally do not want to ban employees from using AI across the board; rather, they set limits as a safeguard against anomalies (e.g., an agent running in a loop all night) and as a signal of where and why consumption is increasing.
According to the article, McKinsey analyst Michael Chui states that although the price per token is falling, the number of tokens consumed and generated is growing even faster, especially in agentic software development and complex tasks using the most advanced models. Nevertheless, according to the survey, nearly a third of companies spend more than 10 percent of their total IT budget on AI, and more than 60 percent of respondents plan to further increase their AI investments next year. Companies contacted, such as Sudolabs and Luigi's Box, confirm this trend and expect that the prices of the most powerful models will continue to rise and that AI agents will spread into less technical departments as well.
Why it matters
For companies deploying or planning to deploy AI agents, this is a signal that operating costs may not follow the falling price per token, because consumption is growing faster — it is therefore necessary to monitor consumption and set limits, not just rely on the unit price of models.
Relevant practical impact
What this means
For a business
Companies must account for the fact that the costs of AI agents are growing faster than the per-token price is falling, and factor this into IT budgets and consumption limits, even though, according to the survey, investment in AI is paying off for companies.
StrategyCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.