April 29, 2026, 10:47 am | Read time: 3 minutes
Many companies have not critically examined their investments in artificial intelligence for a long time. However, a clear effect is now emerging: Ongoing expenses for AI systems are rising so sharply that in some firms, they already exceed the costs for the entire staff. This brings the question into focus of whether the benefits of these technologies still justify the high expenses.
High Costs from Operations and Infrastructure
Companies are investing not only in AI software like ChatGPT but also in the technical foundation behind it. This includes computing power, storage, cloud services, and the ongoing operation of the systems. This combination is precisely what makes AI quickly become expensive in everyday use. The more intensively companies use the technology, the more the total costs increase.
Tokens as Hidden Cost Drivers
A central cost factor is so-called tokens. These are small text units with which language models work, such as word parts or entire words. Each input and each generated response incurs costs.
When companies make many AI requests, these amounts quickly grow into large sums. An example is provided by Uber: According to “The Information,” the company’s Chief Technology Officer (CTO) has already exhausted the entire AI budget for 2026 well before the end of the year. A significant part of these costs arose from intensive token consumption.
Storage Prices Expected to Double by March
Will Samsung Smartphones Soon Become More Expensive and Worse?
AI Expenses Partially Exceed Personnel Costs
The development is particularly evident in large technology companies. Bryan Catanzaro, Vice President of Applied Deep Learning at Nvidia, told Axios that in his area, the costs for computing power now exceed personnel expenses.
Just a few years ago, such a development was hardly imaginable. Nevertheless, market researchers also confirm the trend of rising expenses. For instance, “Gartner” forecasts that worldwide IT spending could rise to $6.31 trillion in 2026 (equivalent to 5.4 trillion euros). This represents an increase of 13.5 percent compared to 2025.
The drivers of this development are primarily AI infrastructure, cloud offerings, and software licenses. Infrastructure includes servers, chips, and networks necessary for the operation of AI.
Also of interest: OpenAI is reportedly working on its own smartphone
Economic Benefit Remains a Central Question
Despite growing investments, many companies still lack a clear assessment of the actual benefits of AI on a large scale. Rising token costs and higher prices from providers can quickly turn supposed competitive advantages into economic risks.
Publicly traded companies, in particular, are under pressure. They must regularly disclose how budgets are used and what returns these investments bring. Shareholders, supervisory boards, and financial officers increasingly demand understandable justifications for high AI expenses. Pure promises of innovation are no longer sufficient. What matters is whether the economic benefit can be clearly measured.