June 13, 2026, 6:25 am | Read time: 3 minutes
Artificial intelligence is intended to make companies more efficient and speed up workflows. Many firms are investing in powerful AI tools and making them available to their employees. However, as usage increases, so do the costs. A recent case shows the consequences of lacking regulations and uncontrolled use.
Lack of Guidelines Drives Up Costs
A large company, whose name is not disclosed, had to pay around $500 million within a month. This amounts to approximately 429 million euros. The sum was solely for the use of Claude, the AI system from Anthropic.
The trigger was neither a technical malfunction nor an exceptionally high load. Instead, there were no clear usage limits. Employees could access the system without restrictions. As a result, the number of requests increased significantly in a short time. Since each input incurs additional costs, many individual operations led to a bill that was substantial even for a large corporation.
Even Simple Requests Can Be Costly
Not all employees used the AI for complex tasks. Developers and office workers, for example, asked the AI about the weather or automated simple routine tasks.
Such requests initially seem unproblematic. However, in large numbers, they can cause noticeable costs. The consumption is particularly high with autonomous AI agents. These systems handle tasks in multiple steps and require significantly more computing power. According to experts, resource consumption can be up to a thousand times higher than with a regular text query.
Companies Question Their Investments
According to industry observers, this is not an isolated case. More and more companies are now examining whether their investments in AI are economically viable.
Also interesting: Revealed! AI for Companies More Expensive Than Human Staff
Several corporations have already responded. Microsoft has reduced the use of Claude in certain areas. At Uber, ongoing expenses are also becoming more of a focus. The expected productivity boost is falling short of expectations in many places. Therefore, executives are increasingly questioning whether the benefits justify the high costs.
Revealed: AI Costs Companies More Than Human Staff
Up to $9,000? GitHub Users Warn of Copilot Costs
Why the Bills Grow So Quickly
Many AI providers charge for their services using so-called tokens. Fees are incurred for inputs and responses. Individual requests often cost only a few cents or even less.
In large companies, however, these amounts add up quickly. When thousands of employees analyze documents, create texts, or evaluate data daily, consumption increases significantly. Many small cost points can thus become millions in a short time. According to experts, a significant risk is that many companies initially focus on the rapid introduction of the technology and establish control mechanisms only later.
Companies Now Focus on Cost Control
During the initial phase of the AI boom, some companies encouraged the most intensive use of their systems possible. According to reports, Amazon had internal rankings that recorded the token consumption of individual teams. This led to what is known as tokenmaxxing. Employees sometimes outsourced even less relevant tasks to the AI to achieve better statistics.
Amazon has since abandoned this approach. Today, the concrete benefit for customers is the focus. Many companies are now following a similar course. They pay more attention to cost-effectiveness and monitor their expenses more closely. Administrators receive stricter guidelines, while departments work with fixed budgets and clearly defined quotas. The goal is a targeted use of the technology instead of uncontrolled consumption.