Companies put AI spending under closer review as usage costs climb
After broad AI rollouts, executives are weighing model costs, employee usage and whether the technology is producing returns.
By Maya Lindqvist · Senior Technology Correspondent
3 min read
Companies that pushed employees to experiment widely with artificial intelligence are starting to rein in costs and demand clearer returns. Fortune reported that executives across industries are now reviewing AI budgets after two years of rapid adoption and generous access to new tools.
The shift matters because AI use can become expensive quickly when companies give large groups of workers access to advanced models without strict limits. Fortune reported that Uber used its full 2026 AI budget in four months, and that the company’s chief operating officer said the spending was becoming harder to defend.
Axios also reported last month that, according to a consultant, one client spent half a billion dollars in a single month after failing to limit employee AI use. Fortune said such cases have pushed companies to ask which tasks need costly frontier models and which can be handled by cheaper systems.
From trials to controls
Peter DeSantis, a senior vice president at Amazon, told Fortune that the current cost pressure resembles earlier phases of cloud adoption. He said companies often gain speed from a new technology before later finding they need better ways to manage budgets and use it efficiently.
That pattern is now showing up in enterprise AI purchasing, according to Fortune. Many companies first encouraged broad experimentation, often by handing out licenses widely, then discovered that heavy usage could create bills that were difficult to predict.
Philippe Rambach, Schneider Electric’s chief AI officer, told Fortune that the company is taking a more careful approach to matching AI tasks with the right model. He said companies do not need the most advanced model for every job and can often use cheaper tools.
Rambach also told Fortune that AI costs increasingly need to be measured, controlled and included in business cases, planning and decisions. Fortune reported that his comments reflect a broader move away from broad experimentation toward more selective deployments.
Return on investment comes under scrutiny
Fortune reported that some employees have used AI for minor tasks, including checking the weather, after being urged to use the technology more often. That kind of usage can add to bills without necessarily improving productivity or revenue.
Many businesses still have not identified where AI produces meaningful returns, according to Fortune. Axios reported that one executive said workers tend to automate tasks they dislike rather than the work that creates the most business value.
The cost review is unfolding as AI remains a major topic at VivaTech in Paris, where Fortune reported that executives were also discussing European concerns over “sovereign AI.” Fortune said those discussions followed the U.S. government’s move last week to cut off foreign access to Anthropic’s powerful Mythos-tier models, a decision that highlighted Europe’s dependence on American AI providers.
For corporate buyers, the immediate issue is more practical: who can use AI, which models they can use and how much those tools should cost. Fortune reported that companies are now trying to ensure AI projects produce business value without letting experimentation overwhelm budgets.
This story draws on original reporting from Fortune.