Tech companies that rushed to put AI tools into the hands of every employee are now hitting the brakes. The reason is not that the technology stopped working but because huge bills are starting to arrive. Amazon, Walmart, Cisco, Uber, and Meta are among early corporate adopters now introducing usage caps, discouraging unnecessary AI use, or pushing employees toward cheaper models to bring AI spending under control, according to a report by the Financial Times.
From Chatbots to AI Agents and a Bigger Bill
Since the launch of ChatGPT, the shift today is dramatic as workers have moved beyond simple chatbot queries toward AI agents—systems capable of performing complex, multi-step tasks autonomously with limited human supervision. Agents are far more useful than chatbots in many workflows, but they also consume significantly more computing power, and that difference shows up directly on the invoice. The cost to run agents has increased because companies like Anthropic and OpenAI have shifted some services from flat monthly subscriptions to token-based billing. This pricing model charges based on the actual units of data processed by the AI model. Every prompt, every automated workflow, and every agent task now has a direct and trackable cost attached to it, in a way flat subscriptions never made visible.
What Top Tech Executives Said About Rising AI Costs
OpenAI CEO Sam Altman acknowledged the shift, saying earlier this month that cost has become a “huge issue” for customers this year. “The issue never came up [last year]. People were totally happy with the amount they were spending,” he said. However, few companies have a problem as stark as Uber. Andrew Macdonald, Uber’s president and chief operating officer, said it has become “harder to justify” the company’s spending on AI tokens. “It's very hard to draw a line between one of those stats and 'OK now we're actually producing like 25 percent more useful consumer features,'” he said on a recent podcast. Uber blew through its entire 2026 AI budget by April—just four months into the year.
Walmart, Cisco, and the Agent Math Problem
Walmart has taken similar steps, capping the number of tokens employees can use on its internal AI agent system. Suresh Kumar, Walmart's global chief technology officer, said usage of the company's internal “Code Puppy” vibe-coding platform had “really skyrocketed.” “This is now an opportunity for us to take a step back,” he said, adding that employees are now being asked to identify the right tool for each specific task. Jeetu Patel, Cisco president and chief product officer, explained that the computing infrastructure required for AI agents is far greater than what chatbots ever needed. “The amount of infrastructure needed for an agent is meaningfully higher than for a chatbot. For every human you might have 10, 100 or on the aggressive side 1,000 agents... They just keep working and that consumes a chunk of [compute],” Patel said.
Amazon and Meta Crack Down on ‘AI for the Sake of AI’
Larger companies are tackling the problem from a different angle: changing internal incentives and messaging around how AI gets used in the first place. Amazon warned employees last month to stop using “AI just for the sake of using AI,” after engineers were found deploying agents simply to climb internal leaderboards tracking AI adoption—a metric that had clearly stopped reflecting genuine usefulness. Meta introduced similar measures in April.
