In an unsurprising move given the rhetoric we’ve seen emerge from Uber recently, the company has cut back on its AI spending, particularly in terms of employee tech usage. While the tech world has rushed to embrace AI in an attempt to lead the pack and capitalize on its advantages, many companies are now experiencing firsthand the exorbitant costs that come with unrestricted internal use of artificial intelligence. Uber is now regulating the use of AI coding tools like Anthropic’s Claude Code or Cursor by placing a monthly $1,500 cap per employee and per agentic coding tool.
Some employees may be allowed to exceed these caps with permission for those higher up in the chain, however, the standardized limit is expected to help Uber plan its AI budget with greater ease. More organizations are likely to follow suit, painting a confusing picture for employees who have thus far been pushed to adopt AI capabilities into every daily task.
After speeding through its 2026 AI budget, Uber is regulating employee AI spending with a $1,500 cap per employee per agentic coding tool. (Image: Pexels)
After Burning Through Its AI Budget, Uber Is Curbing Spending By Setting a Cap on Employee Usage
Just recently, Uber President and COO Andrew Macdonald revealed that the company was revisiting its usage of Claude Code and balancing just how much they could continue to invest towards building useful features that better served its customers. This discourse on the Rapid Response podcast came after the company reported that it had blown through its entire 2026 AI coding tools budget in just four months. “If you’re not actually able to draw a direct line to how [many] useful features and functionality you’re shipping to your users, that trade becomes harder to justify,” Macdonald explained.
Uber’s AI spending reports don’t come as a surprise, as the company has fully embraced the technology and its utility in system operations. During an earnings call earlier this month, company CEO Dara Khosrowshahi reported that approximately 10% of its code is now built by autonomous AI agents. The adoption of AI doesn’t stop there. “We’re seeing uptake of these tools, whether it’s our legal team or marketing team or developers,” the CEO explained, showcasing the technology in a now-familiar positive light. “We think it’s creating employees with superpowers.”
Superpowers or Not, Uber Is Placing Some Restrictions on the Use of AI Coding Tools
As previously stated, Uber is tempering its AI spending by setting a $1,500 monthly cap per employee for each agentic coding tool. As with most operations, employee AI usage can be tracked via a dashboard that details token usage and other aspects of employee AI adoption, which should now be monitored more closely to ensure that workers do not exceed their limit. Across organizations, the use of such dashboards has recently been controversial in its own right. With employees urged to use AI as much as possible in their day-to-day tasks, workers have recently been encouraged to become competitive about their AI usage, battling it out to outperform their colleagues.
This leaderboard approach has not resulted in a direct and immediate boost to productivity or any measurable evidence of improved results, as increased use does not always equate to quality use. Where once employees had to justify spending company money, many organizations have found themselves overcommitting resources towards AI now, ignoring just how shortsighted the strategy could be. Now, businesses continue waiting in hopes that this AI spending pays off.
Reports indicate that the price and expenses associated with AI could decrease in the coming years once businesses establish the initial infrastructure and gain a better understanding of the areas that require further investments in the tech. But the waiting game, alongside rising costs, does have some employers rethinking this wanton adoption of artificial intelligence. While they wait for signs of quality ROI, we return to a message that we have often repeated: without guardrails and careful planning, no strategy can be as effective as intended.
Uber’s Conservative Approach to AI Tools Is Something More Businesses Are Showcasing
Uber isn’t alone in rethinking its AI spending and withdrawing some of its uncontrolled utilization of AI. Microsoft, for example, has reportedly begun cancelling most of its direct Claude Code licenses, urging its engineers to use GitHub Copilot CLI. This comes after the company, just six months ago, opened access to the Claude model and steered its employees towards experimenting with the tech. This back and forth on AI isn’t a sign of companies stepping back entirely on AI usage and denouncing the technology, but is instead a result of employers realizing that some restrictions are better enforced than not.
Increased expenses aren’t the only issue being caused by the use of AI. Earlier this year, Amazon called its employees together to discuss increased supervision of AI tools after issues with its website began popping up. Just this week, Instagram’s AI chatbot reportedly gave hackers access to other user accounts with astonishing ease. While Meta claims the vulnerability has been addressed, there appear to be increasing signs that organizations are unprepared to navigate the technology and have chosen to rush headlong into adoption with little planning or preparation for the risks that come with it.
Reaching the “finish” line and beating out competitors on AI use may sound appealing to some. However, without guardrails, tests, and increased monitoring, businesses are now at risk of spending far more on the tech, whether in money or reputation, than they would by preparing their workforce and planning ahead for change.
Subscribe to The HR Digest for more insights on workplace trends, layoffs, and what to expect with the advent of AI.
