Executives at Uber, Meta, Microsoft, and other companies are trying to rein in "tokenmaxxing" by employees, which led to ballooning AI use costs

Corporate AI spending is spiraling out of control. Some companies have blown through their entire annual AI budgets in just months, while others watched their AI bills double or triple. The culprit has a name: "tokenmaxxing" — the unchecked consumption of AI compute tokens across organizations.

The spending shock is forcing a reckoning. Executives are now scrambling to implement rationing systems and usage tracking as the massive computing costs hit the balance sheet. The era of letting employees freely experiment with AI tools is giving way to hard spending controls.

The core problem is straightforward. Companies rushed to deploy AI across their operations without building adequate cost monitoring. Now the bills are arriving, and leadership is demanding to see actual returns on these ballooning investments. The honeymoon phase between enterprises and generative AI just got a serious reality check.