Remember when your boss told you to use AI for everything? That party is officially over. Executives at Amazon, Walmart, Cisco, Uber, and Meta are now scrambling to rein in AI usage after costs spiraled far beyond projections. The recurring phrase across boardrooms? “We created a monster.”
Consider this: Uber’s entire 2026 AI budget was exhausted by April. Engineers were making individual API calls that cost thousands of dollars per person. That’s not a typo — thousands per engineer. And Uber isn’t alone. Multiple reports from June 2026 show that companies saw AI spending double or triple within the first quarter, forcing emergency budget freezes.
The culprit? Token-based billing models from OpenAI and Anthropic. As employees upped their AI usage — sometimes running complex reasoning agents that chain multiple thinking steps — token consumption exploded. The CBC even coined the term “tokenmaxxing” to describe the ballooning expense. Sam Altman himself acknowledged that sudden AI cost spikes have caught businesses off guard.
Now finance departments are fighting back. Companies are introducing usage caps, meter pricing, and internal chargeback systems to make AI costs visible. Business Insider notes this creates internal drama: employees who relied on AI tools are resisting limits, while finance teams push for stricter governance. The Telegraph describes “out-of-control employees” blowing AI budgets alarmingly fast.
This isn’t just a corporate headache. If enterprise customers rein in spending, the sky-high valuations of AI companies like OpenAI and Anthropic could come under pressure. Industry observers expect pricing model shifts — from per-token to flat-rate or capped plans — to retain enterprise clients. Meanwhile, China’s cheaper energy and more efficient models give it a pricing edge over US labs.
Here’s what the average worker should care about: the AI tools you’ve come to depend on might suddenly disappear behind paywalls or quotas. The era of unlimited AI experimentation is ending. The big question is whether companies can balance cost control with productivity gains — or whether the “monster” will force a brutal reckoning for the entire AI industry.