Why Amazon, Uber, and Meta Are Quietly Curbing Their AI Use (Hint: It's Not About Performance)

Corporate AI adoption is hitting a sobering reality check as major companies like Amazon, Walmart, Cisco, Uber, and Meta curb employee AI usage to control exploding costs. The shift from flat-rate subscriptions to token-based billing by labs like OpenAI and Anthropic has exposed businesses to per-query expenses, especially as workers move from simple chatbots to compute-heavy AI agents. Uber's president noted it's becoming "harder to justify" token outlays without clear ROI, while Deloitte's global AI lead warned that CFOs and boards are now fixated on compute costs. Sam Altman admitted cost emerged as a "huge issue" for customers this year, contrasting with last year's indifference.

Efforts to cap usage—encouraging cheaper models and discouraging wasteful prompts—could weigh on the growth of AI labs planning near-trillion-dollar IPOs. Meanwhile, China has gained a new advantage: cheaper energy and more efficient models have made Chinese AI labs more competitive on token pricing, overtaking US counterparts in token consumption since the start of the year according to OpenRouter data. This cost crunch may slow enterprise AI deployment and force a strategic re-evaluation of which tasks justify the compute expense, potentially reshaping the competitive landscape between US and Chinese AI providers.