Uber Technologies Inc. exhausted its entire 2026 artificial intelligence budget within four months and has begun capping employee use of AI coding tools [1, 2].
The spending surge highlights the volatility of enterprise AI costs and the difficulty companies face when scaling token-based pricing models across global operations.
Company-wide operations depleted the budget by April 2026 [3, 5]. The rapid spend followed heavy encouragement for staff to integrate AI into their workflows, specifically utilizing tools like Claude Code [2]. These token-pricing models eventually broke the financial assumptions made by the company's enterprise finance teams [2, 4].
AI integration has become deeply embedded in the company's technical infrastructure. Approximately 95 percent of Uber engineers now use AI tools on a daily basis [7], and AI has generated at least 10 percent of the company's code [1].
This financial pressure coincides with other corporate restructuring. Uber recently fired 23 percent of its HR staff [7]. While some reports suggest the exhausted AI budget created financial pressure that could drive such cost-cutting measures [2], the company's new president said that AI did not play a role in the HR cuts [7].
Uber is now implementing limits on how much employees can use these tools to prevent further budget overruns [1, 3]. The situation reflects a broader trend where companies like Microsoft and Nvidia have also faced challenges with escalating AI budgets [5].
“Uber exhausted its entire 2026 AI budget within four months”
Uber's budget collapse illustrates the 'hidden cost' of AI productivity. While AI tools significantly accelerate software development, the variable cost of tokens can scale faster than the efficiency gains they provide. This may lead other large enterprises to move away from open-ended AI tool access toward more rigid, quota-based systems or the development of proprietary, fixed-cost models to ensure financial predictability.
