As artificial intelligence becomes more widespread in enterprises, high operational costs are keeping some tech giants清醒. Uber's chairman and chief operating officer, Andrew McDonald, recently publicly stated that the company has already exhausted its entire annual artificial intelligence budget within just four months of 2026.

Faced with such an astonishing rate of spending, the company's management has begun to openly question the reasonableness of this massive investment. Currently, there is no evidence within the company that the high cost of AI tools brings equivalent returns for the business.

High Costs Fail to Bring Equivalent Returns

McDonald admitted in an interview that the usage of Anthropic's AI programming tool, Claude Code, within the company has seen a dramatic increase in token usage. However, this surge in underlying data has not directly translated into more practical and valuable features for end users.

He emphasized that there is currently no clear connection between the current investment and output. Although it may seem that the number of product launches has increased, it is difficult to directly equate these cold numbers with actual improvements in business capabilities.

Reducing Staff to Bear AI Costs

To support the rising AI-related expenses, Uber even adopted an aggressive strategy earlier this year, reducing the scale of staff recruitment. CEO Dara Khosrowshahi previously confirmed that the company is controlling staffing levels to free up funds to cover this significant technology expense.

This phenomenon of "sacrificing human resources for AI" has led to deep considerations among the management. McDonald clearly stated that if the cost of AI tokens remains unclear in terms of their ability to translate into practical services delivered to users in the coming quarters, the trade-off between costs and staffing will be hard to justify.