Set the upper limit to the AI spend because the budget was spent in Uberguaca months

AI is becoming expensive and some companies are reducing its use in an attempt to moderate costs. This cohort includes Uber, which recently instituted internal usage caps to curb its exorbitant AI spending.

Bloomberg reports that the company has instituted a new rule that imposes a monthly cap of $1,500 per employee per agent coding tool, including Claude's Anthropic code or cursor. Usage is trackable via an internal dashboard that every employee has access to, although in some cases caps can be exceeded with permission, the company says.

The news is perhaps not very surprising since in April the company's CTO revealed that the ride-hailing giant had spent its entire annual AI budget in four months. This appears to have happened after Uber encouraged staff to use AI "as much as possible" and even ranked its internal use competitively in internal rankings, The Information previously reported.

Uber COO Andrew Macdonald also recently questioned AI's impact on productivity, noting during a podcast appearance that "it's very difficult to draw a line" between the use of AI and new consumer features.

Uber's cut raises a broader issue currently facing the tech industry: As companies invest money in AI, what exactly is the return on investment? Indeed, AI ROI has until now remained a largely theoretical phenomenon that everyone hopes will eventually materialize – even if some companies are visibly getting a little restless in the meantime.