The era of wild AI spending has given way to token rationing. After early encouragement to max out budgets and even building internal leaderboards for AI usage, companies are now realizing the cost of this tech is harder to predict than they thought.

Consulting firm Accenture recently attempted to stop its employees from using AI for basic tasks like converting PDFs into presentation slides. This move comes on the heels of a recent internal meeting where Justice Kwak, Accenture’s agentic AI strategy lead, stated that leadership is questioning the value received from their AI spending.

The cost of tokens has cast doubt over the viability of the entire AI business model. Companies are now looking for ways to cut costs and ensure they are getting tangible returns on their investments. The recent “AI selloff” among memory chip makers is just one example of how this shift in perspective is impacting various industries.

As AI spending becomes more unpredictable, companies are finding it harder to justify the cost without clear benefits. This reflects a broader realization that while AI may be exciting and innovative, its value must now be proven beyond mere novelty.