Corporate executives are starting to question whether AI spending is delivering real results even as it rises sharply, Axios reported on May 28.

Microsoft cancelled most of its Claude Code licences, citing cost issues, the report said. Uber's COO said AI costs were becoming "increasingly difficult to justify". An AI consultant said one client spent $500 million in a month because it did not set usage limits on employee Claude licences.

Anuj Kapur (아누즈 카푸르), CEO of CloudBees, said layoffs could be "the only way" to offset AI costs.

Ali Ansari (알리 안사리), CEO of model-training company Micro1, said a "healthy correction" was under way away from AI overuse, or 'token-maxing' that consumes as many tokens as possible. He said AI currently works in a practical sense only in coding.

Four obstacles to corporate AI adoption are cited: use cases, cost, people and data.

Sophia Velastegui (소피아 벨라스테기), CEO of Velastegui Ventures and a former chief AI officer at Microsoft, said "most employees choose to automate tasks they dislike first rather than the work that is most valuable to the company."

On costs, one CTO said employees were using AI to check the weather. Josh Pantony (조시 팬토니), CEO of Boosted.ai, said the more companies are reluctant to allow AI agents access to internal data, the worse the agents perform.