"There's no perfect way to forecast AI spend" - Ensono's CFO on governing costs when vendors keep changing the unit
Token retrenchment stories are prolific at the moment. Uber blew its 2026 AI budget in four months and now caps engineers at $1,500 per tool per month. Accenture is trying to stop consultants from converting PDFs with frontier models, and last month launched its own "Tokenomics" service to help other enterprises avoid the same fate. Citi has pushed staff to cheaper models. Walmart has capped its in-house agent, Code Puppy. Amazon scrapped the leaderboard that ranked staff by AI consumption after – predictably – it was gamed.
A lot of coverage (when it isn't breathless and hysterical) treats this as a governance problem. Companies encouraged unbounded AI use, employees took them at their word, and now finance is being handed the bill. The fix, in the standard telling, is closer alignment between the CFO and the Chief AI Officer, better forecasting, and cost discipline modeled on FinOps.
It's a wobbly...
Copyright of this story solely belongs to diginomica.com. To see the full text click HERE
