Meta: AI token costs could soon equal engineer salary
meta's adam mosseri predicts ai token costs per engineer could match their salary within two years, forcing strict budget caps that will limit internal experimentation.
meta exec adam mosseri warns that individual ai token spend could hit salary parity in 1-2 years, necessitating caps proportional to an engineer's roi. the prediction follows cost control moves like canceling microsoft's claude code licenses and uber depleting its 2026 budget early. mosseri compared token management to allocating gpus or payroll, stating meta stopped 'silly' valueless uses like internal leaderboards. this indicates the 'free for all' era of enterprise ai experimentation is ending as companies move toward financially sustainable deployment.