Gartner projects that AI agent inference costs will continue to rise sharply through 2028 due to escalating model complexity and token consumption.

Gartner projects that AI agent inference costs will continue to rise sharply through 2028 due to escalating model complexity and token consumption.

Rising inference economics will force enterprise buyers to prioritize cost-efficient GPU clouds and specialized accelerators, directly impacting utilization rates for neocloud operators.