Uber blew through a year's AI token budget in months; Intel VP says 80% of workloads could run on cheaper tokens

ryanshrout · x · 2026-09-26

Anil Nanduri, VP of AI Products and GTM for Intel's AI Data Center, says Uber blew past a year's AI token budget in a matter of months — and that such overruns are becoming the norm for enterprise AI spend.

A useful data point on enterprise AI cost structures: token consumption is growing far faster than budgets, and billing-model mismatches may be wasting significant money.

Original post →

More from Infra

Infra channel →