PreviewYou're previewing CFO AI Ledger. Sign in to use it.Sign in
← All lessons
L3 · Controlling AI spend and finance data readiness8 min read

FinOps for AI: running AI spend like a utility

FinOps is the practice of managing variable technology spend (originally cloud) with shared accountability between finance, engineering and the business. AI spend now sits squarely in its scope.

Adoption is fast

The FinOps Foundation's State of FinOps 2026 report (1,192 respondents representing more than $83 billion in annual cloud spend) found 98% of respondents now manage AI spend, up from 63% in 2025 and 31% in 2024, and named AI cost management the top skill teams need to develop.

The FinOps loop applied to AI

  1. Inform: see the spend. Tag every AI cost to a team, product and use case; include seats, API usage, cloud GPU and AI features in SaaS.
  2. Optimize: right-size. Reclaim idle seats, route simple tasks to cheaper models, cache repeated prompts, negotiate commitments.
  3. Operate: govern. Budgets per team, caps, alerts, monthly reviews, unit-cost targets.

Roles

  • Finance owns budgets, forecasts and the review cadence.
  • IT / platform owns tagging, tooling, caps and model routing.
  • Business owners own usage and the value it produces.

Quick wins

  • An AI spend register across AP, cloud bills, SaaS and expense claims.
  • A quarterly seat reclaim.
  • Spend caps on every API key.
  • A single monthly AI cost page for the leadership team.

Watch for

AI usage often starts on corporate cards and expense claims. Esker's 2026 survey found 75% of CFOs report increased AI software and subscription costs, versus 45% of other finance leaders: the CFO sees the bill first.

What this means for the CFO

Ask your CIO or FinOps lead for one AI cost view covering seats, APIs, cloud and AI features in SaaS, tagged by team and use case, refreshed monthly. Load it into the spend tracker until a platform exists.

Then set a cadence: monthly spend and spike review, quarterly seat reclaim, annual renegotiation. That rhythm, not a one-off cost-cutting exercise, keeps AI spend in line with value.

Expect the first view to be incomplete. Each month, reconcile it against AP and card spend to find AI costs that slipped through, and tag them back to an owner.

Real example

The FinOps Foundation's State of FinOps 2026 report (1,192 respondents representing more than $83 billion in annual cloud spend) found that 98% now manage AI spend, up from 63% in 2025 and 31% in 2024, and identified AI cost management as the number-one skillset teams need to develop. Source: FinOps Foundation — State of FinOps 2026 ↗

Use it

Quick check · 3 questions

  1. 1.What are the three phases of the FinOps loop?

  2. 2.Who should own usage of a consumption-priced AI tool?

  3. 3.Which is a FinOps quick win for AI?

0 of 3 answered

Sources

Last reviewed 2026-10-01.