Cloud cost management in 2026: FinOps now covers AI spend too
Cloud cost management used to mean finding idle servers. According to the FinOps Foundation's 2026 survey, almost every FinOps team now manages AI spend as well.
Gartner forecast in November 2024 that worldwide public cloud spending would reach $723.4 billion in 2025, up 21.5% from the year before. Bills that grow at that pace get attention from finance, and the discipline that grew up to manage them is called FinOps: finance, engineering and business teams working together to get value from cloud spending.
What changed in 2026
The FinOps Foundation's State of FinOps 2026 survey shows the job widening. According to a summary by Kion, 98% of respondents now manage AI spend in some form. More than half plan to manage software licensing and private cloud costs within the next 12 months, and SaaS spending has become a standard part of the remit.
The top priorities reported were managing AI costs, using AI to make FinOps work itself more efficient, governance at scale, extending FinOps beyond cloud, and aligning the organization around it.
AI changes the cost problem in a specific way. Model and GPU charges scale with usage that is hard to predict, such as tokens processed or inference requests, rather than with the number of servers someone remembered to switch off.
What a cost management practice needs
| Capability | Why it matters |
|---|---|
| Cost allocation | Every dollar mapped to a team, product or customer, usually through tags |
| Budgets and anomaly alerts | Overruns caught in days, not at month end |
| Rightsizing recommendations | Instances and databases sized to actual use |
| Commitment management | Reserved instances and savings plans bought and tracked against usage |
| Unit economics | Cost per customer, per transaction or per AI request, so growth and efficiency can be told apart |
| Coverage beyond one cloud | AWS, Azure, Google Cloud, SaaS and AI services in one view |
Native tools or a platform?
AWS Cost Explorer, Azure Cost Management and Google Cloud's billing reports are free and good enough for a single cloud with clean tagging. Third-party platforms earn their cost when you run more than one cloud, need allocation across complex shared services, or want AI and SaaS spending in the same view.
Whatever the tool, the results come from habits: tagging rules enforced at deployment, a monthly review with each team's owner, and someone accountable for acting on recommendations.
What IT teams should do
- Check what share of last month's cloud bill is allocated to a named team or product.
- Set budget alerts and anomaly detection on every production account.
- Add AI services and model APIs to your cost reports, with a cost per request where possible.
- Hold a monthly cost review with the owners of the five largest spending areas.