FinOps Checklist 2026: It's No Longer Just the Cloud Bill

There is a question worth asking before any conversation about savings: what counts as technology spend under management today? The answer, for most organisations, has changed over the last two years without anyone writing it down.

Is FinOps still just about cloud?

No. And the reference body for the discipline changed its own mission as a result.

The financial management discipline for technology, known as FinOps, was built for one specific thing: bringing the runaway public cloud bill under control. For six years, that was it. The State of FinOps 2026, the sixth annual survey from the FinOps Foundation, a Linux Foundation project, shows that it no longer is. The Foundation changed its stated mission from "managing the value of cloud" to "managing the value of technology," and the survey, with 1,192 respondents representing more than 83 billion dollars in annual spend, explains why.

What changed, in numbers

The series is the most citable part of the report, and it is worth seeing in full before anything else.

98% of practitioners already manage artificial intelligence spend, up from 31% two years ago. 90% manage software as a service, up from 65% the year before. 64% manage licensing, up from 49%. 57% manage private cloud, up from 39%. 48% manage data center spend. And 28% already manage, or plan to manage, labour costs tied to running technology.

Notice what the series describes. It is not a discipline that decided to expand out of ambition. It is a set of costs that already existed, that always existed, and that is only now entering formal management, one line at a time, under the same pressure that pulled the cloud bill under control a decade ago.

The scope checklist

This is the part meant for immediate use. Before asking where to save, it is worth confirming what is actually being managed.

• Public cloud. The historical starting point. Almost always with an owner, a dashboard, and a monthly review.
• Software as a service. 90% of organisations already manage it formally. Licenses by seat, by module, by consumption, often bought by different departments with no cross-visibility.
• Traditional licensing. 64% already cover it. Multi-year contracts, compliance audits, and the old habit of negotiating once and never revisiting.
• Private cloud and hybrid infrastructure. 57% already include it in scope. Reserved-capacity cost that does not follow the elastic logic of public cloud.
• Owned data centers. 48% already manage it as a formal line, not as an untouchable fixed cost.
• Artificial intelligence. 98% already manage it, the line that grew fastest in two years, with the least predictable cost pattern of the whole list.
• Labour costs of the operation. 28% already manage or plan to manage this line, the most recent to enter scope.

Each line in this list typically has a different owner, a different billing cycle, and lands with different departments in different months. That dispersion is exactly what explains the next point.

Why spend outside cloud slips through

The explanation is not a lack of will. It is structural asymmetry.

The public cloud bill almost always has a named owner, a dedicated dashboard, and a formal monthly review. It is the most closely watched line in the entire technology operation, because it was the first to earn that level of scrutiny. The other six lines on the checklist do not have that same discipline built around them. They arrive through separate contracts, approved at different times, against different department budgets. Each one, seen in isolation, looks small next to the cloud bill. Added together, they no longer are.

The practical effect is a partial, unintentional blindness: the organisation knows exactly how much it spends on the one line it has always watched, and has only a rough sense of the rest.

Who decides, and why that changed

The practice has moved up the hierarchy, and that changes who has to ask this question. Where there is VP-level or C-level involvement, real influence rises sharply: cloud service selection, 53% versus 24% without that involvement; vendor selection, 47% versus 16%; the decision between public cloud and an owned data center, 28% versus 12%.

The reading is direct. When FinOps was only about the cloud bill, it could live inside the infrastructure team, a technical matter with a financial echo. With the scope now extended to licensing, artificial intelligence, and data centers, it stops being something that gets resolved without senior leadership. Managing this full list is, increasingly, a decision made at the level of the CTO or the CIO, not below it.

What team size this requires

The wider scope does not require a proportionally larger team, and the number surprises anyone expecting the opposite.

Organisations spending more than 100 million dollars a year in technology operate, on average, with 8 to 10 internal practitioners and 3 to 10 contracted professionals. 60% organise in a centralised model, with a small core and embedded champions inside the teams that spend. 21% use a hub-and-spoke model. Fewer than 10% are fully decentralised.

The report's own explanation is direct: scale comes from enablement and automation, not headcount. A small core with a clear process covers a much wider scope than a large team without that process.

Where to start without stopping the operation

No new tool is required to begin. Three decisions are, in the right order.

First, a written, declared scope: the checklist above, confirmed line by line against the organisation's actual contracts and cost centres, not against a generic list. Second, a named owner for every confirmed line, with an actual name, not a team name. Third, the same review cadence for every line, including the ones that have none today. Cloud stops being the only line audited every month, and becomes one among several, all treated the same way.

None of these three decisions requires a new platform. They require a design decision, made before anyone argues about where to cut.

When this is not worth the effort

The exception is worth stating, because not every organisation needs this full exercise.

If technology spend outside the public cloud is residual, concentrated in a single contract that has already been reviewed and shows no sign of growth, building a formal expanded-scope process solves a problem that does not yet exist at that scale. In that case, the priority stays the discipline already applied to the cloud bill itself. The right distinction is not company size. It is whether three or more lines outside cloud are already growing without a named owner. If so, the exercise pays off. If not, it can wait.

Frequently asked questions

Does FinOps apply to traditional software licensing? Yes. 64% of FinOps practitioners already manage licensing formally within the same process they use for cloud, according to the State of FinOps 2026.

Do we need to buy a new tool to expand scope? That is not the first step. The first step is the written, declared scope and a named owner per line. An aggregation tool helps afterwards, once the process already exists.

Who should own each spend line? One person, with an actual name, not a team. The report shows the decision increasingly rises to the CTO or the CIO as scope grows, but line-level execution still needs a single operational owner.

What is the minimum recommended review cadence? The same one already applied to public cloud, typically monthly. The core rule of this article is that every line in scope shares the same cadence, with cloud no longer the only audited exception.

Can a small team cover this expanded scope? It can, and it is the observed pattern. Organisations spending more than 100 million dollars a year operate with core teams of 8 to 10 internal practitioners, because scale comes from process and automation, not headcount.

What to do with this

The question this article opened with now has a more precise answer than "cloud." It has seven possible lines, a checklist to confirm which apply, and an explanation of why the missing ones cost more than they appear to while they have no owner.

Doing the full inventory, line by line, usually takes less time than expected, and the result is almost always uncomfortable the first time. Book a Clarity Session, no commitment, to see how many of these lines already exist in the operation without a named owner.

On the most common waste patterns inside the cloud bill itself, it is also worth reading the five patterns still destroying the IT budget, and on the scope line growing fastest, why 80% of cloud AI cost forecasts miss.