Six Years in, Cloud Waste Is Still the Number One Job

"First they asked us to fix cloud. Then fix the software mess. Now it’s fix the contract and license mess, now fix the data center."

That is a practitioner quoted in the FinOps Foundation’s State of FinOps 2026 report, and it is a fair summary of what has happened to the discipline. The survey behind it collected responses from 1,192 practitioners who together account for more than $83 billion in annual cloud spend, spread across EMEA, North America, Asia Pacific and South and Central America.

After six annual editions of tooling, certification, dashboards and vendor consolidation, the single top priority reported by those practitioners is still workload optimisation and waste reduction. The same thing it was in 2020.

That deserves a moment of honesty. If the top problem has not shifted in six years, the approach to it probably needs to.

Why Waste Keeps Coming Back

Why does the same problem survive six years of sustained attention? Because waste is not a one-time cleanup. It is a byproduct of how software gets built. Every sprint creates new resources and only some sprints remove old ones. An environment spun up for a load test in March is still running in September because nobody remembers whose it was, and the person who created it moved teams in May.

Dashboards make this visible. They do not make it stop. The organisations that genuinely bend the curve move the decision left, into the moment infrastructure gets defined rather than the moment the bill arrives. In practice that means budget limits enforced in the pipeline, mandatory ownership tags that fail a deployment when they are missing, and automatic expiry on anything created outside production.

Firms that bring in Cloud Engineering Services for this reason are usually not buying another cost report. They are buying the guardrails that make the report boring.

AI Arrived Faster Than Anyone Planned For

Here is the most striking movement in the 2026 data. 98 percent of respondents now manage AI spend. Two years ago that figure was 31 percent.

Have you seen a cost category move that fast before? A line item that barely existed is now near-universal, FinOps for AI is the top forward-looking priority for the coming twelve months, and AI cost management is the skill practitioners most want to build.

AI spend also behaves differently from ordinary compute. It is bursty, it is often billed per token or per request rather than per hour, and the people generating the cost are frequently not the people who understand the pricing model. So the familiar playbook of reserved instances and rightsizing does not map cleanly onto it.

The Remit Keeps Widening

The survey shows scope expanding well past public cloud. 90 percent of respondents now manage SaaS or plan to, up 25 points. 64 percent manage licensing. 57 percent manage private cloud and 48 percent manage data centre spend. 28 percent have been handed labour costs as well.

Is that scope creep or seniority? A bit of both. 78 percent of these teams now report to a CTO or CIO, up 18 points against 2023, and teams with VP-level engagement carry two to four times more influence over technology selection.

Where to Start If You Are Behind

Pick your three largest cost centres and find out who owns each one by name. Then make one change in the pipeline instead of one more change in a dashboard. Ownership tags enforced at deploy time will do more for next quarter’s bill than any number of weekly cost reviews.

And if your AI spend is still landing on a general compute line, split it out now. You will want the history when someone asks why it tripled.

About Author

Shefali Vasave

Shefali manages content marketing at Opus Technologies, a domain-native engineering partner for banks, payment providers, and fintechs, and writes on the various aspects of financial institutions navigating change in a real-time, digital-first world.