Skip to the cards
Filed
Source
Robert Davis
Loans
593

Flexera’s 2026 Cloud Survey: Value Metrics Rise, but Cost Control Still Leads

Flexera reports wider use of value metrics and generative AI, while cloud savings remain the top priority and waste rises to 29%.

Flexera’s 2026 Cloud Survey: Value Metrics Rise, but Cost Control Still Leads

Cloud teams are adding business value measures, but the latest Flexera survey suggests cost control still sets the agenda. Its 2026 State of the Cloud Report, based on responses from 753 cloud decision-makers and users, describes a shift toward measuring business outcomes. Yet cost efficiency and savings remain the leading measure of cloud progress, and estimated waste in infrastructure and platform services has risen.

Flexera reports that 64% of organizations measure “value delivered to business units,” up 12 percentage points year over year. Use of unit economics increased from 40% to 49%. At the same time, 81% still identify cost efficiency and savings as their top cloud progress metric, while 68% say optimizing cloud costs is their leading cloud initiative. Those figures make the “value era” a qualified description: organizations may be adding broader measures, but spending remains central to how they judge cloud performance.

What the survey measures

Several other findings capture the pressures facing cloud teams. Generative AI use through public cloud services reached 58% of respondents. Hybrid cloud remained common, at 73%. Estimated wasted infrastructure and platform cloud spend climbed to 29%, reversing a five-year decline. Flexera also found that 71% of respondents’ organizations have a cloud center of excellence or a similar team, while 63% have a FinOps team.

These results offer a useful view of what people close to cloud decisions report seeing. They don’t, by themselves, establish how every organization uses cloud or whether a particular governance model is working. A survey of cloud professionals is a view from within the discipline. It can describe priorities and perceptions, but readers should be careful about treating reported adoption or maturity as proof of business results.

Hybrid cloud can reflect history, not strategy

The 73% hybrid cloud figure can sound like evidence of a deliberate architecture choice. In practice, a hybrid estate may have grown through acquisitions, inherited private infrastructure, application migrations, or workloads whose economics didn’t suit public cloud. Flexera acknowledges that multicloud environments can arise from mergers, siloed teams, and inherited designs rather than an intentional plan. That context matters when IT leaders use adoption figures to compare their own environments with the wider market.

A mixed estate may be the right fit, but its presence alone doesn’t tell teams whether workloads are placed well or whether operating costs are understood. Some organizations are managing technical debt and historical decisions under the same label used for intentional architecture. The useful questions are more specific: which workloads run in each environment, what drove those placements, and what would it take to change them?

Governance helps, but architecture sets limits

FinOps teams can identify waste, cloud centers of excellence can establish standards, and dashboards can make spending easier to trace. Commitment discounts can also lower bills. Those practices give teams better control, but they can’t repair every design problem. An application that makes excessive cross-region requests, stores data in an unsuitable location, or uses a distributed design without a workload need may remain costly after reporting and governance improve.

The same applies when a steady, heavily used workload has moved to public cloud without a serious total cost of ownership assessment. Better cost visibility can show where money goes; it can’t make an unsuitable placement economical. IT teams should pair cost reviews with architecture reviews so they can distinguish spending that can be tuned from spending caused by a design or placement decision.

Reports that identify operational pain and recommend more visibility or management tools may still offer useful guidance. But readers should consider who funded a report and whether its recommendations align closely with the sponsor’s products. That doesn’t invalidate the findings. It does make it sensible to treat them as one input alongside internal usage, architecture, and financial data.

Test whether value measures change decisions

Measuring value delivered to business units is a reasonable goal, but a survey response isn’t evidence that an organization has connected cloud spend to results. Useful measures might relate costs to transactions, customers, claims, orders, or AI inferences. They become practical when teams use them to change design, capacity, service choices, or investment priorities.

A unit measure that doesn’t reflect how a product or service operates, or that nobody uses to guide decisions, risks becoming an allocation exercise without much operational value. IT and finance teams can check whether their chosen units are understood by the business and whether they show how cloud spending relates to outcomes such as product margin, customer retention, delivery time, or service performance. The measure should help people decide what to do next.

AI adds new cost and placement questions

Flexera’s AI figures deserve careful interpretation. The report says 58% use generative AI public cloud services and that all respondents use generative AI in some capacity. Those categories can cover very different activity: a production service, a developer assistant, a cloud API experiment, or a pilot that may not proceed. Adoption rates are more useful when teams understand what kinds of use they represent.

AI can add costs through specialized compute, inference, data movement, storage growth, model monitoring, vector databases, and duplicated experimental environments. An AI service may be easy for a team to try, while its ongoing cost and data requirements take longer to assess. Teams therefore need to understand who is using these services, what workloads they support, and how usage connects to a credible business case.

Governance is part of that work, but architecture and workload placement matter too. Different AI workloads may suit public cloud, specialized providers, or private infrastructure, depending on their needs. Some experiments may not justify continued spending. As adoption grows, teams should review costs alongside data flows, capacity requirements, and the purpose of each workload.

Use survey findings as a starting point

Flexera surveyed technical professionals and executive leaders from an independent panel. Many respondents were closely involved in cloud decisions, including people in central cloud teams. That makes their experience relevant, but it also shapes the view captured: respondents working in cloud governance or FinOps may be especially familiar with those approaches and their terminology.

Cloud surveys can help IT leaders identify questions to investigate. They’re less useful as a substitute for examining an organization’s own workloads, costs, and outcomes. For now, Flexera’s figures point to a clear tension: teams are adopting value measures and AI services, while cost efficiency remains their dominant benchmark and estimated waste has increased. A practical next step is to test whether each cloud measure leads to a decision, and whether architecture reviews explain the spending that dashboards expose.

Source: Robert Davis · www.infoworld.com

Filed alongside