Cloud reports measure the wrong thing
InfoWorld ·
Flexera’s 2026 State of the Cloud Report agrees with many other cloud reports these days: Cloud is more strategic, cloud is more complex, AI makes everything harder, and better governance is the answer. This report, based on a survey of 753 cloud decision-makers and users, argues that enterprise cloud has entered a “value era” in which organizations are measuring success less by cost savings alone and more by business outcomes. The headline findings are familiar. Flexera says 64% of organizations now use “value delivered to business units” as a metric for cloud progress, up 12 percentage points year over year. Use of unit economics rose from 40% to 49%. Generative AI usage as a public cloud service rose to 58%, making it one of the most widely used public cloud services in the survey. Hybrid cloud remains dominant at 73%. Estimated wasted IaaS and PaaS cloud spend ticked up to 29%, reversing a five-year decline. Meanwhile, 71% of respondents say their company has a cloud center of excellence or a similar team, and 63% say they have a finops team. These reports provide a decent temperature check of what cloud teams are thinking about. The problem is what happens next. A survey of cloud-involved people becomes a story about the market. A set of self-reported answers becomes evidence of maturity. A vendor-funded report becomes a road map. Then CIOs, analysts, vendors, and conference speakers all repeat the same narrative until it sounds like reality. It is not always reality. The findings are not the problem I agree with Flexera that cloud complexity is increasing. I agree that AI is making cloud costs harder to predict. I agree that many enterprises are adding finops teams, governance groups, and centralized cloud functions. The issue is interpretation. The report frames much of this as cloud maturing into a more value-driven discipline. That is partly true, but it is also the optimistic version of the story. The less flattering version is that enterprises are still paying for architecture and workload placement decisions made years ago, and they are now adding governance structures to manage the consequences. Hybrid cloud is a good example. A hybrid cloud adoption figure of 73% sounds like a strategic evolution. In practice, hybrid estates often exist because companies acquired other companies, inherited private infrastructure, lifted and shifted the wrong applications, or discovered that public cloud economics did not work for every workload. Flexera acknowledges that multicloud often happens through mergers, siloed teams, or inherited architectures rather than deliberate strategy. That caveat should be in 40-point type. Accidental hybrid cloud is not a strategy. It is usually technical debt with a better name. Governance won’t fix bad architecture One recurring mistake in these reports is assuming governance can solve problems architecture created. Finops can find waste. A cloud center of excellence can set policies. Better dashboards can show who is spending what. Commitment discounts can reduce bills. All of that matters. But none of it fixes a badly designed system. If an application is too chatty across regions, governance will not save it. If data is in the wrong place, a dashboard will only show you the cost of that mistake. If teams built a distributed architecture because it looked modern rather than because the workload required it, finops will be stuck optimizing a bad design. If a steady, high-utilization workload was moved to public cloud without a serious total cost of ownership analysis, cost reporting will not make the economics work. This is where many vendor and analyst reports fall short. They describe the operational pain accurately, then prescribe more management, more visibility, and more tools. That is understandable. Many of these reports are funded by vendors or analysts whose business models depend heavily on vendor relationships. The answer often maps suspiciously well to what the sponsor sells. The value story is mostly a cost story The “value era” framing is also worth challenging. Flexera says cloud success is increasingly measured by value, not just savings. Yet cost efficiency and savings remain the top cloud progress metric at 81%, and optimizing cloud costs remains the top cloud initiative at 68%. That is not a replacement of cost by value. That is cost control with better language. There is nothing wrong with measuring business value. In fact, it is long overdue. But saying “value delivered to business units” in a survey is not the same thing as proving value. Most enterprises still struggle to connect cloud spend to product margin, customer retention, revenue growth, cycle-time reduction, risk reduction, or measurable operational improvement. Unit economics can help, but only if the units are real. Cost per transaction, cost per customer, cost per claim, cost per inference, or cost per order can be useful. A vague internal allocation metric that nobody uses to change behavior is accounting theater, not unit economics. AI will expose weak spots The AI findings are probably the most important part of the report, but even there, the story should be read with more skepticism. Flexera says generative AI public cloud services are now used by 58% of respondents, and all respondents use generative AI in some capacity. That sounds dramatic, but “using generative AI” can mean many things. It could mean a production AI product or a developer assistant or an experiment with a cloud API. It could mean a pilot that will never survive procurement review. AI will make the cloud cost problem worse. GPU resources, inference costs, data movement, storage growth, model monitoring, vector databases , and duplicated experimentation environments will create bills that many enterprises do not yet understand. The old cloud problem was that teams could provision infrastructure with a credit card. The new problem is that teams can provision very expensive AI services before finance, architecture, security, and data governance know what happened. Reports that say “AI needs governance” are correct. They are also incomplete. AI needs architectural discipline. It needs workload placement discipline. Some AI workloads belong in public cloud. Some may belong with specialized providers. Some may belong on private infrastructure. Some should not exist at all because the business case is weak. Survey data is not market truth The report’s methodology matters. Flexera surveyed technical professionals and executive leaders from an independent panel. Most respondents are deeply involved in cloud decisions, and a large share work in central cloud teams or similar functions. That is a useful audience, but it is not neutral. People inside cloud governance structures are more likely to see cloud governance as the answer. People buying or managing cloud services are more likely to report cloud maturity. People close to finops are more likely to describe the world through finops language. This is a broader issue across the cloud industry. We are drowning in reports. Vendor reports. Vendor-funded analyst reports. Analyst reports that are not technically vendor-funded but depend on vendor ecosystems. Surveys that ask cloud people whether cloud is important. AI reports sponsored by AI vendors. Security reports sponsored by security vendors. Finops reports sponsored by finops vendors. Again, bias does not make the reports worthless. It means they should be read as inputs, not absolute truth.