Flexera 2025 found 84% of organizations struggling to manage cloud spend and budgets overrun by 17%. FinOps stopped being an engineering function when the cloud bill became a CFO problem.
I was reading the Flexera 2025 State of the Cloud report a few weeks ago and I kept stopping at the same set of numbers (https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/). Eighty-four percent of organizations say managing cloud spend is their top challenge. Not security, not migration complexity, not the skills gap, cost. Organizations are exceeding their cloud budgets by an average of 17%. And 28% of respondents expect their cloud spend to grow further in the coming year. If you put those three numbers together, you get a picture of a cost category that is large, growing, consistently over budget, and widely acknowledged as the hardest thing in cloud operations to get right. That is not an engineering problem anymore. That is a financial governance problem.
FinOps did not start as a boardroom discipline. It started as a defensive practice. The name itself is a blend of finance and DevOps, and the early practitioners were often engineers who had been handed a cloud bill that was twice the estimate and told to explain it. The tools were basic: tag your resources, find the idle instances, buy reserved capacity instead of running everything on-demand. The audience was engineering managers and finance business partners. The goal was damage control before the next budget review.
What changed is that the bills got big enough to become a line item that executive teams and boards could not ignore. Cloud spend is now the second or third largest operating cost for many large organizations, after salaries and sometimes after real estate. When a cost category reaches that scale, the question of who owns it and how it is controlled becomes a governance question, not just an operational one. The 17% average budget overrun in the Flexera data is significant not only because of the dollar amount but because of what it signals about organizational systems: if you consistently overspend a major cost category by 17%, the systems for forecasting and controlling that category are not adequate. Boards are beginning to ask that question directly.
The evolution of FinOps as a formal discipline tracks this shift. The FinOps Foundation, which has developed principles, a maturity model, and certification programs for cloud financial management, frames the discipline around three things: visibility, accountability, and optimization. Visibility means knowing, in near real time, what cloud resources exist and what they cost. Accountability means attributing those costs to the business units and teams that generate them, not socializing them as undifferentiated shared infrastructure expense. Optimization means acting on cost data to make better provisioning decisions, right-size workloads, and eliminate waste. In principle this is straightforward. In practice, each of these three things is organizationally hard.
The visibility piece has gotten easier as tooling has matured. AWS Cost Explorer, Azure Cost Management, and a range of third-party FinOps platforms provide dashboards, anomaly detection, and rightsizing recommendations that would have been very expensive to build even five years ago. The information quality, to use the DeLone and McLean IS Success Model (2003) framing, is available. The system quality is commercially mature. The problem is not information availability.
The problem is what DeLone and McLean call "use" and "organizational impact," the organizational processes that turn available information into changed behavior. Visibility without accountability produces dashboards that people look at and then do not act on. Accountability without organizational routines for reviewing and enforcing cost targets produces a reporting structure that exists on paper but does not change behavior. The 84% of organizations that Flexera finds struggling with cloud cost management are not struggling because they cannot see their spend. Most of them have some form of cost visibility. They are struggling because the organizational structures that would allow them to act systematically on what they see have not been built.
This is where the unit economics conversation becomes important. The classic FinOps question is not "how much are we spending on cloud?" It is "what does each dollar of cloud spend produce?" Cost per transaction processed, cost per customer served, cost per model inference, cost per gigabyte of data stored and retrieved: these are the metrics that connect cloud spending to business outcomes. An organization that spends forty million dollars a year on cloud and cannot answer what that forty million produces is missing the foundation for any rational cost optimization conversation. You cannot optimize something you cannot measure in terms of value.
Building unit economics requires investment. It requires resource tagging discipline that most organizations do not have, because tagging is unglamorous work and enforcement is hard without automated tooling and organizational policy. It requires linkage between cloud resource consumption and business process outputs, which means connecting infrastructure metrics to application metrics to business outcome metrics across systems that were not designed to talk to each other. And it requires someone whose job it is to maintain all of this as the cloud environment changes. The 59% of organizations expanding their FinOps teams in the Flexera data are investing in this capacity, and the growth of FinOps as a recognized career path with conferences and certifications reflects genuine market demand for people who can do it.
The "cloud bill shock" phenomenon is worth naming directly because it explains part of why executive attention has increased. This is the experience, common among organizations that have been in the cloud for more than two years, of receiving a monthly bill that is materially larger than expected and scrambling to understand why. The causes are usually a combination of small things: an autoscaling configuration that responded to a traffic spike and never scaled back down, data transfer fees between cloud regions that nobody accounted for in the architecture, a managed service that charges per API call and turned out to be called far more than the project estimate assumed, a development environment provisioned for a sprint and never decommissioned. None of these are individually large. In aggregate across a large organization they compound into the 17% average overage that Flexera is measuring.
From my perspective as an IS researcher, the FinOps situation is a good illustration of what the sociotechnical systems framework (Trist and Bamforth 1951) was trying to capture. The technical side of cloud cost management, the tooling, the data, the dashboards, has improved substantially. The organizational side, the routines, the accountability structures, the incentive alignment, the management attention, has lagged. Sociotechnical theory would predict exactly this kind of gap: you cannot optimize one side without the other, and organizations that invest in technical solutions while neglecting organizational structures will find the technical solutions underperform their potential. The 84% who are still struggling despite years of investment in FinOps tooling are, in my read, organizations where the technical investment ran ahead of the organizational investment.
What surprises me in the Flexera data is how stable the "struggling" percentage is. This is not a new report, Flexera has been running it for years, and the proportion of organizations reporting cloud cost management as their top challenge has not dropped dramatically despite all the tooling investment and discipline building. That stability suggests something more structural is going on. My suspicion is that the organizational changes required to genuinely govern cloud costs, real budget ownership at the team level, incentive alignment between engineers and finance, executive visibility and attention, are the same kinds of organizational changes that are hard in every domain of enterprise IT. The technology is not the bottleneck. The organization is.
---
claims_checked:
- "84% of organizations struggle to manage cloud spend (Flexera 2025)": "https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/"
- "Organizations exceeding budgets by 17% (Flexera 2025)": "https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/"
- "28% expect cloud spend to increase (Flexera 2025)": "https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/"
- "59% expanding FinOps teams (Flexera 2025)": "https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/"
claims_unverified:
- "Cloud spend as second or third largest operating cost: widely discussed in FinOps Foundation materials, directional claim not sourced to a single study"
- "FinOps Foundation maturity model description: based on publicly available FinOps Foundation documentation, not linked to a specific page"
sources_used:
- "https://www.flexera.com/blog/cloud/cloud-computing-trends-tech-spend-pulse/"
word_count: 1100
About the author
Share
More notes
Related notes