FinOps has won the argument for cost transparency.

The State of FinOps 2026 makes that clear. 98% of organizations now manage AI spend. 90% manage SaaS costs or plan to within the next year. 78% of FinOps teams report into the CTO or CIO — technology cost governance has moved closer to engineering leadership than ever before. (Source:
https://data.finops.org)

The discipline has clearly moved up the organization.

But underneath that progress, a structural problem is hiding in plain sight.

In most organizations, FinOps still operates primarily as an advisory function. It produces insights, identifies opportunities, and generates reports. The authority to act on those findings lives somewhere else — in engineering, procurement, product, or platform teams.

As technology consumption expands across AI, SaaS, cloud infrastructure, and private platforms, that advisory model is reaching its limits.

The next stage of FinOps maturity is not better dashboards. It is execution.

Here is what the dangerous middle state looks like in practice.

The FinOps team has strong visibility into technology spend. There are dashboards, cost allocation reports, and anomaly detection. Leadership can see the numbers. The quarterly business review deck looks thorough.

But when a finding surfaces — idle compute, unused SaaS licenses, an AI model burning budget with no ROI signal — no one has clear authority to act on it.

FinOps identifies the opportunity. Engineering owns the infrastructure. Procurement owns the vendor contracts. Product teams own feature design.

Responsibility becomes diffuse. Recommendations accumulate faster than they get implemented.

This is Visibility Theater. The dashboards create the appearance of control without the organizational wiring to produce outcomes. And it is more common than most FinOps leaders are willing to admit publicly, because the dashboards look impressive in a review.

Visibility without ownership is not governance. It is reporting with extra steps.

The advisory model made more sense when technology costs were primarily an infrastructure problem.

That is no longer the case.

Today organizations operate across multiple consumption models simultaneously — each with its own economic dynamics and a different team holding the keys. For example:

Public cloud costs are usage-based, optimized through architecture and commitment vehicles. AI services bill per token or per request, optimized through model selection and caching. SaaS platforms charge per seat or tier, governed through license management. Private cloud and on-premise environments carry capacity-based costs, optimized through utilization and workload placement.

These models look different on an invoice. But they share one trait: they represent technology decisions that create recurring cost structures. And those decisions increasingly determine product margins, platform scalability, and engineering economics.

This is why the FinOps Foundation recently reframed its mission around managing the value of technology — not just cloud costs. The scope of the discipline has expanded because the scope of the problem has expanded. (https://www.finops.org/framework/)

But governance across this many consumption models requires more than a well-designed dashboard. It requires an operating model where analysis reliably turns into action, with clear ownership at every step.

It requires operational execution.

Fixing the execution gap requires working on two fronts simultaneously.

Most FinOps work happens after systems are already in production. Costs are analyzed, anomalies are flagged, and recommendations are made about things that are already running. That is useful, but it is always remedial.

The higher-leverage opportunity is embedding economic thinking earlier in the technology lifecycle — at the point where decisions are still being made.

This does not require new meetings or heavyweight governance processes. It requires adding a small set of economic signals to decision artifacts that already exist: Architecture Decision Records, Product Requirement Documents, architecture reviews, vendor evaluations, AI model selections.

At those decision points, teams capture the expected cost drivers, projected unit economics, acceptable thresholds, and who owns monitoring if assumptions prove wrong. Even in fast-moving organizations, that documentation creates an artifact teams can revisit later — rather than inheriting a cost problem with no traceable origin.

The goal is simple:

Make cost a design input—not just an operational output.

Every organization accumulates inefficiencies over time. Idle compute. Over-provisioned instances. Unused SaaS licenses. Overlapping tools. Expensive AI model choices where a cheaper model would produce equivalent results. Underutilized private cloud clusters with no lifecycle policies.

The challenge is rarely identifying these issues. FinOps tools are good at surfacing them. The challenge is organizational — getting someone with the authority and the bandwidth to actually fix them.

The CloudZero State of AI Cost Report reinforces the urgency: average monthly AI budgets grew from roughly $63K in 2024 to $85K in 2025, while only 51% of organizations reported confidence evaluating AI ROI. Spend is accelerating faster than the governance to match it. (Source: https://www.cloudzero.com/state-of-ai-costs/)

Waste removal matters. But remediation alone is not the end state.

Organizations that build both execution paths — preventing new cost problems and eliminating existing ones — eventually see them converge.

Not into two parallel processes, but into a single continuous operating system for technology economics.

In this steady state, technology decisions and financial outcomes are connected through an ongoing feedback loop:

Design
Architecture decisions establish the expected economic behavior of a system.
Examples include cloud architecture, SaaS platform selection, AI model choice, and private infrastructure placement.

Build
Engineering implementation determines how effectively those decisions are executed through configuration choices, scaling policies, caching strategies, and resource allocation.

Operate
Operational monitoring measures real-world usage, cost patterns, and utilization across cloud, SaaS, AI, and private infrastructure.

Improve
When systems deviate from expected economics, teams implement improvements such as rightsizing compute, reclaiming SaaS licenses, adjusting AI models, or rebalancing workloads.

Learn
Insights from these improvements feed back into future design decisions, gradually improving the economic efficiency of new systems.

In this model, FinOps is not responsible for implementing every improvement. That would recreate the ownership problem from a different direction.

Instead, FinOps acts as the economic control system — maintaining the feedback loop that allows the rest of the organization to make better technology decisions continuously.

The FinOps function provides cost visibility, economic modeling, forecasting, governance frameworks, unit economics definitions, and policy guidance. Execution lives where it belongs: engineering teams implement infrastructure improvements, platform teams manage efficiency, procurement and ITAM govern SaaS licensing, product teams manage AI and feature-level economics.

This is centralized enablement with federated execution — and it is reflected in practice. The State of FinOps survey shows roughly 60% of organizations operate with centralized FinOps enablement combined with embedded champions across engineering and platform teams. Another 21% use hub-and-spoke structuresr.

The pattern is clear. FinOps does not own every cost decision. It maintains the system that makes better decisions possible.

When organizations reach this steady state, something subtle but consequential shifts.

Optimization insights start improving future design decisions rather than just cleaning up past ones. A costly AI architecture leads to better model selection in the next release. SaaS license waste produces stronger provisioning governance. Cloud inefficiencies influence future platform standards.

The organization gradually becomes economically smarter about technology — not because FinOps issued a policy, but because the feedback loop is functioning.

At that point, FinOps stops being the team that explains the bill.

It becomes the system that shapes it.

• dashboards
• cost allocation
• anomaly detection
• savings recommendations

But visibility alone doesn’t change outcomes. Ownership does.

When no team is explicitly responsible for acting on what the data surfaces, recommendations accumulate faster than they get implemented. The dashboards look thorough. The quarterly review deck is impressive. And the waste compounds quietly in the background.

I call this Visibility Theater — the illusion of control through reporting, without the operational wiring to produce results.

Signal: The fix isn't better dashboards. It's building the ownership model that turns signal into action.

A new report from Harness and AWS puts a number on the disconnect.

The FinOps in Focus 2025 study surveyed 700 developers and engineering leaders across the US and UK. The findings are consistent with what many FinOps practitioners already know from experience (and been the focus of this newsletter 😉): visibility and ownership are not the same thing, and most organizations have built more of the former than the latter.

52% of engineering leaders say the disconnect between FinOps teams and developers directly leads to wasted cloud spend. 55% of developers say they largely ignore cost management. Only 35% report that cloud cost efficiency is a key success metric for their teams.

The most interesting finding, though, is not the dysfunction. It's the appetite underneath it: 62% of developers say they want more control and accountability for managing cloud costs.

The ownership gap isn't an engineering culture problem. It's an organizational design problem — and one that we suggest is solvable.

Featured from the FinOps Directory, built and maintained by CloudXray AI.

Company: OpenOps

Category: FinOps Automation and Workflow Orchestration

What They Do: OpenOps is an open-source FinOps automation platform that orchestrates cloud cost and governance workflows across complex, multi-cloud environments.

Why It Matters: Because execution matters over visibility theater

A practical starting point is service-level cost ownership — assigning clear accountability at the level where engineering decisions actually get made.

For every major service, workload, or platform component:

Once ownership exists at the service level, cost improvements stop being FinOps recommendations waiting for a sponsor. They become operational work with an owner.

That is the difference between a team that reports on costs and a team that governs them.