The AI Spending Data Your Board Is About to Ask You About

A survey of 500 Finance leaders at large US and UK enterprises found that AI investment has been institutionalized well ahead of the ability to govern it. Every organization surveyed is already spending on AI, 79% experienced cost overruns in the past 12 months, and only 15% can calculate AI ROI without significant bottlenecks. The most counterintuitive finding: the organizations with the most mature FinOps practices report the highest overrun rates, because mature programs are larger and far better at surfacing problems that less mature organizations never detect. For technology executives, the sharper risk is structural. Accountability for AI spend is split nearly evenly between Technology (55%) and Finance (53%), and the 12-month ROI clock is already running without a clear owner.

AI spending has crossed a threshold that changes the conversation for every CTO and CIO. The question is no longer whether to invest. Every organization in this research already has. The question is whether anyone can prove what that investment is returning before the people who approved it start asking.

That is the central tension in an independent survey conducted by Sapio Research in February 2026, commissioned by DoiT and covering 500 Finance leaders at organizations with 1,000 or more employees across the US and UK. Every respondent's organization is currently spending on AI. The research was designed to capture current practice rather than stated intention, and the picture it produces is of an industry that has formalized AI investment faster than it has built the financial infrastructure to manage it.

Here is the number worth carrying into your next board presentation: 89% of organizations that self-assess as very mature or leading edge on FinOps experienced AI-related cost overruns in the past 12 months. Their mean overspend reaches 30.9%, the highest of any segment in the study.

By contrast, organizations in the early stages of FinOps development show a 69% overrun rate and a 16.1% mean overspend. Read quickly, that looks like an argument against investing in governance at all. It is the opposite. Mature organizations are running larger and more complex AI programs, and they have the instrumentation to surface overruns that less mature organizations simply never catch. The overruns at less mature organizations are not smaller because the spending is better controlled. They are smaller on paper because much of the spending is not being measured.

The practical takeaway reframes how technology leaders should make the case for investment in AI cost governance. The argument cannot be that better tooling will prevent overruns, because the data shows that the best-instrumented organizations overrun the most. The argument is that governance lets you see overruns in time to act on them. Maturity surfaces problems. It does not prevent them, and presenting it as prevention sets up a promise the data will not support.

One of the most consequential findings has nothing to do with tooling and everything to do with perspective. C-suite respondents rate their organization's FinOps maturity at 93% mature or better. Manager-level respondents put that figure at 60%. That is a 33-point gap, and both groups are describing the same organization.

This is not a disagreement about facts. It is a structural visibility problem. Leadership sees investment ambition, governance intent, and the frameworks presented at the board level. Operations sees the projects without named cost owners, the attribution systems that were scoped but never built, and the overruns that land on real budgets rather than in strategy decks.

For technology executives, that gap is a direct exposure. The survey found accountability for AI spend split almost evenly between Technology leadership at 55% and Finance at 53%, with no clear single owner at the operational level. When asked who holds final authority in a spending conflict, C-suite respondents named the CEO three times more often than managers did. At the level where spending actually happens, in other words, the question of who controls the AI budget often has no settled answer. Shared ownership, in practice, tends to mean no ownership, and AI spend is currently one of the most expensive places for that ambiguity to live.

The patience window for AI investment is closing faster than most governance programs are maturing. 83% of Finance leaders expect clear, quantifiable AI returns within 12 months. 81% are already adjusting their AI spend or plan to within the year.

Among C-suite respondents, 65% are acting or will act on AI spend within six months. Only 41% of managers share that urgency. That gap is where programs get cut. When operational teams have not internalized the leadership timeline, their projects become the obvious candidates for restructuring or reallocation at the next budget review, regardless of whether those projects were actually underperforming.

Only 15% of Finance leaders can calculate AI ROI without significant bottlenecks. The leading barriers are the pace of technological change at 40%, Finance and Engineering defining success differently at 37%, and a lack of clear financial attribution at 36%. The definition problem deserves to be separated from the other two, because it is the one barrier that does not require new tooling to fix. It requires Finance and Engineering to agree on what AI success means before any measurement system gets built. C-suite respondents feel this barrier most acutely, with 43% naming it a significant obstacle against 33% of managers. The people best positioned to resolve the definition gap are also the ones most aware of what it costs.

The survey becomes most useful when read at the segment level rather than in aggregate, because the averages conceal more than they reveal. Several cuts matter directly for technology leaders.

By company size, mid-size organizations of 1,000 to 4,999 employees experience overruns more frequently than large enterprises, at 81% against 76%, despite running smaller absolute AI budgets. Their mean overspend is also higher. Large enterprises carry more complexity and longer change cycles, which earns them more board patience, but the research shows their specific bottleneck is attribution complexity rather than execution. Organizations of 5,000 employees or more are markedly more likely to name forecasting difficulty as their hardest FinOps challenge.

By country, US organizations are operating with a 14-point urgency advantage over UK counterparts on AI spend adjustments. They lead on budget formalization at 80% against 67%, on per-unit cost tracking at 56% against 44%, and on unit economics adoption. They also run higher overrun rates, consistent with larger programs and more aggressive deployment. UK organizations reading their lower overrun rates as evidence of stronger governance should be cautious, because lower overruns also correlate with less systematic tracking and smaller projects.

On unit economics specifically, Finance leaders at large enterprises are running ahead of the broader practitioner field. The FinOps Foundation's State of FinOps 2026 found fewer than 20% of practitioners currently applying unit economics to AI spend. In this survey, 26% of Finance leaders have already implemented it and 34% plan to within six months. Among C-suite respondents, 80% expect to reach unit economics within six months. Among managers, that figure is 44%. Organizations that embed per-unit cost tracking before the patience window closes will be in a fundamentally stronger position when the board asks them to justify continued spend.

None of this argues that AI investment is unwise, or that overruns signal strategic failure. Every organization in the study is spending on AI, and the ones with the most sophisticated governance are also running the most ambitious programs. The argument is narrower and more uncomfortable. Without governance, you will not learn about the problems you already have until someone else surfaces them for you, usually at the least convenient moment in the budget cycle.

The highest-leverage action available to technology executives right now is not a new measurement platform. It is alignment. Finance and Engineering need a shared definition of what AI success means, established before tooling decisions are made. Organizations that resolve the definition problem first build measurement infrastructure that produces figures people actually trust. Organizations that buy tooling before resolving it produce numbers no one believes, which is a more expensive failure than having no numbers at all.

The full dataset, segmented by country, seniority, company size, sector, and FinOps maturity, is available in the complete report. It is structured so that you can locate your own organization in the data and see what your position predicts about your exposure.

DoiT Cloud Intelligence brings the same financial discipline to AI spend that FinOps teams already apply to cloud infrastructure, combining AI-driven cost attribution and anomaly detection with the human expertise to translate those signals into decisions a CFO will accept. Book a demo to see where your AI spend stands.