Only 34% of the C-suite knows who owns AI decisions. 57% of their subordinates do.

Pearl Meyer's Q2 2026 Market Intelligence Survey found that only 34% of C-suite executives said it is consistently clear which executive or team makes the calls on AI, the lowest figure of any cohort polled ^[1][5]. Among board members the figure was 53%, and among senior managers and professionals below the C-suite it was 57% ^[2][3].

That is a 23-point inversion of the usual reporting gap ^[1]. The people who sit closest to the decision rights are the least confident those rights exist; the people whose job is to execute against them are the most confident ^[1][3]. One of those readings is wrong, and the expensive interpretation is that clarity looks like clarity only from a distance.

The bill makes the ambiguity matter. Gartner projects total AI spending, including capital expenditure on infrastructure, will reach $2.5 trillion this year, a 44% increase on last year, rising to $3.3 trillion next year ^[6][7]. That implies roughly $1.7 trillion in 2025 and another 32% increase in 2026 ^[2][3]. A survey of 900 CEOs published in May, cited by Fortune, found 80% of U.S. CEOs think their job is at risk if their AI projects wither, and 81% expect a peer to be ousted over an AI failure or crisis ^[8].

Confidence, meanwhile, is untethered from progress. Brad Jayne, a Pearl Meyer principal and co-author of the study, said belief that AI will deliver significant gains within 18 months sits at about 50% among leaders at every stage of maturity, including companies still piloting, companies experimenting, companies deployed at enterprise scale, and companies that have not started ^[10]. If a firm that has done nothing is as optimistic as one running production systems, the optimism is not measuring anything about the firm. Jayne called it an "impact-versus-speed tension": handing out ChatGPT or Copilot licences is quick, while building systems around them and checking they are not producing errors takes far longer ^[11]. Below the C-suite, 78% said their company already has the senior talent to implement and oversee AI enterprise-wide ^[4].

The other gaps in the survey point the same way. Asked whether employees could absorb more organisational change without being stretched too thin, 63% of CEOs said yes, against 33% of the C-suite and 40% of non-C-suite executives, a 30-point spread between the CEO and the layer directly beneath ^[12][4]. Asked whether hitting strategic goals will require significant changes to how the organisation operates within three years, 88% of CEOs and 79% of C-suite executives said yes, but only 42% of directors agreed, a 46-point gap between chief executives and their boards ^[13][5]. Jayne characterised the board position as "We're good. We've made investments, we're structured right, go make changes," while management says it will have to change how it operates to deliver ^[15]. Combined with the change-fatigue answers, he called that "an alarm bell" ^[14].

The study's own conclusion is blunter than most vendor research: ambition for AI outcomes is outpacing the leadership structure needed to deliver it, and additional investment without clear ownership will only widen the gap ^[9].

Caveats worth holding: this is 116 respondents surveyed in May and June, self-reported perception rather than audited decision rights, and it was shared exclusively with Fortune ahead of publication ^[5].

What to watch is the justification cycle Jayne is worried about, roughly a year out, when AI spending has to be tied to outcomes investors recognise ^[16]. His stated fear is finger-pointing ^[16]. In practice the tell will be whether companies name an owner for AI results before that conversation starts or after it goes badly, and whether boards that currently see no need for operating change are the ones asking the questions.