More reports have recently come out questioning the return on investment of AI for corporate customers.
Mostly true True in substance, wrong in a detail.
Factual: Statements about the present or the past, checked against the record: did they hold up?
What would count as true or false
Judged as of 2026-06-02. True: in the weeks or few months before that date, at least two new published reports, surveys or studies (from consultancies, research groups, analysts or major media) questioned or found weak or uncertain returns on investment from AI for corporate customers. Mostly true: such reports appeared, but there was only one, or they were somewhat older than 'recently' suggests (e.g. 3 to 6 months earlier). Mixed: the recent reports were evenly split, or only loosely addressed ROI. Mostly false: recent reports overwhelmingly found strong ROI, and skeptical findings were older or marginal. False: no reports questioning corporate AI ROI came out in that period, as shown by a review of the period's major surveys finding positive ROI.
Set before the research began.
Evidence (4)
-
E1 KPMG report finds enterprise disconnect between AI and its ROI CIO.com · 10 April 2026
“Enterprise CIOs need no convincing that return on investment (ROI) for genAI and agentic AI is elusive, but consulting giant KPMG is reporting that some companies are plowing ahead with the technology anyway.”
-
E2 PwC 2026 Global CEO Survey PricewaterhouseCoopers · 19 January 2026
“Despite widespread experimentation, only one-in-eight (12%) CEOs say AI has delivered both cost and revenue benefits. Overall, 33% report gains in either cost or revenue, while 56% say they have seen no significant financial benefit to date.”
-
E3 2026 AI Adoption and Risk Benchmarking | Gallagher Arthur J. Gallagher & Co. · 20 February 2026
“Nearly two-thirds of organizations are actively measuring ROI, estimating it will take an average of 28 months to realize that return.”
-
E4 The AI Moment? Possibilities, Productivity, and Policy - San Francisco Fed Federal Reserve Bank of San Francisco · 23 February 2026
“So far, most macro-studies of productivity growth find limited evidence of a significant AI effect (see, for example, Imas 2026, Acemoglu 2025, and Aghion and Bunel 2024.)”
Reasoning
E1 (CIO.com, 2026-04-10) is the one clearly recent report. The CIO.com article covering a KPMG report describes enterprise AI ROI as elusive and came out about seven weeks before the claim date [E1]. The other reports that question ROI are older. PwC's CEO survey (E2, January 2026) found that 56% of CEOs had seen no significant financial benefit from AI. Gallagher's benchmark (E3, February 2026) estimated an average of 28 months to realize a return. Both fall in the 3–6 month range that the criteria treat as "mostly true" rather than "recently". The San Francisco Fed letter (E4) is about macro productivity rather than corporate ROI, so only one recent report fully meets the "true" bar.
Cite this claim
Gary Marcus, “More reports have recently come out questioning the return on investment of AI for corporate customers.” (said 2 June 2026). Verdict: mostly true. Reliability API, claim c_9889bea0b032, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_9889bea0b032
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Gary Marcus
-
Mixed
“companies like Anthropic are ending their all you can eat buffet, which will make many customers think twice.”
-
Mixed
“But everybody is building essentially the same technical solution with essentially the same data, so there is no moat.”
-
Mostly true
“The company has not returned any profits, and is not expecting to do so until at least 2029, but has a burn rate in the billions, and still lacks a clear technical moat.”
-
Resolved true
“There is very little moat.”
As of end 2024, AI model makers have very little moat, with many similar models and a price war.
-
Mostly true
“And current-sized models, already expensive, aren’t reliable enough for many jobs.”
Current-sized LLMs are expensive and not reliable enough for many jobs.
Verdicts are made automatically; how verdicts are made.