CEOs Report Little Measurable Productivity Gain from AI

512 Pixels reports that an NBER study of nearly 6,000 CEOs and CFOs across the US, UK, Germany, and Australia found that firms reported zero measurable impact on productivity or employment from AI over the past three years, according to the scraped coverage. The article states average employee AI usage was 1.5 hours per week; the scraped copy did not preserve a complete figure for average CEO AI usage. 512 Pixels also cites Sequoia estimating a $690 billion AI infrastructure buildout while current AI infrastructure revenue is estimated at $50-100 billion in the same report.

512 Pixels collates several commonly cited metrics on enterprise AI outcomes: that only one in five AI investments yields any measurable ROI, that one in 50 produces transformational value, and that 95% of enterprise AI pilots fail to move from lab to production, per the scraped article. These numbers are presented as aggregated industry observations in the piece rather than a single primary-data release.

Editorial analysis: Companies across sectors have pursued heavy AI infrastructure spending even as measured operational impact remains limited in many cases. Observed patterns from prior industry reporting show a disconnect between tooling investment and broad workforce adoption, especially early in large-scale transformations. For practitioners, that gap often translates into integration, measurement, and change-management challenges rather than purely technical model deficiencies.

Editorial analysis: The combination of low average employee usage (1.5 hours per week) and high infrastructure expectations (Sequoia's $690 billion figure as cited by 512 Pixels) highlights stress points where engineering teams, data platform owners, and product managers typically confront ROI measurement, observability, and operator training problems. Industry experience suggests that moving pilots to durable production requires systematic telemetry, reproducible pipelines, and clear success metrics rather than point tooling alone.

Editorial analysis: Observers should track independent, reproducible studies that correlate specific adoption interventions with measurable productivity gains, third-party vendor claims validated by customer case studies, and how organizations shift investment from undifferentiated infrastructure to measurable application-level outcomes. 512 Pixels has not supplied primary data beyond the cited NBER study and third-party estimates; readers should consult the original NBER release and Sequoia publications for primary figures.