89% of Executives Report No AI Effect on Firm Labor Productivity Over Three Years

Global AI spending is forecast to reach $2.7 trillion in 2026, even as 89% of surveyed executives said the technology had not affected labor productivity at their firms over the previous three years.

The worldwide forecast represents a 49.5% increase from 2025. AI infrastructure spending is projected at $1.484 trillion, more than half of the total. These figures cover expected spending during calendar 2026, not expenditures already recorded.

In a survey of nearly 6,000 senior executives across the United States, the United Kingdom, Germany and Australia, 89% said AI had no effect on labor productivity at their firms during the prior three years. The executives also reported using AI for an average of 1.5 hours per week. Their responses capture executives’ assessments of productivity at their firms, rather than a direct measurement of output.

In August 2026, 40.5% of employed U.S. respondents ages 18 to 64 used generative AI for work. AI use accounted for 5.3% of total work hours, while respondents estimated that time saved equaled 2.1% of all work hours. The tracker covers employed adults ages 18 to 64 in a nationally representative survey. Its time-saved figure reflects respondents’ estimates, not additional output measured on the job.

Among 1,719 respondents across 97 countries, 80% said AI improved their individual productivity. Meanwhile, 37% reported a positive effect on their organizations’ earnings before interest and taxes, or EBIT. About 6% met the survey’s high-performer threshold: an EBIT impact of at least 5% and significant value from AI.

The surveys measure different outcomes among different groups. U.S. workers estimated time saved, executives assessed firm-level labor productivity, and the global survey asked about individual productivity and company earnings. Those findings cannot be treated as a direct comparison. Saved time contributes to higher productivity only when it leads to additional output, and faster individual work does not by itself establish an earnings gain. Earlier TokenPost coverage of AI productivity claims and corporate earnings examined the gap between those measures.