U.S. AI Diffusion: What the 28.3% Estimate Actually Measures

How the number was produced

This estimate comes from Microsoft telemetry rather than a survey asking people whether they have ever used generative AI. Stanford used the Microsoft data to compare AI diffusion across countries and to examine its relationship with GDP per capita.

The United States rose from 26.3% in the first half of 2025 to 28.3% in the second half. Even after that increase, it remained in the lower half of the 30 geographic areas shown by Stanford.

What the ranking tells us

The United States leads the world in private AI investment and frontier-model development, but leadership in building AI does not automatically produce leadership in population-level diffusion. Singapore, the United Arab Emirates and several European economies recorded substantially higher estimates in the same dataset.

Stanford found a strong positive relationship between AI diffusion and GDP per capita, but the United States fell below the pattern expected from its wealth. That makes American adoption an important strategic question: the country building much of the technology may not be integrating it through society as quickly as several peers.

What this number does not mean

• It is not a survey finding that only 28.3% of American adults have ever tried AI.
• It does not measure business adoption, paid subscriptions, skill, productivity, or frequency of use.
• It is not directly comparable with adoption percentages produced through self-reported surveys.
• The international ranking is meaningful only within the same Microsoft dataset and reference period.

Why it matters

Diffusion determines whether technical leadership becomes broad economic capability. If AI use remains concentrated among technology firms, highly educated workers, and early adopters, the benefits may spread unevenly while the country underuses a technology it helped create.

The signal to watch is not simply whether the percentage rises, but whether adoption broadens across occupations, regions, income groups and ordinary business functions.