The more important finding is not just growth. It is divergence. Microsoft reports that adoption in the Global North grew nearly twice as fast as in the Global South, leaving richer and better-connected regions further ahead .

The headline numbers

Microsoft’s report frames the divide in North-South terms. In the second half of 2025, 24.7% of the working-age population in the Global North was using generative AI tools, compared with 14.1% in the Global South . A report summary says the gap widened from 9.8 percentage points to 10.6 percentage points between the first and second halves of 2025 .

That turns the report from a simple adoption milestone into a warning about unequal diffusion. Generative AI is spreading worldwide, but it is spreading fastest in places that already have stronger digital foundations.

Developed countries are pulling ahead, too

The Global North and Global South are not perfect substitutes for “developed” and “developing” economies, but the pattern is similar when the data is described that way. One related working-age comparison cited in coverage of the Microsoft report found 27.5% adoption among people aged 15–64 in developed countries, versus 15.4% in developing regions .

The practical takeaway is straightforward: countries with better digital infrastructure, higher connectivity, and more AI-ready institutions are not only starting ahead; they are also moving faster . If that continues, AI adoption could reinforce existing gaps in education, productivity, and access to digital services rather than automatically reducing them.

Why the AI adoption gap is widening

The report and related coverage point to familiar digital-economy constraints. Infrastructure and language are recurring dividing lines, with AI benefits concentrating in a relatively small group of countries . Coverage of the report also highlights limited internet access and English-focused AI models as barriers for poorer and non-English-speaking regions .

Three barriers matter most:

1. Connectivity and devices

Generative AI tools are far less useful when internet access is unreliable, expensive, or unavailable. A Microsoft-focused summary of AI diffusion highlights fast internet, reliable electricity, and device access as basic requirements for participating in the AI economy .

2. Language and local relevance

Many leading AI tools work best in English or in languages with abundant training data and product support. That creates friction for users in regions where local-language access, culturally relevant content, or localized applications are weaker .

3. Institutional readiness

Adoption is not just a consumer behavior. Schools, businesses, public agencies, and local developers need the capacity to turn access into useful applications. Microsoft’s finding that Global North adoption grew nearly twice as fast as Global South adoption suggests that stronger digital foundations help countries absorb new AI tools faster .

What the data does—and does not—show

The report is about diffusion: where generative AI tools are being used. One summary says the study measures adoption as the share of consumers who used a generative AI product during the reported period, using aggregated and anonymized Microsoft telemetry adjusted for factors such as device-market share, internet penetration, and country population .

That means the figures should not be read as a direct measure of productivity gains, economic value, or AI capability. They show where usage is spreading. The longer-term impact depends on whether people, schools, companies, and governments can convert that usage into learning, services, productivity, and locally useful tools.

The policy implication

Microsoft’s report describes a new phase of the digital divide: AI tools are becoming more widely available, but the benefits are not being distributed evenly . Closing that gap will likely require more than simply releasing AI products worldwide.

The evidence points toward investment in reliable connectivity, affordable devices, electricity access, local-language AI support, and policy environments that help institutions use these tools effectively . Without those foundations, generative AI may keep advancing fastest in the countries already best positioned to benefit from it.

The bottom line: generative AI adoption is rising globally, but the global gap is growing with it. The next challenge is not whether AI will diffuse—it already is—but whether developing regions can gain the infrastructure and local support needed to use it on equal terms.

The more important finding is not just growth. It is divergence. Microsoft reports that adoption in the Global North grew nearly twice as fast as in the Global South, leaving richer and better-connected regions further ahead .

The headline numbers

Microsoft’s report frames the divide in North-South terms. In the second half of 2025, 24.7% of the working-age population in the Global North was using generative AI tools, compared with 14.1% in the Global South . A report summary says the gap widened from 9.8 percentage points to 10.6 percentage points between the first and second halves of 2025 .

That turns the report from a simple adoption milestone into a warning about unequal diffusion. Generative AI is spreading worldwide, but it is spreading fastest in places that already have stronger digital foundations.

Developed countries are pulling ahead, too

The Global North and Global South are not perfect substitutes for “developed” and “developing” economies, but the pattern is similar when the data is described that way. One related working-age comparison cited in coverage of the Microsoft report found 27.5% adoption among people aged 15–64 in developed countries, versus 15.4% in developing regions .

The practical takeaway is straightforward: countries with better digital infrastructure, higher connectivity, and more AI-ready institutions are not only starting ahead; they are also moving faster . If that continues, AI adoption could reinforce existing gaps in education, productivity, and access to digital services rather than automatically reducing them.

Why the AI adoption gap is widening

The report and related coverage point to familiar digital-economy constraints. Infrastructure and language are recurring dividing lines, with AI benefits concentrating in a relatively small group of countries . Coverage of the report also highlights limited internet access and English-focused AI models as barriers for poorer and non-English-speaking regions .

Three barriers matter most:

1. Connectivity and devices

Generative AI tools are far less useful when internet access is unreliable, expensive, or unavailable. A Microsoft-focused summary of AI diffusion highlights fast internet, reliable electricity, and device access as basic requirements for participating in the AI economy .

2. Language and local relevance

Many leading AI tools work best in English or in languages with abundant training data and product support. That creates friction for users in regions where local-language access, culturally relevant content, or localized applications are weaker .

3. Institutional readiness

Adoption is not just a consumer behavior. Schools, businesses, public agencies, and local developers need the capacity to turn access into useful applications. Microsoft’s finding that Global North adoption grew nearly twice as fast as Global South adoption suggests that stronger digital foundations help countries absorb new AI tools faster .

What the data does—and does not—show

The report is about diffusion: where generative AI tools are being used. One summary says the study measures adoption as the share of consumers who used a generative AI product during the reported period, using aggregated and anonymized Microsoft telemetry adjusted for factors such as device-market share, internet penetration, and country population .

That means the figures should not be read as a direct measure of productivity gains, economic value, or AI capability. They show where usage is spreading. The longer-term impact depends on whether people, schools, companies, and governments can convert that usage into learning, services, productivity, and locally useful tools.

The policy implication

Microsoft’s report describes a new phase of the digital divide: AI tools are becoming more widely available, but the benefits are not being distributed evenly . Closing that gap will likely require more than simply releasing AI products worldwide.

The evidence points toward investment in reliable connectivity, affordable devices, electricity access, local-language AI support, and policy environments that help institutions use these tools effectively . Without those foundations, generative AI may keep advancing fastest in the countries already best positioned to benefit from it.

The bottom line: generative AI adoption is rising globally, but the global gap is growing with it. The next challenge is not whether AI will diffuse—it already is—but whether developing regions can gain the infrastructure and local support needed to use it on equal terms.