The Pulse
Microsoft Puts Global Generative AI Use at 18.8% in Q2
Microsoft says 18.8% of the world’s working-age population used a generative AI product in June 2026, up from 17.8% in the first quarter. The report also finds a widening gap between the Global North and South and growing use of open-weight
AI.info Team ·
Microsoft says generative AI reached 18.8% of the world’s working-age population in June, marking a one-percentage-point increase from the first quarter of 2026. The figure appears in the company’s latest Global AI Diffusion Report, published September 21 by Juan Lavista Ferres, Microsoft’s chief data scientist.
Microsoft defines AI diffusion as the share of people aged 15 to 64 who used a generative AI product during the reported period. The measure comes from aggregated, anonymized Microsoft telemetry and is adjusted for differences in operating-system and device-market share, internet penetration and population. The company says the estimate does not capture every AI tool or model in use.
The report’s headline number is therefore a measure of observed generative AI usage, not a count of all people who have access to an AI system. Microsoft says it plans to expand the set of tools included in the calculation in its next report, a change it expects to raise measured usage in almost every economy and produce a larger increase in China.
UAE and Singapore Remain Far Ahead
The United Arab Emirates leads Microsoft’s June rankings with an AI user share of 73.3%, followed by Singapore at 64.3%. Ireland ranks third at 49.9%, France fourth at 49.6% and Norway fifth at 49.4%.
South Korea records the largest increase among the leading economies, adding 3.5 percentage points to reach 40.6%. The country rises from 16th place in the first quarter to 12th in the second and has posted the largest gain among major economies for three consecutive reporting periods, according to Microsoft.
Saudi Arabia makes the largest move in the rankings, climbing five places from 30th to 25th and reaching 31.4%. Japan records the largest relative increase, adding 2.2 percentage points from its first-quarter level of 22.5%.
The United States Reaches 33% in Microsoft’s Measure
The United States records a 33% AI user share in the second quarter, up 1.7 percentage points from the first quarter. That places the country 21st in Microsoft’s ranking, behind Canada at 38.5%, Taiwan at 33.8% and Hungary at 33.3%.
The list is not a general ranking of national AI capability or investment. It measures the proportion of the working-age population Microsoft estimates used a generative AI product during the period, which makes the result sensitive to the company’s data sources and the products included in its methodology.
Microsoft’s previous report put worldwide usage at 17.8% in the first quarter, up from 16.3% in the second half of 2025. The new figure indicates that adoption continues to rise, although the quarterly increase is smaller than the 1.5-percentage-point gain recorded in the prior report.
The North-South Gap Widens to 12.6 Points
Microsoft measures AI usage at 28.8% in the Global North during the second quarter, compared with 16.2% in the Global South. The North gains 1.3 percentage points from the first quarter, while the South gains 0.8 points, widening the difference from 12.1 to 12.6 percentage points.
The report also examines how people use Microsoft Copilot in the two regions. Self-directed learning and skills development represent a 29.2% larger share of conversations in the Global South than in the global mix, while schoolwork and academic support account for a 32.3% larger share.
Microsoft cautions that the Copilot analysis covers early adopters and may not describe broader usage as adoption expands. The company also says the data cannot establish why learning appears more often in the Global South, although population differences and the region’s younger demographic may contribute.
Open-Weight Models Gain Ground Through APIs
Microsoft’s report gives significant attention to open-weight models, whose trained weights can be downloaded, modified and run by other organizations. The company argues that these systems can lower access costs and make it easier to adapt models to local languages and specialized needs, although electricity, connectivity, infrastructure and skills remain barriers.
Open-weight usage is difficult to capture through consumer chatbot measurements because developers may access the models through applications, coding agents or APIs. Microsoft’s analysis of public OpenRouter data finds that open-weight models now account for 75% of token usage on that service. The company explicitly says the figure is not a global market share because OpenRouter users are unusually likely to test multiple providers, while direct use of closed models and locally deployed open models is absent from the dataset.
The report also points to a larger supply of AI applications. Hugging Face Spaces, which hosts runnable AI applications and demonstrations, grows from approximately 124,000 at the end of 2023 to 1.46 million by September 2026. Microsoft says the total has increased by about 83% since its first AI Diffusion Report in November 2025.
That growth is one reason Microsoft plans to revise its diffusion measure. The company says many tools introduced since its first report already attract substantial usage but are not included in the current calculation. Its next report is intended to cover more of those tools, especially domestic services in China, where Microsoft expects the revision to have its largest effect.
The full Global AI Diffusion Report identifies 18.8% as the measured worldwide user share for June 2026 and records a 12.6-point difference between the Global North and South.