AI Diffusion Report: Where AI Is Most Used, Developed, Built

Microsoft AI for Good Lab publishes the AI Diffusion Report measuring global AI adoption, frontier development, and infrastructure.

• Main announcement: The report finds 1.2 billion AI users in under 36 months, that AI adoption in the Global North is roughly double the Global South, and it introduces three indices—AI Frontier Index, AI Infrastructure Index, and AI Diffusion Index—using Microsoft telemetry and external data to quantify adoption, frontier model provenance, and infrastructure capacity. It also reports the U.S. and China host ~86% of global data center capacity, identifies seven countries with frontier-level models (U.S., China, France, South Korea, U.K., Canada, Israel), and gives example “months to frontier” gaps (China < 6 months; Israel ~11.6 months).
• Background & details: The report documents the five foundational building blocks—electricity, data centers, internet, digital/AI skills, and language—and provides concrete figures: over 750 million people lack electricity, 18 of the 20 largest electricity access deficits are in Sub-Saharan Africa, Malawi <4% English fluency, and low-resource language countries adopt AI ~20% less than high-resource language peers (controls applied). Data sources include Windows telemetry, IEA data center capacity (GW), World Bank GDP data, Our World in Data, Common Crawl, and Microsoft AI for Good Lab technical reports.