Microsoft’s Report: Inside the Global AI Adoption Divide

The speed at which AI has spread across the globe is creating a widening challenge. 

While the technology has reached more users faster than any previous innovation in history, the gap between those who can access it and those who cannot continue to widen along familiar economic and geographic lines.

Research from Microsoft’s AI Economy Institute, in a report called ‘AI Diffusion Report: Where AI is most used, developed and built’, reveals part of this disparity.

AI adoption in the Global North runs at roughly double the rate of the Global South, with the chasm becoming particularly pronounced in countries where GDP per capita falls below US$20,000. 

The report, which draws on aggregated telemetry from over one billion Windows devices alongside data from major AI platforms including OpenAI’s ChatGPT, Microsoft Copilot, Anthropic’s Claude and Google’s Gemini, exposes the extent of digital inequality in the age of AI.

In the United Arab Emirates, 59.4% of working-age adults use AI tools. Singapore sits close behind at 58.6%. 

Yet in parts of Sub-Saharan Africa and Asia, adoption rates struggle to reach 10%. 

The overall picture shows AI adoption at 23% in the Global North compared with just 13% in the Global South.

CEO of Microsoft, Satya Nadella, says: “While the Global North leads in AI adoption, we are committed to bridging the digital divide and accelerating AI equity worldwide.”

The barriers are both obvious and complex. 

Microsoft’s research identifies five essential building blocks: electricity, data centres, internet connectivity, digital skills and language support. 

Four billion people lack these basics, with over 750 million still without access to electricity at all.

How frontier development concentrates in seven nations

The concentration of AI development tells its own story. 

Only seven countries host models ranking in the world’s top 200. The United States maintains its lead with OpenAI’s GPT-5, but China’s DeepSeek V3.1 follows less than six months behind. 

France, South Korea, the UK, Canada and Israel round out this exclusive club.

What makes this particularly striking is how quickly the performance gap has narrowed. Israel’s most advanced model, AI21 Labs’ Jamba Large 1.7, trails the frontier by just 11.6 months. 

The distance between first and seventh place has shrunk to less than a year, suggesting AI diffusion at the cutting edge is outpacing previous technological revolutions.

Meanwhile, infrastructure remains stubbornly concentrated. 

The US and China together control 86 % of global data centre capacity, according to International Energy Agency figures. 

This matters more than it might seem because proximity to data centres directly affects response times and user experience.

Why language creates unexpected divide

Perhaps the most intriguing finding concerns language. 

Half of all web content exists in English, despite English being spoken natively by just 5 % of the world’s population. 

Countries where low-resource languages dominate show adoption rates 20 % lower than high-resource language countries, even when GDP and internet connectivity are comparable.

Swahili, spoken by over 200 million people, has over 500 times less digital content than German, despite similar speaker numbers. 

Meanwhile, advanced large language models (LLMs) achieve around 80% accuracy in English but drop below 55% for languages like Yoruba, which more than 50 million people speak across Africa.

Meanwhile, Singapore demonstrates what sustained investment can achieve. 

Its 59% adoption rate builds on decades of deliberate policy, beginning in the 1980s when the government began wiring the nation with high-speed connectivity and expanding computer access in schools.

The report draws parallels with South Korea’s semiconductor revolution. 

In the late 1970s, strategic government-private sector partnerships helped transform the country’s economy, which subsequently grew at 6.2% annually. 

The Philippines, starting from a similar position, managed 1.8%.

“We are already a cyber powerhouse and we can and must, also be an AI powerhouse,” says Braverman-Blumenstyk.

“To ensure Israel’s leadership, a national plan is needed to invest in human capital, education and technological infrastructure.”

Microsoft’s CEO says: “AI has crossed from hype to becoming a core part of how every organization operates, innovates and delivers value” and stresses, “We must focus on responsible AI adoption — safeguarding privacy, security and ensuring ethical use remain top priorities across the industry”.​