A new Microsoft research report indicates that Chinese artificial intelligence firm DeepSeek is significantly accelerating AI adoption across developing nations, potentially narrowing the technological gap with advanced economies. The findings reveal that global generative AI tool adoption reached 16.3% of the world’s population by December, marking a notable increase from 15.1% in the previous quarter.
The comprehensive study, released Thursday by Microsoft’s AI for Good Lab, documents a concerning divergence in AI integration rates between developed and developing regions. While adoption rates in the global north continue to expand at nearly double the pace of southern hemisphere nations, DeepSeek’s innovative approach is creating new pathways for digital inclusion.
According to Juan Lavista Ferres, Chief Data Scientist at Microsoft’s AI for Good Lab, “We are observing a concerning divide that threatens to widen without intervention.” The research utilized anonymized telemetry data to track global device usage patterns.
The report identifies early digital infrastructure investors including the United Arab Emirates, Singapore, France, and Spain as leaders in AI adoption rates. These findings align with recent Pew Research Center survey data mapping global attitudes toward artificial intelligence.
DeepSeek, founded in 2023, has driven substantial AI penetration across price-sensitive markets through its free and open-source model architecture. The company’s January 2025 release of its R1 advanced reasoning model—described as more cost-effective than OpenAI’s comparable offering—generated significant attention within the technology sector. The startup’s credibility was further enhanced when leading journal Nature published peer-reviewed research co-authored by DeepSeek founder Liang Wenfeng, characterizing it as a “landmark paper” from China’s AI community.
Microsoft’s analysis notes that DeepSeek operates differently from Western models on certain topics, particularly political subjects, reflecting China’s internet accessibility framework. Despite this, the platform has demonstrated strong performance in mathematical and coding applications.
The company’s business model, which eliminates subscription fees and provides developers with global access to modify its core engine, has dramatically reduced accessibility barriers. This approach has proven particularly effective in regions where Western platforms face restrictions or limited availability, including Russia, Iran, Cuba, and Belarus.
DeepSeek achieved remarkable market penetration in China (89%), with substantial presence in Belarus (56%), Cuba (49%), Russia (43%), Iran (25%), and Syria (23%). African nations including Ethiopia, Zimbabwe, Uganda, and Niger recorded adoption rates between 11-14%.
The report concludes that open-source AI models like DeepSeek’s can function as geopolitical instruments, extending technological influence into markets where Western platforms encounter operational challenges. Despite security concerns that prompted developed nations including Australia, Germany, and the United States to restrict DeepSeek’s usage—and Microsoft’s own internal ban on the platform—the Chinese startup continues to reshape global AI adoption patterns through accessibility and affordability.