Microsoft President Brad Smith issued a stark warning about China's dominance in artificial intelligence (AI) outside the West, noting that Chinese models like DeepSeek are gaining increasing popularity in developing markets, particularly in Africa, thanks to their low cost and accessibility. This warning came in an interview published by the Financial Times, coinciding with an internal company report showing a significant acceleration in the adoption of open-source Chinese technologies in the Global South.
Why is Chinese AI so dominant in developing countries?
Smith explained that the main reason lies in three interconnected factors:
- Consistent government support for Chinese companies, enabling them to offer robust models at very competitive prices, or even for free in some cases.
- A focus on open-source models, a sector largely neglected by American companies like OpenAI and Anthropic, giving China a competitive edge in markets seeking inexpensive and customizable solutions.
- Limited access to Western models in countries under US sanctions (such as Belarus and Cuba), where DeepSeek penetration rates approach 50% in some cases.
A Microsoft study showed that DeepSeek's R1 model significantly accelerated AI adoption in developing countries, with its market share in some African markets (such as Ethiopia at 18% and Zimbabwe at 17%) exceeding that of other Western models.
Jensen Huang's statements align with Smith's.
Smith's statements echo those made by Nvidia CEO Jensen Huang in November 2025, when he said that China would "win" the global AI race due to massive government support and incentives offered by Beijing to its companies, in contrast to the regulatory restrictions and sanctions imposed by the United States on the export of advanced technologies.
Reactions in Africa: Seeking Economic and Local Solutions
Bright Simons, Vice President of the Emani Research Center in Ghana, believes that African countries lack the financial capacity to afford expensive Western solutions, leading them to turn to open-source models, whether Chinese (DeepSeek) or Western (Llama from Meta). He pointed to the emergence of promising local models such as Masakhane (a pan-African multilingual model) and InkubaLM (a South African model) as evidence of the continent's drive to develop independent solutions.
Conclusion: A New Global AI Competition
The report and statements show that the AI race is no longer confined to major US companies but is shifting toward developing markets, where Chinese companies are gaining a competitive edge thanks to lower costs, open access, and government support. This reality poses significant strategic challenges for Western companies, which may find themselves compelled to reconsider their business models and pricing strategies if they want to maintain their market share in the Global South.
