The AI race continues to intensify, with a clear distinction emerging between nations that possess the computing power to build cutting-edge AI systems and those that don’t. According to a New York Times report citing research from Oxford University, the distribution of the world’s most powerful data centers reveals a stark divide.
The leaders in this race are the European Union with 28 data centers, the United States with 26, and China with 22. Together, they house more than half of the world’s AI-capable data centers identified in the study. Only 32 countries, mostly in the northern hemisphere, have the facilities with the necessary computing power to run generative AI systems.
“We have a computing divide at the heart of the AI revolution,” warns Lacina Koné, director general of Smart Africa. “It’s not merely a hardware problem. It’s the sovereignty of our digital future.” The high costs associated with training and running generative AI models are a significant barrier. The cost of graphics processing units needed to train AI models can be tens of thousands of dollars each, with a single data center potentially housing tens of thousands of these AI chips.
For countries without access to such facilities, remotely renting computing power is the only option, which is less than ideal. Kenyan startup Qhala, working on a large language model for African languages, faces this challenge. Their engineers must schedule their work during early morning hours when American programmers are asleep to experience faster speeds.
“Proximity is essential,” says Qhala founder Shikoh Gitau. Kate Kallot, founder of Amini, another AI startup in Kenya, echoes this sentiment: “If you don’t have the resources for compute to process the data and to build your AI models, then you can’t go anywhere.”
The uneven distribution of AI computing power has effectively divided the world into two camps: nations reliant on China and those dependent on the United States. Both countries are poised to build significantly more data centers than others.
While concerns about an AI bubble, the high costs of generative AI models, and their propensity for errors and safety risks are valid, the reality remains that entire economies are being swept up by the AI gold rush. Until the technology is deemed worthless, AI will remain a major lever of power for those who have it over those who don’t.
As Vili Lehdonvirta, an Oxford professor involved in the data center research, notes, “Oil-producing countries have had an oversized influence on international affairs; in an AI-powered near future, compute producers could have something similar since they control access to a critical resource.”
