Africa generates approximately one percent of global AI compute capacity while accounting for seventeen percent of the world's population. This is not a market gap. It is a sovereignty gap. Every African AI model trained on foreign infrastructure is subject to foreign export controls, foreign pricing, and foreign uptime guarantees. The continent's AI development cannot be built on indefinitely borrowed compute.
This project is a rigorous feasibility study for a continent-scale AI compute facility. The study addresses four questions: what capacity is required to serve African AI demand at a ten-year horizon; where the facility should be located, evaluating power grid stability, cooling water availability, and fibre connectivity across five candidate countries; what supply chain partnerships are required to source compute hardware without sanctions exposure; and what public-private financing structure makes the economics viable.
The demand model developed in phase one projects training and inference compute requirements across three scenarios: a base case reflecting current AI adoption trajectories, a growth case reflecting accelerated enterprise AI adoption driven by increasing LLM accessibility, and a sovereign case reflecting government and public sector AI deployment at scale. Each scenario produces a different facility sizing requirement and a different timeline for when capacity constraints become binding.
The site evaluation assessed candidate locations across five countries on power grid reliability, renewable energy availability, water access for cooling, fibre backbone connectivity, and political stability indicators relevant to a long-horizon infrastructure investment. The shortlist from phase one is being used to develop detailed technical and financial models in the current phase.
Phase two is developing the technical architecture specification and the financing structure for presentation to development finance institutions and sovereign wealth funds. Current status: technical architecture in development, financing term sheet in preparation.
