Local AI compute infrastructure in Africa examines what it takes to move from investment in advanced computing infrastructure to sustained national capability and public value. Drawing on UNDP’s experience deploying country-hosted AI compute capacity in Kenya, Malawi, Rwanda, South Africa, Togo and Zambia, the publication captures practical lessons from design and procurement through deployment, activation and operation.
The experience shows that infrastructure alone is not enough. Effective use depends on the systems and capabilities around it, including accessible data, skilled operators and users, clear institutional responsibilities, governance and cybersecurity arrangements, sustainable financing, and defined pathways from priority applications to operational services.
Across the six countries, the programme identified 21 AI applications, while Rwanda has identified 42 priority applications across eight sectors. The emerging evidence shows that progress is strongest when compute capacity is connected to clear institutional ownership, usable data, appropriate skills and concrete public-interest applications.
The publication offers governments, development partners, infrastructure operators and the wider AI community practical considerations for building local compute, putting it to use, governing access and value, and financing and scaling infrastructure over time.