In the hills above Lusaka, a project is underway that its backers believe will reshape how Africa participates in the global artificial intelligence economy. Cassava Technologies, the pan-African technology group, has begun construction on what it describes as the continent’s first network of data centres purpose-built to serve AI workloads—not just traditional cloud computing, but the demanding, power-intensive computations that underpin machine learning, language modelling and automated decision systems.
The project, called Mufungi, is being built in phases across multiple countries, with Zambia as its anchor location. Its ambition is starkly at odds with Africa’s current position in the global AI value chain: the continent generates enormous amounts of data that feeds AI models worldwide, yet captures a negligible fraction of the economic value that data creates. Mufungi is designed to change that equation, keeping African data on African soil and giving continental businesses, researchers and governments access to processing power that has historically required outsourcing to servers in Europe, North America or Asia.
The challenge is formidable. Traditional development finance institutions move too slowly for a sector where compute demand doubles every eighteen months. Cassava’s CEO has acknowledged that the project is therefore being financed substantially through private capital—an approach that makes financial viability non-negotiable, but also means that pricing for Mufungi’s services will need to compete with established international cloud providers. For African businesses accustomed to paying premium rates for infrastructure located thousands of kilometres away, the proposition is compelling. For Cassava, the margin for error is narrow.
The geopolitical dimension of the project is difficult to ignore. AI infrastructure is increasingly viewed through a lens of strategic competition, with the United States and China jostling for influence over the digital foundations that will shape the next half-century of global economic power. Africa’s role in this competition is still being defined. Projects like Mufungi represent a bet that the continent can develop indigenous AI capacity rather than simply becoming a passive consumer of platforms built elsewhere. Whether that bet pays off will depend heavily on whether African governments create the regulatory environments—data governance frameworks, spectrum allocation for connectivity, electricity policy—that allow AI infrastructure to scale.
Connectivity remains a foundational constraint. Many of the regions Cassava hopes to serve lack reliable fibre connectivity, and the energy requirements of AI-optimised data centres are enormous—far beyond what most African national grids can reliably provide. The Lusaka facility is being powered partly through renewable sources, but the broader network will require significant investment in generation and distribution infrastructure that does not yet exist at the necessary scale.
For Africa’s technology entrepreneurs, the Mufungi project is a source of cautious optimism. Access to local compute reduces latency, lowers costs and creates opportunities for AI applications tailored to African contexts—agricultural forecasting, medical diagnosis in low-resource settings, financial inclusion for the unbanked. But it also raises questions about governance and ownership: who controls the infrastructure, who can access it, and who captures the value it creates. These are questions that Africa’s policymakers have yet to fully engage with, and their answers will determine whether Mufungi becomes a landmark in the continent’s digital development or another piece of infrastructure that primarily benefits outside investors.
Construction is expected to continue through 2027, with the first nodes scheduled to come online before the end of next year.