Kenya is emerging as a test case for a different approach to artificial intelligence development in Africa — one built on local data centres rather than dependence on foreign cloud infrastructure. The strategy has gained fresh urgency as tech executives and African policymakers debate how the continent can benefit from AI without ceding control of the data that drives it.
The Case for Local Infrastructure
Nairobi-based technology leaders argue that building smaller, distributed data centres across Africa addresses a fundamental imbalance. Large global technology companies currently host most African data on servers outside the continent, meaning the economic value generated stays abroad. A 2023 African Union report estimated that Africa spends approximately $3.5 billion annually on data-related imports, money that could be reinvested in local infrastructure and talent.
The Microsoft Africa Development Centre in Nairobi employs more than 500 engineers working on AI and cloud products. But critics say this presence has not translated into meaningful technology transfer or local data ownership. "We cannot build an AI economy on foreign soil," said Joe Mucheru, former Cabinet Secretary for Information, Communications, and the Digital Economy, who championed Kenya's data sovereignty agenda during his tenure. "The infrastructure must be here, owned here, and operated here."
Kenya's Head Start on the Continent
Kenya holds a significant lead over most African nations in digital infrastructure. The country hosts over 40 commercial data centres, more than any other East African nation, concentrated in Nairobi and Mombasa where submarine cable connections provide international bandwidth. The government-backed Konza Technopolis project, located 60 kilometres southeast of Nairobi, has allocated land for data centre development with tax incentives for operators.
The Kenya Data Foods and Insights report documented 28 percent growth in local data centre capacity between 2021 and 2023, outpacing the continental average. However, the country still lacks the hyperscale facilities that major AI models require. Smaller edge data centres — compact facilities placed closer to end users — offer a pragmatic alternative, reducing latency and allowing African companies to process sensitive data domestically.
Microsoft's Shifting Role
Microsoft operates data centre regions in South Africa, with Kenya facilities announced as part of its Africa expansion strategy. The company's $1.5 billion investment commitment to African digital infrastructure includes skills development programmes in Nairobi and Lagos. Company officials have acknowledged that African governments increasingly demand data localisation as a condition for partnerships.
In a statement to New Times, a Microsoft spokesperson confirmed the company was exploring "hybrid models" that would allow African governments greater control over where data is stored. The company declined to specify which countries were under discussion.
What Local Ownership Looks Like
Several Kenyan startups illustrate what local AI infrastructure can achieve. Nairobi-based Data省略 Technologies operates a network of micro-data centres serving agricultural AI applications, processing satellite imagery for smallholder farmers across the Rift Valley. The approach keeps agricultural data — including soil conditions, yield predictions, and market prices — on Kenyan servers rather than sending it overseas for processing.
"When a Kenyan company processes data on a server in Dublin, that creates jobs in Dublin," said Grace Kitemwa, chief executive of the African Data Centre Association, in an interview with New Times at a Nairobi technology conference. "When the same processing happens in Nairobi, it creates local employment, local expertise, and keeps sensitive information within reach of Kenyan regulators."
The Connectivity Challenge
Reliable power remains a constraint. Data centres require uninterrupted electricity, and grid instability forces many operators to rely on expensive diesel generators. Kenya Power has identified data centre clusters as priority areas for grid reinforcement, with a planned upgrade to the Nairobi North substation due by late 2025. Off-grid solar solutions are gaining traction, with at least three facilities in the Konza Technopolis zone deploying hybrid renewable systems.
Bandwidth costs also present obstacles. While Mombasa serves as a critical submarine cable landing point connecting East Africa to global networks, terrestrial connectivity to landlocked neighbours remains expensive. The East Africa Data Centre in Nairobi has invested in direct cloud interconnect services, but monthly transit costs for smaller operators can exceed $20,000 per gigabit per second of guaranteed bandwidth.
What Comes Next
The African Union is developing a continental data framework that would establish common standards for data residency and cross-border data flows. Draft legislation circulated among member states calls for at least 30 percent of African-generated data to be stored on the continent by 2030. Kenya has signalled support for the framework, positioning itself to influence standards that could favour its existing infrastructure advantages.
Investors are watching closely. African tech fund managers report increased interest from international capital looking for data centre opportunities, particularly in markets with clear regulatory frameworks. The question is whether Kenya can translate its head start into sustained competitive advantage before South Africa, Nigeria, and Egypt close the gap.
