Africa's data centre boom is accelerating. Driven by cloud adoption, digital transformation and growing interest in artificial intelligence (AI), new facilities are being announced across the continent at an unprecedented pace.
The continent's data centre market is projected to grow from $3.49 billion in 2024 to $6.81 billion by 2030, while it already has approximately 360MW of active capacity, with a further 238MW under construction and 656MW planned, according to the Africa Data Centres Association.
Alongside this growth has come an increasingly urgent debate about electricity supply, renewable energy and sustainability.
These are important considerations. Power matters enormously, particularly in markets where grid capacity and reliability can determine whether a project is feasible.
However, before we ask how to power and cool Africa's next data centre, we should ask a more fundamental question: is the proposed location the right one for that facility and the market it is meant to serve?
This is not an argument against building data centres in Africa. Quite the opposite. It is an argument for recognising that Africa is not a single market, and that what makes a project successful in one country or city may be very different in another.
Location matters more than power
When people talk about Africanising data centre design, they often mean adapting European or American approaches for African conditions. What receives less attention is that projects often need to be adapted within Africa as well.
A facility in Limuru, outside Nairobi, faces very different design considerations to one in Lagos. At roughly 2,230 metres above sea level, free cooling is viable for much of the year in Limuru. In Lagos, where temperatures and humidity are significantly higher, cooling requirements are far more demanding.
The same applies to land availability, connectivity, construction materials, skills and energy infrastructure.
In Nairobi, land is limited and connectivity drops off quickly outside key areas, often pushing developers towards compact or multi-storey facilities. South Africa's industrial nodes create opportunities for larger campus-style developments. In Nigeria, power and connectivity constraints frequently become dominant design considerations.
Each market presents a different combination of opportunities, constraints and commercial realities that should shape how a facility is designed and where it is developed.
Risks of imported templates and 'future-proofing'
We were recently brought into a project in Nigeria where the original consultants had specified a steel structure using imported precast components. On paper, the design worked. In practice, it was poorly aligned with local construction realities and procurement challenges.
By the time the project reached tender stage, costs had escalated to the point where the development was no longer commercially viable. The solution was to redesign the facility around materials, skills and construction methods readily available in the local market.
Another assumption I increasingly challenge is the idea that data centres can be future-proofed with certainty. The industry is changing too quickly for that.
AI has already changed assumptions around density, capacity and infrastructure requirements. Yet in many African markets, demand continues to be driven primarily by cloud adoption, enterprise services and digital transformation.
Facilities can be designed with flexibility, but not certainty. Designing around assumptions borrowed from the world's largest AI markets can result in infrastructure that is overbuilt, unnecessarily expensive and disconnected from actual demand.
The right facility, in the right place, for the right reason
Demand for digital infrastructure across Africa will continue to grow. Cloud adoption, enterprise modernisation and digital transformation remain powerful long-term drivers, and AI applications will increasingly become part of that mix.
The challenge is that Africa's opportunity will not necessarily mirror the trajectory of the United States, China or Europe. Much of the global investment flowing into AI infrastructure is focused on training large language models and supporting hyperscale workloads.
In many African markets, the more immediate need is infrastructure that supports cloud services, inference workloads and practical AI applications tailored to local requirements.
That distinction matters because it changes the conversation around scale, investment priorities and what constitutes a successful project.
The projects most likely to succeed will be those built around genuine demand, strong connectivity, realistic economics and a clear understanding of local conditions.
Access to renewable energy may attract investment, and AI may be fuelling global demand, but neither replaces the need for a compelling business case grounded in local realities.
Africa's data centre opportunity is significant, but success will depend less on how quickly facilities are built and more on whether the right facilities are built in the right places, for the right reasons.
Ultimately, every developer, investor and operator should ask the same question before a project moves forward: does this facility need to exist in this market, for this use case, at this point in time?
It is often the hardest question in the development process, but it is also the most important. Getting that answer right will determine whether Africa's next generation of data centres delivers lasting value or simply adds capacity where it is not needed.
Shaun Versfeld, Director: Mission Critical Facilities at Atana, is speaking at the Pan African DataCentres (PADC) Exhibition & Conference in Johannesburg on 23 and 24 June 2026. He is available for interviews on the themes explored in this article, including AI infrastructure, data centre investment, digital infrastructure growth and the challenges of designing data centres for diverse African markets.
Share
