Oluwaseyi Ayodeji
Ever so often, another headline declares a winner in Africa’s data centre race. South Africa is pulling ahead. Kenya just landed a billion-dollar deal. Nigeria is falling behind. Read enough of these, and you start to believe this is a contest between three countries, decided by whoever signs the biggest cheque or announces the most facilities.
It isn’t. The real contest is each of these countries against its own electricity grid, and the country that wins will not be the one with the most press releases. It will be the one that keeps the lights on.
On paper, South Africa is way ahead. It has 56 data centres (~350 megawatts), more than any other African market, and is widely reported to account for roughly 70% of the continent’s data centre capacity. That lead was built over more than a decade of investment from global cloud and data-centre operators, including Amazon, Microsoft and Google.
Nigeria, by contrast, is still building out its capacity market by market, with Lagos alone currently running around 78.6 megawatts of installed data-centre capacity and a pipeline projected to push past 218 megawatts by 2030. That is genuine, meaningful growth and nothing to be sneezed at. But set against South Africa’s existing base, the gap looks like a straightforward story of one country executing and another catching up.
It is not that simple, and the reason why is the entire point of this piece.
The capital flows tell you where investors currently believe the race is being won. South Africa recently secured a $438 million Equinix investment plan that includes a further 160 megawatts of data-centre capacity, while Kenya has drawn ambitious mega-project commitments, including the Microsoft-G42 initiative centred on a geothermal-powered facility with an initial target of 100 megawatts and an eventual ambition to scale towards 1 gigawatt.
Set against those numbers, Nigeria’s pipeline can look like it’s playing catch-up in a game it is already losing. Investment committees read headlines like these and draw conclusions quickly: the capital is voting with its feet, and it is voting for South Africa and Kenya over Nigeria.
I spent years managing large-scale cloud infrastructure programmes, the kind where a single facility’s timeline depends on dozens of workstreams landing in the right order: land acquisition, power interconnection studies, equipment procurement, construction, commissioning and testing, often running in parallel across multiple continents and vendors. If there is one lesson that experience taught me, it is this: a capital commitment is not a completed facility, and a signed deal is not delivered power.
The gap between an announcement and an operating asset is where infrastructure projects actually live or die, and it is the gap the headlines rarely cover, because a press release is easy to write and a commissioned facility takes years of unglamorous execution.
In 2024, Microsoft and the UAE-based firm G42 announced a $1 billion digital investment initiative in Kenya, centred on a geothermal-powered data centre in Olkaria, near Naivasha. The first phase was planned at around 100 megawatts, with the potential to scale towards 1 gigawatt. The broader initiative also included a new East Africa cloud region, local-language AI development, connectivity investments and digital-skills programmes.
The data-centre project has since stalled. Reporting in 2026 indicates that disagreements over the amount of power capacity the Kenyan government could commit to the project contributed to the breakdown in negotiations, with the original May 2026 launch target no longer achievable. The project has not simply disappeared, but the original timetable and scope are clearly no longer proceeding as initially announced.
This is not a story about Kenya’s failure. It is a story about what happens when even a well-funded, well-intentioned, geothermal-powered project runs into a hard physical and institutional constraint that money alone cannot remove. Kenya has a relatively diversified electricity mix and substantial geothermal resources, and it still found itself having to reconsider how much power could realistically be committed to a single digital infrastructure project.
That should tell every policymaker in this region something important about what actually determines who wins this race. It should also temper how confidently anyone declares a winner based on deal announcements alone.
Nigeria’s problem is not that it has no generation capacity. It is that the amount of power actually available to the grid is dramatically lower than the country’s installed capacity. Nigeria has roughly 13 – 14 gigawatts of installed generation capacity, but available generation is typically in the 4 – 6 gigawatt range, depending on gas supply, plant availability, transmission constraints and outages. South Africa, by comparison, has an electricity system with roughly 48 gigawatts of installed Eskom capacity, although its reliably available output is substantially lower than that figure.
That is not primarily a technology gap or a capital gap. It is a power-security gap, and it is a major reason South Africa’s data-centre lead has proved so durable.
A hyperscale AI facility needs enormous, continuous, reliable electricity, the kind of load that a fragile or frequently disrupted grid simply cannot guarantee, no matter how much capital or ambition sits behind the project. South Africa’s data-centre ecosystem did not emerge because South Africa wanted data centres more than Nigeria did, or because its founders and operators were somehow more capable. It emerged in an environment where the underlying power and infrastructure systems gave global operators greater confidence to make decade-long investment decisions.
This is precisely the constraint I watched shape infrastructure decisions from the inside for years: site selection for large compute facilities almost always comes down to power availability before it comes down to almost anything else, including tax incentives, land cost or even connectivity. A country can offer every other advantage on paper and still lose the deal the moment a serious operator runs the power feasibility studies.
It would be a mistake to treat this purely as a technology competitiveness story. Every gigabyte of data processed abroad, in a facility Nigeria does not host and does not control, represents a small transfer of digital leverage to whichever country does. As African economies build more of their financial systems, government services and AI products on top of cloud infrastructure, the question of where that infrastructure physically sits stops being merely a business decision and becomes a question of national standing and strategic autonomy.
Nigeria’s own financial regulator is already treating it this way. In June 2026, the Central Bank of Nigeria (CBN) issued a circular requiring financial institutions and participants facilitating payments within Nigeria to ensure that payment transaction data generated within Nigeria is stored and managed within Nigeria, with compliance effective from January 1, 2027. The circular explicitly identifies the localisation of critical payment data as part of its effort to strengthen the resilience and oversight of the Nigerian payments system.
That is more than a compliance measure. It is an acknowledgement that critical digital infrastructure cannot be separated from national control. The CBN is telling financial institutions that if Nigerian payment data is important enough to regulate, it is important enough to keep within the country’s jurisdiction. And once that data must remain in Nigeria, the physical infrastructure required to store, process and protect it becomes a national infrastructure issue as well.
A country that cannot reliably power its own data centres will, over time, find itself renting critical digital capacity from foreign providers on terms it did not set. That is not a hypothetical. It is the direction all three countries in this race are confronting, at different speeds, for the same underlying reason.
Closing this gap will not come from another facility announcement or foreign investment press release. Nigeria has spent decades announcing power-sector reforms, generation projects and transmission investments while businesses and households continue to contend with an electricity system that cannot reliably deliver power across the country. The problem is no longer a lack of understanding of what needs to be built; it is the country’s chronic inability to execute, maintain and scale what it already knows it needs.
For the data-centre economy, that means treating reliable electricity as digital infrastructure in its own right. Nigeria needs sustained investment in transmission capacity, enough generation to provide genuine reserve capacity, and a grid with sufficient redundancy to withstand failures without repeatedly taking large parts of the country offline. Generation cannot be the only measure of progress. Power that exists on paper but cannot be transmitted consistently to where it is needed is not useful capacity, and megawatts that disappear every time the grid trips do not give a hyperscale operator the certainty required to commit billions of dollars.
Most importantly, Nigeria needs to stop treating persistent power outages as an unavoidable feature of economic life. A hyperscale data centre cannot be built around the assumption that the national grid will fail and a diesel generator will take over. Nor can an AI economy serving hundreds of millions of people be powered indefinitely through thousands of individual captive-generation arrangements. Those solutions may work at the facility level, and in many cases they are necessary, but they do not solve the national infrastructure problem.
The country does, however, have the ingredients for a different outcome. Nigeria has a large domestic gas resource base that remains significantly underutilised for power generation, alongside growing private-sector interest in embedded generation specifically designed to serve data-centre and industrial demand. The raw materials for a different outcome exist, both in the ground and in the willingness of private operators to invest around a clearly defined need.
What has been missing is the same sustained execution focus that data-centre operators are already applying to their own captive power projects, but applied at national grid scale rather than at the fence line of a single facility. If a private operator can commission a dedicated 100-megawatt gas plant to guarantee power for one facility, the underlying capability to solve this problem clearly exists inside the country. The challenge is applying that same urgency, discipline and coordination to the national system, and sustaining it long enough for businesses to stop planning around the assumption that the grid will fail.
Nigeria does not need to reinvent the technology or wait for some future breakthrough to solve this problem. It needs to make reliable power a national infrastructure priority, execute consistently across generation and transmission, and create enough confidence in the system that data-centre operators can build for the next decade rather than simply insure themselves against the next outage.
So the next time you see a headline declaring that Nigeria is losing the data-centre race to Kenya or South Africa, ask the more useful question instead: losing to whom exactly, and on what terms?
South Africa’s lead is real, but it is largely a product of infrastructure and market development built over decades, not evidence of a technology gap Nigeria cannot close. Kenya’s experience shows that even a hugely ambitious, well-funded project cannot simply wish away the infrastructure and institutional constraints around power capacity. That should reshape how confidently anyone ranks these three countries against each other based on investment announcements alone.
The country that ultimately leads Africa’s AI infrastructure future will not be the one that announces the most gigawatts. It will be the one that delivers them, reliably, at scale, year after year.
That is a power-sector achievement before it is ever a technology one, and it is the metric this region should actually be tracking, far more closely than it tracks facility counts or investment announcements.
What would it take for Nigeria to make grid reliability its top digital economy priority? I’d like to hear where you land.
* Ayodeji is a senior programme leader in AI/Cloud Infrastructure with two decades in the tech industry