• MTN Group plans to develop AI-ready data centres through Africa Data Hub Holding, a venture involving an undisclosed UAE-backed investor.
• The first phase is expected to target about 150MW of capacity across Nigeria and South Africa—not 150MW already operating in Nigeria.
• MTN is expected to be a minority investor while its partner provides most of the capital and technical expertise. The plan is not a new AI feature, a completed facility or a promised consumer benefit.
For most people, AI looks like a chat box, a design tool or a smarter search result. For the person trying to build one, it is also servers, power, cooling, networks and the cost of getting enough computing capacity when an idea finally starts to grow.
That is the less glamorous reason MTN’s newest infrastructure plan deserves attention. The company is not announcing an AI app for every phone. It is talking about the machinery that would sit underneath the apps, services and experiments builders may want to run.
The 150MW figure is a target, not a finished data centre
Reuters reports that MTN plans to develop AI-ready data centres in South Africa and Nigeria through Africa Data Hub Holding, a venture with an undisclosed UAE-backed investor. The initial phase is expected to target about 150 megawatts of capacity across the two countries.
That wording matters. “Across Nigeria and South Africa” is not the same as a 150MW facility already built in one Nigerian city. It is an expected first-phase target for a proposed development, with later expansion expected to depend on demand.
MTN says it would be a minority investor in the venture. Reuters reports that the partner, which has experience building data centres in the United Arab Emirates and other Gulf markets, would provide most of the capital and technical expertise. So this is a joint infrastructure ambition, not a completed MTN-owned AI campus.
Why builders may notice infrastructure before users do
A strong AI product needs more than a clever prompt. Teams need places to host models, process data, store files and run workloads that may become expensive as more people use them. If more serious computing capacity becomes available closer to the markets where African companies operate, it could become relevant to startups, cloud customers and technical teams before it changes what the average phone user sees.
That possibility should not be mistaken for a guarantee. The plan does not promise cheaper AI tools, faster apps, new jobs, better mobile data or a particular service for Nigerian developers. Those outcomes would depend on what gets built, how it is priced, who can use it and whether the wider power and connectivity picture can support it.
But builders are usually the first people to ask practical questions: where will the compute run, what will it cost, how reliable is it and how far away is the infrastructure? A larger local or regional data-centre push can matter because those questions shape what a team can afford to attempt.
AI needs a physical home
It is easy to talk about AI as if it lives entirely in the cloud. The cloud still has an address. It is made of racks, cables, cooling systems, security controls and power-hungry equipment that needs to keep working long after a demo ends.
Nairametrics reports that the venture is expected to focus first on Nigeria and South Africa, with additional expansion tied to demand for computing capacity. That points to the scale of the bet: not just selling connectivity to AI users, but trying to be closer to the infrastructure layer those services require.
MTN Nigeria also has a separate, already launched Sifiso Dabengwa Data Centre with reported 9MW capacity. That existing facility should not be confused with the planned 150MW Africa Data Hub phase. They are different pieces of an evolving infrastructure story.
What to watch before calling it an AI breakthrough
The real milestones will come later: detailed project information, a timeline, capacity actually commissioned and evidence of who can access the infrastructure. Until then, the 150MW figure is an expectation attached to a plan, not capacity that developers can assume is ready today.
Still, the direction is worth watching. Nigeria’s next AI story may not begin with another viral tool. It may begin with the quieter work of creating enough dependable computing infrastructure for local teams to build on.
Also read: Your banking app has a cloud deadline behind it. Why Nigeria’s new push matters
Also read: MTN wants a bigger role in your wallet. Why exploring banking is not an app launch
Do you think more computing infrastructure would make a real difference for Nigerian builders, or is access still the bigger issue? Share your view in the comments.
