MTN has set up a joint venture to build data centres in Africa, starting with South Africa and Nigeria. It has not said how much it is investing, how much capacity it intends to build, or when.
The vehicle is Africa Data Hub Holding Limited, formed with a United Arab Emirates data centre investment platform founded by Tarek Al Ashram. MTN Group Digital Infrastructure announced the partnership on 27 August 2026, describing it as a platform to develop and scale AI-ready data centre capacity for cloud, enterprise, and artificial intelligence workloads.
Bayobab, MTN’s wholesale fibre and subsea business, is a shareholder in the venture and will supply open-access connectivity and go-to-market support through its pan-African network.
What has been committed, and what has not
The named elements are real. There is an incorporated vehicle, an identified partner, a defined starting geography and a stated role for Bayobab.
Everything a reader would use to judge the scale is absent. The announcement carries no investment figure, no megawatt target, no timeline, no named sites and no first project. It does not say whether the venture will build, buy or lease, nor what MTN is contributing beyond connectivity and its balance sheet.
“This partnership is a step forward in execution of our digital infrastructure strategy,” said Mazen Mroué, chief executive of MTN Group Digital Infrastructure.
That is a fair description of what has been disclosed. It is a corporate step rather than a construction announcement, and the distinction matters when the same operator has just committed nearly R20 billion of capital expenditure in a single half and simultaneously announced a share buyback of up to R6 billion. Against numbers of that size, a data centre platform with no disclosed capital is not yet a material commitment.
Why an operator wants data centres
The logic is straightforward. MTN already owns the fibre and the subsea capacity through Bayobab, and it already sells connectivity to enterprises. Data centres are the layer between those two, and the operator that owns the route into a facility captures more of the spend from cloud or AI customers.
Doing it through a joint venture rather than on balance sheet is also readable. Data centres are capital-hungry and slow to fill, and telecoms shareholders have historically punished operators for tying up capital in property. A partner that specialises in data centre investment carries part of that weight and brings development experience MTN does not have in-house.
The two markets chosen
South Africa and Nigeria are the obvious openers, and they are not equivalent problems.
South Africa has the continent’s deepest colocation market, established operators, subsea landings at both coasts and a functioning interconnection ecosystem. It also has a constrained grid, which is why every serious facility there carries its own generation.
Nigeria has the demand and the population, but its data centre growth depends on the power supply more than on capital or land. An AI-ready facility is a dense electrical load. Announcing one in Lagos is easy; energising it reliably is not, and the release does not address power at all.
What would make this a story worth returning to
Three disclosures would turn this from intent into substance: the capital committed and by whom, the first site with a target energisation date, and the power arrangement behind it.
Until then, the useful reading is directional. MTN is signalling that it intends to own compute capacity as well as the network that connects to it, and it has found an investor willing to fund that ambition. Whether it becomes concrete is a question for the next set of results, not this announcement.
Leave a Reply