Tech

Africa's AI boom: Who's building the pipes, and who's paying the price?

Data centres surge across the continent, but the real cost is just becoming clear.

Alex Novak|
Africa's AI boom: Who's building the pipes, and who's paying the price?
Photo by yanping ma on Pexels

Kigali, Rwanda — The hum of servers is the new soundtrack of African ambition. From Nairobi to Lagos, data centres are sprouting like acacia trees after a storm. But beneath the sleek facades and the press releases about 'digital transformation' lies a brutal question: who actually owns this infrastructure, and what are we sacrificing to get it?

The numbers are staggering. According to a 2025 report from the African Data Centre Association, total IT capacity on the continent is set to triple by 2028. Google, Microsoft, and Amazon are leading the charge, pouring billions into facilities that promise to bring cloud services closer to African users. The pitch is seductive — 'low latency,' 'data sovereignty,' 'economic growth.' But peel back the marketing speak, and you find a structure designed as much for extraction as for empowerment.

The ownership puzzle

Let's be clear: most of these data centres are not African-owned. They are built by American or Chinese tech giants, often in partnership with local governments or telecoms. The model is familiar — foreign capital, local land, and a steady stream of revenue that flows offshore. The continent gets jobs, yes. But it also gets a new kind of dependency.

Consider South Africa, the region's data centre hub. More than half of its capacity is controlled by three global players: Amazon Web Services, Microsoft Azure, and Google Cloud. Local providers like Teraco — owned by the British firm DigitalBridge — are the exception, not the rule. The pattern repeats across Kenya, Nigeria, and Ghana. The hardware hums on African soil, but the keys are held elsewhere.

“Africa is being wired into a global grid where the switches are pulled from Silicon Valley or Shenzhen. That's not sovereignty — that's servitude with a faster connection.”

The energy equation

Then there's the power problem. Data centres are gluttons for electricity. A single hyperscale facility can consume as much energy as a small city. In a continent where 600 million people still lack reliable access to electricity, that's not just a technical challenge — it's a moral one.

Rwanda's new 'AI Innovation Hub' in Kigali expects to draw 50 megawatts when fully operational. That's roughly 5 percent of the country's total generating capacity. The government touts the investment as a leap forward. But for every megawatt that powers a server, there's a megawatt that doesn't reach a village clinic or a school. The trade-off is rarely discussed in the ribbon-cutting speeches.

Renewable energy could be the answer. Kenya's Lake Turkana wind farm already powers some data operations. But the reality is that most new centres still rely on diesel generators or coal-fired grids. The carbon footprint of Africa's AI boom is growing faster than the renewable rollout can keep up.

The water war

And it's not just electricity. Data centres need water — lots of it — for cooling. In a continent increasingly parched by climate change, that's a ticking time bomb. The average hyperscale facility uses 1.7 million litres of water per day. In Cape Town, where the last 'Day Zero' water crisis is still fresh in memory, a proposed Google data centre sparked protests over water rights.

Local communities are right to be nervous. In arid regions of Morocco and Egypt, water-hungry data parks compete with agriculture for scarce supplies. The tech giants promise 'water-positive' operations by 2030, but those commitments feel distant when the taps are already running dry.

Who benefits?

The economic arguments for AI infrastructure are overblown. Yes, data centres create jobs — but mostly during construction. Once operational, a modern facility runs with a skeleton crew of engineers and security guards. The real profits are in the data itself: user information, transaction records, and machine learning models trained on African inputs but deployed for foreign gain.

Take the case of Nigeria. The country's fintech boom is powered by cloud infrastructure owned by Amazon and Microsoft. Every transaction, every loan application, every digital ID check generates data that flows through servers controlled overseas. The Nigerian government gets tax revenue (if it can collect it), but the strategic value — the intelligence — stays outside its borders.

“They're mining our data the same way they mined our gold and diamonds. The only difference is this time, the holes are in the cloud.”

The sovereignty mirage

The African Union's 'Digital Transformation Strategy' calls for data sovereignty — keeping African data on African soil. But sovereignty isn't just about geography. It's about control. A server in Nairobi owned by a U.S. corporation is still subject to U.S. law via the CLOUD Act. The promise of 'local data residency' is a comforting fiction.

Some countries are fighting back. South Africa's Protection of Personal Information Act (POPIA) sets strict rules for cross-border data flows. Kenya and Rwanda have passed similar laws. But enforcement is weak, and the tech giants have lawyers who can outlast any regulator. The result is a patchwork of rules that multinationals navigate while small African competitors drown in compliance costs.

The path forward

None of this means Africa should reject AI infrastructure. That would be Luddite folly. But the current model is a raw deal. The continent needs to demand more: local ownership stakes, binding water and energy commitments, and genuine technology transfer — not just server racks and user manuals.

There are glimmers of hope. The pan-African data centre operator Wingu has raised $200 million from African investors to build facilities that are fully continent-controlled. Ethiopia's new data protection law requires any foreign company processing citizen data to partner with a local firm that holds a majority stake. These are small steps, but they point in the right direction.

Africa has a choice. It can continue as a passive participant in the AI gold rush, selling its land and resources for the digital equivalent of beads and mirrors. Or it can demand a seat at the table — not as a supplier of cheap electricity and warm bodies, but as a true partner in building the infrastructure of the future.

The servers are humming. The question is whether the continent will hear its own voice above the noise.

Advertisement
#Africa#AI infrastructure#data centers#digital sovereignty#energy consumption
分享到:XfWB