Open your phone. Every photo, every voice note, every argument you have ever had in a group chat, every song you saved at 2am. Now answer a simple question: where is it?

Not on your phone. Your phone is a window. The thing itself is sitting on a machine in a building, and for roughly 80% of the data Africans generate, that building is not in Africa.

Africa holds less than 1% of the world's data centre capacity. We are 18% of the world's people producing a fraction of a percent of the world's storage. Everything else lives somewhere else, on somebody else's floor, under somebody else's law.

The question is whether that actually matters. And people who have thought hard about it disagree.

What Is Actually True Right Now

Let us start with facts rather than feelings.

As of mid-2025 there were 223 data centres across 38 African countries. South Africa has 56 of them. Kenya has 19. Nigeria — over 200 million people, the continent's biggest music and film industry — has 17.

Meanwhile three companies (Amazon, Microsoft and Google) take more than 60% of global cloud spending. Six firms take over 70%. When a Nigerian startup, a Nigerian bank or a Nigerian social platform decides where to run, it is effectively choosing between a small number of foreign landlords.

This is changing, slowly. Over 36 African countries have passed or are advancing data protection laws. Nigeria's Data Protection Act came into force in March 2025 with annual compliance audits for major processors. The Central Bank of Nigeria has ruled that from January 2027, payment transaction data must be stored inside the country. Kenya now requires government data to stay within its borders. Rwanda enforces data residency and has set up one of Africa's first dedicated national AI agencies.

And infrastructure is being built. Kenya's Servernah Cloud launched in March 2026 as the country's first sovereign-hosted AI platform. Nigeria has Kasi Cloud positioning as an indigenous hyperscale facility. MTN is putting over $240m into a Lagos site. Open Access Data Centres is spending a similar figure expanding capacity. Nigeria's data centre market was worth $288m in 2025 and is projected to pass $1bn by 2031.

So something is happening. The argument is about whether it is the right something.

The Case That This Is the Whole Ball Game

The strongest version of the sovereignty argument is not about patriotism. It is about where value settles.

Data is not like oil, but the comparison survives for one reason: the money is made where the processing happens, not where the raw material comes from. Nigeria has spent sixty years learning what it means to export crude and import refined petrol. Sending raw behavioural data offshore, having it turned into models and ad systems and predictions, and then buying access back as a subscription is the same shape of arrangement in a different century.

Then there is the law problem. Data stored abroad answers to foreign law. A Nigerian citizen's records sitting on a server in another jurisdiction can be reached by processes no Nigerian court is part of and no Nigerian voter can influence. Your government cannot protect what it cannot reach.

And there is the part that hurts most for anyone building here: if African data trains models that are owned elsewhere, we become the training set rather than the beneficiary. The nurse in Enugu, the trader in Onitsha, the student in Ibadan — their behaviour becomes intelligence that is sold back to them, and none of the upside stays.

The Case That Sovereignty Talk Is a Trap

Now the counter-argument, which is stronger than people expect.

First: localisation is expensive, and somebody pays. Building sovereign infrastructure means capital, and it means power. Energy supply is the single biggest constraint on African data centre expansion — not law, not will, not talent. A country that mandates local storage before it can keep the lights on has legislated a problem it cannot fund.

Second: badly written localisation rules can cut you off from the best tools. If a Nigerian startup is legally barred from using the infrastructure its competitors use, sovereignty has made it slower and poorer. There is a version of this policy that protects nobody and simply raises the cost of building anything here.

Third, and most uncomfortable: a data centre in Lagos owned by a foreign hyperscaler is not sovereignty. It is the same landlord with a shorter cable. Moving the building without moving the ownership, the skills or the terms is sovereignty theatre — and it is exactly what a lot of current announcements amount to. The flag on the door is not the same as the deed.

Fourth: fragmentation is doing real damage. Fifty-four countries writing fifty-four different rulebooks does not create leverage, it creates regulatory arbitrage — and global firms are very good at shopping between rulebooks. Individually, no African country has the negotiating weight to set terms with a company that controls a fifth of global cloud.

Where the Real Question Is

I think both camps are answering the wrong question. The argument is framed as where should the servers be, and that is not the thing that decides who wins.

The questions that decide it are:

Who captures the value? Not where is it stored, but who books the revenue, employs the engineers, owns the model weights, and sets the price.

Who sets the terms? A continent negotiating as 54 separate customers gets 54 separate bad deals. The African Union's data policy framework and its continental AI strategy exist precisely because coordination is the only source of leverage available.

Who builds the thing on top? Storage is the least valuable layer. The value is in what runs on it — the models, the products, the platforms. A country can own every server and still import every application, which is just a more expensive dependency.

And there is one uniquely African asset nobody else can build: the languages. Nobody outside is going to do the hard, unprofitable work of building serious models in Yoruba, Igbo, Hausa, Swahili, Amharic, Zulu. If African teams do not, those languages simply become invisible to the systems that will run daily life. That is not a storage problem. That is a build-it-yourself problem, and it is the one place where the advantage is genuinely ours.

The Part Where I Am Not Neutral

I should be straight about my position, because you are reading this on a platform I have a stake in.

CelebIsland is African-owned. The people on it are mostly African. That is not a solved problem — owning a platform is not the same as owning the infrastructure under it, and I am not going to pretend otherwise. But it is a different starting point from building an audience inside somebody else's product and calling it a business.

The honest version is this: every African who builds a platform, a model, a dataset or a tool moves one small piece of the stack onto this side of the ledger. None of us moves it alone, and anyone claiming they have solved digital sovereignty by launching an app is selling you something.

So Where Do You Stand?

Is data sovereignty the fight that decides whether Africa participates in this century or just supplies it? Or is it a distraction — expensive, slow, and less important than simply building things people want, wherever the servers happen to sit?

And the personal one: does it actually bother you that your entire digital life is stored under a legal system you have no vote in? Most people say no until the day it matters.

Developers, students, founders, policy people — comments are open. This one I especially want argued with.

#DataSovereignty #Africa #AI #TechAfrica #Naija #DigitalRights #AfricanTech #DataProtection #CelebIsland #AfricaRising #Innovation #GenZ