Connected, But Not Yet Earning: Africa’s Next Digital Divide

Connectivity is expanding faster than at any point in the continent’s history. Whether getting online changes what people can build, sell and earn is a separate question, and it is the one that should define the next phase of digital inclusion.

A young designer in Kenya can own a 4G smartphone, hold more than one mobile-money account, learn from YouTube and prompt an AI model, and still have no dependable way to earn a shilling from it. Nothing on that list is missing. What is missing sits one layer above the device: a paying customer, a contract, a payment she can trust, and a way to be found by people who don’t already know her.

For most of the past two decades, Africa’s digital-divide debate asked one question: can you get online? The old divide asked whether you could get online. The next one asks what you can do once you are there.

The old divide asked a simpler question

The connectivity gains are real. Globally, 96% of people now live within mobile-internet coverage, and 4.7 billion, about 58% of the world, use mobile internet on their own device. The harder number sits underneath: some 3.1 billion, roughly 38% of the world, live within coverage but do not use mobile internet, a shortfall the GSM Association (GSMA) calls the usage gap in The State of Mobile Internet Connectivity 2025. It is about ten times the coverage gap that remains.

Africa sits at the sharp end. By the GSMA’s The Mobile Economy Africa 2025, about 416 million people use mobile internet across the continent, a penetration rate of roughly 28%, while some 960 million, 64% of the population, live where coverage exists but are not online.

Sub-Saharan Africa is the world’s least-connected region, with the largest usage gap. Cost is much of the reason: by the International Telecommunication Union’s (ITU) Facts and Figures 2024, an entry-level 2GB mobile-broadband plan in Africa cost about 4.2% of income per capita in 2024, the highest of any region and more than double the 2% affordability target. Coverage has run ahead of use.

Coverage, usage, and the questions that come next

But even that distinction stops one step short: being online tells you little about what being online is for.

I find it helps to read the divide in four stages. The first is the coverage gap: people no network reaches. The second is the usage gap: people covered but not connecting. The GSMA tracks both. The two I want to add are not official categories, only a way to extend it. The third is a productive-use gap, between people online mainly to consume and those who use the internet to produce income. The fourth is an economic-participation gap: people connected and capable but still lacking a route to customers, jobs, trusted payment and repeat work.

None of this argues against entertainment; consumption is legitimate, and often the on-ramp to everything else.

Geography did not disappear when the networks arrived

35%	of Kenyans aged three and above had actually used the internet in the 2023/24 CA/KNBS survey – even though 4G now reaches 97.3% of the population.

Kenya shows the gap cleanly, because its access figures are strong. By the count of the Communications Authority of Kenya (CA), 4G reached 97.3% of the population by mid-2025. Yet a joint CA survey with the Kenya National Bureau of Statistics (KNBS), the Analytical Report on ICT Based on the 2023/24 Kenya Housing Survey, found that only about 35% of Kenyans aged three and above had used the internet in the months before it, rising to 56.6% in urban areas against 25% in rural ones, and peaking at 64.7% in Nairobi. Near-universal coverage, far-from-universal use, concentrated in the cities.

Building WorkKE, a small African digital-work marketplace, has made that pattern hard to ignore. In our own data for the year to August 2026, 78.3% of registered users were in Nairobi. That number does not describe Kenya; it describes one early platform shaped by its own signup channels, and should not be read as national evidence. But it raises a fair question: if coverage is increasingly national, why does participation on new digital-work platforms still cluster so tightly in one city?

The same dataset holds a second pattern. Of about 7,140 people who registered as freelancers, 757 sent at least one proposal, 39 won at least one job, and 33 completed paid work. This is one young marketplace over a short window, not a verdict on anyone’s economy; registering is not earning. But that funnel, many registered and far fewer transacting, is exactly what the productive-use and economic-participation gaps are meant to examine.

A transaction between strangers needs more than a connection. It needs identity, reputation, protected payment and a way to settle a dispute.

Access is not the same as economic access

Full technical access is not the same as economic access. The scaffolding a first-time buyer needs, identity that can be checked, a reputation that travels, payment held until work is delivered, a way to settle a dispute, is what separates a market from a directory. For the small sums that dominate early marketplaces it is decisive: on WorkKE the median completed job was worth about KES 2,000, and one bad transaction at that size can end someone’s willingness to try again.

25.7%

of mobile-money accounts worldwide were active in a given month in 2025, even as about $1.4 trillion flowed through them in Africa – roughly two-thirds of the global total.

Payments make the same point at scale. Africa’s mobile-money system is the world’s most developed. By the GSMA’s The State of the Industry Report on Mobile Money 2026, about $1.4 trillion moved through mobile-money accounts on the continent in 2025, close to two-thirds of the global total. Yet worldwide only about a quarter of accounts, 25.7%, were active in a given month, so most sit idle.

Receiving money from a relative is not the same as invoicing a client, holding funds in escrow, issuing a refund or taking payment from abroad. Encouragingly, merchant payments, still a small share, were the fastest-growing use, rising by almost half to $155 billion in 2025, the rail beginning to carry commerce, not only transfers.

From places that got us online to places that help us earn

This is where Africa’s cybercafés become interesting again. They were built to answer the access question when getting online meant leaving home. As home connectivity spreads, that purpose fades, but another may open. A neighbourhood café already has devices, power, a printer and someone who knows how to use them. It is not a large step to a place that helps people build a profile, prove who they are, apply for work and get paid.

WorkKE has more than 180 cybercafés represented across its wider network, and I want to be careful what that means. It is not 180 working digital-work hubs. It is a hypothesis we are testing: that infrastructure built to help Africans get online might evolve into infrastructure that helps them earn, cheap, local and already trusted. Whether it holds at scale is an open question, not a claim.

AI will not close the gap on its own

0.4% → 4%

the IMF’s modelled 10-year effect of AI on Sub-Saharan GDP, under current conditions versus wider adoption. The Fund calls it a diagnosis of today, not a forecast.

Artificial intelligence is about to stage the same argument again. The instinct is to ask who can access the tools; the better question is who can turn access into income. A 2026 International Monetary Fund (IMF) assessment, Unlocking the Potential: AI in Sub-Saharan Africa, put numbers on the stakes: under current conditions it estimates AI would add only about 0.4% to the region’s GDP over a decade, but under wider adoption and stronger infrastructure, close to 4%.

The Fund describes these as a diagnosis of today’s conditions, not a forecast. The technology is identical; the difference is the capacity to use it. Sub-Saharan Africa ranks last on the IMF’s AI preparedness index and hosts only about 160 data centres, roughly 5.5% of the world’s installations and under 1% of its computing capacity, nearly half of them in three countries. The next divide will be about leverage, not login.

Measure income, not just connections

Africa’s investments in networks, mobile money and digital identity have been impressive; the argument is not to slow them but to widen how we judge them. The standard scorecard, coverage, penetration, device ownership, data prices, still matters. But governments and their partners could measure outcomes closer to what counts: the share of people using the internet to earn, the number of first-time digital sellers, where digital earners live, and how many small firms buy digital services.

An International Trade Centre (ITC) study of nearly 5,000 firms in francophone Africa, SME Competitiveness in Francophone Africa 2022, found that four in five of those using digital tools reported lower costs or higher sales, real value where the pathway exists. These are proposed measures, not data anyone yet collects in full. The point is that we manage what we measure, and the next phase is better judged by participation than connection.

The first chapter of Africa’s digital story was about extending connectivity, and the continent has made remarkable progress. But it is not finished; hundreds of millions remain within reach of a signal they do not yet use. The more demanding question now runs alongside the old one: not whether people can get online, but what that creates.

For the next hundred million Africans who connect, that is the measure that will matter, whether the connection changes what they can build, sell, earn and improve. Connectivity is the foundation. It was never meant to be the finish line.

---