Africa Online: The History, the Present Reality, and the Structural Ceiling

Three Eras, Not One Leapfrog

The word *leapfrog* flatters the record. Africa did not skip a stage; it was denied one, and then built a different one.

Era 1 — Scarcity and the state monopoly (to ~1995). Colonial-era telecoms were built to connect capitals to metropoles, not citizens to each other. At independence most countries inherited a single state PTT with a copper network sized for administration. In 1995 there were fewer than three fixed lines per 100 people across the continent, and waiting lists for a landline ran to years. Internet access, where it existed, was a store-and-forward academic link — South Africa's first full IP connection came in 1991; most of the continent connected between 1995 and 2000.

Era 2 — Mobile and the cable (1998–2015). Liberalisation and GSM licensing changed the physics. Prepaid billing removed the credit check, which was the actual barrier: it turned a monthly contract into a daily purchase. Operators such as MTN, Vodacom, Airtel, Orange and Safaricom scaled faster than any utility in African history. Then bandwidth arrived. Until 2009 East Africa had no submarine fibre landing at all and depended on satellite; SEACOM, TEAMS and EASSy landed in 2009–2010 and wholesale bandwidth prices collapsed by well over 90% within a few years. WACS, ACE and MainOne did the same on the west coast.

Era 3 — The platform era (2016– ). Voice and SMS revenue plateaued and the product changed. M-Pesa, launched in Kenya in 2007, matured from a remittance tool into a national payment layer. Nigeria's cashless policy, Ghana's interoperable MoMo switch, and Egypt's InstaPay turned payments into infrastructure. Simultaneously the hyperscale cables — 2Africa, Equiano, Google and Meta-backed — moved the constraint from international capacity to national distribution, power, and demand.

Each era solved the previous era's binding constraint and exposed the next one. That pattern is the key to reading 2026.

The Present Reality: What Is Actually True in 2026

Three numbers frame the present.

First, connectivity is a minority experience. Roughly 38% of Africa's population uses the internet, against a global figure of about 68%. That is an absolute shortfall of somewhere near 800 million people. Within that average sit extremes: Morocco, Egypt, South Africa and Seychelles are above 70–90%; several Sahelian and Central African states remain in the teens.

Second, the coverage gap is no longer the main problem. About 85% of Sub-Saharan Africans live within reach of a mobile broadband signal, but only around 27% use mobile internet. The difference — the *usage gap* — is now roughly three times the size of the coverage gap. Building more towers does not close it. Affordability, devices, skills and relevance do.

Third, the financial rail is genuinely world-leading. Sub-Saharan Africa hosts the majority of the world's registered mobile-money accounts and processes well over $1.1 trillion in annual transaction value. Kenya, Ghana and Tanzania have adult penetration rates that most high-income countries cannot match on any single payment product. This is not a catch-up story; it is an export-grade model.

The honest summary is that Africa is simultaneously the world's most advanced mobile-payments region and its least-connected continent. Both are true, and they are true for the same reason: the mobile phone became the entire computing platform, so anything it could carry scaled explosively, and anything requiring power, bandwidth, literacy or formal identity did not.

The Structural Challenges: Five Ceilings

The gap is not a single problem. It is five, and they compound.

1. Affordability. The UN Broadband Commission target is 2% of monthly income for 1GB of mobile data. Much of Africa clears that on data price alone — but not on devices. An entry-level smartphone can cost 20–50% of monthly income for a low-income household, and in the poorest quintile far more. Device cost, not data cost, is now the primary affordability barrier, which is why handset financing (M-KOPA, FoneYam, operator lease models) has become the highest-leverage intervention available.

2. Electricity. Roughly 600 million Africans lack access to reliable electricity. A connectivity plan that ignores power is a plan for a phone that is charged at a kiosk twice a week. Towers in low-grid markets run on diesel — a large share of operator opex — and every data centre proposal ultimately resolves into a generation and grid question.

3. Skills and relevance. Even where coverage, device and power exist, uptake stalls if there is nothing worth connecting to. Local-language content is thin, digital literacy programmes are underfunded relative to infrastructure, and the majority of content consumed in Africa is hosted offshore. The result is a demand-side ceiling that supply-side subsidy cannot lift.

4. Identity and digital public infrastructure. Hundreds of millions of Africans lack verifiable legal identity. Without ID, KYC fails; without KYC, formal credit, insurance, and government transfers cannot reach the person. India's Aadhaar-plus-UPI stack is the reference model, and the African analogue — Nigeria's NIN, Ethiopia's Fayda, Morocco's national ID, Ghana Card — is being built now. This is the single highest-multiplier layer on the continent.

5. Compute and data sovereignty. Africa has about 17–18% of the world's population and well under 1% of its data-centre capacity. Most African data is stored and processed offshore, which imposes latency, foreign-currency cost and regulatory exposure. Fragmentation makes it worse: too few Internet Exchange Points mean intra-African traffic is often routed via Europe — the *tromboning* problem — paying twice for a domestic conversation.

A sixth, cross-cutting issue sits behind all of them: fragmentation. Fifty-four regulators, fifty-four spectrum regimes, fifty-four data-protection postures. Spectrum auctioned as a fiscal instrument rather than a development one raises the cost of every subsequent megabit. The AfCFTA Digital Trade Protocol is the first serious attempt at a continental answer; its implementation is the thing to watch.

The Counter-Case: Where This Analysis Could Be Wrong

Three honest challenges to the framing above.

1. The data itself is weak. Internet-use statistics for many African markets rely on operator-reported subscriptions and infrequent household surveys. Multi-SIM ownership inflates subscription counts; shared-device use deflates individual-use counts. The true figure for meaningful internet use may be several points either side of 38%, and country-level comparisons should be treated as ranges, not points.

2. Satellite may compress the coverage timeline. LEO constellations are now live in a growing number of African markets and price aggressively against terrestrial backhaul in low-density areas. If terminal costs continue falling, the rural coverage economics that have held since 2010 could change materially before 2030 — though satellite does nothing for device cost, power or skills, which are the larger gaps.

3. Compute may arrive faster than the trend implies. AI demand has pulled forward data-centre commitments across Kenya, Nigeria, South Africa, Morocco and Egypt, several of them anchored to dedicated renewable generation. Capacity that took fifteen years to build could plausibly double in five. The risk in that scenario is the opposite one: capacity concentrated in four or five markets, deepening rather than closing intra-African divergence.

None of these overturn the core reading. They do argue for holding the 2030 numbers loosely.

So What: The Strategic Read for 2026–2030

For investors. The returns have moved down the stack. Subscriber-growth plays are maturing; the underpriced exposures are device financing, distributed power for towers and data centres, local hosting and IXP capacity, and identity/KYC infrastructure. Every one of these is a toll on somebody else's growth.

For operators. ARPU expansion now depends on converting the usage gap, not extending coverage. That means handset lease books, zero-rated onboarding content, local-language interfaces, and partnerships that make the phone worth charging. The operator that solves device affordability at scale captures the next 200 million users.

For governments and regulators. Spectrum priced as a revenue line is a tax on national productivity. The higher-return interventions are digital ID rollout, mandating IXP peering to stop traffic tromboning through Europe, and grid reliability. The AfCFTA Digital Trade Protocol matters more than any single national broadband plan.

For enterprises and founders. Design for the actual median user: an intermittently powered Android device, a shared handset, prepaid data, low literacy in the interface language, and a customer who may not have formal ID. Products that assume a Lagos or Nairobi power user address perhaps 15% of the addressable population.

For development institutions. Fund the layers with no commercial owner — digital literacy, local-language content, and the identity substrate. Infrastructure capital is increasingly available; demand-side capital is not.

The defining question for the rest of this decade is not whether Africa connects. It is whether the compute, power and identity layers get built inside Africa, or whether the continent scales its digital economy on infrastructure it rents from somewhere else.