I get asked the same question in different costumes.
Will LEO satellites eat fibre’s lunch? Will Starlink kill FWA? Will direct-to-device make mobile networks redundant? Will fibre wipe out community Wi-Fi? Is 5G going to finish off fixed wireless?
Every one of these is a supply-side question dressed up as a strategy question. And every one quietly assumes something that is not true: that the customer knows — or cares — which technology carries their traffic.
They don’t. In most African markets, the customer is holding a phone. The connection either works or it doesn’t. It is either affordable today or it isn’t. There is either someone to call when it breaks, or there isn’t. The transmission medium behind it is an implementation detail, invisible at the point of consumption. Nobody has ever asked their neighbour whether their WhatsApp arrived over GPON or LEO.
This matters more than it sounds, because it means the competitive battle is not being fought where most operators and investors are looking.
The number that ends the technology debate
Consumption in Sub-Saharan Africa is overwhelmingly mobile, and the smartphone base is still building: smartphone adoption sits at roughly 54% — the lowest of any region on earth — forecast to reach 81% by 2030 (GSMA, Mobile Economy Africa 2026). Ericsson puts it at 52% for 2025, so the two independent estimates corroborate in the low-to-mid fifties. Usage is modest: the average smartphone in the region consumes around 5 GB a month, heading for about 12 GB by 2031 (Ericsson Mobility Report, June 2026).
Now put the contenders next to each other.
Starlink reached an estimated 500,000 subscribers across Sub-Saharan Africa at the end of 2025 (Dataxis, February 2026), having entered 25 countries. FWA-LTE in the region runs at roughly 18 million connections, heading for about 22 million by 2030, with 5G FWA adding a further 3.5 million today (Omdia). Fixed broadband penetration, meanwhile, is 5.35 per 100 people in South Africa and 0.08 in Nigeria (World Bank, 2025).
So: half a million LEO subscribers. Eighteen million FWA connections. Hundreds of millions of smartphones.
LEO is not eating fibre’s lunch. It is not, at continental scale, eating anything yet. That is not a criticism — it is a category observation. These technologies are not fighting over the same customer.
And Starlink’s own economics prove the point. Plans run USD 30–50 a month, with hardware at USD 435 in Nigeria — against a backdrop where even a 4G-capable handset costs about 26% of monthly GDP per capita in Sub-Saharan Africa, versus 16% across other low- and middle-income countries (GSMA, Mobile Economy Africa 2026). More revealing still: Starlink has reportedly been refusing new subscribers in Lagos and other major Nigerian cities on capacity grounds, while its adoption concentrates in denser, wealthier areas. As Dataxis concludes, it remains genuinely uncertain whether LEO expansion will meaningfully address the digital divide at all.
Read that carefully. The technology supposedly built to serve the remote and unserved is capacity-constrained in the dense cities where fibre wins, and priced beyond the rural poor it is imagined to rescue. It is an excellent product for a specific, valuable and comparatively small set of use cases — enterprise, maritime, backhaul, remote sites, premium households, disaster recovery. That is a real business. It is not a fibre-killer.
Direct-to-device: the development that proves the whole argument
If you want the cleanest possible demonstration that technology is invisible to the customer, look at what is happening right now in direct-to-device (D2D).
D2D lets an ordinary, unmodified handset connect to a satellite. No dish. No router. No app. No new purchase. The phone in the customer’s pocket simply works in a place where it previously showed no bars. The user cannot perceive the technology even in principle — which is precisely the point of this article, rendered in hardware.
And the commercial structure is the real story. African operators now hold 32 D2D partnerships, and D2D services have launched in 18 countries worldwide. Globally there are 123 publicly announced D2D partnerships — 23 launched, 27 in testing or trial, 73 in planning (GSA, July 2026). Starlink accounts for 99 of those partnerships, ahead of AST SpaceMobile (44), Amazon Leo (30) and Lynk (21).
In Africa specifically: Airtel Africa signed with SpaceX in December 2025 for Starlink Direct to Cell across all 14 of its markets — a base that had reached 183.5 million customers by March 2026 — with data and messaging tests recently carried out in Kenya and Airtel Uganda beginning trials. AST SpaceMobile has tied up with Axian Telecom, which operates in 11 African markets, and with Vodacom. Lynk has been trialling with MTN in Ghana and South Africa. MTN has deliberately gone multi-vendor, exploring Lynk, AST, Starlink, Eutelsat OneWeb and Omnispace rather than betting on one constellation.
Now notice who is not in that sentence: a satellite company selling directly to an African consumer.
Here is the line from GSA’s analysis that should end the disruption narrative: most of these partnerships use spectrum from the terrestrial operators, at least initially. The satellite operator needs licensed spectrum, which the MNO holds. It needs billing, distribution, agents, vouchers, brand and trust — which the MNO has spent twenty years building. And it needs the customer to already own a working handset.
So the most hyped “disruptive” technology in the sector has arrived in Africa as wholesale supply flowing through the incumbent’s demand-side relationship. Starlink went from presumed rival to channel partner. The satellite fills the coverage gap; the operator keeps the customer.
Note too what D2D actually launches as: messaging first, then narrow data, with voice on the roadmap. It is a coverage and safety layer, not a broadband substitute. It addresses the 210 million who have no signal — not the billion who have signal and don’t buy. Which brings us to the number that matters.
The gap everyone quotes, and the gap that actually matters
GSMA’s headline finding for 2026 is worth quoting exactly: “Despite widespread coverage, almost 1 billion people in Africa are still not using mobile internet, representing 63% of the population” (GSMA, Mobile Economy Africa 2026).
Now set that against the other gap. In Sub-Saharan Africa, the number of people living outside mobile broadband coverage altogether — the coverage gap — is roughly 210 million (GSMA).
Read those two together, and note they are different scopes: a continent-wide billion who can connect and don’t, against a Sub-Saharan 210 million who can’t. The population that cannot get a signal is a fraction of the population that can get one and still isn’t online.
Every technology argument — LEO, fibre, FWA, 5G, Wi-Fi — is a fight over the smaller number. Meanwhile something close to a billion people are already standing inside the coverage footprint, not buying.
You cannot build your way out of that. There is no satellite constellation, no trench and no spectrum allocation that converts a covered non-user into a paying customer. And GSMA is explicit about what the actual barriers are: device affordability, limited digital skills and a lack of relevant content. Not one of them is a transmission problem. That is a demand-side problem, and it is where the real market sits.
What people actually buy
Strip away the engineering and the purchase decision in an African market comes down to a short, unglamorous list.
Does it work, and does it keep working? Not headline speed — reliability. A connection that drops during a job interview has failed, whatever it benchmarks at.
Can I control what I spend, and when? This is the one most imported business models get wrong. Monthly contracts assume monthly salaries and a bank account. Much of the market earns daily or weekly, in cash. The winning products are daily bundles, per-device passes, vouchers, tokens, top-ups, night bundles, data that doesn’t expire the moment you stop paying. This is not a fringe theory — it is what the market leader is actually doing. Safaricom’s own FY26 strategy cites dynamic pricing for data and voice bundles, “more for more” offers for the mass market, a loyalty programme for high-value customers, and a pilot of tokenised Wi-Fi to unlock affordable, inclusive connectivity. The most successful operator on the continent is engineering flexibility, not headline speed.
How do I buy it? Cash, voucher, token, scratch card, agent, USSD, app, mobile money. If your product can only be bought with a card, you have excluded most of your addressable market before you have started.
Is it mine, and can it move? Mobility is not a feature here; it is the default. The connection lives on a device that travels.
Is it worth more than the megabits? Bundled OTT — Netflix, Spotify, YouTube, Prime — has been around long enough to be table stakes, and its purpose was always to drag the conversation away from a pure price-per-gig comparison. That is a defensive move. The offensive move is a genuine ecosystem.
And when it breaks, is there someone to speak to? Trust is the most underrated line item in this sector, and the hardest to retrofit. In a market where a connection costs a meaningful share of income, the ability to reach a human who resolves your problem is a product feature — arguably the product feature.
Why ARPU is now the wrong scoreboard
If the value sits beyond connectivity, then measuring customers by connectivity ARPU tells you progressively less about your business.
Take Safaricom’s audited results for the year ended 31 March 2026. M-PESA contributed 45.6% of Safaricom Kenya’s service revenue, growing 13.4% to KShs 182.7 billion — while all connectivity revenue combined (voice, data and messaging) accounted for 49.4%. In other words, at Africa’s most celebrated mobile operator, the payments business has drawn almost level with the entire business of selling connectivity. M-PESA now moves 46.4 billion transactions worth KShs 41.7 trillion a year, across a merchant base that grew 71% to 3.1 million.
Airtel Africa tells the same story from a different starting point. For the year ended 31 March 2026, mobile money revenue reached USD 1,355 million — 21.1% of group revenue of USD 6,415 million — serving 54.1 million mobile money customers, up 21.3%, within a total customer base of 183.5 million.
Judge Safaricom on connectivity ARPU and you would miss a business that has nearly matched its entire connectivity line. Globally, mobile money now runs at 2.3 billion registered accounts and about USD 2 trillion in annual transaction value (GSMA State of the Industry 2026).
The customer is no longer a connection generating a monthly bill. They are a relationship generating transactions, data, distribution and partner revenue. The right metric is lifetime value across the ecosystem — and most operators still cannot calculate it, because the systems that would join those datasets were never designed to talk to each other.
The strategic responses — and what the evidence actually says
This thesis explains the two big structural moves in African telecoms right now. It also explains why they don’t always work.
Superapps: the anchor matters more than the app. MTN built Ayoba into a genuine success by user count — it surpassed 30 million monthly users across markets including Nigeria and Ghana. In March 2026, MTN pulled it from app stores and began a phased sunset, folding the ambition into a MoMo-centred platform with Ant International, launching in Nigeria from Q3 2026. Vodacom’s VodaPay, despite heavy investment, has struggled to reach its projected scale.
Contrast that with M-PESA, which now hosts 221 mini apps across the Vodacom group and 9.4 million financial services app users, growing 40% year on year (Vodacom FY2026).
The pattern is instructive, and it is not “superapps work”. It is that superapps built on a habitual, trusted, daily transaction work — and superapps built on engagement don’t. Payments were already a reflex, already trusted, already daily. A chat app with 30 million monthly users still couldn’t become the front door. If your superapp strategy does not sit on top of something the customer already does without thinking, you are building a mall in a field.
Bank MVNOs: connectivity as the door, not the product. Capitec’s own results for the year ended February 2026 are striking. Fintech — value-added services plus Capitec Connect — now contributes 26% of group headline earnings, alongside Personal Banking at 41% and Insurance at 27%. VAS and Capitec Connect grew 38% to R6.1 billion. Capitec Connect is now the country’s largest MVNO, with 1.5 million active clients (Capitec Bank Holdings, FY2026 results, 22 April 2026).
A bank earning roughly a quarter of its profit from a fintech-and-connectivity line is not a telecoms story. It is the whole thesis of this article.
And look at how Capitec competes, because none of it is network engineering. It returned R330 million to clients through lower Connect data prices, and a further R108 million through free Connect data and Live Better rewards. It gave back 3 petabytes of free data, worth R78 million. Most tellingly, it launched affordable device sales to all clients through partnerships with Samsung and Apple, delivered nationwide.
Read that last one again against GSMA’s diagnosis. The single largest barrier to getting the next billion online is device affordability — a 4G handset at 26% of monthly GDP per capita. The organisation moving hardest against that barrier in South Africa is not a mobile operator. It is a bank, using its balance sheet, its credit scoring and its 26 million existing relationships. That is a demand-side institution solving a demand-side problem with demand-side tools — and it is winning.
The synthesis
Technology choice is a supply-side decision that determines whether you can serve a market profitably. Demand-side design determines whether anyone chooses you. Most business cases we review are rigorous on the first and impressionistic on the second — a detailed network model bolted to a demand forecast that is really a hope.
The technology question deserves a real answer: fibre where density and willingness to pay support it, wireless where they don’t, satellite for the hardest geographies and the highest-value niches, community Wi-Fi where the unit economics of the alternatives collapse. All of them have a role. None of them is a strategy.
Because the customer will never know which one you chose. They will only know whether it worked, whether they could afford it today, whether they could buy it the way they get paid, whether it gave them something beyond megabits, and whether anyone answered when it broke.
Where Digitalthings fits
Our work sits deliberately on both sides of this line, because the questions are inseparable.
On the supply side, we do the technoeconomic work: technology selection and comparison across fibre, wireless, satellite and hybrid architectures; network and business-case modelling; total cost of ownership; and the commercial and technical due diligence that tells an investor whether an asset is genuinely buildable and bankable in African conditions.
On the demand side — where most of the industry is thinner — we do market assessment and segmentation, and customer-base value work: how to unlock value from an existing base through partners, ecosystems and loyalty rather than through price. We have built phased digital roadmaps that take an operator from a single app to a partner-integrated marketplace, anchored on the group services a customer already uses.
That combination is the point. We are independent of any technology or vendor, which means we have no incentive to tell you the answer is fibre, or satellite, or anything else. We work for operators, banks and retailers entering connectivity, private-equity and infrastructure funds, development-finance institutions and governments across South Africa, Kenya, the DRC, Namibia, Malawi, Tanzania and beyond.
If you are weighing a technology decision, entering connectivity from an adjacent industry, or trying to work out why a well-built network is not converting into a business, the question is rarely which technology. It is who your customer is, and what would make you relevant to them.
That is a conversation worth having before the capex is committed.
Every figure in this article was verified against its primary source. Operator financials are taken directly from published results: Safaricom PLC audited results for the year ended 31 March 2026 (Board-approved 6 May 2026); Airtel Africa plc FY2026 results for the year ended 31 March 2026 (8 May 2026); Vodacom FY2026 disclosures; and Capitec Bank Holdings’ published results for the year ended February 2026 (22 April 2026). Market data: GSMA (Mobile Economy Africa 2026; State of the Industry Report on Mobile Money 2026), Ericsson Mobility Report (June 2026), World Bank, Omdia, Dataxis (February 2026) and GSA non-terrestrial network research (July 2026). Where scopes differ between Africa and Sub-Saharan Africa, this is stated explicitly. Data is current as of July 2026.