Kenya’s most valuable company gives away most of its flagship product for free. That, it turns out, is the point.
By The Ledger Team
In Kawangware on a Tuesday, a mama mboga sends her supplier KES 50 for a sack of onions. The transaction costs her nothing. Roughly six in ten transactions on M-Pesa now fall into this category, small transfers and merchant payments that Safaricom deliberately zero-rates; the company calls the product M-Pesa Kadogo, and it is, on its face, an odd thing for a listed company to be proud of giving away. The explanation for why Safaricom can afford this generosity, and still post the region’s first $3bn revenue year, is the actual subject here. Kadogo is marketing wearing the costume of philanthropy.
Start with the top line. Safaricom Group’s total revenue for the year to March 2025 reached KES 388.7bn ($3.01bn), a historic first for East Africa, with service revenue climbing to KES 371.4bn, a rise of 10.8% year-on-year, demonstrating resilience amid both regional and global headwinds. Group EBIT crossed KES 104.1bn. None of that is the interesting number, though; the interesting number is how the total splits.
M-Pesa is the business now, not an adjunct to it. Kenyan M-Pesa revenue reached KShs 161.1 billion, contributing 44.2% of service in FY2025 and 61% of the incremental growth in service revenue. Voice, the product Safaricom was originally licensed to sell, limped along at KShs 80.78 billion, growing 1.6% YoY, which in a market this size counts, oddly, as a decent outcome; most telcos would kill for flat voice revenue, let alone growth (the industry, globally, has spent a decade watching this line shrink to nothing). Mobile data added KES 72.9bn, up 15.2%, riding 4G penetration that now covers 97.9% of the Kenyan population. Connectivity and money together, in other words, not connectivity alone; a telecoms company that happens to also run the country’s payment rails, or perhaps a payment rail that happens to also sell airtime. The distinction has stopped mattering commercially, if it ever did.
The customer base underwriting all this is 37.1 million active users in Kenya, up 7.1% year-on-year from 34.6m, and 57.07 million across the Group once Ethiopia is folded in, a market Safaricom entered with an ambition that has, so far, cost considerably more than it has returned. Kenya remains the engine: Safaricom’s domestic business posted a profit before tax of KES 142.95bn, up 11.6% from KES 128.15bn. Group profit before tax grew more modestly, to KES 93.21bn, a 10.1% rise, the gap between the two figures being Ethiopia, almost entirely. Every shilling Kenya generates over and above the group figure is, functionally, a shilling spent keeping the Ethiopian subsidiary alive.
Ethiopia deserves its own paragraph, if only because it is where the “how does this company actually make money” question temporarily stops having a flattering answer. The consortium’s total investment in the market has reached $2.65 billion, with Safaricom PLC’s own equity contribution standing at $1.22 billion, and the birr’s collapse against the dollar, from 57 to the dollar to 146, has done to that investment roughly what currency collapses always do. The company moved from a net cash position of KES 3.54bn to net debt of KES 139.81bn over the period Ethiopia has been under construction; a Safaricom CFO once described the pricing environment there as “untenable,” the company selling data and voice below cost simply to build a subscriber base worth having. Ethiopia’s service revenue for the year came to a comparatively modest KES 7.5bn against 7.3m one-month active customers, a rounding error against the Kenyan business that funds it, and yet the subsidiary still managed to contribute 9.3% to Group service revenue growth, evidence, depending on temperament, either of early traction or of a very large base effect flattering a very small number. Whether the Ethiopian bet resolves into the second M-Pesa or into a very expensive lesson in currency risk remains, at the time of writing, unresolved.
The margin structure explains why Kenya can subsidise Kadogo and Ethiopia can survive its own arithmetic simultaneously. M-Pesa is, structurally, a low-marginal-cost business once the agent network and the switching infrastructure exist; each additional transaction costs the company very little to process, which is precisely why zero-rating the smallest transfers barely dents the segment’s profitability while buying enormous customer loyalty and regulatory goodwill in one stroke (the Competition Authority, predictably, has taken note of Safaricom’s dominance rather more often than it has taken action on it). Connectivity, by contrast, is capital-hungry: fibre now runs to 18,300km, and 1,700 5G sites have gone up in a single year, costs that show up in depreciation long before they show up in customer revenue.
Competitively, Safaricom’s position rests less on network quality, though that helps, than on the securitisation of habit: three-quarters of Kenya’s adult population has arranged its financial life, its savings, its credit, its merchant payments, around a single app, and switching costs for that kind of embeddedness are punishing in ways a cheaper data bundle from Airtel cannot easily undo. Peter Ndegwa, the CEO, put the underlying claim plainly on the company’s 25th anniversary: “We have shown in mobile money that we can actually do much better” than peer markets in the West, a boast that happens, unusually for corporate rhetoric, to be defensible on the data. Fuliza, the overdraft product layered onto M-Pesa, and the various savings and lending tie-ups with KCB and other banks, extend that same habit into credit, a segment that carries margins connectivity can only envy.
The risks sitting beneath this are not exotic. Regulatory pressure on M-Pesa’s dominance is a permanent condition rather than a cyclical one; Kenya’s telecoms authority has floated tighter interoperability and pricing rules more than once, and each cycle chips slightly at the moat. Cyber threats scale with transaction volume, and M-Pesa now processes something in the order of 4,500 transactions every second, a figure that turns any security failure into a systemic one rather than a contained one. Low-earth-orbit satellite services, Starlink chief among them, threaten the rural connectivity business specifically, the segment Safaricom has historically used to justify its universal-service obligations and its favourable regulatory treatment. And Ethiopia, still burning cash, remains the largest single variable on the balance sheet; the company has pencilled in profitability there by FY2027, though pencil is the operative word.
None of which has stopped the dividends. Safaricom returned KES 255bn to shareholders over the last five years, and the FY2025 payout of KES 1.20 per share, a 69% payout ratio against an 80% policy, suggests a board still confident enough to keep paying even as Ethiopia’s losses widen. That confidence rests, ultimately, on the Kawangware transaction costing nothing: a company that gives away 60% of its most visible product and still gets richer has found, in M-Pesa, something closer to infrastructure than to a service line, and infrastructure, once built, tends to keep paying rent long after anyone has stopped noticing it is there.
