We have a habit, in Kenyan fintech conversations, of lining up every new platform against M-PESA as though it were a title fight. It rarely is. But the comparison between M-PESA and Ziya is worth making, not because they are equals, but because they answer the same question in almost opposite ways. The question is the oldest one in money: who gets capital, who gets to move it, and what do we charge them for the privilege?
One of these names you already know. The other you probably don’t. Let’s fix that first.
Meet Ziya
Picture a ‘chama’, (the rotating savings group that has quietly financed African life for generations) and now picture it rebuilt as software, stripped of interest, and handed back to the people who invented it. That is Ziya.
It calls itself a “trust engine,” and the phrase is more than branding. The model is disarmingly simple: ten or more people who already trust each other form a group; capital enters; a few members draw first; they repay in small daily amounts; the next in line draws down; and the circle keeps turning until everyone has had a turn. No interest. No collateral. No credit score. No penalties. The only price is a flat subscription.
What makes Ziya genuinely exciting is not the pitch; it’s the receipts. From a base of roughly KES 7 million, it has rotated KES 37 million through 563 self-managed groups and 6,183 micro-enterprises, at a repayment rate of 99.5% – a number most commercial lenders would frame and hang on a wall. More than 95% of its members are women, and it has put down roots on the Kenyan coast, in Muslim and Nubian communities that mainstream, interest-charging finance has never served on terms they could religiously or financially accept. Its founders left “unicorn-scale” fintech to build it, and they describe what they’re doing, without irony, as a rebellion against financial apartheid.
So this is the matchup: a USD-3-billion national institution against a coastal insurgency with a few thousand users and a radical idea. The interesting part is that the insurgency is winning the argument it picked.
What each one actually does
M-PESA, launched by Safaricom in 2007, is the most successful mobile-money system on earth – around 35.8 million monthly active users in Kenya and a value flow equivalent to more than half of national GDP. Its core job is movement and storage: send, receive, deposit, withdraw, pay a till, pay a bill, buy airtime. Bolted on top are credit products – Fuliza (an overdraft for when your wallet runs short) and M-Shwari (save-and-borrow, run with NCBA). M-PESA is infrastructure first, lender second.
Ziya does two things. The headline act is group-based, zero-interest rotating capital – the circle described above. The quieter, and strategically louder, act is payments: Ziya offers peer-to-peer money transfer, and here is the twist that breaks the easy framing — it doesn’t charge per transaction at all. Once you’ve paid the flat subscription, moving money is simply free. Not “free under 100 shillings.” Free.
That single design choice is where the whole comparison stops being about features and starts being about philosophy. Because it forces a question Kenyans have been joking about for years without quite finishing the thought: what does it actually cost to send money?
What does it really cost to send a shilling?
The old joke goes that M-PESA is just an SMS with delusions of grandeur. It’s funnier because it’s largely true. M-PESA was built on USSD and SIM Toolkit i.e the same humble plumbing that carries a text message. The marginal cost of carrying one more transaction across a network that already exists is close to nothing: fractions of a shilling in compute, a database write, a confirmation message. The system now clears roughly 4,500 transactions per second, which means the per-unit cost of the technology is being driven down toward zero with every passing year, not up.
And yet sending money costs anywhere from about KES 7 to KES 300, with a 20% excise duty stacked on top of every chargeable transaction.
Now, the fair objection is that the marginal cost of a transaction is not the whole cost. Safaricom maintains a vast physical agent network, carries agent float, fights fraud at industrial scale, meets regulatory and capital requirements, and built the thing in the first place. All true. Infrastructure is expensive, and someone has to pay for it. But that explains the existence of a fee, it does not explain its size. The honest way to read the gap between “costs almost nothing to run” and “charges up to KES 300” is this: most of what you pay to send money is not the cost of sending money. It is economic rent (the dividend of dominance) plus a government that has turned M-PESA into a tax-collection rail.
The numbers make the point without needing accusation. M-PESA alone earned Safaricom about KES 161 billion in FY2025 (44% of the company’s Kenyan service revenue) helping power group after-tax profit of roughly KES 93 billion and cementing Safaricom’s standing as the most profitable company in the region. The average M-PESA customer now generates close to KES 395 a month in revenue. None of that is illegal or even surprising; it is what any rational monopolist does with pricing power on a utility people cannot opt out of. But it reframes the entire cost debate. The price of moving money on M-PESA is not set by what it costs to move money; it is set by what the market will bear and what the Treasury adds.
Ziya’s flat KES 100 a month is, in this light, not just cheaper – it is a different theory of pricing entirely. It charges you for the software, once, and then gets out of the way. M-PESA charges you for the act, every time, forever. When the marginal cost of a digital transaction is essentially zero, the flat subscription is arguably the more honest model, and the per-transaction fee is arguably the more extractive one. That is the real argument hiding inside a 100-shilling subscription.
So which is the expensive model?
For moving and paying, M-PESA’s per-transaction fees are small in isolation but, multiplied across a life of daily micro-payments and topped with 20% tax, become a steady toll on participation in the economy – a toll Ziya simply abolishes inside its network.
For credit, the gap is a chasm: Fuliza charges roughly 1.083% per day, which annualises to something near 395%, while Ziya charges zero interest on capital. Borrow KES 1,000 on Fuliza for a month and you can repay close to KES 300 on top; borrow through your Ziya circle and you repay the principal to your neighbours.
M-PESA is cheap per click and brutal as a lender. Ziya is near-free as a lender and free to transact within. On price, this is not close – where the two actually overlap.
Where do these two models overlap?
The first is payments. M-PESA owns the open, universal rail – anyone, anywhere, to anyone. Ziya offers free transfers, but inside its own ecosystem and its own communities. M-PESA’s reach is unmatched; Ziya’s pricing is unmatched. That’s a real, if asymmetric, contest.
The second, and deeper, is credit for the people the formal system ignores. Ziya’s founding statistic is that 70% of Kenya’s 7.5 million MSMEs still can’t get fair credit despite having phones, customers, and grit. M-PESA’s answer is Fuliza and M-Shwari – instant, algorithmic, individual, priced for risk. Ziya’s answer is the circle – slow, relational, collective, priced at nothing. Both are funding the informal economy. They’ve simply made opposite bets about how trust is created: M-PESA infers it from your data; Ziya borrows it from the people who already know you.
And note the dependency hiding underneath: Ziya’s members almost certainly settle in shillings that ride M-PESA rails. The insurgent runs partly on the giant’s roads. They are less enemies than tenant and landlord – which is exactly why the rent question matters.
The impact of the two philosophies
M-PESA’s model is extraction at scale, with extraordinary reach. It turned a basic phone into a bank account for tens of millions who had neither, one of the great inclusion achievements of the century (and not to be romanticised away). But its lending logic monetises desperation: the daily Fuliza fee is invisible in the moment and punishing over time, and Kenya’s credit bureaus carry millions of people negatively listed over mobile loans of less than KES 1,000. The system reaches everyone and, on its credit side, can tax the most vulnerable hardest.
Ziya’s model is rotation with dignity, at limited reach. It refuses interest, refuses to score people, refuses to shame them, and a 99.5% repayment rate suggests social trust may be better collateral than an algorithm for this segment. That a zero-interest model holds up at all is a finding, not a marketing gimmick nor just a slogan. Its cost is scale and speed: a circle of ten neighbours cannot serve a nation overnight, and trust rooted in one community does not copy-paste across borders the way a USSD menu does. Its interest-free design is also not merely economic for its largely Muslim coastal base, riba-free finance is the difference between a product they can use and one they can’t.
So which of these two models can be considered more inclusive?
M-PESA is more inclusive as infrastructure – universal, instant, demanding nothing but a basic handset and an ID.
Ziya is more inclusive as fair credit – it deliberately serves the women, the informal traders, and the faith communities that M-PESA’s lending arm prices out or never reaches on acceptable terms.
If inclusion means “can almost anyone touch it,” M-PESA wins and it isn’t close. If inclusion means “does it extend fair capital to the people the system was built to exclude,” Ziya is doing something M-PESA’s credit products structurally cannot. The gap between those two definitions is the unfinished business of African fintech.
What makes sense for Africa and the future
This feels rather difficult to conclusive crown a winner here. I’d rather say that the useful conclusion is that the future is a layering, and a quiet shift in how things get priced.
M-PESA already won the first argument: you can build national financial infrastructure on cheap phones and human agents. That fight is over. The frontier has moved from access to fair access and, underneath it, from taxing every transaction to pricing the service honestly. When the marginal cost of moving a shilling is effectively zero, per-transaction rent starts to look less like a business model and more like a toll booth on a road that’s already paid for. Flat, bundled, subscription pricing (Ziya’s instinct) is where digital goods tend to end up once competition catches up with the technology. Music did it, Software did it, so Money is next.
So the most desirable future is not Ziya versus M-PESA. It is community-trust credit logic and honest, flat pricing running on top of mobile-money rails: the reach of the giant, the ethics and economics of the insurgent. Maybe that’s Ziya growing up. Maybe it’s Safaricom being pushed (by models like Ziya and by regulators who can do the marginal-cost arithmetic) to make both its lending and its fees less extractive. Most likely it’s the friction between the two that produces it.
For a continent where most economic life is informal, the platform that wins long term won’t be the one with the lowest per-click fee or the biggest user count. It will be the one that treats reach and dignity as the same problem, and refuses to charge the poor 395% for capital, or a tax on every shilling they were always going to have to move anyway.
M-PESA pricing, revenue, user, and profitability figures reflect Safaricom’s published 2025–2026 results and 2026 tariff schedules; transaction costs include Kenya’s 20% excise duty. Fuliza and M-Shwari rates are indicative and subject to change. Ziya’s traction figures (capital rotated, group count, repayment rate, membership) are drawn from the company’s own published statements and are not independently verified here.
