African Fintech: The Plumbing

When a Lagos-based fintech processes a merchant payment, the customer might pay with their OPay wallet while the merchant wants funds in their Providus bank account. The settlement crosses different rails, data formats, and confirmation mechanisms, and the reconciliation step, at many mid-stage companies, happens in a spreadsheet.

Sub-Saharan Africa processed $1.4 trillion through mobile money alone in 2025. That's 66% of global mobile money volume, running through infrastructure that most Western fintech engineers would not recognize.

African fintech's infrastructure problem isn't about scale or connectivity. The payment rails exist. Real-time systems are live in 25 countries.

The problem is that the rails don't talk to each other, the layers between them are thin, and the tools for managing money across them are manual. The consumer-facing products are polished, but the plumbing behind them is still being improvised.

Five rails, one ledger

A fintech operating at any real volume in Nigeria or Kenya touches multiple payment rails simultaneously: mobile money (M-Pesa, MTN MoMo, OPay), bank transfers (NIP, PesaLink), card networks (Visa, Mastercard), USSD sessions, and increasingly stablecoin settlement. Each rail has its own data format, its own settlement cycle (real-time to T+2), its own webhook reliability, and its own failure modes.

Reconciliation across these rails, matching what your internal ledger says happened with what each rail says happened, remains largely manual at early-to-mid scale.

The principle is straightforward: a single webhook is never the final truth. Internal ledgers must model settlement states explicitly, from pending through confirmed, final, and reversible. A transaction can be confirmed on one rail while still pending on another. The ledger has to hold both states simultaneously and reconcile them on different timescales.

In practice, many fintechs at early-to-mid scale still do this manually. A human opens the M-Pesa settlement file, the NIP settlement file, and the card processor statement, then matches them against the internal database row by row. This works at a few hundred transactions per day but breaks at a few thousand.

A handful of companies are building automated solutions. Recital does three-way reconciliation (internal records vs. bank statements vs. payment provider statements) across multiple currencies. Credrails, seed-funded by SoftBank at $2.5 million, handles lending analytics and reconciliation in Kenya and Nigeria.

Peach Payments in South Africa launched a Recon API for high-volume operations. Transaction Junction built TJ Recon PRO. Formance offers an open-source ledger designed for multi-rail architectures.

But the market is wide open. Every fintech with volume needs reconciliation infrastructure, and few have it.

The identity tax

Nothing moves on any payment rail, credit product, or insurance policy until someone answers who the customer is. In most of Africa, that answer costs more than it should.

Nigeria has the NIN (National Identification Number). Kenya has Maisha Namba, the successor to the legally challenged Huduma Namba. Ghana has the Ghana Card. South Africa has its own national ID. None of these systems interoperate. A fintech operating across four countries must integrate with four completely separate identity rails, each with its own data formats, verification requirements, and regulatory expectations.

Smile Identity has emerged as the dominant eKYC layer, completing over 200 million verification checks across 50 countries. It crossed 100 million just 11 months before hitting 200 million. Youverify, VerifyMe, Appruve, and Prembly are in the market, but Smile's scale advantage is widening.

The cost is real. Compliance can cost smaller fintechs up to 10% of revenue. Between 2024 and 2025, deepfake video liveness attacks emerged as a growing fraud vector against digital KYC systems. The identity infrastructure is simultaneously the most critical dependency and the most expensive tax on African fintech operations.

Nigeria's NIMC Act 2026 is the most significant signal. President Tinubu signed it in July, reinforcing the NIN as the country's primary identity credential for public and private services. If enrollment targets hold (3.3 million new registrations per month), it could simplify KYC across banking, lending, and insurance by creating a single trusted verification layer. If they don't, the gap between urban and rural identity coverage widens.

How the rails diverge

The payment architecture across Africa is not one system. It's several systems that evolved independently, built on fundamentally different assumptions.

Nigeria is bank-centric. NIBSS Instant Payment (NIP) processed 11.3 billion transactions in 2024 and underpins over 90% of bank-to-bank transfers. Mobile money came late because banks dominated early digital payments. The CBN didn't license mobile money operators until 2022.

Kenya is the inverse. M-Pesa owns the market. PesaLink connects 80+ financial institutions for interbank transfers, and its leadership is pushing to reposition it as shared infrastructure linking banks, mobile money wallets, SACCOs, and fintechs into one interoperable system. But M-Pesa remains the gravitational center.

WAEMU built a regional system. PI-SPI, launched by the BCEAO in September 2025, is an instant payment platform covering all eight francophone West African countries. A single integration gives a fintech access to the entire zone. By September 2026, 175 participants were authorized.

But 71% of WAEMU transaction value still flows through cheques.

South Africa runs PayShap for real-time low-value payments on top of a mature banking system. Ghana operates GhIPSS with Universal QR. Tanzania has TIPS.

Each of these systems was designed to solve a domestic problem, not to interoperate with the others.

The BaaS dependency

Nigeria's fintech licensing structure forces a specific architecture: fintechs need a partner bank to hold deposits and issue accounts. This created a Banking-as-a-Service market projected at roughly $1.6 billion in 2026 and growing to $3.2 billion by 2031.

Providus Bank became the go-to partner. Its virtual-NUBAN and BIN-sponsorship APIs sit underneath Piggyvest, Paystack, and much of the Lagos fintech scene. Wema Bank (through ALAT, Africa's first fully digital bank) and Stanbic IBTC fill the rest of the market.

When Providus suspended its virtual payment services, it affected fintechs like PiggyVest and Monnify that depended on its infrastructure. The single-point-of-failure risk materialized, and the structural dependency hasn't been resolved. CBN is offering new payment-service licenses and the first wallet-based cross-border corridor between Nigeria and Ghana went live in 2025, but the concentration in a handful of partner banks remains.

Credit without bureaus

An estimated $330 billion in untapped credit demand exists across Sub-Saharan Africa. Formal credit penetration remains below 40% of the adult population. Traditional credit bureaus exist in Nigeria (three licensed), Kenya (three), and South Africa (comprehensive), but coverage is thin. Most adults have no formal credit file.

Alternative credit scoring has moved from pilot to mainstream. Mobile money transaction patterns, telecom data, digital payment history, psychometric assessments, even solar panel installment payment history serve as inputs. In Ghana, formal borrowing has reached 29% of adults, per the World Bank's 2025 Findex data, partly driven by alternative scoring adoption.

But 350 million more remain excluded. The models carry ethical weight. Some score borrowers on typing speed, whether contacts are saved with full names, and social media activity.

Bias amplification is documented, and privacy regulations haven't caught up with data collection practices.

Cross-border: PAPSS and the stablecoin shift

Most African cross-border payments still route through correspondent banks in the US, UK, or EU. This adds intermediary fees, introduces 3-7 day settlement delays, and makes Sub-Saharan Africa the most expensive region for remittances at roughly 8.8% average cost for a $200 transfer.

PAPSS is the most ambitious attempt to fix this. Developed by Afreximbank and the African Union, it enables local-currency settlement without routing through dollars or euros. By late 2026, it connects 28 countries, 24 central banks, and 190 financial institutions.

PAPSSCARD, launched in June 2025, is Africa's first continental card scheme, designed to keep transaction processing, data, and fees within the continent. The African Currency Marketplace eliminates what PAPSS estimates is a $5 billion trade bottleneck by enabling direct currency-pair exchange.

But a parallel rail is forming underneath. Stablecoins are compressing B2B settlement costs from $5-15 to under $1 per transaction. Importers settle invoices, fintechs avoid prefunding positions, and companies run payroll across borders.

Kora launched "One Rail" for stablecoin payments in October 2026. Circle partnered with Onafriq to bring USDC access to one billion mobile money wallets across 40 markets. Sub-Saharan Africa received roughly $205 billion in on-chain value between mid-2024 and mid-2025, up 52% year-over-year.

Remittances to Africa reached $124 billion in 2025. Fees on the London-to-Lagos corridor have fallen from about 7.8% in 2023 to 2-3% in 2026, with settlement in minutes. Those savings are flowing to stablecoin rails and digital channels, not traditional wire services.

What Okra's death tells you

Okra raised $16 million to build open banking APIs for Nigeria. It launched three years before the CBN finalized open banking guidelines. Banks had little incentive to open APIs before regulation required it.

Then the naira devalued from roughly 460 to 1,600 per dollar between 2023 and 2024. Okra's cloud infrastructure ran on dollar-denominated bills. Its revenue came in naira. The unit economics inverted overnight.

Mono and Stitch out-spent Okra on distribution. Mono was eventually acquired by Flutterwave. Stitch raised $55 million in its Series B and became the best-funded open banking player on the continent. Okra returned $4-5.5 million in unspent funding to investors and shut down in May 2025.

The failure captures the full stack of African fintech infrastructure risk: regulatory timing (building before rules exist), currency exposure (dollar costs against volatile local revenue), and the distribution tax (technology alone doesn't win when the market requires country-by-country sales motions). Every infrastructure company on the continent faces some combination of these three.