Today on The Merchant Desk: Kenya's proposed VAT on mobile money threatens to reverse a decade of financial inclusion gains, Paystack restructures beyond payments into a full commerce stack, and Altron's fintech turnaround delivers a R500m special dividend. Plus — the agentic commerce readiness gap is now measurable, and African mobile money operators are hitting structural revenue ceilings.
Kenya's Finance Bill 2026 proposes removing VAT exemptions for licensed payment service providers' commissions — directly overturning an August 2025 High Court judgment that declared PSP fees VAT-exempt. The move stacks an estimated 16% VAT on top of existing 20% excise duties on M-Pesa fees, which processed 46.4 billion transactions in the year to March 2026. Economists warn this could reverse financial inclusion gains and push consumers back to cash. M-Pesa's market share has already declined from 97% in 2023 to 89% by September 2025 as Airtel Money gains on pricing. Industry pushback from Binance and others warns the measures 'set back financial inclusion by almost a thousand steps.'
Why it matters
This is the most consequential African payments policy development this week. The mechanism matters as much as the tax itself: rather than appeal an unfavourable court ruling, Treasury is using legislation to delete the statutory exemption the court recognised. At M-Pesa's scale, even marginal fee increases change consumer behaviour — the existing excise duty already pushed users to cheaper alternatives. A 16% VAT layer would compress operator margins, raise consumer costs, and risk disintermediating the digital payment rails that underpin Kenya's fintech ecosystem. The simultaneous proposal of 5% withholding tax on card transactions and crypto reporting obligations signals a broader fiscal extraction strategy that operators across East Africa must price into their models.
Safaricom updated its consolidated 'My One App' to allow M-Pesa access via rival mobile networks and Wi-Fi, removing restrictions that locked out diaspora users and non-Safaricom subscribers. The app now supports Face ID and fingerprint biometric verification, replacing PIN-only authentication. While initial activation still requires Safaricom connectivity, ongoing transactions can run on any network. M-Pesa generated Sh182.7bn (45.6% of Safaricom's FY26 revenue) — context: M-Pesa's share has already slipped from 97% to 89% as Airtel Money gains on pricing.
Why it matters
This product move directly addresses the competitive squeeze already underway: removing network lock-in trades exclusivity for addressable market, betting that M-Pesa's ecosystem depth (merchant acceptance, Fuliza lending, savings) retains users even without network friction as a moat. The diaspora angle matters against Kenya's $5bn+ remittance corridor, where network restrictions were pushing senders to alternatives. Critically, this lands the same week Vodacom agreed to raise its Safaricom stake from 35% to 55% — the new majority owner inherits a more open-network product, which may inform how M-Pesa governance evolves post-consolidation.
Paystack announced a corporate restructuring under The Stack Group, separating merchant payments, consumer products (Zap), microfinance banking, and AI/infrastructure labs into distinct entities. Simultaneously, it launched Canvas — a fully rebuilt merchant dashboard with an AI-native Command Center for natural language business data queries — its first ground-up redesign in 10 years. The upgrade includes Storefront and Order APIs for deeper commerce integration, expanded local payment rails (bank transfers, Pesalink, Capitec Pay), WhatsApp commerce integration, and published AI governance frameworks. The platform now serves 300,000+ businesses.
Why it matters
This is Paystack's clearest signal yet that it's following the Stripe vertical-integration playbook adapted for African market realities. The corporate restructuring separates regulated activities (banking) from product velocity layers (commerce APIs, consumer app) — a governance architecture that enables faster iteration while maintaining regulatory compliance. The AI Command Center shifts the dashboard from passive transaction display to active merchant intelligence, which directly attacks the retention problem: merchants who can query their data in natural language are stickier than those exporting CSVs. The Capitec Pay integration is notable — it signals Paystack is building local rail coverage that reduces dependence on card networks in South Africa.
Frontier Fintech published a deep financial analysis of Safaricom M-Pesa, MTN, and Airtel's mobile money divisions, revealing divergent maturity stages. M-Pesa has shifted to merchant/business payments (31% of revenue) from withdrawals (20%), with take rates falling from 93bp to 60bp. MTN's advanced services grew to 44% of fintech revenue in three years. Airtel remains withdrawal-dominant at 48% of revenue. Implied valuations diverge sharply: M-Pesa at ~$7B on $1.4B revenue (5x) versus Airtel at $5.2B on similar revenue — the market is pricing ecosystem depth over transaction velocity.
Why it matters
This is the most data-rich breakdown of African mobile money unit economics published this year. The central insight: the agent-dependent, fee-based cash conversion model that built mobile money is hitting structural limits. Operators earning high margins are doing so on the turnstile (cash-in/cash-out), not ecosystem depth — and that model doesn't compound. The valuation gap between M-Pesa and Airtel quantifies exactly what 'ecosystem transition' is worth to the market. For anyone building payments infrastructure in Africa, this analysis defines the competitive endgame: merchant acceptance, lending, and financial services layering are where the margin is, not transaction volume alone.
Lagos-based Sycamore obtained a microfinance bank licence from Nigeria's Central Bank, enabling it to accept customer deposits and connect directly to NIBSS (Nigeria's national payment settlement system). This is Sycamore's third regulatory pillar after FCCPC lending and SEC asset management licences. The company serves 400K+ customers with ₦50B AUM and $5M revenue — all built organically without external venture capital. Direct NIBSS connectivity eliminates third-party bank intermediaries, reducing settlement time and fees.
Why it matters
Sycamore's licensing arc is a masterclass in vertical integration: start in lending (lower regulatory burden), layer on asset management, then add regulated deposits and direct settlement access. Each licence removes a cost layer and deepens customer lock-in. The $5M revenue on zero VC is the real signal — it proves the unit economics of this model work without subsidised growth, which is exactly what the African funding environment now demands. Direct NIBSS access is operationally critical in Nigeria where intermediary bank fees and settlement delays are material cost and friction drivers. This playbook is directly replicable in other markets where fintechs are currently dependent on bank sponsorship for settlement.
Altron declared a R500m special dividend and lifted its final ordinary dividend 44% as its platform businesses — FinTech, Netstar telematics, HealthTech — now account for 95% of operating profit. Altron FinTech was the standout: operating profit up 33% to R561m on R1.5bn revenue, with POS device rentals more than doubling to over 30,000 units. Group operating profit rose 25% to R1.2bn. The transformation from legacy IT services into recurring-revenue platforms is now three years in.
Why it matters
This is the rare SA tech turnaround that's actually producing shareholder returns rather than just restructuring announcements. The FinTech division's 88% annuity revenue base and doubling POS rental footprint demonstrate real merchant acquisition momentum in a market where Yoco, iKhokha, and bank-owned terminals compete for the same countertops. The R500m special dividend signals the board believes the cash generation is structural, not cyclical. For the broader SA fintech ecosystem, Altron proves that patient platform-building with disciplined capital allocation can work — a useful counterpoint to the venture-funded growth-at-all-costs narrative.
Investec's finance director disclosed technology expenditure of £246m (R5.4bn) in FY26 — 20% of total costs — with 50% directed at growth and platform enhancement. The bank has deployed 800 AI agents working alongside 7,700 employees, with several platform upgrades planned over the next 1-2 years. A £282m (R6.2bn) investment through 2028 targets global lending and transactional banking transformation. Business Day also notes that FNB and Discovery Bank are stress-testing Visa's AI agent platform.
Why it matters
The 800 AI agents figure is the standout data point — it's one of the first concrete production deployment disclosures from a South African bank. At roughly one agent per ten employees, Investec is treating AI as operational infrastructure rather than a pilot programme. The R6.2bn three-year investment in lending and transactional banking signals that Investec's 122,000 new-client target (disclosed last week) will be supported by platform capability, not just relationship bankers. For SA fintech operators, this raises the competitive bar: incumbent banks are spending at a scale that most startups can't match on infrastructure alone. The FNB and Discovery Bank references to Visa's AI agent platform suggest the SA banking sector is moving toward agentic commerce infrastructure faster than the merchant ecosystem is preparing for it.
Visa reports nearly half of US shoppers now use AI tools for shopping tasks, with millions expected to use AI agents to complete purchases by the 2026 holiday season. PYMNTS published a concrete merchant readiness checklist built around three infrastructure requirements: machine-readable product data (structured, real-time, normalised), authorization records for agent-initiated purchases (via frameworks like Mastercard's Verifiable Intent), and registration with card networks (Visa Intelligent Commerce, Mastercard Agent Suite) to enable agent verification and fraud controls.
Why it matters
This moves the agentic commerce conversation from 'interesting future trend' to 'operational checklist with deadlines.' The three requirements — structured data, agent authorization frameworks, and network registration — define a concrete infrastructure gap that most merchants haven't addressed. Merchants without machine-readable product catalogs will be skipped by agents that can't parse their inventory. Those without verifiable agent authorization records face chargeback exposure in scenarios where traditional dispute codes don't apply. The holiday 2026 timeline is aggressive but credible given the pace of Google Universal Cart, Stripe's adaptive checkout, and network-level agent programmes already shipping.
allO, a European AI-native POS and payments platform for restaurants, closed a $14M Series A led by Zigg Capital. The platform serves 1,000+ active locations in Germany with 6x location growth and 3.5x revenue growth year-on-year. It compresses onboarding from two weeks to 30 minutes and is rolling out 'digital employees' starting with a Reservation and Ordering Agent. The system integrates POS, scan-to-order, payment processing, and agentic AI into a unified operating layer.
Why it matters
allO represents the emerging competitive template for restaurant tech: AI-native from the ground up rather than AI-bolted-on. The 30-minute onboarding claim is the most commercially significant detail — Toast, Lightspeed, and legacy POS vendors typically require days-to-weeks of setup, which is the primary friction in restaurant tech distribution. If that compression holds at scale, it fundamentally changes merchant acquisition economics. The 'digital employee' framing (reservation agent, ordering agent) is the same pattern Toast is shipping with IQ Grow — but allO is building it into a system designed for agents from day one rather than retrofitting.
PayJustNow's BNPL service is now fully integrated into Peach Payments' Digit Pro POS terminals, allowing South African merchants to offer BNPL and credit products at physical checkout without manual amount entry or printed QR codes. The integration includes embedded fraud prevention, credit checks, and consolidated transaction reconciliation. Peach Payments simultaneously launched Apple Pay support in Mauritius, extending digital wallet capabilities across its SA-Kenya-Mauritius-Eswatini footprint.
Why it matters
This is one of the first native BNPL-in-POS integrations in the South African market — most BNPL in SA still relies on separate apps or QR code flows that add friction at checkout. By embedding PayJustNow directly into the terminal transaction flow alongside fraud and credit checks, Peach is compressing what was a multi-step process into a single merchant action. The Apple Pay Mauritius expansion is smaller but signals Peach's strategy of bundling global wallet standards (Apple Pay, Google Pay) with local payment rails across its regional footprint — the kind of multi-method orchestration that defines modern acquiring.
African startups raised $887M across 84 deals in the first four months of 2026, on pace to exceed $1B in H1. But capital is concentrating into fewer, larger Series A/B/C rounds from mature companies with proven traction, while early-stage funding has dried up as angel investors demand stronger exits and returns before redeploying. Separately, Branch International laid off African staff while reporting ~$30M global profit for FY2025 — profitable operations and headcount reductions happening simultaneously.
Why it matters
The bifurcation is the story: later-stage companies with revenue and unit economics still access capital, but the seed pipeline that feeds the next generation of African fintech is starving. Branch's case is the pattern in miniature — profitability achieved, then costs cut further to optimise rather than expand. TechCabal's parallel analysis of structural exit illiquidity explains why: capital entered Africa in 2021-22 and got trapped with no secondary market deep enough to convert shares back to cash. Until exit infrastructure improves (deeper local markets, more M&A, functioning secondaries), the funding reset will persist and African fintech growth will be self-funded or not at all.
Pick n Pay returned to profit before tax (R360m vs R237m loss in FY25) but pushed core supermarket break-even one year further to FY29. Boxer posted 12.3% turnover growth versus the core supermarket's 1.6% decline. Online sales surged 32.7% and like-for-like company-owned store sales improved to 3.9%. Labour costs remain at 41.4% of supermarket trading expenses. This follows last week's second Boxer share sale (R4.7bn via accelerated bookbuild) — the two moves together confirm Boxer is funding the core supermarket runway while the turnaround extends.
Why it matters
The FY29 break-even push-out adds a year to a timeline already under scrutiny — the core supermarket remains a project while Boxer and its disposals provide the capital. The consumer signal is structural: Boxer's discount format outperforms as cost-of-living pressure from the diesel crisis and food inflation persists. The 32.7% online growth signals accelerating digital payment adoption across Pick n Pay's base, relevant for payment mix, digital loyalty, and omnichannel merchant tech integration.
Governments are taxing digital payments harder — and faster than operators can adapt Kenya's proposed 16% VAT on mobile money, stacking on existing 20% excise duties, represents a pattern visible across African and emerging markets: governments under fiscal pressure treat digital payment rails as tax collection points. The Kenyan Treasury is literally overriding a court ruling via legislation to extract revenue. This creates a structural margin squeeze for operators and risks pushing transactions back to cash — directly undermining the financial inclusion that made these markets attractive.
African payment platforms are restructuring from payments-only into full commerce stacks Paystack's corporate restructuring into The Stack Group (payments + consumer + banking + AI labs) and Sycamore's three-licence vertical integration in Nigeria show mature African fintechs following the Stripe/Toast playbook: embed payments into broader commerce infrastructure to capture more margin and reduce churn. The competitive battleground is shifting from transaction processing to merchant operating system control.
Mobile money operators are hitting structural revenue ceilings as take rates compress Frontier Fintech's analysis shows M-Pesa's take rate falling from 93bp to 60bp while merchant payments rise as a revenue share. Airtel remains withdrawal-dependent at 48% of revenue. The cash-out revenue model that built mobile money is saturating — the next margin capture requires merchant acceptance, financial services layering, and ecosystem depth that most operators haven't built.
Agentic commerce readiness is now a measurable merchant infrastructure gap Multiple data points this week — Visa's merchant checklist, SA-focused Accenture analysis, and Google's Universal Commerce Protocol rollout — converge on the same finding: merchants without structured product data, machine-readable policies, and reliable payment infrastructure will become invisible to AI purchasing agents. The gap between prepared and unprepared merchants is widening and will show up in conversion rates within 12 months.
African fintech funding is concentrating into fewer, later-stage bets while the seed pipeline dries up $887M across just 84 deals in four months of 2026 means larger cheques going to proven operators, while angel and seed investors demand exits before redeploying. Branch International's profitable-but-still-cutting pattern and TechCabal's structural analysis of the exit problem both point to the same conclusion: African fintech's capital allocation is rationalizing, but the missing exit infrastructure means the funding reset will persist longer than operators hope.
What to Expect
2026-06-05—Trident Digital / Ghana Revenue Authority MSME platform commercial launch — first test of sovereign-scale digital tax formalization in West Africa
2026-06-15—SARB payments regulatory framework comment deadline — third draft of activity-based authorisation regime that opens direct non-bank payment licensing in South Africa
2026-06-30—SA Treasury CFMR crypto regulations extended comment deadline — forced-disposal clause and crypto-to-fiat tightening still in the file
2026-Q3—SARB payments regulatory framework expected final publication — will define competitive boundaries for SA merchant acquiring and payment initiation
2026-H2—Kenya VASP Regulations expected to take effect — formalises $18B digital asset market and establishes stablecoin framework for cross-border settlement
How We Built This Briefing
Every story, researched.
Every story verified across multiple sources before publication.
🔍
Scanned
Across multiple search engines and news databases
685
📖
Read in full
Every article opened, read, and evaluated
216
⭐
Published today
Ranked by importance and verified across sources
12
— The Merchant Desk
🎙 Listen as a podcast
Subscribe in your favorite podcast app to get each new briefing delivered automatically as audio.
Apple Podcasts
Library tab → ••• menu → Follow a Show by URL → paste