THE BOTTOM LINE
Africa will not own the platform layer of its digital economy. Hyperscale cloud and foundation models belong to global players with the capital and R&D scale to win them. But Africa can own — and in several verticals already dominates — the four layers where most of the distributable value actually accrues: the payment and identity rails, the vertical software franchises (telematics, insurtech, logistics), the sovereign and regulated perimeter, and the integration and transformation services that stitch it all together. The next five years will be decided by who can integrate, not who can build.
Our last piece argued that AI has turned Africa’s connectivity story into a compute-and-power story. That remains true. But infrastructure is a means, not an end. A 40 MW AI-ready hall in Johannesburg is only worth what the workloads running inside it are worth.
So this piece goes up the stack. What is actually happening at Sub-Saharan Africa’s application and services layer — cloud, AI and digitisation, telematics, fintech, superapps, online retail, and the integration and transformation services that connect them? Who is winning? Is there room for African champions, or does this layer get colonised by international platforms the way so many others have?
The answer is more interesting than the usual binary. Every figure below is sourced, and every operator financial has been checked against the primary published results.
1. The stack, and where the value is migrating
It helps to be precise about what “the layer above infrastructure” actually contains. In practice there are five, and they are repricing at very different speeds.
| Layer | What it is | Who owns it today in SSA | Margin & defensibility |
|---|---|---|---|
| L1 — Physical | Subsea, fibre, towers, data centres, power | Mixed: African operators + global capital (Teraco, Cassava, IHS, Helios, Maziv) | Low margin, high capex, high defensibility (asset moat) |
| L2 — Platform | Public cloud, sovereign cloud, GPU-as-a-service | Overwhelmingly international (AWS, Azure, Google, Oracle, Huawei) with local challengers | High margin, winner-takes-most |
| L3 — Enablement | Payments rails, identity, data, APIs, connectivity-as-a-service | Strongly African (M-PESA, MoMo, Moniepoint, Flutterwave, Onafriq) | High margin, network-effect moat |
| L4 — Applications | Superapps, e-commerce, telematics, insurtech, agtech, healthtech | Contested — genuine African leadership in several verticals | Variable; distribution is the moat |
| L5 — Services | Strategy, integration, managed services, digital transformation, AI adoption | Fragmented: global SIs, telco enterprise arms, local boutiques | Asset-light, people-constrained, chronically undersupplied |
The pattern worth internalising: Africa has been most successful where the constraint was local — distribution, trust, regulation, informality — and least successful where the constraint was capital intensity plus global R&D scale. That single sentence predicts most of what follows.
2. The demand base: still small, still steep
The macro numbers still frame everything.
- Sub-Saharan African mobile internet usage sits at roughly 38%, versus around 68% of the world using the internet (GSMA, Mobile Economy Africa 2026; ITU, 2025).
- Nearly 1 billion Africans — about 63% — still do not use mobile internet, even under coverage (GSMA, Mobile Economy Africa 2026).
- Mobile already contributes $240 billion, or 7.8% of Africa’s GDP (2025), rising to ~$290 billion by 2030 (GSMA).
- Digital technologies broadly could contribute up to $1.5 trillion to GDP by 2030 (Brookings, Foresight Africa 2025–2030).
Globally, public cloud spending will pass $1 trillion in 2026 and total ICT spend will reach $4 trillion, with software the largest category for the first time (IDC). Africa’s share of that is a rounding error today. That is precisely the arbitrage.
3. Layer by layer: what is actually happening
3.1 Cloud — the layer Africa is most at risk of losing
This is the least ambiguous story on the continent, and it is not a comfortable one.
South Africa’s cloud market reached $6.44 billion in 2025 and is projected at $40.1 billion by 2034 — a 21.8% CAGR — driven by hyperscaler localisation, enterprise migration and POPIA-driven data residency (IMARC; single-source estimate, methodologies vary). In managed services, the public-cloud deployment model already holds 76.7% of the African market and is growing at 12.8% CAGR; South Africa commands 22.8% of regional share while Egypt grows fastest at 13.0% CAGR (Mordor Intelligence).
Who captures it? Microsoft Azure, Google Cloud, AWS, Oracle and Huawei — increasingly by deploying into local high-tier colocation in Kenya, Nigeria and Morocco rather than building their own shells (Africa Data Centre Association, Data Centres in Africa 2026). Microsoft and G42 committed $1 billion to a Kenyan green data centre plus an Azure East Africa region. Kenya launched an AI & Cloud Pioneer Initiative with Huawei Cloud, and Safaricom partnered with iXAfrica to deliver East Africa’s first AI-ready data centre services.
Read that carefully. African players are supplying the real estate, power and fibre. International players are supplying the platform and booking the software margin. That is the L1-versus-L2 value split, playing out in real time.
The counter-move is sovereignty. Cassava Technologies became Africa’s first NVIDIA Cloud Partner, deploying AI Factories in South Africa with expansion to Nigeria, Kenya, Egypt and Morocco, explicitly to “keep data within African borders” and build models in Swahili, Zulu and Afrikaans.
And — to be precise about a partnership that is frequently mis-stated — Cassava and Accenture announced a formal strategic collaboration in September 2025, under which Accenture’s AI Refinery platform is deployed on Cassava’s NVIDIA-backed GPU-as-a-Service, hosted in Cassava’s own data centres. It starts in South Africa and extends to Egypt, Kenya, Morocco and Nigeria, targeting financial services, mining, telecoms, agriculture and healthcare. Read the structure carefully, because it is the template for this whole layer: African infrastructure and sovereignty positioning, international platform IP and delivery capacity. Neither side could sell the proposition alone.
Does data localisation actually favour local players? Less than everyone assumes.
Regulation is doing real work at this layer — but a different kind of work, and less of it, than most investment memos claim.
The instruments are real enough. Kenya’s Data Protection Act allows the Cabinet Secretary to prescribe in-country processing on grounds of “strategic interests of the state or protection of revenue”. Nigeria’s Data Protection Act 2023 restricts cross-border transfers by default, with penalties of ₦10 million or 2% of annual gross revenue (ITIF). South Africa’s National Policy on Data and Cloud (2024) makes SITA the state’s cloud champion. Nigeria’s National Cloud Policy 2025 goes furthest of all: it designates indigenous system integrators as the primary entities for provisioning government cloud services, while creating a Strategic Investment Waiver that gives hyperscalers a conditional entry path in exchange for verifiable in-country infrastructure commitments.
But here is the part that gets mis-stated. Data residency on its own is a weak moat. POPIA, Kenya’s DPA and Nigeria’s NDPA are GDPR-derived instruments that restrict cross-border transfer — not foreign ownership. If the bytes rest in Azure Johannesburg or AWS Cape Town, the transfer never occurs and the obligation is met. A local provider gains nothing from the residency rule itself. And the hyperscalers are closing the remaining gap fast, with regional data boundaries, customer-held encryption keys and National Partner Clouds operated by domestic entities. Since Davos 2026 the debate has already moved from where data sits to who can govern, port and withdraw it (Detecon).
So where is the local advantage genuinely real? Four places — and only one of them is residency law.
- Geography. Sub-Saharan Africa has full hyperscaler regions in essentially one country: AWS Cape Town, Azure Johannesburg and Cape Town, Google Cloud Johannesburg. Kenya and Nigeria have interconnect and colocation, not regions. In forty-odd markets, “keep it in-country” cannot be satisfied by a hyperscaler at all, and the compliant option is local colocation or a local CSP. That is a real moat — but it is a moat made of absence, it surrounds small markets, and it evaporates the day a region lands.
- Jurisdiction, not location. Residency is not sovereignty. A US-incorporated provider remains reachable by US legal process wherever the data physically sits. Irrelevant for most commercial workloads; decisive for a narrow set — defence, intelligence, central-bank core systems, national identity, tax and voter registries. Small in volume, very large in political salience. It is exactly why Cassava sells sovereign AI rather than cheap AI.
- Procurement policy, not privacy law. Nigeria’s indigenous-SI mandate is industrial policy wearing a compliance costume. It is the strongest pro-local lever at this layer — and it is granted rather than earned, which means it can be withdrawn.
- Currency. Hyperscalers bill in dollars. In Nigeria, Egypt, Ghana and Ethiopia, FX scarcity is a live procurement constraint, and a provider invoicing in naira or cedi against a local balance sheet solves a treasury problem rather than a compliance one. This almost certainly wins more mid-market West African deals than POPIA ever has, and it is almost never modelled.
Verdict: Internationals win the platform. Africans win the substrate, the state and systemically-important perimeter, the markets with no region at all, and the local-currency mid-market. Local pure-play IaaS competing head-on with a hyperscaler on price is not an investable thesis — and neither is “we are compliant because we are local.” Underwrite the procurement mandate, the FX position and the sovereignty use case, not the residency rule.
3.2 AI and digitisation — a readiness gap, not a technology gap
Africa’s AI constraint is not model access. It is absorptive capacity.
Government AI readiness on the continent is led by Egypt, Mauritius, South Africa, Rwanda, Senegal, Seychelles, Tunisia, Kenya and Nigeria (Oxford Insights, via Mastercard). Egypt’s National AI Strategy 2025–2030 targets ICT at 7.7% of GDP, 250+ AI companies and 30,000 trained AI professionals by 2030. That is a policy blueprint other markets are copying.
But globally, McKinsey finds nearly nine in ten organisations use AI while most have not embedded it deeply enough to see enterprise-level benefit — 88% report using AI in at least one function, yet only 39% attribute any EBIT impact to it (McKinsey, State of AI 2025) — and Microsoft’s diffusion data shows a widening, not narrowing, global divide (Microsoft AI Economy Institute). In markets with thin data estates, legacy core systems and scarce MLOps talent, the pilot-to-production gap is wider still.
This is the most under-priced opportunity in the whole stack, and it is a services opportunity, not a product one. The bottleneck is not “can we get GPT-class models in Lagos” — it is data readiness, process redesign, governance, change management and integration into core banking, billing, ERP and network OSS/BSS.
3.3 Fintech — the layer Africa unambiguously owns
Here the continent is not catching up. It is the global reference case.
- Sub-Saharan Africa accounts for roughly two-thirds of global mobile money activity — and by some measures as high as 74% (GSMA State of the Industry 2026, which puts the SSA share at ~66%; BCG cites ~74% on a transaction-volume basis).
- African fintech revenues are projected to grow ~13x to ~$65 billion by 2030 — the highest growth multiple of any region globally (BCG). McKinsey’s framing: a $312 billion total African financial-services market by 2028, of which $47 billion is fintech.
- Eight of Africa’s nine unicorns are fintechs (IFC, 2025).
The operating numbers are what convince investment committees:
| Player | Scale metric | Source |
|---|---|---|
| M-PESA (Safaricom) | 45.6% of Safaricom Kenya’s service revenue (KSh 182.7bn); 40.99m one-month-active customers; 46.4bn transactions worth KSh 41.68tn (FY26) | Safaricom FY26 |
| Moniepoint | ₦412 trillion (~$297bn) processed and 14bn transactions in 2025; ~$22bn monthly; 10m businesses and individuals | Moniepoint; TechCabal |
| Flutterwave | Series E at a $3.25bn valuation | Reported |
| MTN Group | 307.2m subscribers across 19 markets; R218.5bn service revenue, +22.9% (FY2025), with MoMo a core growth engine | MTN Group FY2025 |
| Airtel Africa | FY2026 revenue up 29.5% to $6.4bn, at a record 49.3% EBITDA margin; Airtel Money at 54.1m customers | Airtel Africa FY26 |
The caveat is capital concentration. In 2025 African startups raised roughly $3.2–3.9 billion across ~506 deals, but Kenya alone took nearly $1 billion — almost a third of the continental total — and Egypt $614 million (Africa: The Big Deal, 2025 in Review). The Big Four (Nigeria, Kenya, South Africa, Egypt) absorb close to 80% of startup funding (UNCDF). In H1 2026, the continent raised ~$1.4 billion, with fintech at 41% (Africa: The Big Deal).
That concentration is the opportunity, not the objection. Perfectly good businesses in Ghana, Tanzania, Zambia, Senegal, Côte d’Ivoire and Malawi are underfunded for reasons of investor coverage, not unit economics.
3.4 Telematics and mobility data — the quiet African champion
If you want proof that African firms can build global-grade software businesses, telematics is the sector to point at — and Cartrack is only the most visible name in a genuinely deep local field.
Cartrack passed 2 million subscribers in South Africa in Q4 FY2026, with South African subscription revenue up 20% to ZAR 3,468 million and group SaaS ARR of ZAR 5,179 million, up 18% (38% in USD terms) (Karooooo FY26, SEC filing). Group subscribers reached roughly 2.6 million, up 16%, with subscription revenue at 98% of total revenue and adjusted free cash flow up 90% to ZAR 809 million.
That is a Johannesburg-born, NASDAQ-listed, high-retention SaaS business with African operations in Botswana, Kenya, Malawi, Mozambique, Namibia, Nigeria, Rwanda, Tanzania and Zimbabwe — competing globally.
It is not one company. It is a cluster.
This matters, because a single success is an anecdote and a cluster is an industry. South Africa’s fleet-management installed base reached roughly 2.3 million active systems at the end of 2023 and is forecast to hit 3.8 million by 2028 at a 10.3% CAGR, with penetration of business fleet vehicles rising from 47.3% to 70.0% — well above international norms (Berg Insight). The market is led by Cartrack, Tracker, MiX by Powerfleet, Ctrack and Netstar — every one of them South African in origin, and together roughly 70% of systems in use.
What is genuinely instructive is that these five have taken four completely different ownership paths out of the same market — and each one is a different answer to the local-versus-international question.
| Player | Ownership path | What it demonstrates |
|---|---|---|
| Cartrack (Karooooo) | SA-founded, NASDAQ-listed, still African-run and African-controlled | The full export path — global capital markets without losing operational control |
| MiX Telematics | SA-founded; all-stock merger into US-listed Powerfleet (April 2024), MiX holders taking 65.5% of the combined group, which then acquired Fleet Complete | Absorption into a global platform — African shareholders won, but the head office moved |
| Ctrack | Bought back from NASDAQ-listed Inseego by Convergence Partners (African PE) in 2021 — then bought Inseego’s remaining international telematics business for $52 million | Reverse flow: African PE acquiring European, UK and Australasian operations |
| Netstar | Wholly owned by JSE-listed Altron; pioneered the SA category in 1994 | The corporate cash engine — kept African, funding the parent’s transformation |
| Tracker | Long-standing independent SVR incumbent with deep insurer distribution | Insurance-channel distribution as the durable moat |
The Ctrack story is the one we would put in front of a sceptical investment committee. An African private-equity firm bought a telematics business back from a NASDAQ-listed American owner, and then used it as the platform to acquire that owner’s operations in the United Kingdom, Belgium, France, Germany, the Netherlands, Australia and New Zealand. Capital flowing from Johannesburg to Birmingham and Brisbane, in a software category, is not the direction the standard narrative predicts.
Netstar makes the profitability case just as plainly. Altron’s Platforms division — Netstar, FinTech and HealthTech — delivered R4.4 billion of revenue (+12%), R1.8 billion EBITDA (+21%) at a 41.4% margin, and R1.2 billion operating profit (+45%) in the year to February 2026, with 91% annuity revenue. Netstar itself grew revenue 9% to R2.5 billion, with South Africa up 11%. Those platforms now generate around 95% of Altron’s operating profit, funding a R500 million special dividend and effectively bankrolling the group’s exit from legacy IT services (Altron FY2026 results).
Read the two together and the pattern is unmistakable: telematics is the one layer where African companies have achieved genuine consolidation power, foreign acquisition capacity and 40%-plus EBITDA margins simultaneously. No other segment in this article can claim all three.
The adjacency is already visible in logistics fintech. Korridor — Centurion-based, ECP-backed — runs a cashless cross-border expense platform for over 2,000 customers and more than 35,000 trucks across 13 countries, handling fuel, tolls, border fees and en-route payouts, and is building its own truck stops in Zambia. That is telematics data converging with payments and physical infrastructure — a combination no international platform is positioned to assemble.
Why did Africa win telematics? Because the demand driver was local: vehicle theft and recovery, informal fleet operations, insurance risk pricing in markets with poor actuarial data, and financing of vehicles for gig drivers. No Silicon Valley product team was going to solve that. The adjacencies are now obvious — usage-based insurance, asset finance and collateral monitoring, cold-chain and logistics visibility, and fleet electrification analytics.
3.5 Superapps — four armies marching on the same customer
Africa’s superapp race is not a startup story. It is being fought by incumbents who already own the customer, and they are converging from four directions at once: telcos moving up from payments, banks moving out from the account, retailers and insurers moving out from footfall and loyalty, and mobility platforms moving out from the ride.
Telcos are the largest force — and increasingly they are building on Chinese architecture. Vodacom’s VodaPay, built on Alipay’s mini-programme model from launch, reached 11.9 million registered users by FY25 (Vodafone Annual Report 2025). At group level Vodacom now serves 237.3 million customers and 103 million financial-services customers, with financial-services revenue up 19.6% to R16.8 billion in FY2026 (Vodacom FY26 results).
The bigger signal is MTN. In June 2026, MTN Group Fintech announced a strategic partnership with Ant International — the operator of Alipay — to rebuild MoMo as a super-app with a mini-app platform, launching first in Nigeria (MTN; ITWeb Africa). MTN’s Q1 2026 fintech revenue rose 22.4% to $461.3 million, on $163 billion of quarterly transaction value, off a base of 69.5 million active MoMo users at the close of FY2025.
Pause on what that means. Two of Africa’s three largest telco wallets are now being rebuilt on Alipay-derived mini-app architecture. The distribution is African. The platform pattern is imported. That is the L3-versus-L2 split appearing inside the wallet itself.
Banks are building the same thing from the other side. Standard Bank’s super app grew 12% to 3.5 million users, with digitally active retail clients in South Africa above 4.5 million (Standard Bank; Euromoney). Discovery Bank is explicitly repositioning as a “super bank”, and in December 2025 became the first South African bank to offer in-app crypto trading, through Luno (Reuters).
Retailers are the dark horse, and are underestimated. Shoprite’s Money Market account — backed by African Bank, available on USSD, WhatsApp and the Shoprite app, with cash-in and cash-out at 25,000 tills, linked to the Xtra Savings rewards base and open to foreign nationals without a traditional bank account — is a distribution network no bank in the country can replicate. Nedbank runs Avo, a bank-owned super-shop. Checkers Sixty60 turned grocery delivery into a daily-use app.
Insurers hold the deepest behavioural data. Discovery’s Vitality platform now reaches over 42 million people across 41 markets, and Vitality Health International is exporting the behaviour-linked pricing model into Ghana, Kenya, Nigeria and Zambia.
| Attacker | Asset they start with | What they are missing |
|---|---|---|
| Telcos (MTN, Vodacom, Safaricom, Airtel) | Wallet, agent network, USSD reach, billing relationship | Merchant depth and commerce UX — hence Ant International and Alipay architecture |
| Banks (Standard Bank, Discovery, Nedbank) | Balance sheet, credit data, regulatory licences | Daily-use frequency and informal-economy reach |
| Retailers (Shoprite, Pick n Pay, Nedbank’s Avo) | Physical footfall, cash handling, loyalty base, trust | Financial-product depth — solved by bank partnerships |
| Insurers (Discovery / Vitality) | Behavioural data, incentive design, cross-sell economics | Transactional frequency outside claims and rewards |
| Mobility (Yango, Bolt, Uber, inDrive) | Daily-use habit, driver and courier supply, logistics | Regulated financial licences and float |
And the mobility platforms are still coming. The African ride-hailing market is $2.53bn (2025) growing to $3.25bn by 2031 at 4.25% CAGR (Mordor Intelligence) — a modest number that badly understates strategic value, because the app is a distribution channel for far more.
Watch Yango. The Dubai-based group is committing $150 million to enter 10 new African markets in 2026, deliberately targeting secondary cities and Francophone West and Central Africa rather than fighting Uber and Bolt in the Big Four, and bundling ride-hailing, food delivery, parcel logistics, payments and vehicle financing (Bloomberg; Dealroom). This is the clearest live test of the local-versus-international question: well-capitalised internationals attacking exactly the underserved geographies where local players assumed they were protected.
3.6 Online retail — the hardest unit economics on the continent
Africa’s e-commerce market is estimated at $40.5 billion in 2025 growing to $56 billion by 2029 (8.5% CAGR), with over 60% of transactions on mobile, Nigeria at 26% of continental revenue and South Africa the most mature market (Lengow / Statista; single-source estimate, methodologies vary). Social commerce alone is forecast to reach $9.43 billion by 2030 at 16.2% CAGR (Research and Markets).
But this is now the most internationally contested layer on the continent — more so than cloud.
Amazon launched in South Africa in May 2024 and recruited over 10,000 sellers in its first year. On ECDB’s 2025 ranking, the leading online retail platforms in Africa by GMV are Amazon, Takealot and Jumia — followed by Shein, Temu and Noon. AliExpress has been quietly localising the boring but decisive part: local-currency payment, adding M-PESA in Kenya and OPay and Verve cards in Nigeria, then extending rand, pound, dirham, birr and shilling settlement across South Africa, Egypt, Morocco, Algeria, Ethiopia, Tanzania and Uganda.
The Chinese fast-fashion platforms have been the most disruptive. In Nigeria, Temu and Shein together account for roughly 37% of the clothing, textiles, footwear and leather e-commerce market. Jumia has responded by onboarding Chinese merchants directly onto its own platform rather than trying to out-price them (Rest of World) — while continuing to shrink its footprint, exiting South Africa and Tunisia and winding down Algeria by Q1 2026.
And here trade policy, not technology, has been the decisive weapon. South Africa’s revenue service removed the de minimis concession from 1 July 2024, applying the standard 45% import duty plus 15% VAT to low-value parcels. Independent transaction data shows Shein’s South African market share falling materially since the change (Reuters; BusinessTech). One customs ruling did more for local retailers than a decade of platform investment.
Against that backdrop, profitability has finally diverged between the two flagship African cases:
| Company | Latest full year | Outcome |
|---|---|---|
| Jumia | FY2025: revenue growth to GMV $818.6m (+14%) | Still lossmaking — operating loss $63.2m, adjusted EBITDA loss $50.5m (Jumia IR) |
| Takealot Group | FY2026 (to Mar 2026): revenue +19% to R16.4bn (~$1bn) | Profitable for the first time — aEBIT R180.4m (2025: −R213.2m) (Naspers / Takealot) |
Takealot’s turnaround is the more instructive story. It held share against Amazon’s entry and reached its first full-year profit on the back of gross-margin discipline, retail media revenue and a subscription programme — TakealotMORE now drives 27% of takealot.com GMV — while takealot.com revenue rose 19% to R14.86bn and Mr D grew revenue 11% to R2.26bn at a stable R65.6m aEBIT.
The lesson is not “e-commerce doesn’t work in Africa.” It is that the retail margin is thin, and the logistics, payments, subscription and retail-media layers underneath it are where the durable value sits. Jumia’s own disclosure that over half of Nigerian orders now come from secondary cities, served by 350+ pickup stations and 67 logistics partners, is the tell: the moat is physical last-mile plus payment trust, not the storefront.
Does the distribution moat actually hold? South Africa has just run the experiment in public.
This deserves evidence rather than assertion, because the test case is unusually clean.
Amazon launched amazon.co.za in May 2024. In the financial year that followed, Takealot.com did not merely survive — GMV rose 13%, revenue 17% and order volumes 15%. Prosus and Naspers Group CFO Nico Marais said so on the record: “Their performance in the last year was ahead of our expectations… we actually saw Amazon moving, probably not at the speed that we originally expected, which was to our benefit” (Reuters). It is rare to get the incumbent’s own CFO conceding that the attacker under-delivered.
Twenty-five months in, the scoreboard has barely moved. Similarweb’s June 2026 South African e-commerce ranking reads takealot.com first, makro.co.za second, temu.com third, amazon.com fourth, woolworths.co.za fifth — and note that the fourth entry is the global Amazon domain, not the local site, which does not make the top five at all. Takealot was running roughly 12.3 million visits in May 2026, with 96% of traffic originating in South Africa.
The physical asymmetry explains why. Amazon’s South African fulfilment footprint is two nodes — Eastport Logistics Park in Kempton Park and Montague Gardens in Cape Town. That covers Johannesburg and Cape Town well and the rest of the country badly. Against it, Takealot runs a multi-warehouse network, its own last-mile fleet through Mr D, its own pickup-point estate, and integration with Pargo’s 4,000+ click-and-collect points sitting inside spaza shops, pharmacies and filling stations. That collection layer was assembled franchise by franchise over a decade, and it is the part that reaches customers without a reliable street address.
Amazon’s own accounts confirm this is intrinsically slow work rather than a failure of intent. Its fulfilment costs reached 16.4% of net sales in 2022, and the modest improvement to 15.8% in 2023 was only the second time that ratio had fallen in fourteen years (Harding Loevner). The company took roughly a decade to move from outsourced logistics to owning its network end to end, and it has been building in India since 2013 without becoming the uncontested leader.
So the defensible claim is not that Amazon cannot build this. It is that distribution is a five-to-ten-year capital programme rather than a single capex cycle — and Amazon entered South Africa with two warehouses against fifteen years of incumbent head start and a collection network embedded in the informal economy.
One caveat, stated plainly, because it is the strongest objection: Temu ranks third in South Africa with no local distribution whatsoever. It ships cross-border from China and still outranks Amazon. Where price and assortment are extreme enough, consumers will wait two weeks. Which sharpens rather than weakens the conclusion — the logistics moat protects Takealot from Amazon; it is customs policy that protects it from Temu and Shein. Two different threats, two different defences, and only one of them was built by the company itself.
3.7 Integration and digital transformation services — structurally undersupplied
The least glamorous layer, and the one with the cleanest supply-demand imbalance.
Globally, system integration services are a $763.8 billion market by 2030 and managed services $705.2 billion by 2031 at 8.9% CAGR (MarketsandMarkets; single-source estimates). In Africa, managed data-centre and hosting leads at 37.0% of the market, managed security is the fastest-growing at 12.3% CAGR, and SMEs are 55.7% of demand (Mordor Intelligence).
Meanwhile the supply side is consolidating upward and away from the mid-market: Dimension Data — founded in Johannesburg in 1983 — has now fully rebranded as NTT DATA across the Middle East and Africa (NTT DATA). And IDC records a global shortage of roughly 4 million software engineers.
Who else is actually delivering this work in Africa?
NTT DATA is the most visible name, but it is far from alone. The African integration and managed-services bench is deeper than most investors assume — it is simply fragmented and poorly covered.
- Datatec / Logicalis — JSE-listed, group revenue above $5 billion, with Logicalis International running integration and managed services across 21 countries with more than 4,000 employees (February 2026), plus Westcon in distribution and Mason Advisory in consulting. The most credible African-headquartered global SI.
- Liquid C2 (Cassava Technologies) — the cloud and cyber-security arm built on Africa’s largest fibre network. It became the first Google Cloud Interconnect provider on the continent, launched Liquid G, an AI-driven Google Cloud distribution programme for African resellers, and runs Secure CloudConnect into Microsoft Azure. Its Egyptian business, Cloudmania, won Microsoft’s 2025 Partner of the Year award for Egypt.
- Altron — a South African technology group since 1965, now refocused on managed services, digital platforms and fintech infrastructure after divesting non-core assets.
- Telco enterprise arms — BCX (Telkom), MTN Business and Vodacom Business, each bundling connectivity, cloud and managed security into the enterprise and public-sector base they already bill.
- A long tail of specialists — Egypt, Kenya, Nigeria and Morocco each have credible mid-market integrators, and Egypt in particular is exporting delivery capacity as a near-shore centre for Gulf and European clients.
And what about the Big Four and Accenture?
They are the most powerful competitors at the top of this layer, and their role is widely misunderstood.
The scale is simply not comparable to anything African. Deloitte reported $70.5 billion of global FY2025 revenue, PwC $56.9 billion, EY $53.2 billion and KPMG $39.8 billion — over $220 billion between them — with Accenture at roughly $65 billion and making 23 acquisitions in 2025 alone. Consultancy.africa assessed over 800 firms with an African presence for its 2025 ranking, and the leaderboard is dominated by exactly these names.
Their advantage on the continent is not local knowledge. It is four things that are genuinely hard to replicate:
- Global alliance economics. Accenture is a Leader in Gartner’s 2025 Magic Quadrant for Public Cloud IT Transformation Services and holds the highest partner tiers with AWS, Azure, Google, SAP, Oracle, ServiceNow and NVIDIA. Those tiers come with co-funding, rebates and vendor-sourced pipeline a local firm cannot access.
- Reusable IP and accelerators. AI Refinery, industry data models, pre-built migration factories. The work is priced as a product, delivered as a service.
- Global delivery arbitrage. Work is staffed from India, Egypt, Poland and Manila and fronted locally. That structurally undercuts a purely local cost base on volume work.
- Procurement comfort. For a bank board or a sovereign fund, the Big Four brand is risk transfer. Nobody is criticised for hiring them.
Is that a barrier for local contenders? Partly — but not where it matters most. These firms optimise for large, standardised, repeatable programmes. They are structurally expensive for the mid-market, thin on genuine regulatory and political context outside the Big Four economies, and heavily reliant on junior pyramids that clients increasingly resent paying for. And the Cassava–Accenture deal shows the actual equilibrium: the global firm needed an African partner with infrastructure, sovereignty positioning and market access.
The defensible local positions are therefore specific: senior-only advisory where judgement rather than leverage is the product; regulated and sovereign work where local standing is a licence to operate; mid-market delivery below the global cost floor; and sub-scale geographies the tier ones will not staff. That is a real market — it is just not the one the Big Four are measuring.
A large African bank, utility, mining group, government agency or telco that wants to modernise faces a genuine gap between global-tier-one pricing and local capacity. That gap is the single most reliable services opportunity on the continent.
4. Local versus international: the honest scorecard
No single answer holds across the stack. Here is where we would actually put money on Africans winning — and where we would not.
| Segment | Who wins | Why |
|---|---|---|
| Hyperscale public cloud (IaaS/PaaS) | International | Global R&D scale, capex depth, developer ecosystem lock-in |
| Sovereign / regulated cloud & AI compute | Local, but narrower than assumed | Residency alone is satisfied by an in-country hyperscaler region; the real edge is state and systemically-important workloads, procurement mandates, the 40+ markets with no region, and local-currency billing |
| Payments and mobile money rails | Local, decisively | Agent networks, regulatory licences, informal-economy distribution, ~two-thirds of global mobile money activity |
| Vertical SaaS (telematics, insurtech, agtech, logistics) | Local, decisively — and now acquisitive abroad | Problem definition is local; all five leaders in SA fleet management are African in origin, Cartrack proves the listing path and Ctrack has bought European and Australasian operations outright |
| Superapps and mobility | Contested — African distribution, imported architecture | Telcos, banks, retailers and insurers all converging; MTN on Ant International, VodaPay on Alipay’s model, Yango’s $150m secondary-city push |
| Online retail marketplaces | Contested — and tilting international | Takealot has held Amazon off for two years on last-mile, pickup networks and subscription — but Temu outranks Amazon in SA with no local logistics at all, so customs policy is doing as much work as the moat |
| Foundation models | International | No realistic path to frontier-model competition; the win is fine-tuning and African-language adaptation |
| Integration, managed services, transformation | Split: Big Four take the top, locals take the middle | Global alliance economics and reusable IP win large programmes; proximity, senior judgement and cost structure win the mid-market and the sovereign perimeter |
The synthesis: Africa will not own the platform layer. It can absolutely own the rails, the verticals, the sovereign perimeter, and the services — and those four together are where most of the distributable value in this stack actually accrues.
5. Where we would deploy capital, and what we would underwrite
Five theses, in rough order of conviction:
- Sovereign and regulated cloud/AI compute. Be precise about the moat, because residency law alone is satisfied by a hyperscaler region — do not pay for it. Underwrite the procurement mandate, the classified or systemically-important workload, the absence of an in-country hyperscaler region, the anchor public-sector tenancy, the power contract, and the ability to invoice in local currency.
- Vertical SaaS with a hardware or data moat. Telematics is the proven template; extend into UBI insurance, asset finance monitoring, cold chain, and energy/mini-grid analytics. Underwrite churn, ARPU and net revenue retention — not TAM.
- Second-wave fintech beyond payments. Payments are crowded; lending, embedded finance, treasury, FX and B2B rails are where BCG’s 13x growth actually sits. Underwrite credit quality and funding cost, not GMV.
- Non-Big-Four geographies. With ~80% of capital in four markets, competent operators in Ghana, Tanzania, Zambia, Senegal, Côte d’Ivoire and Malawi are mispriced. Underwrite FX, regulatory stability and exit path.
- Digital transformation and managed services platforms. Asset-light, cash-generative, roll-up-friendly, and structurally short of supply. Underwrite talent retention and delivery capacity above all.
And four risks that kill deals at this layer:
- FX and repatriation. Revenue in naira, kwacha or shilling against dollar-denominated cloud and hardware costs. This has destroyed more African tech returns than product failure.
- Regulatory whiplash. Localisation and local-content rules are the strongest pro-local lever at this layer — and they are political grants, not earned advantages. Nigeria has already written a Strategic Investment Waiver for hyperscalers into its own 2025 cloud policy. What is legislated in can be waived out.
- Talent. Against a global 4-million-engineer shortfall, African firms compete with remote-first employers paying dollar salaries.
- Exit. Thin local capital markets and a narrow strategic-buyer universe. Design the exit at entry.
6. A 2010–2030 view of how this layer evolved
| Era | Defining shift | Value capture |
|---|---|---|
| 2010–2015 | Mobile money and voice/SMS monetisation; first subsea capacity wave | Telcos capture nearly everything |
| 2016–2020 | Smartphone and 4G scale; VC-funded app layer; e-commerce land grab | Value fragments to startups; most burn cash |
| 2021–2025 | Fintech consolidation and profitability; hyperscaler regions land; data-residency law arrives | Split: rails stay African, platform goes international |
| 2026–2030 | AI absorption, sovereign compute, vertical SaaS export, services consolidation | Decided by who can integrate — not who can build |
7. How Digitalthings assists
This is the same terrain we have worked for over a decade — now one layer up.
Digitalthings is a boutique, investment-and-strategy-focused advisory firm serving the TMT and digital-infrastructure ecosystem across Africa, combining deep local insight with global best practice and assigning senior practitioners to every mandate. We work with investors, operators, governments, technology companies and DFIs across the full business lifecycle.
At this layer specifically, we support clients with:
- Transaction and investment advisory — commercial and technical due diligence on cloud, SaaS, fintech, telematics and platform assets, red-flag reviews, and investment-grade techno-economic models. Our technology due-diligence work spans fleet-telematics and logistics-platform businesses alongside fibre and data-centre assets.
- Cloud, AI and digitisation strategy — target operating model, cloud and data-residency strategy, build-versus-buy-versus-partner, AI use-case prioritisation and value-case construction, and realistic adoption roadmaps that survive contact with legacy systems.
- Market entry and expansion — market sizing, competitive mapping, partnership and channel strategy, and regulatory pathway assessment for operators and investors entering new African markets.
- Business planning and financial modelling — subscription and platform economics, cohort and churn modelling, ARPU build-ups, and capital structuring across FX-exposed markets.
- Programme and interim management — hands-on delivery leadership for transformation and rollout programmes, including interim executive roles where capacity is the constraint.
- Regulatory and policy advisory — data protection and localisation, licensing, open-access and universal-service frameworks, and engagement with regulators and policymakers.
- DFI and institutional advisory — supporting development finance institutions and infrastructure investors on digital-economy strategy and portfolio diligence.
If you are underwriting a cloud, AI, fintech, mobility-data or platform asset in Sub-Saharan Africa — or you are an operator deciding how far up this stack to move — the difference between a good and a bad outcome is senior, local, investment-grade judgement. That is what we do. Let’s talk.
Prepared by Digitalthings. Every operator financial in this piece has been checked against the primary published results: Safaricom FY26 (6 May 2026), Airtel Africa FY26 (8 May 2026), MTN Group FY2025, Vodacom FY26, Karooooo FY26 and Altron FY2026. Market-sizing estimates vary materially by methodology, scope and definition of “Africa” versus “Sub-Saharan Africa”; single-source forecasts are flagged as such and should be validated against primary sources before transaction use. Figures reflect the most recent publicly available data as of mid-2026.
Sources
Primary company results
- Safaricom PLC — FY2026 audited results (7 May 2026)
- Airtel Africa plc — FY2026 results (8 May 2026)
- MTN Group — FY2025 annual results (year ended 31 December 2025)
- Vodacom Group — FY2026 annual results (11 May 2026)
- Karooooo / Cartrack — Q4 & FY2026 results (SEC filing)
- Jumia Technologies — Q4 & FY2025 results
- Naspers / Takealot — FY2026 results
- Altron — audited results for the year ended 28 February 2026
- Vodafone — Annual Report 2025 (VodaPay); Standard Bank — digital customer disclosures
Market and macro research
- GSMA — The Mobile Economy Africa 2026
- GSMA — State of the Industry Report on Mobile Money 2026
- ITU — Facts and Figures 2025
- BCG — Beyond Payments: Unlocking Africa’s Second FinTech Wave
- McKinsey — The State of AI 2025
- IMARC — South Africa Cloud Computing Market 2026–2034
- Berg Insight — Fleet Management in South Africa (3.8m units by 2028)
- Africa: The Big Deal — 2025 in Review
- Africa Data Centre Association — Data Centres in Africa 2026
- MTN Group Fintech — Ant International partnership (June 2026)
- Brookings — Foresight Africa 2025–2030; IDC — global public-cloud & ICT spend; IFC — Venture Capital and the Rise of Africa’s Tech Startups; Mordor Intelligence — Africa Managed Services / Africa Ride-Hailing; Oxford Insights / Mastercard — AI readiness; Microsoft AI Economy Institute; MarketsandMarkets — System Integration / Managed Services; ITIF — Nigeria cross-border data regulation; Lengow / ECDB — African marketplaces; Rest of World — Jumia vs Temu/Shein; Reuters / BusinessTech — de minimis removal; Similarweb — SA e-commerce ranking (June 2026); Harding Loevner — Amazon fulfilment cost structure; Pargo — click-and-collect network; Consultancy.africa / CRN — Big Four & Accenture; Datatec / Liquid C2; RSA National Policy on Data and Cloud 2024 / NITDA National Cloud Policy 2025; Detecon — digital sovereignty; Bloomberg / Dealroom — Yango.
Links are provided for sources verified against their primary publication. Remaining sources are cited by name; their landing pages vary by paywall and region and should be confirmed before transaction use.