Every South African bank that has launched a mobile virtual network operator (MVNO) – Capitec, FNB, Standard Bank, Nedbank and soon Absa – has built its mobile business on phone numbers it does not control.
The international mobile subscriber identity (IMSI) embedded in every Sim card belongs to the host network. The phone number – the MSISDN – belongs to the customer, who can port it between networks at will. The bank controls neither. This is not a design flaw; it is a structural feature of South Africa’s numbering framework. And at the scale these businesses have reached, its commercial consequences have gone largely unexamined.
An MVNO sells mobile services under its own brand without owning the network. Capitec Connect, FNB Connect, Standard Bank Connect, Nedbank Connect and the forthcoming Absa offering control billing, pricing, the customer relationship and the app. None controls the IMSI range from which each subscriber’s network identity is drawn. In South Africa, Icasa assigns IMSI ranges only to spectrum licence holders, and MVNOs hold no spectrum. The range behind Capitec Connect is registered to Cell C.
The MSISDN is portable – mobile number portability gives every subscriber the right to move their number to another network – but that right belongs to the customer, exercised one number at a time. The MVNO cannot port on the customer’s behalf. So Capitec owns the billing layer and the relationship, Cell C owns the underlying network identity and the customer owns the number. Any migration depends on each subscriber choosing to act.
The trap
Capitec Connect is the clearest illustration because it discloses the most. It generated R442-million in net income in the year to February 2026, more than double the prior year’s R193-million, with 1.5 million active subscribers – one of the most successful product launches in South African banking history. It is also built on network identities the bank does not control and numbers it cannot move without the individual consent of every customer.
To switch hosts, Capitec would have to issue every subscriber a new Sim or reprovision an embedded Sim (eSim) on the new network, then coordinate a portability request for each one. Mobile number portability exists; bulk IMSI portability does not, and it is not on Icasa’s agenda. The move would happen subscriber by subscriber.
Read: Free calls, dead voice and Shameel Joosub’s Spanish ghost
Standard Bank’s switch from Cell C to MTN in June 2024, covering about 300 000 subscribers, showed that migration is possible. It also showed why no banking MVNO has attempted it at greater scale: at 1.5 million subscribers, the cost, time and passive churn from customers who never complete the port make it commercially prohibitive as a near-term answer to host repricing. A conservative 10-15% passive churn would mean 150 000 to 225 000 lost customers. Every host operator knows this arithmetic. So does every banking MVNO board.
In April 2026, Capitec made calls between Capitec Connect numbers free. The logic is obvious: bring your family onto the network and call them for nothing. The effect is a closed user group – a switching cost built into the relationship itself.
It also binds Capitec to Cell C more tightly than any contract clause. Every subscriber who joins for free family calls is one who will resist moving to a new Sim on a different host, because the move breaks the on-net benefit. The more successful the proposition, the larger the base locked in and the more prohibitive migration becomes. Capitec’s customers are doing the work Cell C’s lawyers would otherwise have to.
Capitec Connect carried 768 million voice minutes in the year to February 2026. Those minutes are now free to make – but not free to carry. At renewal, Cell C negotiates from structural strength. Capitec’s only real lever is future growth: slowing new activations on Cell C while the existing base stays put. The base cannot move. That is a far weaker hand than the language of mutual dependency suggests.
The dependency runs deeper
Cell C hosts 13 of South Africa’s 23 MVNOs, and its return to profitability after years of restructuring now leans heavily on that wholesale business. Yet Cell C owns no towers: its coverage comes from roaming agreements with MTN and Vodacom, with a multi-operator core network (MOCN) overlay rolling out. MTN has said it wants to be South Africa’s leading MVNO wholesaler; Vodacom has entered the hosting market. Both now compete with Cell C for MVNO clients – and both are the infrastructure on which Cell C itself depends. The dependency is a stack, and decisions at the top of it ripple down through every layer.
Nor is Capitec alone. FNB Connect’s numbers sit with Cell C, Standard Bank Connect’s with MTN, Nedbank Connect’s with MTN. No South African bank owns a single subscriber MSISDN. Each faces the same renewal dynamic: the host negotiates from strength, and the bank’s leverage reaches only its future growth, never its existing base.
Read: Capitec blows up MVNO pricing with free on-net calls
FNB signalled this route as far back as 2023, signing MTN as a second supplier for FNB Connect alongside Cell C – though there is little public evidence its base has yet been provisioned on MTN.
Dual-hosting caps future concentration risk without forcing a re-Sim of the existing base. But it does not dissolve the dependency – it manages it – and it fractures the very feature that builds the strongest switching costs. Capitec’s free on-net calls work precisely because every subscriber sits on one Cell C core. Split the base across two hosts and the free-calling proposition breaks where it was meant to deliver.
The longer-term fix is regulatory, but with sober expectations. Decoupling IMSI assignment from spectrum licensing would solve the problem at root. The International Telecommunication Union’s E.212 framework expressly allows mobile network codes to be assigned to full MVNOs without spectrum, as is standard in the US, UK, Germany and most of Europe. The case is technically and comparatively sound. The South African reality is not: IMSI portability is not on Icasa’s agenda, the Electronic Communications Amendment Bill does not address it and the realistic path from sustained lobbying to implementation runs 7-10 years.
Read: MVNOs take centre stage in legislative shake-up
The banking MVNO is a bank’s weapon in the convergence race against operators pushing into financial services from the other side. It yields behavioural data between transactions, multiplies switching costs and – as Capitec shows – earns real money. But the mitigations reduce the risk without removing it. Dual-hosting manages future exposure while the existing base stays dependent. Regulatory reform arrives too late to protect anyone now. Neither touches the dynamic at the next wholesale renewal – conducted in private, on undisclosed terms, between a host that controls the infrastructure and a bank whose most successful product has made migration less viable than the day it launched.
The connectivity product is the door into the most valuable relationship in South African retail finance. Most banks are busy building a better door. The terms belong to the host.
- The author, Pambos Soteriades, is a telecoms and strategy consultant with 28 years’ experience in African mobile markets, including executive roles at Vodacom Group and Telkom Kenya. He has worked in the South African MVNO market on both the operator and MVNO side. He is not affiliated with, employed by or invested in any institution, operator or MVNO mentioned in this article
- Subscribe to TechCentral’s daily newsletter
- Get breaking news alerts on WhatsApp
