What Revolut can and cannot take from South Africa’s banks

Revolut’s application for a South African banking licence is progressing, and the company is still targeting a 2028 launch. That is a long wait in a market that keeps moving. By 2028 the app-quality gap Revolut exploits elsewhere will have narrowed, the newer entrants will have dug in and reform of the National Payment System will have changed what a licence is worth.

The European model works like this: free accounts bring customers in, a share of them convert to paid subscriptions, and that revenue funds credit, savings and trading. It works where banking is fragmented, where foreign exchange fees are extractive and where money crosses borders with little friction. South Africa has the extractive fees. Its banking market is concentrated and its cross-border flows are controlled.

Read: Revolut gears up to disrupt South African banking

Exchange controls apply. An individual may move up to R2-million/year under the single discretionary allowance without tax clearance, and a further R10-million for foreign investment with it. Better design does not remove those ceilings. They cap the volume and the velocity that make the European model pay.

Revolut can still beat the banks on the experience of moving money, inside the same regulatory envelope, by partnering with an authorised dealer. The authorised dealer network is dominated by the large banks. How closely Revolut can reproduce what it does in Europe – live rates, transparent pricing, fast transfers – depends on what a partner will enable and at what price. Its biggest advantage would rest on commercial decisions taken by the institutions it is trying to take customers from.

Revolut’s target customer already banks with FNB, Standard Bank, Absa, Nedbank or Investec, and usually holds credit there, too – a home loan, vehicle finance, a business account – along with a rewards programme. A currency wallet is easy to add and easy to drop. A 20-year home loan is neither.

Discovery Bank is the closest competitor. Its paid subscription model, Vitality integration and actuarial hooks aim at the same high-income, internationally mobile customer, and it passed a million clients in August 2024, two years ahead of its own target. Price is not what holds those customers in place.

Concentration is high by any measure. World Bank figures put the three largest banks at about 79% of banking assets in 2021, and the Reserve Bank’s own risk assessment puts six large banks at roughly 93% of sector assets. The positions Revolut broke into in Europe looked nothing like this.

The one route to something more durable

There is a version of this in which Revolut builds more than a second wallet. It turns on a decision the company has not made publicly.

In Europe, Revolut bundles a global eSim into its top tier as a retention perk for customers it already has. South Africa would need more: a local mobile proposition, through an MVNO or a wholesale deal, giving the customer a South African number they use every day. A traveller will carry a second Sim. Making it the everyday number means beating the inertia of an existing number, a configured handset and an existing mobile relationship.

The reasoning is that connectivity becomes an acquisition and engagement tool among customers banked elsewhere. A travel eSim produces a thin dataset clustered around trips abroad. A local number produces recurring signals: when someone recharges, how heavily they use data, when they change handsets, when they roam. Combined with Revolut’s own payments data, and if those signals prove predictive, they could support a credit or savings offer timed to the moment a customer’s main bank is failing them. I have written before about what banks give up when they build on someone else’s network, and the same caution applies here.

Capitec Connect shows the logic holds. Its MVNO produced R442-million in net income in the year to February 2026, up 129%, with 1.5 million subscribers active in the final three months of the period. Capitec runs that on top of roughly 25 million banking clients and their deposits. Revolut would be attempting it from outside, with no deposit base, funding any credit book from capital markets or a partner at higher cost and thinner margin.

Regulation bites here, too. Combining connectivity data with transaction data for credit scoring needs a lawful basis under Popia, a compatible purpose and, where the decision is automated, compliance with the act’s provisions on human review. A banking licence applicant must hold at least R250-million in capital. And the activity-based authorisation framework under National Payment System reform, still being built, will decide whether a lighter way in exists for a foreign applicant at all.

A staged entry – payments first, a full licence later – would change the arithmetic. Arriving earlier with a thinner product builds transaction history, brand familiarity and habit before the bank opens. That is the version that gives the incumbents most to think about, and it depends on decisions nobody has taken yet.

What comes after the wallet

Revolut will find customers here. The travel and foreign exchange product is good, and the banks have been complacent about international money for years.

The more difficult part is what follows. Selling credit, savings and subscriptions requires being the customer’s main bank, and a currency wallet does not produce that on its own. The moments when people actually move their banking – a new job, a house purchase, a service failure – are invisible to a travel wallet. Revolut can respond when a customer comes to it. It cannot yet see the moment coming.

Revolut should take the FX wallet. The home loan stays where it is, unless it builds the local data and distribution layer it has not committed to building.

• The author, Pambos Soteriades, has spent 28 years in mobile telecommunications, including executive roles at Vodacom Group and Telkom Kenya. He writes on strategy and technology in African mobile markets at African Comms Intelligence. He has declared no shareholding in, and no client, advisory or prospective business relationship with, any company named in this article