Loop is trying to do something most payment startups avoid: it is walking straight into the taxi rank, where cash, trust, and speed still run the business. In a minibus-taxi sector that carries an estimated 67% of daily commuters, even a small change in how money moves can change how vehicles are managed, how drivers are paid, and how operators prove they are making money.
The hard part is that fare collection is the easy bit. The real prize is a usable record of what happened on the road, because once a trip is logged properly it can support reconciliation, fleet oversight, lending, insurance, and route decisions. That is the line Loop is testing, and it is a much tougher proposition than dropping a card reader into a taxi and calling it innovation.
What Loop is building
Loop, a South African company in Google’s 2026 accelerator cohort, is building a payment system for the minibus-taxi economy. The company is focused on commuter payments, but the longer-term value sits around the payment rail rather than inside it.
A taxi that collects fares digitally creates a trail. That trail can show which routes were busy, when vehicles ran, how much was taken in, and how that money should be split between driver, owner, association, and any other party with a claim on the day’s takings. For an industry that still runs largely on paper memory and cash-up arguments, that is a serious change.
The use case is not a neat retail checkout. Taxi operations are messy by design. Passengers board quickly. Fares are low. Connectivity is uneven. Routes shift. A vehicle can be full, half-full, then gone in seconds. Any system that cannot survive that environment is decorative.
Loop’s bet is that the payment layer can be made useful enough to sit inside the operating model of the business, not just on top of it.
Fact block
- Company: Loop
- Market: Minibus-taxi payments and mobility operations
- Status: Part of Google’s 2026 accelerator cohort
- Core aim: Digitise commuter payments and build transaction records that support operations and finance
- Main users: Taxi owners, drivers, associations, commuters
- Operating logic: B2B2C, with fleet and association-level adoption feeding commuter usage
Why cash still rules the rank
Cash is not popular in the taxi economy by accident. It is fast, familiar, and works when the signal is bad. It also creates the exact problems that formal businesses spend years trying to remove.
Owners struggle to see whether a vehicle is performing well or simply moving money around. Drivers and owners argue about takings. Association fees and rank charges are easier to collect when everyone is already dealing in cash, but they are harder to reconcile later. The result is a system with weak records, weak proof of income, and weak visibility into where the money went.
That lack of proof has a knock-on effect. A taxi business without reliable records is a poor candidate for bank finance. It is also harder to insure on sensible terms, because an insurer can only price risk properly when it has data to work with. Maintenance decisions become guesswork. Route planning becomes habit. Performance tracking becomes a monthly fight.
The security issue is obvious. Drivers carrying cash are exposed to theft and robbery. So are owners who collect takings in bulk and then still have to count, bank, and protect it. The industry pays for cash twice, once in risk and again in handling cost.
How Loop has to work in practice
A payment tool for taxis cannot behave like a restaurant card terminal. It has to keep working when the network drops, when boarding is rushed, and when the fare is too small to tolerate heavy fees.
Loop’s model is built around that constraint. The system needs offline capability so a trip can be logged even without live connectivity, then synchronised later when a signal is available. It also has to support quick tap-and-go use, whether through NFC-enabled cards or QR code scans, because no driver wants to hold up a queue while passengers fumble through a slow payment flow.
The hardware matters too. Taxi terminals have to be rugged, battery-powered, and quick enough to match the pace of the rank. If the device slows boarding, drivers will stop trusting it. If it breaks easily, operators will treat it like a toy.
Low-value fares are another trap. In normal retail, a transaction fee can look small. In taxis, the margins are thin and the volume is high. A tiny fee multiplied across thousands of trips becomes material. That means the pricing model matters as much as the software. If Loop’s take per ride is too high, adoption will stall. If the fees are too opaque, the system will be blamed for eating into daily earnings.
The simplest version of the product is a commuter taps, money moves, record is stored. The real test is whether that can happen without slowing the taxi down.
Who pays and who gets paid
Loop’s customer model is not just commuter-facing. It is built more like a business service wrapped around a consumer payment layer.
Taxi owners and associations are the direct customers. Drivers use the terminal. Commuters are the people paying the fare. That matters because taxi adoption usually happens through authority, not through individual riders. If the owner or association does not trust the system, it will not spread. A single commuter app does not solve a fleet problem.
The revenue structure is likely to be layered. A small transaction fee is the starting point, but the more interesting income comes from services built on the data, such as fleet management dashboards, operational reporting, and eventually lending or insurance partnerships. That is the point where the company stops being a payment processor and starts acting like infrastructure.
The transaction cost question is the one that will decide whether this gets real traction. In practice, the fee can sit with the operator, be passed through to the commuter, or be partly subsidised by Loop to accelerate rollout. Any of those can work in a pilot. None of them can hide forever at scale.
Loop’s inclusion in Google’s 2026 accelerator cohort gives it visibility and credibility, but taxi operators will care about the economics, not the logo on the slide deck.
What the data changes
A digital fare is useful. The record behind the fare is more useful.
Loop can capture fare values, timestamps, routes, vehicle IDs, driver IDs, origin and destination points, and aggregate passenger flow. That creates a working picture of how a vehicle or route actually performs over time. For an owner, that is not abstract analytics. It is the difference between guessing and knowing.
The data can support:
- route planning based on actual demand
- driver reconciliation using traceable daily records
- maintenance planning from usage patterns
- lending applications backed by income history
- insurance pricing based on real operating behaviour
- visibility into association deductions and rank levies
There is also a governance question. If the platform owns the raw aggregated dataset, but operators own the business-level information tied to their fleet, the terms have to be clear from the start. Commuter privacy has to be protected, and data handling has to sit inside POPIA expectations. If operators think the platform is quietly extracting value from their routes while giving little back, trust will collapse fast.
The most useful version of Loop is not a payment app with some charts bolted on. It is a transaction system that becomes the source of truth for a business that has lived for decades without one.
The real test for Loop
The taxi industry does not need another shiny layer that fails when the rank gets busy. It needs tools that fit how the business already works and then make that business easier to run.
Loop’s challenge is to prove that digital fare collection can do more than move money. It has to reduce argument, reduce cash exposure, and produce records that an owner can use to run a fleet like a business rather than a daily gamble. If that happens, the company will have done something larger than payments.
If it does not, it will be remembered as another app that misunderstood how informal transport actually moves.
The difference between those two outcomes sits in the details, who pays the fee, how offline mode behaves, who controls the data, and whether a driver can use it without slowing the vehicle down. In this market, the smallest failure becomes the loudest objection.
