Africa's electric-vehicle transition will not be constrained by the number of chargers that can be purchased. It will be constrained by whether charging sites can receive dependable power, operate at useful utilisation, survive failures and be expanded without rebuilding the system from scratch.
That distinction matters because a charger is easy to photograph. A transformer upgrade is not. A software roaming agreement is not. A depot cable route, spare-parts inventory, utility interconnection study or maintenance response time rarely appears in an announcement. Yet those less visible pieces decide whether vehicles actually leave the depot charged the next morning.
This article is an infrastructure and systems analysis, not a forecast that every African market will follow the same path. Sources were reviewed through 21 September 2026. Where I describe a design rule rather than a published statistic, I state it as an engineering recommendation rather than as established market fact.
The charger is the visible layer of a much larger system
Electric-mobility infrastructure is usually discussed in charger counts: how many units, how many kilowatts, how many locations. Those metrics matter, but they are outputs of a deeper system. A charging site has to combine utility capacity, civil works, switchgear, transformers, protection, metering, networking, charger hardware, backend software, payment or fleet authorization, monitoring, field maintenance and an operating model.
The International Energy Agency reported more than 7 million public charging points globally at the end of 2025, after more than 33% growth during that year.3 That global scale can make charging look like a hardware rollout problem. It is not. The IEA also warns that higher charging speeds and concentrated charging demand can make grid-capacity constraints more pronounced, while smart charging can shift demand away from peaks.4
If one of these layers is weak, the site can be technically installed and still operationally useless.
Grid connection is part of the mobility product
For a fleet, electricity is fuel. That makes the distribution network part of transport operations whether the transport operator wants it to be or not.
Rwanda provides a useful African example because the power-system implications have been modelled publicly. A World Bank study on Rwanda's electric-mobility transition says Kigali's peak power demand is projected to rise substantially even before EV load is added, and warns that without upgrades the number of overloaded lines could be four times higher by 2030. The same work recommends coordinated grid upgrades, smart charging and planning around transit hubs.1 A companion World Bank analysis describes line loading above safe levels and explicitly says power lines, transformers and substations must be upgraded as battery-electric buses scale.2
The lesson is broader than Rwanda. A transport operator should not select a depot, order chargers and then ask the utility what power is available. The power study belongs near the beginning of site selection.
Start with the load profile
Model vehicle energy demand by hour, not only daily kWh. Ten buses arriving at once create a different infrastructure requirement from the same ten buses arriving across six hours.
Get utility constraints in writing
Available capacity, connection voltage, transformer requirements, protection standards, metering rules and upgrade lead time should be treated as project inputs.
Design controlled charging
Not every vehicle needs maximum charger power at the same moment. Scheduling and dynamic load allocation can reduce the required peak connection.
Plan expansion before concrete is poured
Ducting, switchboard capacity, transformer sizing and physical space should anticipate the next fleet tranche even if chargers are installed gradually.
A 120 kW charger is not a 120 kW project
Nameplate charger power is only one line in the electrical design. A site with four 120 kW chargers can present a theoretical 480 kW load before auxiliary loads, losses or expansion headroom are considered. Whether the utility connection must support that entire peak depends on load management and the operating schedule, but the electrical system still has to be engineered around credible maximum demand and protection requirements.
That can mean a dedicated transformer, new medium-voltage equipment, cable runs, protection coordination, earthing, civil works, utility approvals and sometimes upstream reinforcement. These items are not accessories to the charger. They are what make the charger possible.
This is also why charger procurement should follow an electrical single-line design rather than lead it. If hardware is bought first, the rest of the project gets forced around decisions that may have been made without knowing the grid, site or fleet constraints.
Depot design is operations design
A bus depot or logistics yard is not a public charging forecourt. Vehicles have schedules, cleaning windows, inspections, driver handovers, parking patterns and departure deadlines. Charging has to fit inside those operations.
The site should answer practical questions before charger count is decided: Which vehicles arrive first? Which must leave first? Can a failed charger strand a route? Can a vehicle reach another connector without blocking the yard? Can maintenance staff isolate equipment safely? Does one damaged cable take an entire charging island offline? Is there enough turning radius for the vehicle class being served?
For fleet infrastructure, I prefer designing around energy required by departure time rather than maximum charging power. A vehicle that parks for eight hours may not need to occupy a high-power charger for eight hours. That opens room for lower-cost chargers, sequential charging, dynamic power sharing or a mixed fleet of charger sizes.
Do not ask only, "How fast can this charger charge?" Ask, "Can every required vehicle leave with the required state of charge even when one charger, one network link or one maintenance action is unavailable?"
Tariffs can make a technically good site economically bad
Electricity price is not one number. A project can face energy charges, demand charges, time-of-use rates, fixed charges, taxes and connection fees. The tariff structure determines whether smart charging is merely technically interesting or financially necessary.
The World Bank's Rwanda work recommends cost-reflective, time-of-use tariffs so operators are encouraged to charge when the electricity system can serve them more efficiently.1 That is important because the same fleet can have materially different charging cost depending on when and how its load is scheduled.
For a depot model, I would stress-test at least three electricity cases: expected tariff, adverse tariff and outage-plus-backup case. If diesel generation or battery storage is required during outages, that cost belongs in the energy model. If a tariff penalises peak demand, uncontrolled simultaneous charging belongs in the downside model too.
Uptime is an end-to-end property
A charger can report 99% device uptime while the service delivered to vehicles is much worse. A site is unavailable if the charger is healthy but the grid is down, the backend cannot authorize a session, the payment path fails, the cellular connection is dead, the connector is damaged or the site is inaccessible.
That means operators need an operational definition of uptime. For a commercial fleet, the metric that matters may be closer to: percentage of scheduled charging windows in which the required energy was successfully delivered before departure.
That metric forces the organisation to monitor the full chain. It also changes incident priority. A failed decorative screen and a failed contactor should not create the same severity. A software alert that says "charger offline" is useful only if someone owns the path from detection to restoration.
Interoperability should be decided before the network is large
Open protocols solve different boundaries. OCPP is designed for communication between charging stations and charging-management systems.7 OCPI is designed for exchange between market platforms such as Charge Point Operators and eMobility Service Providers, including locations, authorization, tariffs, sessions, charge detail records and remote commands.8
Those boundaries matter because an operator should be able to change hardware, backend software or customer-facing mobility services without replacing the entire network. Proprietary integrations can be reasonable during a pilot, but they become expensive once dozens of sites, multiple hardware vendors and external mobility partners depend on them.
For African markets, interoperability is especially valuable because networks are likely to grow unevenly across operators and countries. A driver should not need a new proprietary integration for every charging network, and a charging operator should not have to rebuild its backend each time an aggregator wants access.
Maintenance capacity is infrastructure too
Hardware deployed across multiple cities needs people, spare parts, diagnostic access and escalation procedures. Mean time to repair can matter more than theoretical charger efficiency.
A practical maintenance design includes remote diagnostics, a fault taxonomy, spare high-failure components, vendor escalation, trained local technicians, safe isolation procedures and a clear service-level target. If a replacement component has to be flown in after every failure, the network does not have a maintenance model. It has a shipping dependency.
Software should support this by preserving fault history, firmware version, charger identity, connector state, intervention notes and repeat-failure patterns. Maintenance data eventually becomes procurement data: it tells the operator which hardware actually survives the local environment.
Finance the whole system, not just the charger invoice
The African Development Bank's Green Mobility Financing Facility for Africa was designed around the reality that green mobility faces high upfront costs, technology risk, limited long-term financing and fragmented market readiness. Its 2026 funding package combines concessional debt, senior debt and technical assistance, and explicitly includes charging infrastructure and battery swapping among supported technologies.5
That financing logic matches the engineering reality. A charging project can have separate capital needs for land, utility connection, transformer and switchgear, chargers, solar or storage, civil works, software, vehicles and working capital. Bundling all of that into "charger capex" makes the project harder to reason about.
I would separate the model into at least four layers: site and grid connection, charging equipment, digital operations, and vehicle or fleet demand. That makes it easier to see which risks can be financed independently and which require long-term utilization commitments.
A practical deployment framework
For a serious fleet or public charging project, I would not approve procurement until the following questions have written answers.
What exact vehicles and duty cycles are being served?
Vehicle count, battery size, efficiency, route kilometres, dwell time and departure state of charge create the demand model.
What electrical capacity exists today?
Connection voltage, usable spare capacity, transformer condition, short-circuit level, protection and upgrade lead time should be verified.
What happens under a single failure?
Model charger failure, network failure, utility outage and maintenance isolation. Decide which failures are tolerable and which stop operations.
What does one delivered kWh really cost?
Include tariff structure, demand charges, losses, backup energy, maintenance, software, finance and utilization.
Can the site interoperate?
Document charger protocol, backend ownership, APIs, OCPI readiness, data portability and exit requirements before vendor lock-in becomes expensive.
Who restores service at 2 a.m.?
Monitoring, escalation, technicians, spares and vendor support need named owners and response targets.
Africa needs more EV infrastructure. But the useful unit of infrastructure is not the charger. It is the reliable charging service that results when energy, hardware, software, operations and finance are engineered together.
Williams Ashibuogwu
Williams is the founder of Zivora Labs and a backend and platform engineer working across production software, financial infrastructure and electric-mobility systems, including charging-network interoperability and platform architecture.
Reference basis
- World Bank, September 2025. Rwanda power-system readiness for electric mobility, smart charging, tariffs and grid upgrades.
- World Bank, October 2025. Grid-to-road analysis covering line loading, transformers, substations and bus electrification.
- IEA, Global EV Outlook 2026. Charging deployment, charger power, utilisation and network outlook.
- IEA, May 2026. Grid impacts of concentrated charging demand, load shifting and bidirectional charging.
- African Development Bank, July 2026. Green Mobility Financing Facility for Africa and blended-finance structure.
- World Bank, June 2026. African transport electrification, financing constraints and infrastructure needs.
- Open Charge Alliance. OCPP as the open protocol between charging stations and charging-management systems.
- EVRoaming Foundation. OCPI as the open interface between CPO and eMSP platforms.
