Build the Fleet First: Why Africa's EV Infrastructure Should Follow Real Demand

The infrastructure question is not "where are the gaps?"

It is where is the repeatable demand?

The strongest recent research is moving in that direction. A 2026 World Bank analysis across 40 developing countries describes a "segment-led transition pathway" in which the economic case for electric mobility is strongest in high-utilisation segments such as two- and three-wheelers, urban buses and selected commercial fleets.¹ In a separate June 2026 Africa-focused note, the Bank identifies buses, taxis, ride-hailing fleets and two- and three-wheelers as market segments where electrification is already economically viable in several African contexts.²

That makes intuitive engineering sense. A private car may move for an hour or two and spend most of its life parked. A commercial bus, delivery van, taxi or motorcycle earns money by moving. High daily kilometres turn fuel savings into a bigger operating advantage, and predictable routes make energy demand easier to model.

The strongest objection is the chicken-and-egg problem. The World Bank's June 2026 Africa note explicitly argues that charging infrastructure should also be treated as a public good, including in secondary towns and along freight corridors, because operators will not switch at scale without reliable access.² I do not read that as an argument for indiscriminate charger construction. It is a reminder that some network coverage has strategic value before utilisation is proven. Fleet-first is therefore a capital-allocation rule for the first wave of sites, not an argument against public charging, corridor coverage or government support where the wider system would otherwise fail to emerge.

Why commercial fleets are a better first anchor

This does not mean public charging is unnecessary. The IEA reports that the global stock of public charging points exceeded seven million at the end of 2025 after growing more than 33% in one year, but China alone held more than 65% of that stock.³ Africa is starting from a very different base. The IEA estimates electric-car sales on the continent rose from roughly 4,000 in 2023 to about 25,000 in 2025, with Egypt, Morocco and South Africa accounting for nearly 70% of 2025 sales.⁴ The IEA charging page does not give one continent-wide African public-charger stock figure, so I would not invent one. The sequencing question is therefore practical: in an early market with limited vehicle stock, should the first wave of capital optimise for geographic coverage or dependable utilisation?

The economics improve when vehicles work harder

The World Bank's 2026 cross-country study says the strengthening case for electric mobility is driven in part by high vehicle utilisation, lower energy and maintenance costs, and the particular economics of two- and three-wheelers and buses.¹ That is a useful way to think about African deployment: electrify the kilometres that are already being driven intensively.

Global bus economics are also moving quickly. The IEA's 2026 outlook says electric-bus sales reached almost 70,000 in 2025, up 12% year on year, with China still accounting for around 60% of sales.⁵

There is an infrastructure lesson hidden in that growth. Vehicles and charging cannot be planned separately. In June 2026, Ireland's National Transport Authority told the Oireachtas Public Accounts Committee that 98 battery-electric buses it had purchased had not yet entered operational use. It said 56 Dublin buses were expected to enter service as additional grid capacity and charging infrastructure came online, while 42 Galway buses depended on new depot charging infrastructure scheduled across late 2026 and early 2027. The vehicles existed; the enabling infrastructure was still catching up.⁶

For Africa, the same rule applies in both directions. Buying vehicles without energy infrastructure fails. Building charging infrastructure without vehicles can also strand capital. The productive unit is the fleet plus energy plus site plus operating schedule.

Africa already has examples of demand-led deployment

Commercial motorcycles: energy infrastructure follows daily work

Ampersand reported in May 2025 that it was powering more than 5,700 electric motorcycles travelling an average of about 950,000 kilometres per day.⁷ Those figures are company-reported and are now more than a year old, so I would not treat them as a current market baseline. They still illustrate the operating pattern: battery-swapping infrastructure attached to commercial riders with intense, repeatable daily demand.

Rwanda: buses and power infrastructure planned together

Rwanda has targeted electrification of 20% of its bus fleet by 2030. The World Bank notes that private operators in Kigali have already deployed electric buses and charging stations, while its energy-sector work models the power infrastructure required as the fleet grows.⁸ That is the right planning unit: route demand, buses, depot charging and power capacity considered together.

Dakar: measured utilisation matters more than design capacity

Dakar's electric Bus Rapid Transit system runs on a defined 18-kilometre corridor, but design capacity is not the evidence I would use for this argument. A 2026 performance report says the system was averaging about 80,000 weekday passengers and had recorded more than 34 million passenger trips over its first 20 months of operation.⁹ That is the useful metric: actual repeated passenger demand on an operating service, not a theoretical maximum.

Nigeria: the better pilots are already asking operational questions

An earlier Lagos electric-bus proof of concept between Oando Clean Energy and LAMATA was explicitly structured to gather operating data before larger deployment. By day 50 of the three-month 2023 proof of concept, the two buses had carried 41,678 passengers and travelled 22,129 kilometres, according to the UN Global Compact's Africa regional hub.¹⁰ Those are the kinds of measured operating numbers that should inform the next stage.

In Abuja, a 2026 partnership reported by Punch paired plans for 400 to 500 electric ride-hailing vehicles with ten solar-powered charging stations.¹¹ That headline ratio is roughly 40 to 50 vehicles per site, but it is impossible to judge adequacy from site count alone. The relevant variables are charger count per site, charger power, vehicle efficiency, daily kilometres, dwell time, simultaneous arrivals, solar and storage capacity, grid backup and target state of charge. The project should be judged on those operating numbers, not on the announcement.

The difficult part is not only the charger

The vehicles themselves still have to be financed

A fleet-first model does not magically solve the higher upfront cost of electric vehicles. The World Bank argues that financing has to be adapted to operator realities, especially for motorcycle taxi riders, minibus operators and smaller fleets that rely on informal credit. It specifically points to structures built around operating savings, blended finance and fleet aggregation that creates bankable transaction sizes.² In July 2026, the African Development Bank also announced new funding for its Green Mobility Financing Facility for Africa, using concessional debt, senior debt and technical assistance to crowd in private capital.¹²

Grid reliability, tariffs and foreign exchange can reverse the economics

The fleet may be predictable while the electricity is not. A serious site model needs the local tariff, demand charges where applicable, transformer and connection cost, outage profile, cost of backup energy, solar and storage economics, and the foreign-exchange exposure on imported vehicles, batteries and power electronics. The World Bank's 2026 Africa analysis also makes the macro point that petroleum imports and currency depreciation expose transport systems to foreign-exchange shocks.² EV economics therefore have to be modelled in local currency under stress, not only from a dollar-denominated vehicle brochure.

In Nigeria, EVs are competing with CNG for transport capital

That competition should be acknowledged rather than waved away. In March 2026, the Presidency expanded the Presidential CNG initiative to include electric vehicles, explicitly positioning both CNG and EVs inside the national clean-mobility strategy.¹³ For a fleet operator, the relevant question is not which technology is ideologically cleaner. It is which powertrain, energy network and financing structure produce the best route-level economics, reliability and foreign-exchange exposure for that duty cycle.

Charging and swapping should not be collapsed into one standard

Fixed-route buses may favour depot charging. High-throughput motorcycle fleets can make battery swapping attractive because downtime is expensive. Delivery fleets may combine overnight AC charging with a smaller number of fast chargers. The infrastructure standard should follow the vehicle class and operating pattern. Interoperable payments, connector standards and open data matter, but forcing one energy-delivery model across every fleet can create a different kind of stranded asset.

Anchor demand creates its own concentration risk

A site that depends on one fleet can be highly utilised and still be fragile. If the anchor customer fails, relocates or changes vehicle technology, revenue can collapse. The financing model should therefore test contract length, minimum-volume commitments, customer credit quality, secondary users and whether the site can serve other compatible fleets if the anchor disappears.

How I would deploy the next charging capital

I would start with anchor demand and force every proposed site to answer six questions before money goes into hardware. Treat this as a decision gate: a site should not pass investment approval until each answer is supported by measured or contract-backed evidence, and the utilisation case clears the site-specific break-even model rather than an arbitrary continent-wide percentage.

Corridor charging can follow the same rule. A logistics corridor should begin with actual freight operators, scheduled vehicle movement and known energy demand, not only the observation that two cities are far apart. Once anchor fleets create base utilisation, the same infrastructure can support third-party vehicles and private users.