Insight · E-Mobility
A diesel generator charging an electric motorcycle is not a contradiction. It is a ledger entry. Where the grid is unreliable, adoption is decided by which party absorbs the volatility, and whether they can survive holding it.
In Lagos this year, some electric-vehicle dealerships and battery-swap stations kept their chargers running on petrol and diesel generators. Nigeria waived taxes on nearly four thousand electric vehicles in the first half of 2026 (as reported), added VAT exemptions and zero import duties, and watched the vehicles arrive into an electricity system of roughly four thousand megawatts serving more than two hundred million people.
The easy response is irony. Electric vehicles charged by diesel. A transition that is not a transition.
The easy response is wrong, and expensively so. A generator behind a swap station is not a failure of the technology. It is a ledger entry. It tells you exactly who, in that particular business model, has been assigned the job of absorbing the gap between the electricity a fleet needs and the electricity a grid delivers. Read that ledger correctly and you can predict which electric-mobility models will hold in a market and which will quietly stall. Read it wrongly and you will forecast adoption off price parity, which is the single most reliable way to be surprised.
There is a useful idea in the sociology of infrastructure, from Star and Ruhleder in the mid-1990s: infrastructure is not a thing, it is a relation. Nothing is infrastructure in itself. It becomes infrastructure for a particular set of people, doing a particular kind of work, under particular conditions. And it is invisible until it breaks.
Applied to electricity, this dissolves a question that market-entry decks ask badly. "Is the grid good enough?" has no answer, because good enough is not a property of the grid. It is a property of the relationship between the grid and a specific user with a specific tolerance for interruption.
A household charging a car overnight can tolerate a four-hour outage without noticing. A commercial motorcycle rider whose entire daily income is a function of hours on the road cannot tolerate forty minutes. Same grid. Two completely different infrastructures.
Which means the real question is not whether the grid is reliable. It is: when the grid fails, who pays for the failure?
Somebody always does. The volatility does not disappear because a policy waived import duty on it.
There are only four parties available to absorb the gap, and every electric-mobility model is, whether or not its founders describe it this way, a decision about how to distribute the burden between them.
Absorbs it as downtime, lost fares, unpredictable range, and the cognitive load of planning a day around an unreliable input. The cheapest option for everyone else and the most destructive to adoption, because it lands on the party with the least capital and the shortest time horizon.
Absorbs it as capital expenditure: generators, on-site solar, stationary storage, oversized battery inventory. Converts an unpredictable operational risk into a predictable, financeable cost. Expensive, but the kind of expensive a balance sheet can hold.
Absorbs it as asset risk, priced into interest rates, tenor, and collateral requirements. Often absorbs it invisibly, by declining to lend at all, which shows up in the market as low adoption rather than as a financing failure.
Absorbs it through generation, transmission, and interconnection investment, which is slow and lumpy and rarely arrives on the timeline a fleet operator needs.
Nigeria's electric-mobility market is now visibly sorting itself along this logic. Motorcycles and three-wheelers are emerging as the viable near-term segment, and operators including MAX and Spiro are running battery swapping rather than home charging. That is not a preference for a technology. It is a structural decision to take the gap off the rider, who cannot hold it, and put it onto an operator balance sheet and a financier's risk model, which sometimes can.
The generator behind the swap station, in other words, is the price of moving the burden. It is a bad outcome measured in emissions and a rational one measured in whether the fleet exists at all next year. Both things are true, and an analysis that only sees the first will misjudge the market.
The International Energy Agency's 2026 Southeast Asia outlook is the same problem written with more zeros. Regional electricity demand is already growing at twice the rate of overall energy use. Electric-vehicle sales more than doubled in 2025 to roughly half a million, and the electric share of two- and three-wheeler sales could approach sixty per cent by 2035 under current policy.
The constraint is no longer generation. Transmission and distribution networks would need to more than double in length by 2050. Annual grid and storage investment needs to rise from around thirteen billion dollars to fifty billion, including an estimated twenty-seven billion dollars of ASEAN Power Grid interconnection by 2040. Regional clean-energy investment has climbed sixty per cent since 2015 to more than one hundred billion dollars in 2025, and the region still attracts around three per cent of global energy investment while holding about nine per cent of the world's population.
Read through the ledger, that is a decision to place the burden on the state and the multilateral financier rather than on riders and operators. It is the more efficient allocation, because interconnection is a shared asset and generators are not. It is also the slower one, and the gap between the vehicles arriving and the network arriving is exactly where individual operators will be forced to hold volatility they did not budget for.
Two markets, two answers, one question. Southeast Asia is attempting to socialise the gap. Nigeria, for now, is privatising it. Both are legitimate paths. They produce very different businesses, very different unit economics, and very different people bearing the cost of the transition.
Aggregate infrastructure statistics are not how riders experience any of this. In the Lab's fieldwork on electric two-wheelers in Nairobi, the calculation riders make is daily and concrete: what does this machine earn me today, net of what it costs me today, and how many hours of that day am I certain to have it.
That last clause is the one financial models tend to drop. An hour of unplanned charging is not an inconvenience. It is a fare, and then a second fare, and then a repayment instalment that has to come from somewhere else. Uncertainty about when the hour will disappear is worse than the hour itself, because it cannot be planned around and so it is priced as a permanent discount on the asset.
This is why swap networks change adoption curves more than battery chemistry does. They convert an uncertain, rider-held risk into a fixed, visible fee. The rider is paying more per kilometre and buying something they value more than the margin: a day whose shape they can predict.
Any market-entry study that measures total cost of ownership without measuring variance in productive hours has measured the wrong thing. This is the sort of question field research exists to settle, and it cannot be settled from a spreadsheet in another country. It is also the shape of question our Market & Expansion Research service is built for.
Three questions, in order, before the vehicle specification is even discussed.
Which party in your model is holding the electricity volatility? Name them. If the answer is not immediately obvious, it is the customer, and it is not on your balance sheet only because you have not looked.
Can that party actually hold it? A commercial rider on daily margins cannot absorb a variance they cannot forecast. A depot operator with access to working capital can. A financier can, if someone has quantified the risk well enough to price it, and will refuse if nobody has.
What would have to change for the burden to move? Interconnection timelines, tariff reform, swap density, or a local storage market. This is the question that decides whether your model is a permanent structure or a bridge, and the answer usually sits with institutions rather than with your engineering team.
Markets where the grid cannot be assumed are not simply harder versions of markets where it can. They are structurally different, and the models that win in them are the ones that redistribute a burden rather than the ones that ignore it. Our reading of how state capacity and niche success interact points the same way: the technology rarely decides the outcome on its own.
The Lab works on this question in e-mobility and transport and at the boundary with energy access, in the places we work in where the answer is not obvious from the outside.
If you are weighing an entry decision and the electricity question keeps getting deferred to a later slide, tell us what you need to know. We will design the study around the decision, not around the template.
This is an independent insight piece by Transitions Lab. For the underlying method, see Field Research and the BRW framework. For the field case behind the argument, see the mobility case. See also Own the Battery, Rent the Shopfront on network architecture in the same market, The Customers Who Can Leave on households as the shock absorber of last resort when the larger customers step out of the loop, Stacking, Not Switching on the reliability calculation that decides which fuel a firm still burns after the panels are on the roof, and A Thousand Cars, One Risk on the same technology-and-context question when a driverless car meets a city that is institutionally, infrastructurally and economically unready.