Africa's Next Infrastructure Won't Look Like Roads

Every infrastructure conversation in this country starts with roads. Then power. Then, if there is money left over, connectivity.

I understand why. A road is visible. You can cut a ribbon on a road. A minister can stand on it and be photographed. Nobody has ever been photographed standing on a data layer.

But the sequencing is inherited, and it is wrong for where we actually are.

A road moves maize. It does not tell you whether the maize was worth planting.

Consider a cotton farmer in Eastern Province. She has perhaps two hectares. The nearest tarred road is real, and it matters. But her margin is not lost in transport. It is lost months earlier, in decisions made without information: when to plant against a rainfall pattern that has shifted, whether the stand is stressed while there is still time to act, what the buyer will actually pay at gate. She is not short of movement. She is short of sight.

Now consider the mini-grid at Ikelenge. It generates more power than the community around it consumes. The generation exists. The poles exist. What is missing is a customer large enough to make the economics work, so the tariff stays high, connections stay slow, and the asset sits underused. Again, not a movement problem.

This is the pattern almost everywhere I look. The physical asset is either present or fundable. What is absent is the layer that makes it productive, and that layer is made of data, power and payment rails rather than concrete.

There is a second reason the road-first instinct fails us, and it is about capital. A national road programme is a sovereign undertaking. It requires a loan, a decade and a political consensus. A modular data centre next to an existing hydro plant is a container, a transformer and a contract. One of those is available to a Zambian company this year. The other is not.

So the infrastructure that gets built here over the next twenty years will not look like the infrastructure that got built in the twentieth century. It will be smaller, privately financed, placed wherever the underlying resource already sits, and duplicated rather than scaled. Containers, not corridors. Constellations, not cables. Fleets of small things that each pay for themselves, rather than one enormous thing that pays for itself in 2050.

I want to be careful here, because this argument is easy to overstate. Roads matter. A satellite will not get a truck to Mongu. Ports matter. Transmission matters, and Zambia's transmission constraints are real and expensive. I am not arguing that the physical layer is obsolete.

I am arguing against the assumption that development has to arrive in the same order it arrived elsewhere, and that we must therefore wait our turn. We did not wait for landlines. We should not wait for the rest of it either.

Build the layer you can finance this year, on top of the asset that is already standing.

That is the whole strategy. It is not romantic. It happens to be available.