The best infrastructure in Africa isn’t public anymore.

The obvious conclusion is that public infrastructure is failing.

A better reading is that it is being outrun.

Africa’s urban population has grown from 248 million in 2000 to 829 million today. Much of that growth sits around the edges of cities, where people build first, and services arrive later.

Sometimes much later.

And wherever the built-out version fell behind, a lighter one grew into the gap.

The grid did not reach new neighborhoods, so off-grid solar delivered 55% of new electricity connections in Sub-Saharan Africa between 2020 and 2022.

Bank branches did not arrive, so the phone became the branch. Mobile money moved $1.43 trillion across the region last year.

Public transport did not scale fast enough, so two-wheelers did. Uganda sold 30,000 electric bikes last year, while Rwanda stopped licensing new petrol taxis.

Fixed broadband never reached most homes, so it came from orbit instead. Starlink is now live in 27 African markets and faster than every terrestrial provider in 22 of the 23 measured. In Nigeria it became the second-largest ISP in two years.

Different sectors but same pattern.

These products did not win simply because they were better.

They won because they could be bought and used in the conditions that already existed. That is why pay-as-you-go solar grew while cash sales fell: households could buy power the same way they build homes. Gradually, when money comes in, without waiting for a formal address, a credit file, or a utility connection.

These products did not work around missing infrastructure. They removed the part that had to reach the customer.

A solar kit deletes the grid connection, not the grid. Mobile money deletes the branch, not the bank. Starlink deletes the last mile, not the network.

None of it is infrastructure-free. The winners stood on the one layer that did get built, and cut out the line to the door.

After working across Ghana, Malawi, and Nigeria, I’ve seen the same question decide these things, and it is never the one in the deck.

Not how large the unmet need is.

Whether your product needs someone to run a line to the door.