In August 2026, Congo's Council of Ministers approved a collaboration convention worth about $1.26 billion to rehabilitate the Dilolo–Sakania line: roughly a thousand kilometres of track running from the Angolan border, across the Congolese copper belt, to the Zambian border.
It is good news. Copper and cobalt from Katanga have spent decades travelling thousands of kilometres by road to ports in South Africa, Tanzania and Mozambique. Trucking is slow, expensive, and exposed to every border queue between the mine and the ship. A working railway to the Atlantic cuts that journey from something like forty-five days to under ten.
I want to sit with a smaller detail. That line is the Congolese leg of the Lobito Corridor, and the corridor's spine is the Benguela Railway. The Benguela was chartered in 1902, when the Portuguese government granted a ninety-nine-year concession to Sir Robert Williams, a Scottish mining magnate and an associate of Cecil Rhodes. Construction started in 1903. The line reached the Belgian Congo border in 1929.
So the flagship infrastructure project of Africa's 2026 critical-minerals moment is a rehabilitation of a route designed in 1902 to move Katanga copper to a European-facing port.
The route was correct then, for the people who commissioned it. The question worth asking is whether it is still the route we would draw today, and what it means that we are mostly repairing rather than redrawing.
The test
Here is a test you can run on any colonial-era African railway, using nothing but a map.
Find the two endpoints. One of them will be a mine, an oil field, or a plantation belt. The other will be a port. Draw the line between them and you will notice it runs more or less perpendicular to the coast — inland to seaward — and that it does not stop anywhere particularly useful along the way except to pick up more of the same cargo.
Then look for what is missing. Look for lines running parallel to the coast, connecting one colony to its neighbour. Look for a line joining two interior cities that both had populations, markets, and reasons to trade.
You will find very few. Not because the engineering was harder, but because nobody was buying that.
Dakar–Niger ran groundnut country to the Atlantic. The Uganda Railway ran Lake Victoria to Mombasa. The Benguela ran Katanga to Lobito. In Cameroon, the trunk line the Germans began runs inland from the port at Douala, and even the northern extension to Ngaoundéré, finished in 1974 well after independence, was justified largely by what it could bring down to that same port.
The pattern is not subtle once you look for it. These were not transport networks. They were extraction pipes with stations attached.
The map was never neutral. It was a business plan laid in steel, and we are still shipping on it.
The gauge problem
There is a second-order effect that compounds the first, and it is almost comic in its stupidity.
Colonial powers did not coordinate track gauges. Why would they? A railway in French West Africa had no commercial reason to interoperate with one in British territory. Each system was a closed loop between somebody's asset and somebody's harbour.
So the continent inherited a rail map that is not merely badly shaped but physically unable to join up in places. Where two colonial networks meet at a modern border, the rolling stock frequently cannot continue. Freight gets unloaded, cleared, and reloaded — with every handling step adding cost, time, and an opportunity for something to go missing.
Sixty-odd years after independence, most of those loops are still loops.
What it costs
Now the number.
Intra-African trade sits at roughly 15 to 16 percent of the continent's total trade. Afreximbank's 2026 outlook puts it at about $210 billion in 2025, with an optimistic path to $230 billion this year. The comparison that matters: intra-Asian trade runs near 60 percent of Asia's total, and intra-European trade higher still.
African countries trade with the rest of the world far more readily than with each other. There are many reasons for this — tariffs, rules of origin, currency fragmentation, customs systems that do not speak to one another. Average customs dwell time on the continent has been measured around 126 hours, and logistics costs run close to double the global average.
But underneath all of those sits a physical fact. It is often genuinely easier to move a container from Douala to Marseille than from Douala to Kinshasa. The infrastructure was built to make the first journey cheap. Nobody was ever paid to make the second one cheap.
That is what I mean by the routing still being on the ground. Tariff reform is a policy question and it moves at the speed of ratification. Steel in the earth moves at the speed of capital projects, and it has a century of head start pointing the wrong way.
The part that breaks the determinism
I should stop here and say what this argument is not.
It is not a claim that geography sealed our fate. If colonial routing were destiny, nothing would ever change, and things have changed.
The clearest counter-example is TAZARA — the line from the Zambian copperbelt to Dar es Salaam, built in the 1970s specifically so that Zambian copper would not have to transit white-minority-ruled Rhodesia and South Africa. That railway was a political decision to redraw a route rather than inherit one. It got built. It has had a hard operating history since, and I am not going to pretend otherwise, but the point stands: the map is editable when someone decides to edit it.
The more current example is not steel at all.
The Pan-African Payment and Settlement System entered force in 2025 and is projected to cut foreign exchange costs on intra-African transactions by something like 20 to 30 percent. What PAPSS does is let two African countries settle in their own currencies rather than routing the payment through a correspondent bank in New York or London.
Read that again, because it is the same shape as the railway. The old financial plumbing also ran inland-to-seaward. Value from a Nigerian buyer to a Kenyan seller went out to a Western clearing system and came back, paying a toll each way, because that is how the pipes were laid.
PAPSS is a lateral line. It is the connection between two interior points that the original designers had no commercial reason to build.
And we built it in software, in years, at a fraction of what a railway costs.
Where this leaves us
The capability is not in question. African institutions designed and shipped a continental settlement system, and the continent runs the world's largest mobile money market by transaction value. When the route is ours to draw, we draw it.
The constraint is that the physical layer still mostly follows the 1902 logic, and the current wave of investment is largely reinforcing it. The Lobito Corridor is real, it is running, and it will make Congolese copper cheaper to export. It is also, in its essential geometry, a mine-to-port line with European and American strategic backing, arriving at a moment when everyone wants cobalt.
That is not a scandal. It is a fact about what gets financed. Minerals-to-coast has a bankable revenue model and always has. A line connecting two mid-sized African cities so they can trade cement and tomatoes does not, at least not on the spreadsheets currently being used.
Which means the interesting question for the next decade is not whether the corridors get built. They will. The question is what gets built alongside them, who is doing the underwriting, and whether anyone is measuring success by intra-African tonnage rather than export tonnage.
If we rehabilitate the colonial map faster than we extend it, we will have spent a great deal of money becoming more efficient at the one thing the map was always good at.
There is a version of the next twenty years where the corridors become genuine trade infrastructure — where the freight running to Lobito is joined by freight running along the continent, and the customs harmonisation being negotiated around the corridor gets reused everywhere else. The Lobito agreement includes exactly that kind of cross-border facilitation work. That precedent is worth more than the track.
There is another version where we get a very good extraction pipe with better rolling stock.
The difference between those two futures is not capacity. It is what we decide to finance, and who we let hold the pen on the map.
Day 3 of 60. This series names one real capability and one real constraint in every piece. Sources for this one: Afreximbank African Trade and Economic Outlook 2026; Lobito Atlantic Railway operational statements, June 2026; Congo Council of Ministers approval reported August 2026; GSMA State of the Industry Report on Mobile Money 2026.
