Why Satellites Matter More Than Fibre

The title is deliberately unfair, so let me be precise about the claim. Fibre is better than satellite for almost everything it can reach. Where fibre exists, use it. I am not arguing about bandwidth.

I am arguing about what is scarce, and bandwidth is not the scarcest thing here.

Zambia is 752,000 square kilometres. Ask a simple question about it. How much maize is standing in Central Province this week, and how stressed is it? Which stretches of the Kafue have dropped below last year? Where was charcoal cut in the last thirty days? How many hectares of the cotton crop were actually planted against what was contracted?

These are not hard questions. They are unanswered questions, which is a different thing.

We answer them today with sampling, self-reporting and extension officers covering areas the size of small countries. The result is a number that arrives late, at national resolution, with an error bar wide enough to hide a bad season inside it. Policy gets made on it. Loans get priced on it. Insurance gets refused on it.

Fibre does not fix this, because fibre transports information from one place that has it to another. If nobody generated the observation in the first place, a faster pipe moves nothing.

Earth observation generates it. A satellite pass does not care that the road to the district is impassable in February. It does not need an officer with fuel. It returns on schedule, covers everything, and produces a measurement that can be compared to the same measurement from last year.

This is why SpaceZ exists, and why the first products are not glamorous. Crop area and condition. Water extent. Land use change. Flood and drought signal. These are the boring measurements that everything else depends on, and right now they are largely imported, delayed, or absent.

The second reason to build here rather than buy is that the analysis layer is where the value sits, and it is the layer that has to be local. A global model will tell you a field is stressed. It will not tell you that this is the third year of a lease dispute, that the cooperative switched seed varieties, or that the road washed out and the inputs never arrived. Interpretation requires being from here.

I get asked whether Zambia should be launching its own satellites. Eventually, for specific purposes. But that is the wrong first question and it wastes a decade. The immediate opportunity is that enormous quantities of imagery already exist, much of it free, and almost nobody in the region is turning it into decisions people will pay for. The constellation is not the moat. The interpretation is.

Buy the pixels. Own the meaning.

Build fibre. Build towers. Build the terrestrial layer as fast as capital allows. But understand that connectivity is a distribution problem, and observation is a creation problem, and we have been treating the second one as if it were somebody else's job.