Read enough startup post-mortems and the same explanations recur: the market was not ready, the team fell out, the round did not close. Those are the reasons founders give. They are not always the reasons the company died.
A substantial share of failures across the continent trace back to operating costs that would be rounding errors elsewhere. Diesel for the generator. Redundant connectivity because the primary link cannot be trusted. Staff time lost to outages. None of it appears in a pitch deck, and all of it comes out of the same runway.
The generator line item
Companies running any kind of physical operation — a warehouse, a fulfilment centre, a hardware assembly line — routinely spend more keeping the lights on than on the engineering team. That ratio is survivable while a company is small. It becomes fatal at the point where scale is supposed to improve margins and instead makes the power bill grow in step with revenue.
Bandwidth is not the constraint people think it is
Raw bandwidth has improved markedly in coastal markets over the last decade. Reliability and price have not improved at the same rate, and landlocked markets remain dependent on transit through neighbours, with the pricing power that implies.
The practical consequence is that latency-sensitive products get built for the markets where the infrastructure already works, and the rest of the continent waits. That is a policy failure, not a market signal, and it deserves more attention than the funding charts get.