Solar in Africa doesn't have a hardware problem
The panels work fine. It's everything around them that breaks.
Panel prices have fallen off a cliff. A watt of solar costs a fraction of what it did a decade ago, the efficiency is fine, the manufacturing is mature and boring. The hardware is solved.
So why is there still a village two hours from a city centre running on a generator?
Because the panel was never the constraint. I say this as someone selling the panels.
The mismatch nobody prices in
A solar system is a large payment today in exchange for small savings spread over years. A household running on kerosene and phone-charging kiosks is paying small amounts, constantly, forever.
Those are the same total money and completely different products. One requires having the whole amount in hand at once. The other requires having a small amount on most days.
Every serious attempt at solar access in Africa is, underneath, an attempt to solve that mismatch. Pay-as-you-go works because it reshapes the payment into the cadence people actually have income in. The technology that made PAYG possible wasn't photovoltaic. It was mobile money and a cheap GSM module that can lock the unit remotely.
That's the thing to sit with: the innovation that unlocked solar adoption was a billing innovation.
Batteries are where it actually dies
Ask anyone who bought a system four years ago and stopped using it. It won't be the panel. Panels degrade slowly and visibly and mostly just keep working.
It'll be the battery.
Cheap flooded lead-acid in a hot room, cycled too deep, with no temperature compensation on the charge controller — that's a battery with an eighteen-month life being sold into an expectation of five years. The customer doesn't experience this as "my battery wore out." They experience it as "solar doesn't work," and they tell their neighbours that.
The honest options cost more upfront. GEL tolerates heat and neglect better than flooded. LiFePO4 costs more per unit and considerably less per usable cycle, tolerates depth of discharge that would kill lead-acid, and lasts long enough that the arithmetic works — if the customer can finance the gap.
Which puts you right back at the financing problem. It's financing all the way down.
Heat is not a footnote
Datasheet numbers are measured at 25°C. A rooftop in the afternoon is not 25°C. Panels lose output as they heat, inverters derate or shut down, and battery chemistry ages faster in a hot enclosure than in a cool one.
A system specified off datasheet numbers and installed under a tin roof with no ventilation will underperform its spec, and the installer will blame the customer's usage. I've seen enclosures that were effectively ovens with batteries inside them. The equipment was fine. The install killed it.
Why donated systems fail
Someone funds fifty systems for a community. There's a ribbon. Eighteen months later, a third are dead.
Not because the equipment was bad, usually. Because nobody was paid to come back. No spare parts in the region, no technician whose income depends on the units running, no relationship that survives the handover. A donation creates an asset with no owner.
A business creates a much less inspiring thing — a person whose rent depends on that system working next year. That person will drive out to fix it. The charity, by then, is funding a different community.
I'd rather build the boring version.
What this means if you're building
Stop leading with the panel. Everyone has panels; they're commodities from the same handful of factories.
Lead with the parts that are hard to copy: the financing structure, the technician network, the parts depot, the collections process, the willingness to specify a battery that costs more and lasts. That's the actual company. The hardware is the easy part, which is exactly why it's the part that isn't the moat.