East Africa sits on solar irradiation of roughly 4.5 to 6.5 kWh/m²/day. That is a stronger and, more importantly, a far more consistent resource than most of the markets where solar first scaled. Germany built a solar industry on about half of it. So the interesting question was never whether the sun is there. It is why deployment still lags the resource, and what actually has to change on the ground.
Having designed and delivered solar PV and backup systems for agricultural and commercial clients, and run energy audits and site feasibility assessments across those deployments, my answer has shifted. The binding constraints are rarely technical. They are sizing discipline, financing structure, and what happens to a system in year three.
Oversizing is the most expensive habit in the industry
The single most common failure I encounter is not a faulty panel or an undersized inverter. It is a system specified from a wish rather than a measurement. A client describes their load, a supplier quotes generously, and the array arrives 40% larger than the site will ever use. The capital is spent, the payback stretches out, and the customer concludes that solar is expensive.
An energy audit before specification changes the economics more than any equipment choice. Measuring actual load profiles, separating baseload from peak, and understanding when demand occurs relative to the irradiance curve routinely reveals that a smaller, better-matched system delivers the same service for materially less money. Tools like PVsyst and HOMER Pro make this analysis cheap. Skipping it does not.
The cheapest kilowatt-hour on any site is the one you correctly established you never needed to generate.
Off-grid is where the returns are clearest
Where solar economics become genuinely obvious is anywhere it displaces diesel. Deploying off-grid solar pumping for smallholder farmers across semi-arid Kenya, each installation replaced generator-driven pumping with a PV array and submersible pump. The comparison is not solar against grid tariffs. It is solar against diesel at pump prices, plus transport to remote sites, plus engine maintenance, plus the days when fuel simply is not available.
That last factor rarely appears in a financial model and matters enormously in practice. Reliability of water access changes what a farmer is willing to plant. An irrigation system that works every day supports a different crop choice than one that works when fuel arrives.
The same logic extends past pumping. A solar-powered evaporative cooling unit built for off-grid smallholder farms extended tomato shelf life by an average of seven days in field trials, using low-cost, locally sourced materials. No grid connection, no compressor, no refrigerant. Post-harvest loss is an energy problem that does not look like one, and it is often solvable with less power than people assume.
Grid integration is the next real bottleneck
As grid-tied capacity grows, the constraint moves from generation to integration. Distribution networks in the region were designed for one-way power flow. Feeders serving dense commercial solar begin to see voltage rise, reverse flow and protection coordination issues that were never part of the original design case.
This is where tariff structure and engineering meet. Whether a commercial system is worth building depends heavily on how export is treated, what demand charges apply, and how net billing is settled. Modelling a Kenyan installation without encoding KPLC tariff logic produces a number that looks precise and is not decision-grade.
- Measure load before specifying capacity, not after
- Model against local tariff structure, not generic assumptions
- Treat EPRA compliance as a design input rather than a closing formality
- Budget for operations and maintenance from day one, not as an afterthought
- Design for the technician who will service it, not the engineer who specified it
What actually determines whether a system survives
Commissioning is a milestone, not an outcome. The systems that are still performing years later share unglamorous traits: accessible mounting, documented as-built configurations, an owner who understands the basics, and a defined path for fault reporting. In work supporting installation, commissioning and performance monitoring of household biogas systems, structured after-sales technical training and fault tracking cut system failures by around 30%. The hardware did not change. The support structure did.
This is the part of the industry that does not attract funding announcements. It is also the part that determines whether the next customer believes the technology works. Every failed installation in a community is a marketing cost paid by everyone who comes after.
Where this goes
The trajectory I would bet on is not utility-scale megaprojects. It is a dense mesh of correctly sized distributed systems: commercial rooftops sized against measured load, agricultural installations displacing diesel, and productive-use applications where energy unlocks value beyond lighting.
What makes that mesh investable is measurement. Monitoring across a portfolio of active sites, with faults identified and escalated rather than discovered during a complaint call, is what converts scattered installations into an asset class. The engineering is largely solved. Proving performance at portfolio scale is the work.
Working on something in this space? Start a conversation.