There is an irony at the heart of Nigeria's biggest energy access programme that nobody seems to be talking about.
DARES — the $750 million Distributed Access through Renewable Energy Scale-up project — is fundamentally a performance-based programme. The World Bank structured it so that mini-grid developers earn grants based on verified results. You commission a site, connect customers, prove reliable service for 12 months, and the money flows. Up to $600 per verified connection, with a significant holdback contingent on proven reliable operations. The Independent Verification Agent visits your site. The data either backs your claim or it does not.
This is pay-for-outcomes logic. Incentives are aligned, capital flows toward results, and underperformers are filtered out naturally.
And yet every piece of operational software serving the mini-grid sector in Nigeria charges flat monthly subscriptions. Per-site fees. Per-seat licenses. Annual contracts denominated in dollars. The programme itself is performance-based. The tools that operators depend on to deliver that performance are not. They get paid whether the operator succeeds or fails.
Something is broken here.
Why SaaS Pricing Fails Nigerian Mini-Grid Operators
Let me walk through the economics that most software vendors either do not understand or choose to ignore.
A mini-grid developer deploying under DARES has typically borrowed capital at Nigerian commercial lending rates — which hit 29.79% on average in January 2025, with some manufacturers paying 35%. They have invested that capital into site construction before receiving performance-based grants. The grant structure is milestone-driven, with the critical final tranche held back pending 12 months of demonstrated reliable service.
The developer is cash-negative from day one, servicing expensive debt while waiting for grant disbursements that depend on verified performance data.
Now layer in the naira. Nigeria's currency depreciated 41% against the dollar in 2024, hitting an average of ₦1,479 per USD. Annual inflation surged to 31%. Operational costs for businesses rose 67% year-on-year. Revenue is collected in naira from rural communities. Grant disbursements flow through processes that can take months.
Into this environment walks a monitoring platform saying: "That will be $150 per site per month, payable annually in advance, in dollars."
For an operator running 30 sites, that is $54,000 a year in fixed software costs — before the platform has demonstrated it will save a single naira. At current exchange rates, software costs can eat 3-5% of gross revenue. For a sector where margins are already razor-thin, that is an existential line item.
The deeper problem: the operator cannot evaluate whether the software delivers value until they have used it for months. But the pricing model demands commitment before proof. The vendor's revenue is decoupled from the operator's outcome.
This is exactly the misalignment that DARES was designed to eliminate from the grant structure. Why are we tolerating it in software?
What Diesel Waste Actually Costs
Let me make the savings concrete, because this argument lives or dies on the numbers.
The SEforALL Mini-grid CAPEX and OPEX Benchmark Study (August 2024) — covering Nigeria, Burkina Faso, and Sierra Leone — found that fuel constitutes 51% of ongoing operational expenditure for diesel-hybrid mini-grids. For the typical Nigerian hybrid site spending ₦500,000 per month on diesel, that is the single largest controllable cost centre.
AI-optimised diesel dispatch — intelligently scheduling genset runtime based on load forecasting, battery state of charge, and solar irradiance prediction — captures 15-30% more value than rule-based systems. Stem demonstrated this at utility scale with their Athena platform across 30 GW of assets. The technology is not theoretical.
For that ₦500,000/month diesel site, 15-30% optimisation means ₦75,000 to ₦150,000 in monthly savings. Per site. Every month.
Scale that across a fleet. DARES targets 1,350 mini-grids. Even if half run diesel-hybrid configurations, the aggregate diesel waste from suboptimal dispatch runs into billions of naira annually. Money burned — literally — because dispatch decisions are made by timer-based rules instead of intelligent systems.
O&M more broadly represents 33% of total tariff cost build-up. Equipment failures that go undetected for days compound into premature battery degradation, shortened inverter lifespans, and cascading costs from reactive rather than predictive maintenance.
The value is there. The question is who captures it, and how.
What If Operational Software Was Priced Like DARES Grants?
Imagine a different model.
An operational platform connects to your fleet — Victron VRM, SparkMeter Koios, FusionSolar. It monitors your sites, optimises diesel dispatch, automates your Odyssey reporting for DARES compliance, detects faults before they cascade. And you pay based on verified outcomes. A percentage of measured diesel savings. A fee tied to accelerated grant disbursements. If the platform does not save you money, you do not pay.
This is not radical. The energy sector has precedents.
Omnidian guarantees that every solar asset under its management will produce at least 95% of projected output — and pays the difference if it falls short. Revenue tripled between 2022 and 2024. The model works because Omnidian only profits when asset owners profit.
ESCOs have operated on shared-savings contracts for decades. The ESCO finances the efficiency upgrade and splits measured savings with the customer. If savings do not materialise, the ESCO absorbs the loss. The IEA documents these models as proven globally.
Schneider Electric's EaaS and Bloom Energy's zero-upfront power solutions convert capital expenditure into performance-guaranteed fees. Customers pay for outcomes, not for licenses.
The thread connecting all of these: the vendor's revenue depends on the customer's outcome. This is precisely how DARES structures its grants. Why should the software layer operate on a different philosophy?
Addressing the Objections
I know what the sceptics are thinking.
"How do you verify savings?" The same way DARES verifies connections — with data. Baseline diesel consumption before optimisation versus measured consumption after. Grant disbursement timelines before automated compliance reporting versus after. Mini-grid systems generate telemetry continuously. The measurement infrastructure already exists. A vendor whose revenue depends on verified savings has a stronger incentive to ensure metering accuracy than one collecting flat fees regardless.
"What if the operator games the system?" The same controls that prevent DARES gaming apply. Baselines are established during a measurement period. The IVA model — independent verification by a third party — can apply to software savings just as it applies to connection counts.
"Can a software company survive on performance fees?" Ask Omnidian, whose revenue tripled in two years. Ask the ESCO industry, which the IEA values in the tens of billions globally. If the platform genuinely delivers value, the vendor's share of verified savings can exceed what flat subscriptions would have earned. The difference is they have to earn it. Every month.
"What about free pilots?" This is where the model accelerates adoption. A 90-day pilot with no upfront cost removes every barrier to trial. After 90 days of real data showing real savings, the conversation shifts from "will this work?" to "how do we structure the ongoing relationship?" You are negotiating from demonstrated value, not projected ROI in a pitch deck.
Trust as a Business Model
There is a dimension to this that goes beyond spreadsheet economics.
Anyone who has done B2B sales in Nigeria knows that relationships and demonstrated value come before contracts. The operator burned by a vendor promising transformation and delivering a dashboard is not signing another annual contract on faith. The CFO who watched the naira halve while locked into a USD-denominated SaaS agreement is not enthusiastic about the next one.
Performance-based pricing is, at its core, a trust mechanism. It says: we believe in our technology enough to bet our revenue on your results. If we fail, we eat the cost. If we succeed, we share the upside.
Operators deploying sites with borrowed capital at 30% interest rates, collecting revenue in a depreciating currency, staring at a 12-month performance holdback that determines whether their project breaks even — they do not need another fixed cost delivering uncertain value. They need a partner structured the same way the grant programme is structured: outcome-first.
The World Bank's GPRBA exists because pay-for-outcomes drives better results than pay-for-activities. DARES embodies this. It is time for the software layer to catch up.
The Sector Needs a New Default
The mini-grid sector in Nigeria is entering its most critical phase. Over 900 sites are under construction. DARES is disbursing. Nobody is managing a fleet of 50 sites with WhatsApp and Excel — not if they want to hit their performance milestones, collect their grant holdbacks, and keep their diesel costs from eating their margins alive.
But the pricing model matters as much as the technology. A flat-fee SaaS subscription asks a cash-strapped operator to absorb risk the vendor should be sharing. A performance-based model distributes risk correctly: both parties profit only when outcomes improve.
The difference between a developer who collects their 12-month performance holdback on time and one who misses the deadline because their data was not in order can be tens of thousands of dollars per site. Multiply that across a fleet. Multiply that across the sector.
If the grant programme itself is structured around verified performance, why should the operational tools that enable that performance be priced any differently?
I would like to hear from operators, finance leads, and investors in this space. Does performance-based pricing for operational software make sense for your portfolio — or are there structural reasons it cannot work that I am not seeing?