← The catalogue
Technical paper · Extrafemi Research · Nov 2026

From ladder to exit: Nigerian households, generators, solar and the grid

Olorunfemi Obasa, MBA

First published on this site in November 2026. Appendix B records which figures were checked against a source and which were not.

In plain words

Nigerian households do not simply climb from kerosene to solar; they keep every old source and add new ones. The paper prices each step per unit of electricity and finds that a small petrol generator costs five to eleven times the grid, while solar and a battery cost less than the generator in every case tested. Once a household owns solar, the grid is the dearer source, so the logical end of the ladder is leaving the grid, which only households that can find about ₦3 million can do.

Published
17 November 2026
Reading time
33 min
Data
20 charts

Abstract

A popular account of Nigerian household electricity describes a ladder: kerosene lanterns, then rechargeable lamps, then a small generator, then a larger one, then an inverter with solar panels. This paper tests that account against published survey evidence and a transparent cost model, and asks where the ladder ends. It does not assume the grid recovers; every comparison uses the grid as measured. Four findings follow. First, today's households stack sources instead of replacing them: in a 2021 survey of 3,599 grid-electrified rural and peri-urban households, 37%, 18% and 66% used a generator for lighting in the three zones sampled, alongside grid connection of 85% to 98%, so the ladder is better described as a pile. Second, the rungs differ sharply in cost. At November 2026 prices a small petrol generator costs about ₦1,000 to ₦2,400 per kilowatt-hour in fuel alone, 4.8 to 11 times the Band A grid tariff of ₦209.5, while a solar-and-battery system costs about ₦170 to ₦560. Third, the end of the ladder is leaving the grid, not keeping it: once a household owns the system, the grid is the dearer source for anything the system can supply, and unless the grid delivers more than about 22 hours a day (16 to 24 across the scenarios tested), solar and battery alone is cheaper than the grid with a petrol generator behind it. Surveys put supply at 6.6 to 8 hours. Fourth, few can make the move: a mid-sized system costs about ₦3.0 million, 43 months of the minimum wage, and in the 2024 Demographic and Health Survey only 1.7% of the poorest wealth fifth had electricity of any kind. The paper also traces the best-known figures on generator spending, which range from $3.8 billion to $14 billion a year with no primary source for the highest, and tests the post's two predictions: rooftop wind turbines are very unlikely to compete, and estate-level interconnection is plausible but sits in a legal gap.

1. The claim

In November 2026 a post on X (by @cchukudebelu) described the household path to electricity as a ladder: kerosene lanterns, then rechargeable lanterns, then a small generator, then a 2.5 kVA generator, then an inverter with solar. It added that the government had done nothing tangible in 30 to 50 years, and predicted that households will next "install wind turbines on your house roof" and later link estates into mini-grids when the national grid fails.

The post is an opinion, not evidence. It is useful because it states three testable claims:

  1. The ladder claim. Households move up a sequence of sources, each replacing the one below.
  2. The state claim. Nothing tangible was delivered by the public system over several decades.
  3. The forecast. Rooftop wind and estate interconnection come next.

This paper takes each in turn. It builds on the desk's earlier essay on household solar and on a companion piece, Why Nigeria Has No Power, which gives the history of the grid. Where this paper uses a figure from that history it says so and flags whether it has been checked (Appendix B).

2. The setting

Nigeria's grid delivers far less power than its installed capacity suggests. The Nigerian Electricity Regulatory Commission's quarterly report for Q3 2023, as cited in a 2024 study for the Lagos State Government by Sustainable Energy for All (SEforALL), puts installed grid capacity at 12.2 GW, peak available capacity at about 5.4 GW and average available generation at 4.2 GW. The national grid collapsed 12 times in 2024 (Guardian timeline, February to December). The 2025 edition of Tracking SDG7, from the World Bank and partner agencies, puts Nigeria at the top of the world's list of people without electricity: about 86.8 million in 2023, an access rate of about 61%.

2.1 Collapses and output over time

I could not find a systematic public record of national grid collapses before 2010, so the series below starts there and cannot cover the whole period since independence. The System Operator's own records, as published by the ICIR in October 2019, count 206 collapses between 2010 and October 2019, 146 of them total and 60 partial.

System Operator records show 206 grid collapses between 2010 and October 2019, 146 of them total

Hover, focus or click a row. Use the legend to show or hide a series.

Unit: collapses. The article gives no partial count for 2019, shown as 0. TCN's later figures use a different definition and are not comparable (see the table below).

Source: ICIR, "Nigeria's power grid records 206 collapses in 9 years" (20 October 2019), citing the System Operator

Counts for later years come from other sources, with other definitions, and they disagree.

YearCountSource and what it counts
201715 total and 9 partialSystem Operator records, via ICIR (2019)
20178The Republic's story "Why Nigeria Has No Power", citing TCN event logs (not independently checked)
20204TCN, as summarised in a search result for a Leadership report; the original was not read
20212same
202412Guardian timeline of collapses, 4 February to 11 December 2024
2025not establishedA partial collapse on 29 December was called total by some media and partial by the System Operator; I found no full-year count

Reading the two together, collapses were frequent in 2010 to 2016, fewer on TCN's own figures in 2019 to 2021, and back to 12 in 2024. The sources are not consistent enough to call a trend, and the point for this paper is narrower: a household reading this record would not conclude that the grid is about to become reliable.

01020304050050100150200TWhkWh per person200020052010201520202025

Hover, touch or use the arrow keys on the chart to read every line at a point. Click a name above to hide or show it.

Unit: left: TWh; right: kWh per person. Total generation as reported by Ember. It does not include the roughly 42 GW of small generators estimated by Dalberg and A2EI (2019), which are not metered. Per-person figures use the same source’s population.

Electricity generated in Nigeria rose 2.8 times between 2000 and 2025, from 14.7 to 41.5 terawatt-hours, according to Ember and Energy Institute data. Over the same period the population grew by about 90%, so generation per person rose by about half, from 117 to 175 kilowatt-hours a year. For comparison, the PeopleSuN paper cites South Africa at 4,198 kilowatt-hours per person and Ghana at 351.

Only about a third of Nigeria's installed grid capacity reaches the grid: 2,000 of 6,000 MW in 1999 and 4,200 of 12,200 MW in 2023

Hover, focus or click a row. Use the legend to show or hide a series.

Unit: MW. 1999 delivered is the midpoint of the 1,800 to 2,200 MW the story gives; 2023 delivered is the average available generation reported by NERC.

Source: 2023: NERC Electricity on Demand report Q3 2023, as cited by SEforALL and Lagos State Government (2024); 1999: The Republic's story "Why Nigeria Has No Power", citing Ministry of Power and NERC data; not independently checked

The long record the companion essay gives for installed capacity runs from 132 MW at independence to 805 MW by 1972, which on UN population figures is about 3 and 14 watts per person. I have not checked those figures, and the measures change over time (installed against delivered), so I do not chart them.

Household surveys give lower and more varied figures. The Demographic and Health Survey (DHS) reported 59.4% of Nigerian households with electricity in 2018 and 55.2% in 2024. The National Bureau of Statistics (NBS) Nigeria Residential Energy Demand-Side Survey 2024, covering nine states and 8,100 households, reported 58.2% of surveyed households connected to the grid and an average supply of 6.6 hours a day among those connected (as republished by Premium Times; I have not seen the NBS report itself).

Supply is poor even where there is a connection. In the 2021 PeopleSuN survey of grid-electrified rural and peri-urban households in the North West, North Central and South South, about half of households and enterprises received fewer than eight hours of grid electricity a day.

3. Stacking or climbing

3.1 What the survey evidence shows

The best published household evidence on how sources combine is the PeopleSuN survey (Pelz and colleagues, 2023, Scientific Data; preprint on arXiv). It interviewed 3,599 households and 1,122 small and medium enterprises across three geopolitical zones in 2021, drawn to represent grid-electrified rural and peri-urban areas. It asked which sources households use for lighting.

Source used for lighting (share of households)North CentralNorth WestSouth South
National grid connection98%85%93%
Electric generator37%18%66%
Rechargeable battery35%26%30%
Dry-cell battery or torch59%68%32%
Solar home system1%6%1%
Solar lantern1%2%0%
Local mini-grid0%0%4%

The story

A ladder: each step replaces the one below

  1. Kerosene lantern
  2. Rechargeable lamp or torch
  3. Small generator
  4. 2.5 kVA generator
  5. Inverter and solar

The survey

A pile: households use several sources at once

North Central
  • National grid98%
  • Generator37%
  • Dry-cell battery or torch59%
  • Rechargeable battery lamp35%
  • Solar home system1%
  • Solar lantern1%
  • Local mini-grid0%

Shares add up to 231%: about 2.3 sources of light per household.

North West
  • National grid85%
  • Generator18%
  • Dry-cell battery or torch68%
  • Rechargeable battery lamp26%
  • Solar home system6%
  • Solar lantern2%
  • Local mini-grid0%

Shares add up to 205%: about 2.0 sources of light per household.

South South
  • National grid93%
  • Generator66%
  • Dry-cell battery or torch32%
  • Rechargeable battery lamp30%
  • Solar home system1%
  • Solar lantern0%
  • Local mini-grid4%

Shares add up to 226%: about 2.3 sources of light per household.

Share of households using each source for lighting, 2021 PeopleSuN survey of 3,599 grid-electrified rural and peri-urban households (Pelz and colleagues, 2023). Households can use several sources, so each column adds to more than 100%. The survey covers places that already have a grid connection and reports lighting, not all electricity use.

The shares add to far more than 100% in every zone. The authors note this themselves: the column "adds to more than 100% due to fuel stacking", and they report primary, secondary and tertiary lighting sources for each household. In the South South, two-thirds of households use a generator for light while 93% are connected to the grid. In the North West, 68% use dry-cell batteries or torches and 85% are connected.

Two limitations matter. The survey covers grid-electrified places only, so it says nothing about the 40% or so of households with no connection. And it reports lighting, not total electricity, so it understates the use of generators for fans, fridges and televisions.

The NBS 2024 survey, as reported, found 4.8% of households with solar access, 90.9% of them on solar home systems, with solar users reporting an average of nine hours of supply a day. That figure is low beside the import data: Ember's analysis of Chinese customs data found Nigeria imported 1,721 MW of solar panels in the 12 months to June 2025, the second most in Africa after South Africa's 3,784 MW, though imports are not installations and Ember notes a government import-ban plan cut first-quarter 2025 shipments sharply.

3.2 What this says about the ladder

The ladder describes an order of adoption. The surveys describe a pile. A cross-section can show that a household uses several sources at once, and cannot show the order in which it acquired them. Settling the order would take panel data that follows the same households over several years, such as the NBS General Household Survey Panel microdata, which require registration and which I could not access.

What the evidence does support is a weaker claim: new sources are added far more often than old ones are dropped. In the survey, generator use coexists with grid connection, battery lamps and torches. That matters for cost, because every retained rung keeps carrying its own fixed and running costs.

It does not follow that households will keep the grid. Stacking describes where households are today; it says little about where they end up. A connection that costs nothing to hold is a free option, and people keep free options. I did not find whether distribution companies charge a standing fee on prepaid meters; if they do not, that would explain why households keep a connection they barely use. Section 5 asks what happens to that option once a household owns enough solar and battery to do without it.

4. What each rung costs

4.1 Method

The model compares the cost of one kilowatt-hour of electricity from three sources: the grid at its highest-quality tariff (Band A), a small petrol generator, and a household solar-and-battery system. Lighting from kerosene is treated separately because a kilowatt-hour of lantern fuel is not comparable to a kilowatt-hour of electricity.

All inputs are named constants in drafts/analysis/energy-ladder.py, each with its source (Appendix A). Three scenarios move the uncertain inputs together: one tilted in favour of the generator, one central, and one against it. The tilt is deliberate. Where I do not know an input well, the "favourable to the generator" scenario uses the value that makes the generator look best and solar look worst, so a conclusion that survives it is robust.

Prices are November 2026: petrol ₦1,596 per litre, diesel ₦3,277 per litre and household kerosene ₦2,972 per litre (NBS, May 2026, via ThisDay); Band A at ₦209.5 per kilowatt-hour; the naira at about ₦1,354 to the dollar (official market, mid-November).

4.2 The petrol generator

Fuel use comes from a published specification, not a guess. The Honda EU2200i inverter generator has a 3.6-litre tank and runs 3.2 hours at its rated 1,800 W and 8.1 hours at a quarter load (450 W). That is 0.625 litres per kilowatt-hour at full load and 0.99 litres at a quarter load. A household running a fridge, fans and a television rarely loads a generator fully, so I take 0.99 as the central value. Honda's inverter model is among the most efficient small generators, so this is generous to the generator; the unfavourable scenario uses 1.5 litres per kilowatt-hour for an older conventional unit, which is an assumption and is labelled as one.

Capital cost uses 2026 retail listings for 2.5 kVA units (₦291,000 to ₦550,000) spread over an assumed life of three to eight years at a household discount rate of 10% to 30%. Maintenance is set at 13.4% of fuel cost, the ratio in the SEforALL Lagos study (₦192 billion on maintenance against ₦1.43 trillion on fuel).

4.3 Solar and battery

The reference system is a 5 kVA hybrid inverter, four 550 W panels (2.2 kW) and a 5 kWh lithium iron phosphate battery. Component prices come from 2026 Nigerian retail listings: panels ₦115,000 to ₦165,000 each, battery ₦1.45 million to ₦1.9 million, inverter ₦420,000 to ₦850,000, installation ₦120,000 to ₦200,000. Battery life is taken as eight years, because manufacturers rate lithium iron phosphate units such as the EcoFlow DELTA 2 for about 3,000 cycles to 80% capacity, which is about eight years at one cycle a day. Panels last 25 years and the inverter 10. Annual output uses a panel yield of 3.5 to 4.5 kWh per kilowatt-peak per day (a range reported for Nigeria; the central 4.0 is an assumption), 85% system efficiency and a 10% allowance for cloudy days. Operation and maintenance adds 1.5% of capital a year.

4.4 Results

Cost per kilowatt-hour (₦, November 2026)Favourable to the generatorCentralUnfavourable to the generator
Grid, Band A tariff210210210
Petrol generator, fuel only9981,5772,394
Petrol generator, fuel, capital and maintenance1,2041,9673,120
Solar and battery, all costs557319169
Generator all-in cost as a multiple of solar2.26.218.5

A small petrol generator costs about seven times the grid tariff per kilowatt-hour in fuel alone

Hover, focus or click a bar for its detail. Click to pin it and compare the others against it.

Unit: ₦ per kWh. NERC Band A tariff (₦209.5) assumes about 20 hours of supply a day; most households in the surveys cited here get much less. Generator fuel use is from the Honda EU2200i specification at a quarter load.

Source: Extrafemi Research calculation; inputs and sources in Appendix A

Three results stand out.

Fuel alone costs 4.8 to 11 times the Band A tariff. That holds in every scenario. Even the best case for the generator, a Honda inverter at full load and petrol at ₦1,596, costs about five times the grid. This is why the grid, when it works, is the cheapest electricity a household can buy, and why unreliable supply is so expensive: every hour the grid is off, the household buys electricity at five to eleven times the price.

Solar and battery is cheaper than the generator in every scenario, by a factor of 2.2 in the scenario that is most unfavourable to solar and 18.5 in the one most favourable to it. In the central case the break-even petrol price is about ₦323 a litre, a fifth of today's. In the stress case where solar is dearest, break-even is ₦890 a litre, still well below the May 2026 price. Solar and battery is dearer than Band A, by 1.5 times in the central case, so it competes with the generator and not with a working grid.

The costs favour the top rung, yet most households cannot buy it. The central system costs about ₦3.0 million up front. Against the generator the payback depends on how much electricity the household uses:

Household use (kWh a day)Fuel and maintenance avoided per year (₦)Simple payback (years)
1653,0004.6
21,305,0002.3
31,958,0001.5
53,263,0000.9
7.54,895,0000.6

A household using three units of electricity a day pays back a ₦3.0 million solar system against petrol generation in about a year and a half

Hover, focus or click a bar for its detail. Click to pin it and compare the others against it.

Unit: years. Savings are fuel and maintenance avoided at November 2026 petrol prices. A household that already owns a generator has sunk its cost; this chart does not count it.

Source: Extrafemi Research calculation; inputs in Appendix A

These paybacks assume the household would otherwise buy all that electricity from a petrol generator. Ember's published estimate for full-system payback in Nigeria, 18 to 36 months, is consistent with the 1.5 to 2.3 years above.

4.5 Lighting

Lighting is where the ladder is clearest, and where the numbers are oldest. Studies of fuel-based lighting by Mills (Lawrence Berkeley National Laboratory) put the cost of light from a hurricane kerosene lamp at about $0.40 per thousand lumen-hours against $0.15 for a solar LED lamp in a Tibetan study, and estimate worldwide spending on fuel-based lighting at $38 billion a year; a GTZ test found solar lanterns near par with kerosene lamps. These are older studies from other countries, so I do not convert them to naira. At ₦2,972 a litre for kerosene, the direction of the gap has certainly not narrowed. The desk's earlier essay lists a Sun King Home 200 solar lighting kit at ₦105,000, or about ₦2,000 a month on instalment.

5. When a household leaves the grid

The ladder story ends with solar. The question for the grid is what happens next. Three tests, each on the central scenario unless stated, say that a household with a system sized for its load does better without the connection, and that the case for staying connected weakens as supply gets worse.

5.1 Once a system is owned, the grid is the dearer source

A solar-and-battery system's cost is almost all up front. After it is bought, the next kilowatt-hour is cheap: about ₦18 for solar used directly (operation and maintenance), and about ₦144 for solar passed through the battery, which adds ₦126 of battery wear (a ₦1.7 million battery over about 13,500 kilowatt-hours of use). Band A is ₦209.5. For any energy the system can supply, the grid is the more expensive source, even at its best tariff. A household keeps the connection only for the hours when the system falls short, and only while holding it costs nothing.

5.2 How close the battery price is to the crossover

01002003004005008001000120014001600180020002200Band A grid tariff (₦209.5)

Hover, touch or use the arrow keys on the chart to read every line at a point. Click a name above to hide or show it.

Unit: ₦ per kWh. Central assumptions for everything but the battery: panels ₦140,000 each, inverter ₦600,000, installation ₦160,000, 4.0 kWh per kWp a day, battery life eight years. Today’s 5 kWh lithium batteries list at ₦1.45 million to ₦1.9 million. A higher discount rate reflects borrowing costs and the pull of other uses of money; it favours cheaper equipment with less up-front cost.

The chart finds the battery price at which the whole system costs the same per kilowatt-hour as Band A. At a 10% discount rate it is ₦1.53 million, inside the ₦1.45 million to ₦1.9 million range at which 5 kWh batteries list today. At 20% it is about ₦728,000, 57% below the central battery price. At 30% it is about ₦227,000. Which rate applies depends on how a household values money now against savings later, and for many it will be high. I do not model where battery prices go. The point is that for a household with access to cheap finance the crossover is already here, and for others it is within reach of a price fall.

5.3 The grid with a generator behind it

Most households that keep the grid also need backup for the hours it is off. If the backup is a petrol generator, the cost of a kilowatt-hour is the grid's tariff for the hours it is on and the generator's all-in cost for the rest.

05001,0001,5002,00005101520Band A tariff

Hover, touch or use the arrow keys on the chart to read every line at a point. Click a name above to hide or show it.

Unit: ₦ per kWh. Load is assumed spread evenly over the day, the generator’s all-in cost is ₦1,967 a kilowatt-hour and the solar system is sized for the whole load. The break-even is 15.6 hours in the scenario least favourable to solar and 24.3 (above 24, so solar alone always wins) in the one most favourable to it. Surveys put actual supply at about 6.6 to 8 hours a day for connected households. Households that already own a generator have sunk its cost, which this chart does not.

With 8 hours of grid supply a day the blended cost is about ₦1,381 per kilowatt-hour in the central scenario, against ₦319 for solar and battery alone. Even with Band A's promised 20 hours it is ₦502. Solar alone is cheaper unless the grid supplies more than 22.5 hours a day. In the scenario least favourable to solar, where its equipment is dear and money costly, the threshold falls to 15.6 hours; in the one most favourable to it, solar alone wins at any supply. Surveys report 6.6 to 8 hours for connected households. On these numbers the grid adds a small saving on top of a system that is already doing most of the work.

5.4 What this does and does not show

It shows that exit is the cost-minimising path for a household that can finance it, in all three scenarios, and that the case for staying connected weakens with every hour the grid is off. It does not show that households have left. The surveys find connection and generators together, and solar use was only 4.8% in the NBS 2024 survey, mostly small home systems that sit beside a connection. Households may stay connected for reasons the model leaves out: the option is free, air conditioners and other heavy loads need a larger system than the 7.5 kWh a day assumed here, tenants cannot always install equipment, and finance is scarce. The companion essay reports that the Presidential Villa announced in 2026 that it would disconnect from the grid; I have not checked that report.

For the grid, the consequence is a hypothesis this paper does not test. The customers best placed to leave are those on the highest tariffs and the heaviest loads, which are the customers who pay the most. If they go, the distribution companies lose their best revenue and the connections that remain belong to the households least able to leave. I found no data on connection cancellations, which is the evidence that would settle it.

6. Who can climb

The cost model says the upper rungs are cheaper per kilowatt-hour. It does not say who can afford them.

In 2024, 1.7% of the poorest fifth of Nigerian households had electricity, against 96.5% of the richest

Hover, focus or click a bar for its detail. Click to pin it and compare the others against it.

Unit: % of households. The DHS wealth index is built partly from household assets, and electricity is one of them, so some of this gradient is built in. The 2018 survey gave 5.7% for the poorest fifth and 97.7% for the richest.

Source: The DHS Program (USAID), Nigeria DHS 2024, indicator HC_ELEC_H_ELC

The DHS reports the share of households with any electricity: 55.2% in 2024, down from 59.4% in 2018. Urban access was 79.8% and rural 29.9% in 2024. The gradient by wealth is steep, and the poorest fifth has essentially no access. For these households the first rung is a connection, not a lantern.

0204060801002005201020152020

Hover, touch or use the arrow keys on the chart to read every line at a point. Click a name above to hide or show it.

Unit: % of households. The DHS wealth index is built partly from household assets and electricity is one of them, so some of this gradient is built in. Whether solar or generator power counts as “electricity” in the DHS question has not been checked here. National access was 52.2% in 2003 and 55.2% in 2024, down from 59.4% in 2018.

The capital barrier compounds this. The central system costs about ₦3.02 million. That is 43 months of the ₦70,000 minimum wage, and 60 years of the ₦4,156 average monthly electricity spending the NBS 2024 survey reports. Pay-as-you-go lighting kits cost 1.5 months of minimum wage and put an entry rung within reach, but they run a few lights and a phone charger; they do not run a fridge.

The price of each step

One square is one month of the ₦70,000 minimum wage

  • Solar lighting kit (Sun King Home 200)₦105,000 · 1.5 months

    3 lights and phone charging

  • 2.5 kVA petrol generator₦400,000 · 5.7 months

    ₦291,000 to ₦550,000 on 2026 listings; fuel extra

  • Portable battery station (EcoFlow DELTA 2)₦767,000 · 11 months

    1 kWh; panels sold separately

  • 5 kVA solar-and-battery system₦3,020,000 · 43 months

    4 panels, 5 kWh battery, hybrid inverter, installation

Who has any electricity at all

  • Poorest fifth1.7%
  • Second fifth13.2%
  • Middle fifth52.2%
  • Fourth fifth85.8%
  • Richest fifth96.5%
Prices are 2026 retail listings in naira (the desk’s earlier essay on household solar, and Nigerian retailers). Minimum wage ₦70,000 a month. Access is the share of households with electricity in the 2024 Nigeria Demographic and Health Survey; the survey’s wealth index includes electricity as one of its assets, so part of that gradient is built in.

The implication is that exit is a choice for middle- and upper-income households. The households left on a weak grid with little else are disproportionately the ones least able to leave it, which is the equity cost of the pattern in Section 5.

7. Auditing the numbers we quote

Discussion of Nigerian self-generation leans on a small set of headline figures: how many generators there are and how much is spent on them. Tracing them back, they do not rest on one source.

FigureAttributed toWhat I found
22 million small generatorsDalberg and the Access to Energy Institute (A2EI), 2019SEforALL (2024) cites "22 million small gensets (0–4 kVA)", total capacity about 42 GW, from Dalberg and A2EI 2019. BusinessDay and Daily Maverick repeat it, the latter with a wider range of 22 to 60 million.
$12 billion a year on generator fuelBloomberg, 2019Carried by The Cable and Nairametrics. No primary study located.
$14 billion a yearThe African Development Bank, the IMF, the Central Bank of Nigeria or Stears, depending on the outletNo primary report located. Four different organisations are credited in different articles.
$10 billion a year (about ₦7.6 trillion) on fuel and maintenance for small petrol generators"Nigerians reportedly spend…", SEforALL, 2024Stated as "reportedly" in the report. Source not stated in the text I read.
₦1.6 trillion a yearNESREA, 2018, citing the Centre for Management DevelopmentSecondary citation only.
₦1.56 trillion ($3.76 billion) a yearPeopleSuN survey paper, 2021Cited in the paper's introduction to an earlier source I could not read.
₦1.43 trillion on fuel, ₦192 billion on maintenance (Lagos households)SEforALL and Lagos State Government, 2024Modelled from household counts and genset size classes.
₦1.11 trillion (2024) and ₦1.34 trillion (2025) spent by manufacturers on alternative powerManufacturers Association of Nigeria; NairametricsIndustry only, not households.

Published estimates of what Nigerians spend each year on generator fuel run from $3.8 billion to $14 billion, and the highest has no traceable source

Hover, focus or click a bar for its detail. Click to pin it and compare the others against it.

Unit: US$ billion a year. The $14 billion figure is credited to the African Development Bank, the IMF, the Central Bank of Nigeria or Stears depending on the outlet; I found no primary report. The $10 billion figure is stated as "reportedly" in the SEforALL report. Estimates are for different years and definitions and are not strictly comparable.

Source: PeopleSuN survey paper (2023), which converts ₦1.56 trillion at the 2021 rate; SEforALL and Lagos State Government (2024), "Beyond Gensets"; The Cable and Nairametrics, citing a Bloomberg report (2019)

The estimates for national annual spending run from $3.76 billion to $14 billion, a spread of 3.7 times, and the highest has no primary source I could find. The range does not come from different exchange rates alone, because the naira lost about three-quarters of its dollar value between 2018 and 2026. A fair reading is that spending is large, probably several billion dollars a year, but that no figure can be cited as measured.

Two internal checks are worth making on the Lagos study. It estimates about 19 GW of generator capacity in Lagos State against 4.2 GW of average available grid generation, a ratio of about 4.5 to 1, and about 9 billion litres of fuel a year burned by Lagos households. Whether 9 billion litres is plausible depends on national petrol consumption, which I did not retrieve; if national consumption is of the order of 15 to 25 billion litres a year, a figure of 9 billion for one state's households would be too high, and that should be checked against the regulator's published consumption data before the number is used.

The companion essay gives self-generation as 20,000 to 40,000 MW from about 22 million generators at $0.22 to $0.28 per kilowatt-hour. The capacity range is consistent with the SEforALL estimate (about 42 GW nationwide). The price is not consistent with this paper's result for a small petrol generator: at ₦1,596 a litre and 0.99 litres per kilowatt-hour, fuel alone is about ₦1,577, or $1.16 a kilowatt-hour at ₦1,354 to the dollar, four to five times that range. It is consistent with a large diesel set, which burns about a quarter of a litre per kilowatt-hour. The two figures describe different machines, and the essay should say which one it means.

8. The post's predictions

8.1 The state claim

"Nothing tangible" is too strong, and the history in the companion essay shows why. Dispatch rose from about 1,800 to 2,200 MW in 1999 to about 4,100 MW in May 2023 (as reported there, citing Ministry of Power and NERC data), roughly a doubling. Population roughly doubled over the same years, from about 120 million to about 230 million, so dispatch per person was close to flat at about 17 to 18 watts. ``chart title: Grid dispatch per person was about 16 watts in 1999 and 18 watts in 2023, up one tenth in 24 years subtitle: Average power delivered to the grid per head of population, watts. Dispatch of about 2,000 MW in 1999 and 4,100 MW in May 2023. unit: watts per person decimals: 1 source: Dispatch figures as reported in The Republic's story "Why Nigeria Has No Power", citing Ministry of Power and NERC data (not independently checked) | https://therepublic.extrafemi.com/stories/why-nigeria-has-no-power source: Population: UN World Population Prospects 2024 (123.0 million in 1999; 227.9 million in 2023) | https://population.un.org/wpp/ note: The 1999 dispatch is the midpoint of the 1,800 to 2,200 MW range the story gives. data: 1999 | 16.3 *2023 | 18.0 ``

A more accurate statement is that the public system did not keep pace with demand, and that the missing capacity was supplied privately, at a scale of about 42 GW of small generators by the Dalberg estimate, about ten times the grid's average available generation. The household ladder is how that private capacity was built.

8.2 Rooftop wind

The post predicts households will install rooftop wind turbines when panel space runs out. The evidence points the other way.

The Warwick Wind Trials, which monitored 26 building-mounted turbines from five manufacturers in Great Britain in 2007 and 2008, found an average capacity factor of 4.15%, about 230 kWh per turbine per year, with the best installation reaching 7%. A study of Lagos found mean wind speeds of 4.67 m/s at Ikeja, 5.03 m/s at Lagos Island (roof) and 3.72 m/s at the marine site, and the same study warns that wind atlases tend to overstate urban wind.

To see what that means for a household, take the central solar array, which delivers about 2,457 kWh a year in this model:

Capacity factor of the roof turbineRated capacity needed to match the array
4.15% (Warwick average)6.8 kW
7% (Warwick best)4.0 kW
20% (optimistic, not shown for any urban rooftop in the sources above)1.4 kW

To match four solar panels a rooftop turbine would need 4 to 7 kilowatts of rated capacity at the capacity factors measured on British rooftops

Hover, focus or click a bar for its detail. Click to pin it and compare the others against it.

Unit: kW of rated capacity. The 20% row is an optimistic case that the trials did not find at any urban rooftop.

Source: Warwick Wind Trials (Encraft, 2009): 4.15% average, 7% best; Extrafemi Research calculation

A household would need several kilowatts of rated rooftop turbine to equal four solar panels, in conditions where the best measured urban rooftop installation reached 7%. I did not find a Nigerian retail price for small turbines, so I make no cost comparison, but the size needed makes it very unlikely. Wind belongs on exposed sites and at scale.

8.3 Estate and mini-grid interconnection

This forecast is more plausible and the law is moving. The Electricity Act 2023 lets states create their own electricity markets and regulators; ThisDay reported in April 2026 that 15 states had assumed regulation, and a later report put the number at 16. The Nigerian Electricity Regulatory Commission issued Mini-Grid Regulations on 10 April 2026 that raise the size cap for isolated mini-grids to 5 MW per site and for interconnected ones to 10 MW, from 1 MW before. The Distributed Access through Renewable Energy Scale-up (DARES) programme, funded by the World Bank, had 391 projects under grant agreements by early 2026.

I found nothing on the legal status of a gated estate that links its own solar systems or generators and sells to neighbours. That is a gap: whether such a scheme needs a licence, whether it counts as a mini-grid, or whether it is a private arrangement outside the regulations. This is the part of the post's forecast that needs a lawyer, not an engineer.

9. Discussion

The ladder is incomplete, and its top rung points somewhere the story does not follow. Reading the evidence together gives three stages. Households begin on a ladder, because the first steps are cheap and each fixes a visible problem. They then live in a pile, because the old rungs stay: the generator keeps running when the batteries are flat, the torch stays in the drawer, and the connection costs nothing to keep. And for households that can finance the top rung the cost data point to an exit, because once a system is owned the grid is the dearer source and the pile is the dearer way to live.

Nothing here assumes the grid recovers. Every comparison uses the grid as measured: Band A at ₦209.5 for a service that surveys find delivers 6.6 to 8 hours a day. The result that solar and battery alone beats the grid with a generator behind it at any supply under about 22 hours does not depend on the grid getting better or worse; it depends on what the household would otherwise buy for the hours the grid is off.

The poorest are not on the ladder. A rung you cannot afford is not a rung. Households with no electricity, or a connection that gives a few hours a day, stay on the lowest steps. The exit is a choice open to the households that can find about ₦3 million or finance it, which is the equity cost of the whole pattern.

The private system is now bigger than the public one. On the best available estimate, generator capacity exceeds grid availability by about ten times, and the households that can afford to are replacing it with solar. Policy that treats the grid as the electricity system will miss most of the capacity people actually use, and tariff policy that raises Band A will move more of its best customers toward the exit.

10. Limitations and what would settle the open questions

  • No evidence yet of exit. The cost model says exit is rational; no source I found measures disconnections or connection cancellations, so the claim is a prediction from costs, not an observation.
  • No panel data. Order of adoption can't be established from cross-sections. The NBS General Household Survey Panel would answer it. The Rural Electrification Agency and the NBS signed an agreement in December 2025 for a nationwide energy survey; its results should be used when available.
  • Retail prices from listings and vendor sites. Many 2026 equipment prices come from commercial price guides of uneven reliability. The model is a range, not a quotation, and anyone applying it should get itemised local quotes.
  • Assumptions. Generator life, the 1.5 litres per kilowatt-hour for older units, discount rates and the central panel yield are assumptions, labelled as such, and the scenarios test them. No lifetime is assumed for the grid.
  • Fuel prices move. Petrol rose 55% in the year to May 2026 and diesel 86%, so any cost per kilowatt-hour has a short shelf life. The script should be rerun with current prices before each citation.
  • Scope. The cost model is for a household with a 5 kVA system. It says nothing about businesses, which have different load patterns and which the Manufacturers Association reports spent over ₦1 trillion a year on alternative power.

11. Conclusion

The popular story has the direction right and the ending wrong. Households do climb, but they climb into a pile before they climb out of it. Today they hold the grid, a generator, batteries and sometimes solar all at once, at a growing total cost. The generator is the expensive rung: its fuel alone costs five to eleven times the grid tariff. Solar and battery is cheaper than the generator in every scenario tested, and once a household owns it the grid is the dearer source. At the supply surveys report, a household does better without the grid than with the grid and a generator, so the rational end of the ladder is exit, for the households that can pay. Affordability is the binding constraint, not technology, and it leaves the poorest on the weakest rungs. The numbers most often quoted about the problem are weaker than they look, with annual spending estimates spanning a factor of nearly four and the largest having no traceable source. The post's forecasts split: rooftop wind is very unlikely to compete, while estate interconnection is plausible but sits in a legal grey area. The most useful next steps are data: a panel survey to see the sequence, records of connection cancellations to see the exit, and a measured national figure for what Nigerians spend on keeping the lights on.


Appendix A. Inputs and sources

InputValue usedSource
Petrol price₦1,596.25 per litre (May 2026)NBS, via ThisDay, 30 June 2026
Diesel price₦3,277.47 per litre (May 2026)same
Household kerosene price₦2,971.94 per litre (May 2026)same
Grid tariff, Band A₦209.5 per kWhLeadership (national); Premium Times (Enugu order cites ₦209 as the prior rate; the state cut it to ₦160 from August 2025)
Exchange rate₦1,354 per US dollarOfficial market, mid-November 2026, per CBN/NFEM reports in Legit.ng and Techeconomy
Generator fuel use0.625 L/kWh at rated load, 0.99 L/kWh at a quarter loadHonda EU2200i specification (3.6 L tank; 3.2 h at 1,800 W; 8.1 h at 1/4 load)
Generator price₦291,000 to ₦550,0002026 retail listings (Jumia, naijatechguide)
Generator maintenance13.4% of fuel costSEforALL and Lagos State Government (2024): ₦192 bn against ₦1.43 tn
Solar panel (550 W)₦115,000 to ₦165,000, central ₦140,000Finelib, October 2026
5 kWh lithium battery₦1.45 million to ₦1.9 millionKara.com.ng listings
5 kVA hybrid inverter₦420,000 to ₦850,000Commercial price guides, 2026
Installation₦120,000 to ₦200,000Commercial price guides, 2026
Battery life3,000 cycles to 80% capacity, about eight yearsEcoFlow DELTA 2 specification, as listed by retailers
Solar yield3.5 to 4.5 kWh per kWp per day, central 4.0Nigeria range of 3.2 to 4.8 reported in a Nigeria PV study; central is an assumption
Minimum wage₦70,000 a monthThe desk's earlier essay on household solar
Average monthly grid spending₦4,155.8NBS Residential Energy Demand-Side Survey 2024, as republished by Premium Times
Roof turbine capacity factor4.15% average, 7% bestWarwick Wind Trials, Encraft, 2009

Appendix B. Verification status

Checked against a source I read directly: the PeopleSuN lighting table, survey design and stacking statement; the DHS electricity figures (pulled from the DHS API); the Honda tank size and runtimes as published on Honda pages (the 1,800 W rated output used to turn them into litres per kilowatt-hour is from my knowledge of the model and was not in the pages I retrieved, so check it); the NBS May 2026 fuel prices (ThisDay report of NBS data); SEforALL's genset counts, capacity, the ₦1.43 trillion and ₦192 billion Lagos figures, the 22 million count and its attribution to Dalberg and A2EI; the Premium Times report of the Enugu tariff order.

Reported by a news source and not checked against the original: the Band A rate of ₦209.5 (Leadership; I found a conflicting report of ₦206.80 and could not resolve it from a regulator document); the NBS Residential Energy Demand-Side Survey figures (read as republished); Tracking SDG7 figures (86.8 million, 61%); the 12 grid collapses in 2024; Ember's import figures (Ember page and press coverage); DARES counts; the state count under the Electricity Act (15, later 16, reported inconsistently); the NERC Mini-Grid Regulations 2026 caps; the Warwick Wind Trials figures (secondary summaries); the Lagos wind speeds (a study surfaced by search whose full text I did not read).

Not verified and flagged: the Nigeria PV yield range; the 1.5 L/kWh for older generators (an assumption); generator lives and discount rates (assumptions); the PeopleSuN source for ₦1.56 trillion; the NBS fuel-purchase figure of 4.8% (it looked inconsistent in the republished article and is not used); any claim in the companion essay not listed above, including the self-generation survey it names (the 22 million and capacity figures trace to Dalberg and A2EI, 2019, and not to a 2024 survey); the 9 billion litre Lagos fuel figure.

Taken from the companion essay and not independently checked here: the 1999 and 2023 dispatch figures, the grid-collapse counts by year and the 2015 to 2025 subsidy totals.

References

Pelz, S., Chinichian, N., Jütte, C. and Blechinger, P. (2023). Electricity supply quality and use among rural and peri-urban households and small firms in Nigeria. Scientific Data (preprint: arXiv:2302.08800).

Sustainable Energy for All and Lagos State Government (2024). Beyond Gensets. SEforALL.

National Bureau of Statistics (2024). Nigeria Residential Energy Demand-Side Survey Report 2024, as reported by Premium Times, 25 October 2024.

The DHS Program (USAID). Nigeria Demographic and Health Surveys 2003 to 2024, indicator HC_ELEC_H_ELC. api.dhsprogram.com.

World Bank, IEA, IRENA, UN Statistics Division and WHO (2025). Tracking SDG7: The Energy Progress Report 2025.

Ember (2025). Africa's solar imports surge 60%, giving the first evidence of a take-off in solar in Africa.

Encraft (2009). Warwick Wind Trials Final Report.

Mills, E. Fuel-based lighting studies, Lawrence Berkeley National Laboratory (light.lbl.gov).

Premium Times (21 July 2025). Enugu State Electricity Regulatory Commission Order EERC/2025/003.

ThisDay (30 June 2026). Fuel inflation persists as diesel soars 86%, petrol climbs 55%, kerosene up 37% in one year.

The Republic (2026). Why Nigeria Has No Power: From the Marina Power House to the DisCos. therepublic.extrafemi.com/stories/why-nigeria-has-no-power.

Cite this

Obasa, O. (2026). From ladder to exit: Nigerian households, generators, solar and the grid. Extrafemi Research. https://research.extrafemi.com/read/energy-ladder