Research

Grows where the grid hasn’t. Yet.

Near the grid, not on it

Most people in Africa who still live without electric light are not far from a power line. Many are close enough for the grid to arrive within years, but not soon enough to wait for it. That is where a mini-grid built to join the grid later can do the most good.

617 Mpeople living without night-time light across the 39 countries screened
79%of them, about 490 million, live within 15 km of an existing or predicted power line
300 Mare 2 to 15 km away: close enough to plan for the grid, too far for a quick connection
127 Mlive more than 15 km away, where standalone mini-grids and solar still fit best

Country pagesOpen the atlas full screen

Why it matters

Power first, grid next

Mini-grids are usually planned as if the grid will never come. When it does, the investment can be stranded and customers can end up with weaker supply than before. We think there is a better way: plan each mini-grid as a future part of the grid, with a clear handover when the line arrives.

Local solar and batteries at the far end of a long rural line can also help the grid itself, by cutting losses and holding up the voltage where it is weakest.

What a network model shows

On a typical long rural line (a model, not a real network), adding local solar with batteries at the far end:

  • cut the line's total losses by about 14%, and losses on the 33 kV feeder by three quarters
  • lifted the voltage at the end of the line from 0.78 to 0.87 per unit
  • eased the substation's evening peak from 93% to 77% of its rating
  • solar alone barely helped in the evening, when demand peaks: storage makes the difference
New: Compacts layer

What the compacts say, beside what the map shows

The Compacts tab in the atlas lines up all 35 published Mission 300 national energy compacts side by side. For each, it shows the access target, the planned split between the grid, mini-grids and standalone solar, and commitments on productive use, e-waste, planning, mini-grids, clean cooking, tariffs, private investment and reported progress. Every value links to the page it is quoted from. Nothing is scored or ranked.

To read them as text, country by country, see Mission 300 energy compacts, side by side.

Select up to five countries to compare them, next to where their people without night-time light live today. Each card also sets the compact's mini-grid target beside the people living 2 to 15 km from a line, and estimates the battery and panel waste its solar plans would create each year.

The full compacts table is a free CSV download on the open data page.

How the numbers are made comparable

  • The planned split is shown as shares, so compacts that count people, households or connections can sit side by side.
  • Access targets are shown as the percentage-point gain to 2030.
  • Where a compact counts households or connections, new access in people is an estimate: we multiply by the UN DESA average household size and treat one connection as one household.
  • Where a compact gives two figures for the same thing, we use its main target table and log the other.
  • Progress comes from the World Bank's compact progress reports, which are country-reported and use national methods.
  • Solar waste is a low-high estimate from published battery and panel sizes, weights and lifetimes (ESMAP, GIZ, IRENA).
New: Night lights

Lit, but off the map

The Night lights tab shows what satellites see after dark, with every power line we could find on a map drawn on top. Zoom in to see single towns and villages.

Of about 34,000 places lit at night in 2024 across the 39 countries, 19,900 (59%) are 2 km or more from any mapped line, and two thirds of those were dark in 2014. Blue dots mark places 2 to 10 km from a line, usually a line nobody has mapped. Pink dots mark places 10 km or more away: a minigrid, a mine or diesel town, or a long unmapped line. About 102 million people live in or within 1 km of these off-map places. Checked against public lists of mines, minigrids, diesel sites and power plants, only about 640 are explained. Where full utility maps exist, most of the places this method flagged turned out to sit on a real distribution line missing from other maps: 95% in Kenya, about 8 in 10 in Zambia and Uganda; the rest are unmapped lines or off-grid power no list covers. No site is named. A new option colors each place by its chance of an unmapped grid line. The model behind it was trained on places in Kenya, Zambia and Uganda checked against their utilities' full maps: there, 98% of places rated high and 90% rated medium were on a real line.

Download the data. Every lit place off the map, 20,208 in all, with its distance to the nearest mapped line, chance of an unmapped grid line, people nearby and an OpenStreetMap link: CSV, all countries (2.7 MB) · one CSV per country (zip, 0.5 MB) · GeoJSON (zip, 0.6 MB) · read me. Free under the Open Database License (ODbL 1.0), like the OpenStreetMap data it is built on: credit Aloe Energy and OpenStreetMap contributors, and share databases you adapt from it under the same license. The file leaves out the ECOWAS 2017 map (CC BY-SA 4.0), whose license cannot be combined with the ODbL, so about 300 West African places appear in the file but not on the map.

How it is made

  • Lights: annual VIIRS night-light composites for 2014 and 2024 (Earth Observation Group, CC BY 4.0), a place counts as lit at 1 nW/cm²/sr or more in at least two of 2022 to 2024. That level matches national access rates best: for the typical country the lit share is within 1 point of the World Bank's 2024 access rate, though single countries differ by about 12 points on average, and countries with many solar home systems sit below.
  • Lines: OpenStreetMap (November 2026), Aloe's SAPP, EAPP and Madagascar grid atlases, the World Bank's 2017 Africa grid map, and utility distribution maps for Kenya, West Africa, Uganda, Zambia, Liberia, Mauritania and Kano (energydata.info). Predicted lines are left out on purpose.
  • Gas flares from the World Bank flare list are set aside.
  • People: Google Open Buildings (323 million buildings), scaled so each country matches its 2024 population. Living near light does not mean being connected.
  • Light is not access: villages on solar home systems or with low use stay dark.
New: grid arrival

When will the grid arrive?

Plans say when the grid should reach a place; history says how late it usually is. Every rural electrification target we could check in the region was missed: Kenya's first Last Mile phase was due in 2019 and its closing date moved to 2022, Tanzania's "all villages by 2021" reached 98% of villages in 2024, Rwanda's universal access target moved from 2024 to 2029, and Zambia's rural access was 3.5% in 2009 against a planned 9.2%.

For Zambia, the Plan a connection tool now shows the nearest grid-extension site in the Rural Electrification Master Plan 2025-2030, its planned year and two likely arrival bands. Of the plan's 3,438 grid sites, 618 are likely to see the grid before 2030; if delivery runs at Zambia's past pace, 1,342 wait until after 2035.

How it is made

  • Low case: the plan year plus 3 years, the regional record (Kenya 3 years or more, Tanzania 3 to 4, Rwanda 5, transmission projects a median 5).
  • High case: delivery at about 40% of the planned pace, Zambia's own early record under its 2008 plan.
  • Bands: before 2030, 2030 to 2035, after 2035, or not in any public plan.
  • Site positions come from matching plan names to public place names: 2,287 of 3,438 sites placed with high or medium confidence. Treat them as approximate.
  • A line arriving is not homes connected, and the delay evidence is national and program-level, not site by site.
  • Sources: Sida (2011), AfDB IDEV (2022), Daily News Tanzania (2024), REG Rwanda (2024), World Bank (2025), Uganda Ministry of Finance (2018), Aloe Energy Promised vs built (2026).
Reference

Method, sources and updates

Read it with care

“Without night-time light” is a proxy, not a measured access rate. It misses homes with low use or solar home systems, so it overstates the unconnected population in places like Ethiopia, Uganda and Zimbabwe. Distances are straight lines to mapped or predicted lines. The atlas does not model how strong each network is.

Sources

Population: Google Open Buildings v3 (CC BY 4.0), scaled to World Bank 2024 population (WorldPop for Chad). Night lights: annual VIIRS VNL V2, 2022 to 2024 (Earth Observation Group, CC BY 4.0), lit in at least two of the three years. Grid: gridfinder (Arderne et al. 2020, CC BY 4.0), which combines OpenStreetMap lines with predicted medium-voltage lines. Boundaries: Natural Earth. Countries: the 36 Mission 300 compact countries as of mid-2026 and the 12 Southern African Power Pool members.

The position paper “Power first, grid next” (November 2026) builds on this atlas. Note on Zambia: the paper counts far fewer people 2 to 15 km from the grid than this atlas does, because it uses a stricter Zambia-only method (mapped medium-voltage transformers from public utility data and a threshold for what counts as lit), while this atlas uses one method for all 39 countries. Both are kept, and each number says which method it comes from.

What changed 12 updates

Every month we check for new Mission 300 energy compacts, new releases of the grid, population and night-light data behind the map, and update it. Each change keeps its source.

  • 16 November 2026 - Compacts tab: a new "Also holds a UN Energy Compact" view marks the eight countries that the UN-Energy Energy Compacts Annual Progress Report 2026 lists as holding both a UN Energy Compact and a Mission 300 compact. The two are separate; UN Energy Compact figures are not used in the comparison.
  • 16 November 2026 - The "Constraint screens" layer now covers all of Malawi (8,465 candidate villages) as well as the Southern Province pilot in Zambia.
  • 13 November 2026 - Night lights tab: a new "Constraint screens" layer shows 106 candidate mini-grid villages in two pilot areas (Southern Province, Zambia and the Lower Shire, Malawi), each colored by the most serious flag from our desk constraint screen, with the IFC and World Bank standards the flags point to.
  • 12 November 2026 - Grid arrival: for Zambia, the Plan a connection tool now shows the nearest site in the 2025-2030 electrification plan and when the grid is likely to arrive, adjusted for how late rural electrification has run in the region.
  • 12 November 2026 - The lit-but-off-the-map data is now a free download (CSV and GeoJSON, ODbL 1.0): 20,208 places with chance bands and OpenStreetMap links.
  • 11 November 2026 - Headline numbers updated with 2024 night lights: 617 million people without night-time light (was 701 M), 300 million at 2 to 15 km (was 326 M).
  • 11 November 2026 - Night lights: every off-map place now shows its chance of an unmapped grid line (high, medium or low), from a model tested against utility maps in Kenya, Zambia and Uganda.
  • 10 November 2026 - Night lights: Kenya Power's full network added (11 kV to 220 kV, energydata.info). 95% of the Kenyan places the layer had flagged sit on one of these lines. Off-map places across the 39 countries: 19,900 (59%).
  • 9 November 2026 - Night lights: method improved. Lights now count at 1 nW lit in two of the last three years, and people come from Google Open Buildings. For the typical country the lit share now comes within about 1 point of its national access rate. Utility distribution maps for West Africa, Uganda, Zambia, Liberia, Mauritania and Kano added: 21,400 lit places sit 2 km or more from any mapped line.
  • 9 November 2026 - Night lights tab added: satellite night lights with every mapped power line, and the 12,400 places lit in 2024 that sit 2 km or more from any mapped line, with those newly lit since 2014.
  • 8 November 2026 - Compacts layer added: all 35 published Mission 300 energy compacts side by side, with the page each value is quoted from, mini-grid targets beside the map and a solar waste estimate.
  • 6 November 2026 - First edition published: 39 countries, the 36 Mission 300 compact countries and all 12 SAPP members. Monthly checks start on 8 December.

Common questions

How many people in Africa live without electricity near a power line?

Across the 39 countries screened, about 617 million people live without night-time light. About 79% of them, some 490 million, live within 15 km of an existing or predicted power line.

What is a grid-ready mini-grid?

A mini-grid built from the start to join the national grid later, with equipment and a handover plan that let it connect when the line arrives. It brings power first and avoids stranded investment when the grid comes.

Which countries are covered?

39 African countries: the 36 Mission 300 energy compact countries as of mid-2026 and the 12 Southern African Power Pool members.

How accurate is it?

It is a screening atlas. "Without night-time light" is a proxy, not a measured access rate, and it overstates the unconnected population where solar home systems are common, such as Ethiopia, Uganda and Zimbabwe. For the typical country it comes within about 1 point of the World Bank's 2024 access rate, though single countries differ by about 12 points on average. Distances are straight lines.

Where does the data come from?

Population from Google Open Buildings v3 (CC BY 4.0), scaled to World Bank 2024 population (WorldPop for Chad); night lights from annual VIIRS VNL V2, 2022 to 2024 (Earth Observation Group, CC BY 4.0); grid lines from gridfinder (Arderne et al. 2020, CC BY 4.0); Natural Earth boundaries. Utility line maps from energydata.info are credited on the map, including the ECOWAS 2017 distribution grid (CC BY-SA 4.0).