Uganda has a bold ambition: connect 10 million households to the most suitable energy source by 2030. That target, enshrined in the National Electrification Strategy (NES), is not just a policy goal; it is a lifeline for millions of Ugandans for whom a reliable power supply remains out of reach. Today, only about 27% of Uganda’s population has access to electricity, and in rural areas, that figure drops to around 10%.
The Ministry of Energy and Mineral Development is at the forefront of closing this gap. To do it at scale, across hundreds of districts and thousands of villages, requires something the country has not had before: precise, actionable, data-driven intelligence about where people live, how their communities are structured, and which energy solution best fits each settlement.
That is exactly what Sunbird AI, working in close partnership with the Ministry of Energy and Mineral Development, set out to build, powered by Google Earth AI.
The work started in Lamwo District, in Northern Uganda. Not as a contained pilot to be filed away, but as a living case study designed to teach us how to do this well, efficiently, and in a way that can be rolled out across every district in Uganda. Lamwo was chosen deliberately: it is a district of genuine complexity, diverse communities, refugee populations from South Sudan, a mix of terrain, and an existing but partial solar mini-grid presence, making it the ideal environment to stress-test an approach before scaling it nationally.
For electrification to happen, government planners need to know: How large is a given settlement? What types of buildings does it contain? Are there schools, health centres, or commercial buildings with higher energy demands? How dispersed are the structures — and is a mini-grid even physically viable?
Collecting this data through field surveys is expensive, slow, and in many areas, near impossible. By the time a survey is complete, the information is already outdated. For example, Roads in Lamwo require four-wheel-drive vehicles and become impassable during the rainy season.
The answer, it turns out, is visible from space.
Sunbird AI’s approach to solving this problem is built on Google Earth AI, AI-generated building footprints derived from satellite imagery across Africa. For every settlement in Lamwo, the system can now identify building locations, their shapes, their sizes, and their spatial density.
From these building footprints, our machine learning algorithms classify structures: permanent versus non-permanent buildings, commercial versus residential, facilities like schools and health centres. Settlement layout, density, and building composition together paint a detailed picture of electricity demand and of whether a solar mini-grid, grid extension, biogas energy, or individual solar home systems would be most appropriate for a given village.
The Workflow: From Pixel to Plan
The end-to-end AI workflow operates in several stages:
1. Settlement Detection and Mapping Using the Google Open Buildings dataset, the system automatically maps every detectable structure across a district. Building footprints are extracted, georeferenced, and overlaid on administrative boundaries at the village level.
2. Building Classification Machine learning algorithms classify buildings by type and permanence. The size, shape, and arrangement of structures provide strong signals about a settlement’s energy demand profile. A cluster of large permanent buildings near a road, for instance, suggests a trading centre with different needs than a dispersed residential settlement deeper in the village.
3. Electrification Recommendation Engine The system cross-references building data with supply-side variables, solar irradiance, wind potential, proximity to the existing grid, vegetation cover for biomass assessment, and generates ranked electrification recommendations for each village: grid extension, solar mini-grid, or solar home systems.
4. Conversational Interface An AI-powered assistant makes the platform accessible to a broad range of users. Ministry planners can query the system conversationally, request comparative analyses across villages, and receive plain-language summaries of technical recommendations. This dramatically lowers the barrier to using sophisticated geospatial intelligence in day-to-day planning decisions.
What Change Looks Like on the Ground
Behind every data point is a community. In Lamwo, the arrival of solar mini-grids — guided in part by AI-assisted planning- has already begun to change daily life.
Nancy Amito, a restaurant owner in Palabekogilli Trading Centre, put it plainly: “I’m able to continue serving my customers till as late as 10 pm. This extension has doubled my daily income.”
For Denis Ongola, a local electrician who has helped install mini-grids across 25 villages in Lamwo, the change is just as visible: “Unlike before, community members are now running their businesses 24/7 — millet and maize grinding mills, salons, welding shops, retail stores.”
Evening study hours have started for children. Clinics feel safer. Mobile money agents are operating. These are not abstract development indicators; they are the tangible result of better planning decisions made possible by better data.
This is what happens when the right technology reaches the right hands at the right moment.
The Road Ahead: Lamwo as Blueprint for Uganda
Lamwo was always meant to be a beginning, not an endpoint.
The architecture of the platform, built on Google Earth AI, designed for conversational access and validated by the Ministry, is designed to scale. The same workflow that mapped Lamwo’s settlements can be applied to any district in Uganda.
Uganda has 146 districts. The National Electrification Strategy demands answers for all of them. The Lamwo case study has shown that those answers, precise, data-driven, and accessible, can be generated at a fraction of the time and cost of traditional approaches.
Sunbird AI, in partnership with the Ministry of Energy and Mineral Development, is now working to expand this system nationally. The goal is not just to electrify Uganda faster, but to do it smarter, ensuring that every investment in rural energy infrastructure is guided by the best available intelligence, and that the benefits reach those who need them most.
Written by Nimpamya Janat Namara, Communications and Engagement Lead at Sunbird AI.