Geospatial Tools Power Off-Grid Solar in Sub-Saharan Africa

Google has spent the past three years helping Ignite Energy Access expand off-grid solar access in sub-Saharan Africa through its Startups for Sustainable Development program, combining Google's geospatial, operations research, and machine-learning tools with Ignite's field deployment. The collaboration targets a gap that leaves nearly half of sub-Saharan Africa's population, or more than 600 million people, without electricity, forcing many households to rely on kerosene and candles. Ignite aims to connect 100 million people across Africa to electricity by 2030. The program focuses on three barriers: identifying remote homes without formal addresses, delivering solar equipment across unmapped rural terrain, and verifying installed micro-solar systems from afar. To locate potential customers, researchers merged satellite imagery with Google's Open Buildings dataset and routing algorithms, while developing geospatial models to represent unmapped rural paths. For last-mile logistics, Google's Operations Research models optimize Ignite's supply chain by determining warehouse and mini-hub locations and balancing transportation costs, delivery routes, vehicle capacity, and inventory expenses. The initiative also addresses the challenge of tracking widespread rooftop micro-solar installations, which appear as only about one pixel in standard satellite imagery. Google researchers are building computer vision models based on Gemini for Science and training them with ground-truth drone imagery from Ignite to identify solar setups from space. By testing numerous model variations to improve precision and recall, the teams aim to create a scalable mapping tool that helps operators find underserved villages, allocate resources efficiently, and plan future electrification rollouts. The work forms a repeatable blueprint for using AI and data to extend clean energy access to dispersed households across Africa.