Roughly 770 million people on Earth still live without reliable access to electricity. The communities affected are overwhelmingly in Sub-Saharan Africa and South Asia, often in rural areas far from existing power grids. For these households, the absence of electricity shapes everything: whether children can study after dark, whether clinics can refrigerate vaccines, whether small businesses can operate beyond daylight hours.
The challenge is not a shortage of solutions. Solar technology has become remarkably affordable. The challenge is knowing exactly where to put it. And that is the problem Atlas AI was built to solve.
Seeing Poverty from Orbit
Atlas AI, founded in 2018 by researchers from Stanford University, uses satellite imagery and geospatial data combined with machine learning to predict economic conditions and energy access at high resolution. Their core insight is that you can learn a great deal about a community's wealth by looking at it from above. The brightness of nighttime lights, the density and material of rooftops, the condition of roads, the patterns of agricultural land use — all of these correlate strongly with household income and access to basic services.
Traditional methods of measuring poverty rely on household surveys, which are expensive, slow, and often years out of date. Atlas AI's models can predict household wealth and consumption at the village level from satellite data, producing estimates that are updated continuously and cover areas where no survey team has ever set foot. The models are trained on ground-truth data from Demographic and Health Surveys and Living Standards Measurement Studies, learning to generalize across regions and countries.
From Maps to Solar Microgrids
Understanding where energy poverty exists is only the first step. Atlas AI's platform helps solar companies and development organizations identify where off-grid solar and solar microgrids will have the greatest impact. A solar microgrid is a small, localized power system that operates independently from the national grid. For communities that may never receive a grid connection, microgrids are often the fastest path to electrification.
But installing a microgrid in the wrong location can mean years of underutilization and wasted capital. Atlas AI's platform layers together population density, economic activity, solar resource potential, distance from existing grid infrastructure, and local demand patterns to produce site-level recommendations. Analysis that used to require months of field assessments now takes days. A solar company can screen thousands of potential sites and narrow to the most promising candidates before sending a single team into the field.
Partners on the Ground
Atlas AI's work has attracted partnerships with some of the largest organizations in international development. USAID has used the platform to inform electrification planning in East Africa. The World Bank has drawn on the data to guide investments under its Lighting Global program. Major solar companies operating across Africa and South Asia use the platform to plan distribution networks and identify underserved markets.
In Nigeria, where roughly 90 million people lack electricity despite the country being Africa's largest economy, Atlas AI mapped energy access at a granular level across the rural north, identifying communities where solar microgrids could serve thousands of people who had been invisible to previous planning. In India, where the government declared 100 percent village electrification in 2018, the models revealed that actual household-level access remained spotty — many families still relied on kerosene because grid connections were unreliable or too expensive.
The Technology Behind the Maps
The models draw on multiple streams of satellite data. Daytime imagery from Sentinel-2 provides information about land use and building density. Nighttime light data from the VIIRS satellite reveals where electricity is and is not being used. Radar data penetrates cloud cover, which is critical in tropical regions. The machine learning architecture combines convolutional neural networks for image analysis with gradient-boosted models that integrate non-image features like distance to roads, elevation, and climate data.
A key technical challenge is resolution. Satellite pixels cover areas of 10 to 30 meters, but economic conditions can vary within a single village. Atlas AI uses super-resolution techniques to produce estimates at finer granularity than raw satellite data would normally support, distinguishing between a prosperous market town and an underserved settlement just kilometers away.
Aligning with SDG 7
The United Nations Sustainable Development Goal 7 calls for universal access to affordable, reliable, and modern energy by 2030. At the current pace, that target will not be met. The International Energy Agency estimates that 660 million people will still lack electricity in 2030 under current policies.
Atlas AI's contribution to SDG 7 is not the solar panels themselves, but the intelligence layer that makes every dollar go further. When a development bank allocates funding for rural electrification, well-targeted projects can mean the difference between thousands of households connected or left in the dark. The platform also enables post-installation monitoring by tracking changes in nighttime light intensity, assessing whether projects are actually expanding energy access as intended.
What Comes Next
Atlas AI is expanding beyond energy to cover agriculture, financial inclusion, and climate resilience. The same satellite-based approach that maps energy poverty can identify crop stress, predict food insecurity, and estimate economic impacts of climate events. But energy remains the foundation. Without electricity, almost every other development intervention is harder to deliver and sustain.
The scale of the energy access gap is still enormous. But for the first time, the organizations working to close it have a continuously updated, high-resolution picture of where the need is greatest. That picture does not build solar panels or string wires. What it does is make sure that when those panels are built, they end up in the places where they will change the most lives.
Sources: Atlas AI platform documentation, Stanford Sustainability and Artificial Intelligence Lab, International Energy Agency World Energy Outlook (2025), World Bank Lighting Global program, USAID Power Africa reports, Nature Energy (2024), UN Sustainable Development Goals Progress Report (2025).
