Facebook has created a map of Earth that can pinpoint any man-made structures to a resolution of 5 meters, to help company's Internet provision efforts
Context & Ripple Effects
This 2017 report is the seed of what became one of Facebook's most durable infrastructure programs: a global map of man-made structures at 5-meter resolution, built so the company can decide where its Internet provision efforts are worth deploying. Rather than a one-off dataset, the coverage shows it compounding — by 2019 the effort had matured into Map With AI, which uses commercial satellite imagery to map previously unmapped places across the developing world.
The same pipeline then broadened beyond connectivity targets: AI-derived population-density maps covering much of Africa followed in April 2019, and weeks later Facebook opened population, movement, and network-coverage maps to nonprofits and universities tracking infectious-disease spread. What started as an internal siting tool has become a public-facing geodata operation.
First-order effects
- Facebook gains a precise, self-built layer for deciding where to spend on Internet provision — every new structure detected sharpens its targeting of connectivity investment toward underserved areas.
Second-order effects
- Governments, nonprofits, and universities become downstream consumers of Facebook's mapping output rather than official or academic cartography, giving the company influence over how development and public-health resources get allocated.
Third-order effects
- If the pattern holds, platform companies evolve into de facto cartographers of the developing world — private datasets standing in where state mapping capacity is thin, with the governance questions that implies left largely unaddressed.
The trend: Connectivity-driven tech giants are turning satellite imagery and AI into proprietary global basemaps that double as public goods and strategic infrastructure-siting assets.