The resources and timelines required to launch a geospatial imaging satellite have dropped dramatically; instead of tens of millions of dollars and years of preparation, the cost has fallen to around $100,000 to place a small satellite as a secondary payload into a low-earth orbit. Hence, companies can obtain much higher-frequency coverage (currently about daily) of specific locations using entire fleets of satellites.

Use cases include the monitoring of economic and commercial activity that can be captured using aerial coverage, such as agricultural and mineral production and shipments, construction of real estates or ships, industrial incidents such as a fire, or car, and foot traffic at locations of interest. Related sensor data is contributed by drones that are used in agriculture to monitor crops using infrared light.

Several challenges may need to be addressed before satellite image data can be reliably used in ML models. These include accounting for weather conditions and in particular, cloud cover and seasonal effects, around holidays, and the irregular coverage of specific locations that may affect the quality of the predictive signals.