Every day, Earth Observation satellites stream down petabytes of imagery: more data than any team of experts could ever open, let alone analyze. The bottleneck is no longer capturing the data, it is making sense of it fast enough to matter. Volume, velocity, and veracity have turned space imagery into a big data problem that outpaces human capacity, and this is where agentic AI changes the game: instead of passive pipelines waiting for an analyst, agentic systems reason over the data, prioritize what matters, and turn raw pixels into decisions on their own.

In this talk, Mohamed Boukhebouze shares how his team at EarthLab Luxembourg builds agentic GeoAI platforms that meet this challenge head on, and why the same principles will power the next generation of space exploration. Because when the volume is overwhelming and ground control is too far away, the machines will have to decide for themselves. The question is no longer whether we can teach them to see, but whether we can trust them to act.