Xoople Raises $130M to Build a Satellite Constellation Feeding AI Training Data

Spanish satellite startup Xoople has closed a $130 million Series B funding round, bringing its total raised to $225 million. The company is building a constellation of spacecraft designed specifically to supply AI training and inference pipelines with continuous, high-resolution Earth observation data — a vertical that didn't meaningfully exist five years ago.

The round included a hardware announcement: a manufacturing partnership with L3Harris, the US aerospace and defense contractor, for the satellite sensor payloads. L3Harris brings aerospace-grade imaging optics and defense supply chain rigor to a startup that's positioning itself at the intersection of Earth observation and AI infrastructure.

Xoople's thesis is that the AI industry's hunger for training data will extend far beyond scraped text and synthetic generation — that real-world, continuously-updated spatial data from orbit will become a premium input for foundation models applied to logistics, agriculture, climate, urban planning, and defense. The company is selling data subscription access to AI labs, government agencies, and enterprise customers.

The funding round signals a broader trend: AI infrastructure investment is now reaching the hardware and orbital level. It's no longer enough to build data centers and GPU clusters — the supply chain for unique, non-replicable training data is becoming a capital-intensive moat. Xoople's bet is that satellite-sourced Earth data is one of the few data categories that can't be synthetically generated or scraped from the web.

The L3Harris partnership also opens a defense market path that pure software AI companies can't easily access, giving Xoople a customer base with long contract cycles and large budgets — a structural advantage as the commercial satellite market faces commoditization pressure from SpaceX and others.