Artificial intelligence models understand text, images and code astonishingly well. What they still struggle to grasp is the physical world.

That gap—and the growing economic value of closing it—is what Spain’s Xoople is betting on.

The Madrid‑based startup has raised $130 million in a Series B funding round, bringing total capital raised to $225 million. The investment backs an ambitious plan: building a satellite constellation designed from the ground up for artificial intelligence, not human analysts.

Unlike traditional Earth‑observation companies, Xoople is not primarily selling pictures. Its goal is to create what CEO Fabrizio Pirondini calls a new system of record for the physical world—a continuously updated data layer of surface‑level change, streamed directly into machine‑learning systems.

At the core of this strategy is EarthAI, Xoople’s platform built on Microsoft Azure and distributed through Microsoft and Esri’s geospatial ecosystems. EarthAI ingests raw satellite data—currently sourced from government missions such as the European Space Agency’s Sentinel constellation—and converts it into AI‑ready datasets optimized for change detection, prediction and autonomous decision‑making. ‍

This design choice matters. Traditional satellite workflows are shaped by how humans parse images: snapshots, analysis cycles and reports. AI systems, especially agentic and autonomous models, require something else entirely: continuity, structure and reliable ground truth.

After seven years of development, Xoople believes it is ready to commercialize that approach.

The new funding will allow the company to move beyond third‑party data and start selling information generated by its own Earth‑observation constellation. While Xoople has not disclosed how many satellites it plans to launch or their orbits, it has announced a critical industrial partner: L3Harris Technologies.

Under the agreement, the U.S. aerospace and defence group will develop optical sensors for Xoople’s satellites, sensors the company says will deliver data “two orders of magnitude better than existing monitoring systems”. For Xoople, this is less about hardware prestige and more about control: owning the entire data‑generation pipeline from photons to predictions. ‍

Investors appear convinced. The Series B was led by Nazca Capital, with participation from private‑equity firms and Spain’s government‑backed innovation fund CDTI, a reminder that geospatial intelligence now sits at the intersection of commercial opportunity and strategic policy.

The market Xoople is entering, however, is crowded and mature. Companies such as Planet, BlackSky and Airbus already operate satellites and increasingly market AI‑focused analytics. Xoople’s differentiation is not resolution, revisit rate or coverage—but data architecture.

Its bet is that in an AI‑driven economy, competitive advantage flows to whoever defines the cleanest, most trusted and most machine‑readable representation of reality.

That has implications far beyond space.

Xoople’s customers include government agencies and large enterprises looking to optimize infrastructure planning, agricultural forecasting, supply‑chain resilience, insurance risk models and urban development. In each case, the value is not in seeing what happened, but in enabling algorithms to act on what is changing—continuously and autonomously.

This framing explains why the company talks less about “imagery” and more about an “Earth Record System.” As Pirondini puts it, every computing era produces new systems of record—and the companies that define them become economic centers of gravity.

Cloud platforms did this for software. Enterprise SaaS did it for business processes. Xoople is arguing that AI requires the same foundation for the physical world.

Whether it succeeds will depend less on launch schedules than on trust. Earth‑scale data drives consequential decisions: infrastructure investment, insurance pricing, disaster response. For AI agents making or influencing those decisions, data provenance and transparency are not optional.

Xoople is positioning itself as the layer where that trust is engineered.

If it works, the next generation of AI may not just read the world—it may observe it live, through a constellation designed to speak machine from the start.

‍