Newsroom

Through Horizon Europe, EUSPA drives the uptake of space-based solutions across Europe. The ThinkingEarth project is a prime example — and it’s now inviting innovators, developers, and Earth science enthusiasts to its first Hackathon, taking place 29–30 September in Riga, Latvia, as part of the BiDS25 satellite events.

Earth Observation (EO) is entering a new era where models are no longer tied to a single data type. From optical to radar, multispectral to hyperspectral, each modality provides a distinct environmental signal but leveraging them together in a unified framework has remained a challenge. Recent advances such as DOFA (Domain-Oriented Feature Adaptation) and DOFA-CLIP are addressing these gaps, offering promising directions for large-scale, adaptive EO understanding — directions highly relevant to the goals of ThinkingEarth.

The ThinkingEarth project sat down with Duccio Piovani from the World Food Programme to get an inside look on how Earth Observation and sentinel data can impact food insecurity across the globe.

Food insecurity is becoming one of the most urgent global challenges of our time. In 2023, alone, more than 345 million people were projected to be food insecure—more than double the number just three years earlier. The causes are complex: poverty, conflict, economic shocks, and especially climate change, which disrupts food systems through droughts, floods, and extreme weather events that reduce crop yields and damage infrastructure.

In April 2025, a widespread blackout hit Spain and Portugal, disrupting essential services, public transportation, and daily life. The incident highlighted a growing vulnerability in modern energy systems: the challenge of managing high shares of renewable energy—especially solar—without sufficient real-time oversight and coordination.

As the volume, diversity, and complexity of satellite Earth observation (EO) data continue to surge, the need for intelligent models that can seamlessly integrate multimodal information has never been more pressing. At ThinkingEarth, our mission has always centered on unlocking the full potential of EO data, through adaptive, scalable, and generalisable AI solutions.

At ThinkingEarth, we harness the power of Self-Supervised Learning (SSL) and Graph Neural Networks (GNNs) to construct task-agnostic Copernicus Foundation Models and a graph-based representation of the Earth. By integrating diverse Earth Observation (EO) datasets, we aim to improve environmental monitoring, climate action, and sustainable development.

At ThinkingEarth, we are developing advanced Artificial Intelligence (AI) techniques to better understand our planet as a complex system. Using Deep Learning (DL), eXplainable AI (xAI), and physics-aware Machine Learning (ML), we enhance Earth observation (EO).

Mapping land cover is important for tracking environmental changes, managing natural resources, and making decisions about cities, farms, and climate. Traditional methods need a lot of manual work and labelled data, which can be slow and costly. ThinkingEarth's self-supervised learning (SSL) technique can offer a better way to map land cover accurately and efficiently, using large amounts of unlabelled satellite data.