Geospatial Foundation Models Workshop

Geospatial Foundation Models (GFMs) are rapidly reshaping how we extract meaning from Earth observation, maps, trajectories, and multimodal geospatial data. The space is moving fast—often in parallel across academia, startups, and industry—which makes collaboration unusually valuable right now.

CARTO and the BSC AIF are hosting a one-day workshop in Barcelona to gather a small group of people actively building or researching foundation models for geospatial problems. The goal is to share what’s real, what’s working, and what’s missing and then create space for candid discussion and potential collaboration.

What to expect

  • ​Lightning talks + short technical talks from invited participants
  • ​Discussion-oriented sessions (less “conference”, more “working room”)
  • ​Time for collaboration: shared challenges, benchmarks, datasets, eval, deployment patterns
  • ​A community-first vibe: pragmatic, technical, and honest

Who this is for

  • ​Research groups working on GFMs EO foundation models geospatial VLMs
  • ​Startups building foundational geospatial AI (EO, mobility, mapping, risk, climate, etc.)
  • ​Applied teams at labs or companies pushing the frontier (modeling, eval, infra, products)

Event details

Date: February 04, 2026

Time: 9:30h to 17:30h (Madrid, CTE),

Location: Hybrid | Via Laietana, 26, Ciutat Vella, 08003 Barcelona

Language: English

Participation: this is a curated / capacity-limited workshop to keep discussion high-quality.

Agenda

TimeSessionSpeaker
09:30 – 10:00Check-in & Breakfast
10:00 – 10:10WelcomeJavier Pérez, CARTO
10:10 – 10:30BSC AI FactoryAlbert Cañigueral, BSC
10:30 – 10:50Invited Talk – Geospatial Foundation Models: Opportunities and Open ChallengesLucía García-Duarte & Miguel Álvarez, CARTO
10:50 – 11:20Invited Talk – Inference-Centric vs Embedding-Centric GeoAI: Two Visions for Earth IntelligenceBruno Sanchez-Andrade, LGND AI
11:20 – 11:45Coffee Break & Discussion
11:45 – 12:45Lightning Talks, Part 1Review of embedding models based on vector geo data – Kamil Raczycki, Dataplace.ai
Validating LLM spatial reasoning with deterministic geometry – Sergio Contreras, Arup
Trustworthy and Interoperable Spatial Knowledge Graphs as Infrastructure for Foundation Models Using Emerging Metadata Standards – Nathan McEachen, TerraFrame Inc.
Integrating AI inference into Sentinel Cloud‑Based Pipelines – Marcin Kluczek, CloudFerro
12:45 – 13:30Discussion Sessions
13:30 – 14:30Lunch
14:30 – 15:00Invited Talk – Lifestyle Embeddings as Geo-Foundational Models of Human ActivityEsteban Moro, Northeastern University
15:00 – 16:15Lightning Talks, Part 2Planaura: Geospatial Foundation Model for Bi-Temporal Land-Cover Change Detection – Mikhail Sokolov, Natural Resources Canada
Crop Classification with Crop-Calendar-Aligned TESSERA Embeddings – Yapkan Choi, NEO
Geospatial foundation models for predicting population dynamics – Francisco Rowe, University of Liverpool
Climate Responsive Urbanism – Lakshmi Narayanan, IAAC
New approach to Air Quality maps – Endika Aguirre, Tracasa Global
16:15 – 16:45Discussion Sessions
16:45 – 17:15Invited Talk – Population Dynamics Foundation Model EmbeddingsJoydeep Paul, Google Research
17:15 – 17:30Final Remarks

🔗 Registration

In case you would like to join online, please join the session following link here .