# Foundation Models for Earth Systems

Labs increasingly apply the promptable foundation-model approach to satellite, weather, and geospatial data, extending general models from text and vision into physical-world forecasting with direct public-safety stakes.

- Conviction: 40 / 100 (forming)
- Horizon: Emerging (watchlist)
- Tracking since: 2026-07-12T00:00:00.000Z
- Last updated: 2026-07-12T05:33:08.133Z
- Canonical: https://polylog.news/ai/trends/ai-geospatial-earth-systems
- Publisher: Polylog
- Affected regions: Global

## Recent evidence

- [confirms] Google Consolidates Its Geospatial Models Under Earth AI (2026-07-12): Google consolidated its geospatial models under Earth AI, grouping population, environment, and crisis-forecasting models to apply foundation models to physical-world data. A hyperscaler organizing a unified Earth foundation-model program directly advances the thesis of extending general models into physical-world forecasting with public-safety stakes.