Building tools for a more connected planet.

Intelligence grounded in the living world.

For us, planetary intelligence means connecting people, technology, and nature to better understand our planet and respond together to its challenges. It is not just about a larger AI model. It is about a better relationship between what we observe, what we understand, and what we do next.

Satellite imagery offers one perspective. Field measurements, environmental sensors, scientific models, and the knowledge of people who live and work in a place offer others. Bringing these perspectives together can help us ask more useful questions: What is changing? What might happen next? Where do we need more evidence?

The ambition is planetary; the work has to be specific. A useful tool should help someone understand a particular landscape, investigate a change, or weigh a decision—not just produce another map.

Practical questions. Testable tools.

Our focus is the applied work between promising research and something people can actually use. That means starting with a real question, building a focused tool, and testing whether it helps.

1. Start with a decision.

   Work backward from what someone needs to understand or do. Let the problem guide the choice of data and methods, rather than the other way around.

2. Build on the research.

   Explore how Earth-observation models, geospatial data, and existing scientific tools can be connected and adapted to useful workflows. A new model is not always the missing piece.

3. Check against reality.

   Compare outputs with observations and domain knowledge. Make data gaps, uncertainty, and limitations visible, and keep human judgment central.

Questions about ecosystem change, water, land use, and climate resilience are examples of where this approach could matter. These are directions for exploration, not claims that we already have all the answers.

Part of a wider conversation.

We’re inspired by a growing community connecting people, technology, and the living world. These ideas and projects help frame the opportunity, from long-term thinking and a broad vision of planetary intelligence to practical representations of Earth-observation data.

• The vision / Planet

  What Is Planetary Intelligence and How Could It Move AI From the Internet to the Real World?

  Will Marshall, Robbie Schingler, and Andrew Zolli explore how Earth sensing and AI could move beyond describing the world toward understanding change and anticipating its consequences.

• Long-term thinking / Long Now

  The Long Now Foundation

  A foundation fostering long-term thinking and responsibility. Its perspective encourages us to build tools with future generations—and the longer life of the planet—in mind.

• People, nature & AI / Long Now Labs

  Interspecies Protocols

  A lab in development exploring protocols for interactions between humans, nature, and synthetic intelligences. It raises questions about how these different constituencies can shape a shared world.

• The vision / Ryan Orbuch

  Natural General Intelligence

  A case for a nature model grounded in the state and dynamics of the planet itself—and for connecting observation, prediction, human judgment, and evidence about what happens next.

• Open AI research / Marin

  Marin

  A community building AI in the open, sharing code, model weights, training data sources, and experimental results. Marin offers a model for making research transparent and reusable.

• Open web data / Common Crawl

  Common Crawl

  A nonprofit providing a free, open repository of web crawl data. Common Crawl illustrates how shared data infrastructure can support research and make large-scale analysis accessible to more people.

• Earth representations / GeoTessera

  Tessera

  Earth-observation representations learned from Sentinel-1 and Sentinel-2 data over time. Tessera illustrates how satellite time series can become reusable features for understanding landscapes.

• Earth foundation models / Clay

  Clay Foundation Model

  An open-source model that turns satellite imagery, location, and time into representations that can support tasks such as land-cover classification and change detection.

• Earth foundation models / IBM, NASA & Jülich

  Prithvi-EO 2.0

  A multi-temporal Earth-observation foundation model. Another example of the broader effort to build reusable models that can be adapted to specific environmental questions.

• Earth system forecasting / Microsoft

  Aurora

  A foundation model for Earth system forecasting, with versions for weather, air pollution, and ocean waves. Aurora shows how shared models can be adapted to different environmental forecasting tasks.

These are sources of inspiration, not a list of affiliations or endorsements. Earth-observation foundation models are important building blocks; they are not, on their own, a complete understanding of the planet.