Earthmover advanced its positioning as an infrastructure provider at the intersection of Earth observation, agriculture, and AI this week by expanding its geospatial data marketplace and highlighting core technology progress. The company added Taylor Geospatial’s “Fields of the World” global field-boundary dataset, completing an agricultural AI stack that also includes Sentinel-2 median mosaics and Google DeepMind-based AlphaEarth Foundations embeddings.
Claim 55% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The new datasets, offered in cloud-native formats such as Zarr V3, GeoParquet, and PMTiles, are designed for large-scale, analysis-ready workloads spanning smallholder plots in Ethiopia to large farms in the U.S. and Brazil. Collaborations with researchers from Arizona State University, Washington University in St. Louis, Clark University, Microsoft’s AI for Good Lab, NASA Harvest, and Wherobots RasterFlow underscore alignment with cutting-edge academic and industry efforts.
Earthmover also highlighted engineering work on Icechunk, a new array format intended to integrate seamlessly with existing Zarr and Xarray workflows. The company emphasized rigorous validation using property-based and stateful testing with Hypothesis, along with running upstream test suites to surface complex bugs and strengthen reliability for data-intensive users.
This focus on testing and interoperability points to a strategy of building trusted, open-source–oriented infrastructure rather than only application-layer tools. If widely adopted, Icechunk could deepen Earthmover’s role in scientific and large-scale data processing ecosystems, potentially increasing platform stickiness for enterprise and research users.
On the commercial front, CTO Dr. Joe Hamman engaged with energy and commodity trading firms at the Commodity Trading Summit in Switzerland, promoting Earthmover’s “AI-ready” weather data. Targeting quantitative and AI-driven trading strategies, the company is seeking high-value enterprise clients that rely on sophisticated environmental and satellite data.
Earthmover also showcased a case study with climate risk firm Kettle, which uses its platform to manage more than 100 TB of satellite, weather, and geospatial data for wildfire underwriting. The array-native architecture, built on Zarr and Icechunk, supports automatic versioning of NASA vegetation index updates and on-demand spatial queries feeding three proprietary AI models.
Collectively, these developments indicate a week focused on expanding Earthmover’s data catalog, reinforcing its technical foundation, and deepening engagement with high-value verticals such as agriculture, commodity trading, and climate risk analytics. The moves appear to strengthen its long-term positioning as a specialized data infrastructure provider for AI-driven applications.