Earthmover Raises $7.2 Million Seed Round to Transform Earth Science Data Management

Earthmover, a cloud platform for scientific data to solve humanity’s greatest challenges, led by Ryan Abernathey, Joe Hamman, Margaret Francis, and others, has secured $7.2 million seed round led by Lowercarbon Capital, with participation from Costanoa Ventures (pre-seed investor in Earthmover) and Preston-Werner Ventures (GitHub co-founder Tom Preston-Werner).

Earthmover was founded to tackle a major challenge: critical teams addressing climate change, healthcare, and other global problems were slowed by inefficient data management. Scientific datasets in obscure formats weren’t designed for the cloud or AI, forcing teams to build bespoke systems or overload platforms like Snowflake and Databricks—leading to high costs and slow performance. Leveraging experience from building open-source tools like Xarray, Earthmover provides a solution to this bottleneck.

Since its founding three years ago, Earthmover’s technology has proven to be a perfect fit for the AI-driven transformation in weather, climate, Earth observation, and computational science. Customers include RWE AI Research Laboratory, which builds customized weather models for renewable energy forecasting; Kettle, developing AI-powered wildfire predictions for insurance underwriting; and Sylvera, using AI and remote sensing to deliver faster, more accurate carbon credit ratings. By leveraging Earthmover, these teams can keep GPUs running efficiently and data scientists focused on innovation instead of data management.

Earthmover’s platform is a foundational data infrastructure, designed to meet the needs of both startups and enterprise customers. Open source is central, with Icechunk, a high-performance cloud-native tensor storage engine, at its core. Icechunk marks a significant pivot from the original design and reinforces the commitment to open science. The platform is now onboarding new startups weekly while successfully navigating enterprise and government procurement, with implementations at NASA, NOAA, the UK National Oceanography Center, and MeteoSwiss. Looking ahead, Earthmover plans to expand access to analysis-ready datasets from trusted partners, offer reactive array views for user-defined computations, and integrate SQL to bridge tensor and tabular data workflows.