Earthmover Raises $7.2M to Power Physical AI With Scientific Data Infrastructure

• Earthmover has raised $7.2 million in seed funding to scale its cloud-native platform for AI-ready Earth system data.
• The company enables faster model training and simulation development across climate, weather, and other scientific domains.

Earthmover, a New York-based public benefit corporation founded by climate scientists and open-source developers, has secured $7.2 million in seed funding led by Lowercarbon Capital, with participation from Costanoa Ventures and Preston-Werner Ventures. The funding will support its mission to modernize infrastructure for Earth science and other physical-world data. The platform helps organizations train AI models more efficiently and deliver results faster across a range of domains.

The platform is built around multidimensional tensor data and includes three core products, including Icechunk, a cloud-native array storage format; Arraylake, a data management and governance layer; and Flux, a high-speed geospatial API gateway. These tools help teams reduce compute costs, accelerate data retrieval, and speed up model deployment when working with massive climate, weather, and geospatial datasets.

Earthmover plans to expand into additional science and sensor-rich fields such as bioimaging, neuroscience, and robotics. With deep roots in the open-source scientific Python ecosystem, the company is positioning itself as the infrastructure layer for organizations building physical AI from Earth-scale data.

🌀 Tom’s Take:

As physical AI systems scale, they need stronger infrastructure to manage the data behind simulations and digital twins. Earthmover is building that foundation.

Source: Business Wire / Earthmover