Earthmover, a company at the forefront of transforming how we handle climate, weather, and other complex scientific data. Founded by a team of experienced climate scientists and pioneers in open-source software, Earthmover has secured $7.2 million in seed funding. The seed funding round was led by Lowercarbon Capital, with additional investments from Costanoa Ventures and Preston-Werner Ventures, signaling strong confidence in Earthmover’s vision and technology.
This significant investment will propel their mission to accelerate data exploration, advance the development of artificial intelligence and machine learning models, and streamline the delivery of data products essential for comprehending, modeling, and forecasting events in the physical world around us. This funding announcement comes on the heels of several key achievements, including the recent public release of their innovative open-source data format called Icechunk 1.0, the launch of their geospatial API gateway product known as Flux, and an ongoing partnership with NASA that has already yielded a remarkable one hundredfold improvement in the retrieval of Earth observation data.
The urgent reality of climate change is bringing about accelerated shifts, leading to increased instability and unpredictability across vital sectors, including energy, commodities, agriculture, insurance, and real estate, among many others. A recent report from Gallagher Re highlighted the severe financial impact of these changes, detailing $402 billion in damages from twenty-seven extreme weather events in 2024 alone. This figure represents a twenty percent increase over the ten-year inflation-adjusted average, underscoring the growing crisis. In response to these escalating challenges, a diverse array of organizations, ranging from inventive AI startups to major energy traders and expansive government agencies, are increasingly turning to artificial intelligence to enhance their capabilities.
The company’s goal is to model the physical world in real time, enabling more informed decision-making by leveraging vast quantities of high-velocity Earth-system data. However, these ambitious efforts are often hindered by outdated technologies and inefficient data management systems. Earthmover directly addresses this critical need by making physical-world data readily available and optimized for AI in the cloud, empowering teams to effectively train and deploy advanced models on petabyte-scale datasets encompassing weather, climate, and geoscience information.
The Earthmover platform is thoughtfully constructed around three primary components. First is Icechunk, an optimized array storage format designed for cloud environments. Second is Arraylake, a dedicated data management layer tailored for teams. Finally, Flux provides a suite of high-performance API access methods. When a customer’s complex multidimensional array data is ingested, it is stored in the Icechunk format within cloud object storage, ensuring compatibility across all major cloud platforms.
Inspired by the widely adopted Apache Iceberg table format, Icechunk is an open-source, cloud-native transactional storage engine specifically engineered for multidimensional array data, making it ideal for high-performance analytics and demanding AI workloads.
Arraylake then offers a robust, version-controlled data catalog, management tools, and a comprehensive data governance layer for analytics and AI teams. Earthmover’s newest product, Flux, delivers highly performant geospatial query APIs, significantly accelerating the delivery of critical data to downstream applications. Collectively, these integrated features drastically reduce the time teams need to establish, execute, and deliver their own machine learning or AI models, transforming a process that once took months into mere days. This efficiency is achieved while meticulously preserving data provenance and simultaneously reducing costly computational expenses.
The founders of Earthmover are recognized leaders within the open-source software and data ecosystem, having made substantial contributions to the Python stack used for scientific data analysis. Before establishing Earthmover, Abernathey served as a distinguished professor at Columbia University, where his research focused on vital areas, including ocean circulation and climate. Meanwhile, Hamman contributed his scientific expertise at the National Center for Atmospheric Research before co-founding CarbonPlan, a prominent data-driven climate think tank. The founders’ paths converged through their collaborative work on Xarray, a fundamental open-source software package crucial for analyzing weather, climate, and geospatial data.
This foundational tool has been downloaded over six million times and is actively employed by renowned organizations, including NASA, ESA, Google, NVIDIA, and Planet. The Earthmover team also stands as the primary contributor to Zarr, another impactful open-source data format designed for compressed, chunked array storage. They were instrumental in launching the open-source project Icechunk, a cloud-native transactional storage engine specifically designed for tensor data.
Earthmover’s aspirations extend considerably beyond the confines of the geospatial category. The extensive experience of the team within the Scientific Python community has illuminated a crucial insight: the same core data structures, known as tensors, appear consistently across virtually all scientific and computational disciplines. For example, the Zarr format is widely adopted in fields as diverse as bioinformatics, microscopy, neuroscience, and even fusion research. As autonomous multi-modal sensor platforms, ranging from self-driving vehicles to advanced ocean-going robots, continue to gather an ever-increasing volume of data about the physical world, the demand for scalable, cloud-native tensor data infrastructure is only poised to grow exponentially.
KEY QUOTES:
“Earthmover’s open-source leadership and deep domain expertise is the recipe for success in data infrastructure companies.”
Tom Preston-Werner, Earthmover investor and co-founder of GitHub
“The lack of purpose-built cloud data infrastructure for physical-world AI acts as a drag on our entire space, limiting the value of Earth System data and ultimately slowing climate change adaptation and mitigation efforts. We are helping people get answers from data in seconds rather than days.”
CEO and co-founder Ryan Abernathey
“Where other platforms see ‘unstructured data’, Earthmover brings a native understanding of the scientific data formats, data structures, and modeling techniques common in weather, climate, and geospatial data analytics. This unlocks incredible performance as well as cost savings for our customers.”
CTO and co-founder Joe Hamman
“We evaluated multiple datalake architectures, but none of the tabular solutions could meet the demands of our AI weather models. Earthmover’s array-native approach was the only one that scaled with our data and aligned with how we actually do science.”
Galen Yacalis, Lead Scientist at the RWE AI Research Laboratory
“Industry leaders and CTOs want cloud agnostic, maximally flexible infrastructure data platform choices, especially given the speed at which AI is forcing them to operate.”
Tony Liu, Partner at Costanoa Ventures.
“Ryan and Joe are the perfect founders to deliver on a next generation platform for n-dimensional array data, as evidenced by the support Earthmover has received from the Chan Zuckerberg Initiative for their work on open source tools critical to science, including potential applications in bioimaging, radar, and AI models of the physical world.”
Tony Liu, Partner at Costanoa Ventures
“Earthmover provides the fastest, cheapest, and most advanced cloud-native infrastructure to manage climate and weather data — an increasingly essential asset for critical industries building physical AI.”
Shawn Xu, Partner at Lowercarbon Capital
