Earthmover wants to become the Snowflake of weather and geospatial data

Few things generate as much data as simply observing the Earth from above. But Ryan Abernathey and Joe Hamman very quickly realized that all that data still wasn’t enough for their startup to thrive. Their data-centric, climate tech startup, Earthmoverwould need to pivot.

The pivot isn’t entirely away from climate tech, though. Instead, the company is shrinking the time scale, focusing on how climate affects daily life — in other words, the weather.

“What makes a compelling use case for our platform? Data that change frequently,” Abernathey, Earthmover’s co-founder and CEO, told TechCrunch. “That’s where there’s a lot more urgency around solutions. That [data] goes to weather, goes to fire, goes to new observations that are being generated.”

Climate outputs, he noted, are “important, but they’re kind of static,” with new data emerging every few years.

Earthmover’s core product remains its data structure, which was built to handle large, complex data sets. “In geospatial they call it a raster. In AI, they call it a tensor. In old school Fortran, they just call it an array,” Abernathey said. On top of that, the startup has built a range of tools to help customers tease insights out of their data.

The pivot has helped Earthmover find paying customers — more than 10 so far, per Abernathey — and land a $7.2 million seed round, the company exclusively told TechCrunch. The round was led by Lowercarbon Capital, with participation from Costanoa Ventures and Preston-Werner Ventures. The company is using the cash to build new tools atop its data storage platform.

Earthmover is built on open-source software, including Xarray, Pangeo and Icechunk, and runs on major cloud providers, including Google Cloud, AWS and Microsoft Azure, as well as on-premise servers. Both Abernathey and Hamman, Earthmover’s CTO, have deep ties to the open-source world, having worked on Pangeo and Xarray.

Part of the decision to use those tools was motivated by their facility with Earth observation data, which can quickly demand terabytes or even petabytes of storage. Typical Earthmover customers have between tens to hundreds of terabytes, Abernathey said.

But the decision to go open source was also motivated by a business reality: Building tools atop open-source projects helps give customers peace of mind, Abernathey said.

“If you’re a serious multinational company, you’re going to have a dependency on a startup,” he said. But by using open source tools, “the customer’s risk, in a way, is mitigated because if we pivot or go out of business or whatever, they’ve still got all their own data in their own storage.”

Earthmover’s customers include insurance startup Kettle, which uses the platform to assess wildfire risk, and RWE, the German multinational energy company. Abernathey said renewable companies, which are exposed to the vagaries of the weather, use Earthmover’s tools to forecast supply and demand.

The goal is to make weather and geospatial data accessible to more users. “If you’re a trading desk, you want to see the map of the latest forecast, and you want it on a dashboard,” Hamman said. “You don’t want to run a Python script or something to do it.”