Turn environmental hazards into a product, not a data project
Climate risk is inherently N-dimensional — every hazard, every scenario, every decade, every asset. Earthmover gives risk teams a tensor-native data platform so science teams ship products instead of fighting NetCDF files, ETL, and warehouse bills.
Trusted by climate-risk and insurance innovators
Climate risk doesn't fit in a spreadsheet — or a tabular data warehouse
Physical risk is multidimensional by nature: every hazard, across every climate scenario, across every decade, against every asset in a portfolio. Most teams flatten that structure into static files and warehouse rows — and pay for it in ETL, storage, and lost science.
- The 2D trap: risk data is naturally multidimensional (hazard × scenario × time-horizon × space), but flattening it into NetCDF files or warehouse tables strips the structure that makes it analyzable.
- The dual-store tax: NetCDF-in-S3 plus Snowflake or Postgres for query means paying twice to store the same data — and writing ETL to keep the two in sync.
- Versioning chaos: risk models re-release constantly, and naïve versioning multiplies your storage to retain the reproducible, auditable versions downstream customers depend on.
- Science teams stuck doing DevOps: small teams burn their best scientists on data plumbing instead of model quality — right when a product launch is on the line.
One climate-risk analytics vendor we work with traced roughly $100K a year of warehouse spend to this dual-store pattern alone. We call it the "data engineering tax" — and in risk modeling, it's paid in slower model releases, fewer scenarios, and delayed product launches.
A data platform built for multidimensional risk
Store risk the way it actually exists — and serve it straight to your customers.
Model in N dimensions natively
Store hazard × scenario × horizon × space as tensors in Zarr and Icechunk — no flattening, no information loss.
Kill the dual store
One Icechunk dataset on your own S3 replaces the file-plus-Snowflake split. Science teams write results directly — less ETL, lower cloud bill.
Ship interactive risk maps day one
Flux serves OGC-standard tiles plus point and zonal extraction, so your frontend team builds customer-facing maps without a custom tile pipeline.
Portfolio-scale extraction
Query 10,000+ assets against gridded hazards through a single API — the workload insurers and reinsurers actually run.
Auditable, versioned hazards
Icechunk's diff-based versioning delivers a multi-version retention SLA without multiplying storage — every model release is reproducible and traceable.
Elevate model quality, not infrastructure
Earthmover is a tensor-native data platform: Arraylake, the governed, versioned data plane that runs on your own cloud; and Flux, the delivery gateway that turns your data into OGC-standard maps and APIs — both built on Icechunk, the open-source transactional storage engine our team authors.
- Data sovereignty: your data stays in your own S3 bucket, in open Zarr and Icechunk formats. Earthmover holds only catalog metadata and the control plane — never your bytes.
- No lock-in: because the formats are open standards we wrote, you could read your data with a single line of xarray.open_zarr(...) tomorrow, with no Earthmover involvement.
- Built for vendor-risk reviews: open formats and customer-owned storage are exactly what your own bank and insurer customers ask for in due diligence.
Store risk in its native dimensions. Serve it as maps and APIs. Ship products, not plumbing.
It's like the Databricks of geospatial data. You don't need to hire twice — you can just focus on creating the data.
Paolo Melissari
Data Engineer, Eoliann
Eoliann moved from a 2D GeoTIFF stack to N-dimensional Zarr at 30-meter European scale — and Flux let their frontend team start building Airis, their customer-facing climate-risk dashboard, before internal engineering was even ready. When they needed Flux to scale ahead of launch, Earthmover's team resolved it within a week. Airis shipped on schedule.
Earthmover is the one investment we made that I'm absolutely most certain about the return.
— Giovanni Luddeni, CTO & Co-Founder, Eoliann
Risk teams shipping faster on Earthmover
From GeoTIFF Chaos to Cloud-Native Climate Risk: How Eoliann Built Airis on Earthmover
Eoliann replaced a 2D GeoTIFF stack with N-dimensional Zarr at 30 m European scale — and Flux let their frontend team build the Airis risk dashboard before internal engineering was even ready.
How Kettle Uses Earthmover to Power Wildfire Risk Modeling at Scale
A lean data team manages 130+ TB and runs 10,000 annual wildfire simulations on Arraylake to price parcel-level risk across the American West.