Earthmover

Name: Earthmover
Brand: Earthmover

AI Native

The cloud platform for scientific data teams

Est 2022|Valuation $26M|Raised $13.3M|26 Ref Customers

LinkedIn 6K|Website|Web traffic 8K|FTEs 11|HQ New York, New York, US

MarketsData Warehouses & Lakehouses Data Catalog

Published Sep 26, 2026

01Common complaints

- You need extensive independent customer proof: mainstream review platforms show few or no reviews, making peer validation difficult. - Your primary need is a classic SQL warehouse/lakehouse: Earthmover is specialized for arrays/tensors; Databricks/Snowflake/BigQuery may be more appropriate. - You require formal compliance attestations up front: SOC 2 / ISO 27001 are not publicly documented in the surfaced sources. - You want a fully managed data+compute+catalog solution with minimal architecture decisions: vendor-managed platforms such as Google Earth Engine or Planetary Computer may provide a faster start, albeit with different tradeoffs. - You have strong in-house expertise and prefer a DIY approach: teams proficient with Zarr, Kerchunk, STAC, TiTiler/xpublish, and cloud IAM may choose to build a custom stack to reduce vendor dependency, subject to ongoing engineering costs.

Deal disqualifiers (poor fit)

Primarily tabular BI/SQL analytics use cases
Hard requirement for publicly documented SOC 2 / ISO certifications (unless provided under NDA)
Requirement for abundant third-party reviews/references prior to adoption

Limitations acknowledged in vendor technical content

Vendor posts describe limitations of generating every map tile dynamically from high-resolution data (a performance/cost tradeoff). This implies that for some visualization workloads, precomputation or caching strategies may be required.

Product maturity / roadmap signals

Arraylake launched as a private beta in October 2023.
The vendor published a preview of Icechunk 2, indicating ongoing development and potential breaking changes for early adopters experimenting with preview builds.

Operational limitations

The status page shows scheduled maintenance affecting API/web app/Flux endpoints; this is typical of SaaS offerings but relevant for production reliability planning.

Compliance documentation gaps

SOC 2 / ISO certifications were not publicly documented in the reviewed sources; buyers should request security documentation under NDA.

02Why they buy

Unique selling points supported by evidence in vendor materials

Icechunk as an open-source transactional storage engine for Zarr on object storage (GitHub and vendor blog posts).
Flux as a delivery layer supporting WMS, OGC EDR, and OPeNDAP/DAP.
Architecture that enables data to remain in customer-managed cloud storage while backend services run in the vendor's AWS region.

Where competitors have stronger advantages

Review footprint, procurement familiarity, and enterprise compliance evidence are typically stronger with larger incumbents.

03Field reports

Review platform availability and limitations

A major constraint in assessing voice-of-customer is that Earthmover/Arraylake appears under-reviewed on mainstream B2B review platforms: - G2: the Arraylake page shows few or no reviews and indicates there are not enough reviews to provide buying insight. - AWS Marketplace: Earthmover listing shows few or no customer reviews. - Gartner Peer Insights / TrustRadius: searches did not surface an accessible, product-specific review page for Earthmover/Arraylake; this may mean the product is not listed, is listed under a different market name/category, or the pages are not easily discoverable/indexed. As a result, there are no independently verifiable review quotes available from those platforms at the time of research.

Community and social signals

Searches for "Earthmover" surface many irrelevant results (e.g., the term used for music or games). This is a discoverability issue: community sentiment specific to the vendor and product is difficult to isolate by keyword.

However, the open-source components (Icechunk and related tools) provide developer adoption signals:

Icechunk is open source on GitHub, and the vendor highlights external contributors and repository activity in its blog.

Why customers choose Earthmover (from official case studies and vendor narratives)

From vendor case studies and blog narratives, typical reasons include:

Avoiding repeated downloads / re-ingestion of large scientific datasets
Shortening iteration time for analysis/ML workflows by using cloud-native, chunked formats
Delivering data products through standard protocols rather than bespoke services

Why some customers might avoid / leave (inferred)

Because independent reviews are scarce, churn reasons are not well documented publicly. Likely risk factors (inferred from the category and vendor messaging) include: - Preference for more established vendors (TileDB, Databricks ecosystem, hyperscalers) - Internal capability to build with open source (Zarr + Kerchunk + STAC + custom API) - Concerns over compliance posture if SOC 2/ISO evidence is required - Dependency on a relatively young platform for mission-critical production forecasting

Sylvera case study (vendor-published)

Earthmover published a case study describing Sylvera (carbon ratings provider) evaluating Arraylake by connecting an existing pipeline, querying data, and avoiding a re-download step. Critical evaluation: - The case study describes workflow improvements but does not provide independently auditable metrics (e.g., precise cost, latency, or throughput numbers) in publicly surfaced excerpts. - As a vendor-authored case study, it is consistent with vendor-selected favorable outcomes and should be validated in procurement conversations.

University / research usage (vendor site)

Earthmover highlights university and research team usage (e.g., UW-Madison ALIVE) on its website.

Critical evaluation:

These references support fit for research workflows, but they do not substitute for independent reviews about production hardening, procurement friction, and multi-year TCO.

NASA collaboration / performance improvement claim

Seed funding announcements reference collaboration with NASA and a reported substantial performance improvement in retrieval of Earth observation data.

Critical evaluation:

The claim appears in funding/press materials and lacks publicly accessible methodology details in the surfaced excerpts (hardware baseline, data volume, cache effects, workload definition). Buyers should request a technical write-up or benchmark replication details.

04Best fit

Company size and stage (practical ICP estimate)

Given quote-based pricing, specialized domain focus, and need for cloud data maturity, the practical ICP often looks like: - Startups to mid-market climate/EO companies with substantial scientific data volumes and dedicated data/ML teams - Research organizations and universities modernizing data distribution and reproducibility - Enterprises in sectors such as energy, insurance, commodities, aviation, and defense contractors with scientific data pipelines—if security and compliance requirements can be met A reasonable employee range is often small to mid-sized organizations up to larger enterprise teams depending on data maturity; smaller teams may adopt if they are highly technical and data-intensive.

Industries / verticals

Weather forecasting and reforecast pipelines
Climate risk analytics
Carbon markets / MRV (monitoring, reporting, verification)
Remote sensing / Earth observation analytics
Energy and commodities trading

Roles and personas

Data engineering leads for scientific data
Climate / geospatial ML engineers
Research software engineers
GIS engineers needing standards-based APIs
Platform engineering / cloud teams supporting architectures where data resides in customer-managed storage

Qualification criteria (good fit)

Core data assets are arrays/raster/tensors (not primarily tables)
Need for versioning/time travel and reproducibility
Need to serve data products through EDR/WMS/OPeNDAP to heterogeneous consumers
Preference for open standards and minimizing storage-layer lock-in

05The pitch

What the product does (objective value proposition)

Earthmover is a cloud-native data platform designed for multidimensional array (tensor) data, with an initial focus on weather, climate, and geospatial / Earth-observation workloads. The platform is intended to replace file-and-folder–based scientific data workflows (NetCDF/HDF/GRIB/TIFF collections, often at large scale) with a system that supports: - Cloud object storage as the primary persistence layer - Transactional / versioned array repositories (Git-like concepts such as snapshots / time travel are implemented) - Team collaboration, governance, and controlled sharing - End-stage data delivery via standards-based APIs for common geospatial and scientific tooling The vendor states the platform aims to reduce data engineering overhead for scientific data teams, to support machine-learning pipelines and to shorten iteration cycles for forecasting and modeling.

Target market and positioning

Earthmover targets a niche that overlaps: - Climate / weather data engineering and climate-tech product teams - Earth observation and remote sensing pipelines (satellite, radar/lidar) - Quant/energy trading teams working with high-velocity weather data (Earthmover has an "Energy Trading" solution page) - Scientific computing groups moving to cloud-native formats and needing governance, sharing, and standardized APIs The company differentiates from general-purpose lakehouses by focusing on multidimensional arrays rather than tabular-first architectures.

Get a comprehensive analysis of Earthmover including market position, competitive landscape, adoption trends, and peer benchmarks.

Download full report →

Published Sep 26, 2026

01Common complaints

You need extensive independent customer proof: mainstream review platforms show few or no reviews, making peer validation difficult.
Your primary need is a classic SQL warehouse/lakehouse: Earthmover is specialized for arrays/tensors; Databricks/Snowflake/BigQuery may be more appropriate.
You require formal compliance attestations up front: SOC 2 / ISO 27001 are not publicly documented in the surfaced sources.
You want a fully managed data+compute+catalog solution with minimal architecture decisions: vendor-managed platforms such as Google Earth Engine or Planetary Computer may provide a faster start, albeit with different tradeoffs.
You have strong in-house expertise and prefer a DIY approach: teams proficient with Zarr, Kerchunk, STAC, TiTiler/xpublish, and cloud IAM may choose to build a custom stack to reduce vendor dependency, subject to ongoing engineering costs.

Deal disqualifiers (poor fit)

Primarily tabular BI/SQL analytics use cases
Hard requirement for publicly documented SOC 2 / ISO certifications (unless provided under NDA)
Requirement for abundant third-party reviews/references prior to adoption

Limitations acknowledged in vendor technical content

Vendor posts describe limitations of generating every map tile dynamically from high-resolution data (a performance/cost tradeoff). This implies that for some visualization workloads, precomputation or caching strategies may be required.

Product maturity / roadmap signals

Arraylake launched as a private beta in October 2023.
The vendor published a preview of Icechunk 2, indicating ongoing development and potential breaking changes for early adopters experimenting with preview builds.

Operational limitations

The status page shows scheduled maintenance affecting API/web app/Flux endpoints; this is typical of SaaS offerings but relevant for production reliability planning.

Compliance documentation gaps

SOC 2 / ISO certifications were not publicly documented in the reviewed sources; buyers should request security documentation under NDA.

02Why they buy

Unique selling points supported by evidence in vendor materials

Icechunk as an open-source transactional storage engine for Zarr on object storage (GitHub and vendor blog posts).
Flux as a delivery layer supporting WMS, OGC EDR, and OPeNDAP/DAP.
Architecture that enables data to remain in customer-managed cloud storage while backend services run in the vendor's AWS region.

Where competitors have stronger advantages

Review footprint, procurement familiarity, and enterprise compliance evidence are typically stronger with larger incumbents.

03Field reports

Review platform availability and limitations

A major constraint in assessing voice-of-customer is that Earthmover/Arraylake appears under-reviewed on mainstream B2B review platforms:

G2: the Arraylake page shows few or no reviews and indicates there are not enough reviews to provide buying insight.
AWS Marketplace: Earthmover listing shows few or no customer reviews.
Gartner Peer Insights / TrustRadius: searches did not surface an accessible, product-specific review page for Earthmover/Arraylake; this may mean the product is not listed, is listed under a different market name/category, or the pages are not easily discoverable/indexed.

As a result, there are no independently verifiable review quotes available from those platforms at the time of research.

Community and social signals

However, the open-source components (Icechunk and related tools) provide developer adoption signals:

Icechunk is open source on GitHub, and the vendor highlights external contributors and repository activity in its blog.

Why customers choose Earthmover (from official case studies and vendor narratives)

From vendor case studies and blog narratives, typical reasons include:

Avoiding repeated downloads / re-ingestion of large scientific datasets
Shortening iteration time for analysis/ML workflows by using cloud-native, chunked formats
Delivering data products through standard protocols rather than bespoke services

Why some customers might avoid / leave (inferred)

Because independent reviews are scarce, churn reasons are not well documented publicly. Likely risk factors (inferred from the category and vendor messaging) include:

Preference for more established vendors (TileDB, Databricks ecosystem, hyperscalers)
Internal capability to build with open source (Zarr + Kerchunk + STAC + custom API)
Concerns over compliance posture if SOC 2/ISO evidence is required
Dependency on a relatively young platform for mission-critical production forecasting

Sylvera case study (vendor-published)

Earthmover published a case study describing Sylvera (carbon ratings provider) evaluating Arraylake by connecting an existing pipeline, querying data, and avoiding a re-download step.

Critical evaluation:

The case study describes workflow improvements but does not provide independently auditable metrics (e.g., precise cost, latency, or throughput numbers) in publicly surfaced excerpts.
As a vendor-authored case study, it is consistent with vendor-selected favorable outcomes and should be validated in procurement conversations.

University / research usage (vendor site)

Earthmover highlights university and research team usage (e.g., UW-Madison ALIVE) on its website.

Critical evaluation:

These references support fit for research workflows, but they do not substitute for independent reviews about production hardening, procurement friction, and multi-year TCO.

NASA collaboration / performance improvement claim

Seed funding announcements reference collaboration with NASA and a reported substantial performance improvement in retrieval of Earth observation data.

Critical evaluation:

The claim appears in funding/press materials and lacks publicly accessible methodology details in the surfaced excerpts (hardware baseline, data volume, cache effects, workload definition). Buyers should request a technical write-up or benchmark replication details.

04Best fit

Company size and stage (practical ICP estimate)

Given quote-based pricing, specialized domain focus, and need for cloud data maturity, the practical ICP often looks like:

Startups to mid-market climate/EO companies with substantial scientific data volumes and dedicated data/ML teams
Research organizations and universities modernizing data distribution and reproducibility
Enterprises in sectors such as energy, insurance, commodities, aviation, and defense contractors with scientific data pipelines—if security and compliance requirements can be met

A reasonable employee range is often small to mid-sized organizations up to larger enterprise teams depending on data maturity; smaller teams may adopt if they are highly technical and data-intensive.

Industries / verticals

Weather forecasting and reforecast pipelines
Climate risk analytics
Carbon markets / MRV (monitoring, reporting, verification)
Remote sensing / Earth observation analytics
Energy and commodities trading

Roles and personas

Data engineering leads for scientific data
Climate / geospatial ML engineers
Research software engineers
GIS engineers needing standards-based APIs
Platform engineering / cloud teams supporting architectures where data resides in customer-managed storage

Qualification criteria (good fit)

Core data assets are arrays/raster/tensors (not primarily tables)
Need for versioning/time travel and reproducibility
Need to serve data products through EDR/WMS/OPeNDAP to heterogeneous consumers
Preference for open standards and minimizing storage-layer lock-in

05The pitch

What the product does (objective value proposition)

Cloud object storage as the primary persistence layer
Transactional / versioned array repositories (Git-like concepts such as snapshots / time travel are implemented)
Team collaboration, governance, and controlled sharing
End-stage data delivery via standards-based APIs for common geospatial and scientific tooling

The vendor states the platform aims to reduce data engineering overhead for scientific data teams, to support machine-learning pipelines and to shorten iteration cycles for forecasting and modeling.

Target market and positioning

Earthmover targets a niche that overlaps:

Climate / weather data engineering and climate-tech product teams
Earth observation and remote sensing pipelines (satellite, radar/lidar)
Quant/energy trading teams working with high-velocity weather data (Earthmover has an "Energy Trading" solution page)
Scientific computing groups moving to cloud-native formats and needing governance, sharing, and standardized APIs

The company differentiates from general-purpose lakehouses by focusing on multidimensional arrays rather than tabular-first architectures.

Get a comprehensive analysis of Earthmover including market position, competitive landscape, adoption trends, and peer benchmarks.

Download full report →