Terradot

Select Language

Home

Technology

Science

Team

Press

Terradot

REMOVAL YOU CAN
TRUST

REMOVAL YOU CAN TRUST

Terradot’s subsurface intelligence platform finds the best projects, optimizes deployment, and verifies every ton.

LEADING SCIENCE,
BACKED BY DATA

Five years of earth science at scale.

A team of 20+ PhDs spent half a decade learning where ERW works, how fast, and at what cost.

Data quality starts in the field. Great models need great measurement — so ours begins with the most rigorous field-data program in ERW.

At our Sentinel Site, we ran the most comprehensive scientific monitoring ever deployed at a commercial ERW project — the ground truth behind everything we build.

THE SUBSURFACE
INTELLIGENCE PLATFORM

Proprietary datasets, informing subsurface intelligence models.

Remote Sensing and Environmental Data

Satellite imagery, historical land-use records, and environmental management data give us a baseline read on every site before a single sample is taken.

Proprietary Soil Data

Five-plus years of ground-truth soil sampling across Brazil and the US, the largest dataset of its kind in ERW, with 75k+ distinct soil samples.

Process-Based Reactive Transport Model

Simulates weathering chemistry ion by ion, from the soil profile to the watershed, grounding every prediction in real geochemistry.

Enhanced by Our AI Model

TerraNova predicts removal and weathering rates directly from environmental data, at a fraction of the compute and calibration that modeling alone demands.

Remote Sensing and Environmental Data.

Satellite imagery, historical land-use records, and environmental management data give us a baseline read on every site before a single sample is taken.

Proprietary Soil Data

Five-plus years of ground-truth soil sampling across Brazil and the US, the largest dataset of its kind in ERW, with 75k+ distinct soil samples.

Process-Based Reactive Transport Model

Simulates weathering chemistry ion by ion, from the soil profile to the watershed, grounding every prediction in real geochemistry.

Enhanced by Our AI Model

TerraNova predicts removal and weathering rates directly from environmental data, at a fraction of the compute and calibration that modeling alone demands.

Remote Sensing and Environmental Data.

Satellite imagery, historical land-use records, and environmental management data give us a baseline read on every site before a single sample is taken.

Proprietary Soil Data

Five-plus years of ground-truth soil sampling across Brazil and the US, the largest dataset of its kind in ERW, with 75k+ distinct soil samples.

Process-Based Reactive Transport Model

Simulates weathering chemistry ion by ion, from the soil profile to the watershed, grounding every prediction in real geochemistry.

Enhanced by Our AI Model

TerraNova predicts removal and weathering rates directly from environmental data, at a fraction of the compute and calibration that modeling alone demands.

TerraNova

Developed with

Microsoft Research and

Google AI for Nature

TerraNova, developed with Microsoft Research and AI for Good, is the first frontier model built for carbon removal, with research published at ICLR 2026. Through Google’s AI for Nature program, we extend this with Earth Engine’s planetary-scale remote sensing to find the best sites on Earth.

Read about our AI work: Microsoft, Google, and ICLR 2026.

TerraNova

Developed with

Microsoft Research and

Google AI for Nature

Read about our AI work: Microsoft, Google, and ICLR 2026.

IDENTIFY. OPTIMIZE. VERIFY.

One platform, driving every project from site selection through verified delivery.

Identify

The platform maps where rock, soil, climate, and hydrology combine to engineer projects that remove carbon durably and predictably.

Optimize

It determines the optimal project design, weighing feedstock, grain size, application rate, and timing to maximize removal and agronomic benefit for each field.

Verify

It quantifies the carbon removed across the whole system, soil, water, and the fate of the weathering products, ready for independent validation.