Climate startup Terradot and Microsoft have developed TerraNova, an AI model intended to estimate processes involved in removing carbon dioxide using crushed rock, Microsoft announced.
Built on Aurora, the model was adapted to predict chemical changes in soil using Terradot’s field data. It could support estimates of rock reactions and site selection. Findings remain preliminary, with publication planned ahead of peer review.
The important distinction is between a prediction and verified carbon removal. A faster model can help plan measurements, but its output should not automatically be treated as proof of what actually happened in a specific field.
Commercial usefulness depends not only on predictive accuracy but also on traceable inputs and credible verification procedures. Lower modelling costs alone cannot establish the reliability of a claimed climate benefit.
