Stratum AI

Stratum is a mining tech company that produces resource models that are >25% more accurate than all resource estimation methods employed today. For example, for an end-of-life gold asset, their models decreased quarterly grade tonnage reconciliation error (from ground truth) by 70%, tested blind over 4 years. They also found ~160M USD of additional in-situ value, equivalent to a 2-year mine life extension. This has been used as guided drilling programs for multiple gold asset clients.

Stratum’s technology uses just the data that a mine already collects – there is no need to collect new data. Stratum’s protocols ingests this data, processes the data with the sources of error taken into account, and produces a block model that has been battle tested against ground truth, and is understandable to the geologist.
Stratum has clients in copper, gold, zinc, and tin assets with companies in every major mining jurisdiction including Kazakhstan, Peru, Canada, and Australia, PNG, South Africa, etc. Stratum’s models deliver similar value described above to all clients and address challenges unique to those chosen mine sites.

Use cases include better stockpiling for copper mines (and reduced grade tonnage deviation by 58%), and modelling metallurgical parameters (sulfide/oxide zones) critical to optimal recoveries, modelling deleterious elements, like arsenic, for feasibility stage assets to de-risk investment, and expansion-related resource definition.