Earth AI builds proprietary labs to slash critical mineral analysis time
By processing rock samples in-house, the firm aims to reduce turnaround from months to days, targeting copper, platinum and palladium deposits in Australia.
Earth AI has announced a strategic shift to vertically integrate its operations by establishing its own laboratories for processing rock samples. This decision comes as the startup faces significant delays at third-party facilities, where backlogs have escalated to five months due to a surge in industry demand for critical mineral exploration.
The move is designed to cut the processing time for rock sample analysis from five months to just five days. By bypassing external bottlenecks, the company aims to accelerate its drilling and exploration cycles, ensuring that its artificial intelligence models receive timely data to optimise drill locations in Australia.
CEO Roman Teslyuk confirmed that the initiative specifically targets the exploration phase to ensure the AI models can function effectively. While the company's philosophy relies on the speed and quality of data to drive iterative drilling decisions, final economic validation for any potential mines will still rely on established third-party validators.
The urgency of this pivot was highlighted by recent operational hurdles. Following a ramp-up in drilling capacity, Earth AI found itself approximately 7,000 metres behind on unprocessed samples. Teslyuk noted that while external labs typically handled workloads with a two-month backlog, surging interest in new sources has more than doubled those wait times to five months.
The new in-house facilities will focus on minerals such as copper, platinum and palladium in Australian regions previously considered low-potential. The strategy acknowledges that while the AI models are effective at highlighting promising sites, physical drilling remains the only method to confirm mineral concentration and distribution.
However, the long-term cost implications of building and maintaining proprietary labs versus the cost of waiting for external results are not yet quantified. Furthermore, it remains unclear if the in-house labs will maintain the same analytical accuracy and certification standards as established third-party validators while achieving the five-day processing target.
