Machine learning meets mining: our investment in EARTH AI

Roman Teslyuk, EARTH AI founder

Intelligent software will disrupt every industry. Some industries will take longer to transform than others, but it is simply a matter of time.

As a venture fund based in Australia, the AirTree team spends a lot of time thinking about industries where Australia punches above its weight globally to which intelligent software might be applied. Examples include agriculture, education, financial services, mining, energy and healthcare. We are looking to partner with extraordinary entrepreneurs who have a unique insight into one of these industries and an ambition to disrupt the status quo.

We found just that in Roman Teslyuk, and are delighted to be investing in EARTH AI’s seed round along with our pals at Blackbird Ventures.

Roman Teslyuk

Roman is exactly what we look for in a founder. As a PhD candidate in Geosciences, he joined our friends at the Sydney University Incubate accelerator (shout out to James Alexander!) to test out an idea he had. The idea stemmed from his work as a geologist in Kyrgyzstan, Vladivostok and the Taigonos Peninsula where he had experienced first hand the challenges of mineral exploration.

Roman is the perfect combination of hacker and hustler. In just a few short months he and his friend Igor had hacked together v1 of the EARTH AI platform. He then hit the phone books, literally cold calling a list of mining exploration companies he found on the internet. One of the first companies he called, ActiveX, became EARTH AI’s launch customer in March 2017. The company now has dozens more mining firms using its software.


Mineral exploration has historically been conducted relying on the intuition of geologists and using only basic heuristics, models and data sets. The financial upside to more accurately locating mineral deposits is self evident and would be a huge boon for mining companies.

EARTH AI is an AI-powered mineral exploration platform. It passes a variety of data from geological maps, satellite imagery, geophysical and geochemical datasets through a neural net and auto-generates suggested locations for likely mineral deposits.

Each of EARTH AI’s customers uploads their own data (soil samples etc.) which increase the accuracy of the system’s predictions. Over time, this effect compounds, leading to more and more accurate predictions for all customers.

At AirTree, we love data network effects like this. They result in continual product improvement and provide natural defensibility against competitors.

We are excited to be backing Roman and the EARTH AI team on their journey and can’t wait to work with them as they disrupt one of Australia’s greatest industries.