Roman Teslyuk, co-founder of EARTH AI.

Investment Notes: EARTH AI

Date Published:
September 4, 2017

EARTH AI is a company helping mineral explorers identify promising areas.

This prediction of the future was made in 1956 at Dartmouth University. It is the fledgling thesis that would form the basis for Artificial Intelligence as we know it today.

73 years is a long time to wait but we’re finally starting to see some of this promise coming to fruition.

EARTH AI are helping mineral explorers identify promising areas. The company does this by ingesting data from multiple sources and using a machine learning algorithm to identify areas where minerals are likely to be found.

Blackbird is investing in EARTH AI’s seed round.

Artificial Intelligence

Plummeting computing costs mean that the concepts of artificial intelligence invented in the 1950s are now ripe for implementation.

We are truly in a grand age, where large parts of society will be transformed by AI. But naturally, every startup pitch now claims AI at the centre of their plans and the term has rapidly lost any semblance of meaning, like other hall of fame buzzwords such as ‘gamification’, ‘big data’ and ‘Internet of things’ before it.

At Blackbird, when evaluating such startups we focus very carefully on the type of problem being solved and if there are data network effects that develop over time.

The initial wave of AI is broadly good at two things: spotting patterns in data and interpreting and classifying remote sensing imagery. With that in mind, EARTH AI is profiling geological data and patterns to discern where promising exploration targets might be.

The second aspect to EARTH AI’s business is the data network effects we hope will appear over time. Initially, the company synthesises public data over satellite imagery. Then each client uploads the results of their soil samples, adding depth to the public data.

EARTH AI’s algorithms get better as more private soil sample data is uploaded, which then increases the value and accuracy of its software to new customers.

Rock Licking

EARTH AI develops algorithms that match the patterns of current mineral discoveries to then look for similar geological patterns in unexplored areas. The software is relevant to minerals of all kinds from Iron Ore and Gold through to Lithium.

EARTH AI is co-founded by Roman Teslyuk, a PHD student at the University of Sydney who recently finished up at the Incubate accelerator, where we first got to know him.

Roman is a rare combination of product insights paired with extreme hustle. He has spent his career as a geologist in the most remote parts of the world working on mines in Siberia, through to the outskirts of Far East and remote Australia. He’s a “rock licker” through and through.

After putting together the first product with his childhood friend and fellow co-founder, Igor Gerechko, he downloaded a list of mining exploration companies to cold call alphabetically. He didn’t need to go past the letter A to land their first customer, ActivEX Limited.

This combination of hustle and smarts is why we’re backing EARTH AI from an early seed round.


Mining exploration has historically been a tedious and “gut-feel” industry. EARTH AI is symbolic of our belief in the power of software to disrupt this sort of work. Roman and Igor are exactly the kind of founders we love to back.

It’s our hope that they’ll help to bring humanity closer to some of future predicted at Dartmouth in 1956.

If you want to receive these notes direct to your inbox, join the Blackbird community or follow us Facebook, Linkedin or Twitter.