Earth AI vertically integrates critical mineral exploration

A model is only as good as its data. For Roman Teslyuk, the data wasn’t retrieving fast enough.

“I hate delays,” Earth AI founder and CEO Teslyuk told TechCrunch.

For the past few years, Earth AI has been searching for important minerals like copper, platinum and palladium in parts of Australia where no one thought they existed. The startup’s AI model suggested some spots that proved promising, but identifying the rocks with the highest mineral concentrations was slower than Teslyuk would have liked.

The problem is with the lab, he said.

“Ever since we ramped up our drilling capacity, we started experiencing these significant delays,” he said. Teslyuk said laboratories that process rock samples for evidence of important minerals typically have wait times of about two months. But recently, due to increased interest in the development of new sources, the delays have more than doubled. “We are 7 km behind. There is no data on the 7,000 meter sample.”

So Earth AI is instead setting up its own lab and hopes to cut that time from five months to five days, the startup exclusively told TechCrunch.

Earth AI’s models are good at highlighting areas that could develop into mines, but even once they’re identified, the startup still needs to drill to see what minerals are underneath and how they’re distributed, Teslyuk said. Although underground exploration has made great strides, there is still no substitute for drilling.

Once the drill core is extracted, it must be processed in a laboratory. “I don’t know if I’ve reached the gold medal. I can’t see it with my eyes,” he said.

For final decisions regarding the economic value of a mine, including those that may affect a sale, Earth AI will continue to use third parties to verify its findings. However, during the exploration process, a rapid in-house lab can significantly reduce costs by ensuring the drill is sent to the right location to obtain the best data for the model.

“If you don’t find the answer by the deadline, you have to wait 5 months to find the answer.The next question is [of where to drill] “To minimize drilling, you need to effectively ask the right questions and get the information in time to narrow down exactly where to go,” Teslyuk said.