Startup harnesses the power of AI in the ongoing hunt for minerals

It is said that during the gold rush, people were so focused on gold that they forgot the “goldmine” that was shovels. In the AI and clean energy boom, where infrastructure and applications are the focus, one company is betting on the most fundamental of shovels: raw materials. Introducing Earth AI, an Australian-founded, US-headquartered mineral exploration company that uses artificial intelligence to discover critical mineral deposits faster and more efficiently.

The global AI and clean energy race is triggering an unprecedented buildout of physical infrastructure.

We have AI data centers packed with advanced chips, sprawling solar farms, EV batteries, grid-scale storage systems, and much more. A unifying characteristic across these is that they are all built with critical minerals such as lithium, copper, nickel, cobalt, graphite, and rare earth elements. With additional demand from consumer electronics, telecommunications, and military technology, global trade in critical minerals is expected to triple by 2030 and quadruple by 2040, from approximately $2.5 trillion in 2023, according to the UN.

The problem is that while demand projections for these minerals keep climbing, major new mineral discoveries have become increasingly rare. Studies and industry reports have shown that the rate of significant discoveries has fallen sharply over the past few decades, even as exploration spending has surged. Many of the easiest-to-find deposits have already been uncovered, leaving companies searching deeper, farther, and at greater cost with low success rates. Basically, mining companies searching for new deposits using traditional methods are consistently coming up short despite increased time, effort, and resources.

Earth AI believes artificial intelligence can change that reality.

In a rather fitting twist, the same AI revolution driving explosive demand for critical minerals is now being used to hunt for them. Rather than relying heavily on traditional exploration methods and geological intuition alone, the company trains AI models on vast geological datasets to identify overlooked regions with high mineral potential. Basically, the discovery and exploration company uses predictive software and proprietary, relatively minimal, mobile low-disturbance drilling technology to find and verify sites. It then sells the rights to successful projects to mining companies. The AI is trained on a combination of several decades of historical mining data and satellite remote sensing to identify greenfield prospects.

“Despite the tremendous global need for mineral resources for everything from the energy transition to day-to-day life, new mineral deposit discoveries are notorious for being expensive and time consuming,” says Roman Teslyuk, founder and CEO of the firm. “Earth AI is changing this paradigm using its proprietary AI and drilling technology that has already made verified discoveries in greenfield regions.”

According to the company, this approach has been wildly successful.

Earth AI says it has already identified multiple previously unknown deposits across Australia, including discoveries containing copper, cobalt, and gold mineralization, while dramatically accelerating the targeting process compared to conventional exploration timelines. The company boasts an outstanding discovery rate of 75%. For context, the industry average is less than 1%! Among its discoveries are a massive underground trove of nickel and palladium on Australia’s east coast, and a deposit of indium, a rare-earth metal critical for AI semiconductors, also in Australia.

Earth AI’s solution is especially vital at a time when industry experts warn that material shortages are no longer hypothetical but a real, ongoing situation. Apparently, investors agree. In early 2025, the company announced that it had raised $20 million (£14.79 million) in a Series B funding.

More recently, just last month, the company announced the opening of its first-of-its-kind, in-house geochemical analysis lab. For the past few years, Earth AI’s process has been as follows: the company uses its AI to search for deposits across Australia. After identifying a potential site, the company’s geologists collect rock samples and send them to labs for testing. If the samples test positive, then the company drills to validate the discovery, secures rights to the site, and sells the rights off to a mining company.

However, Earth AI found this process too slow. The problem? Outsourcing sample testing to external labs. These labs typically had a processing time of two months due to backlogs. With increased interest in discovery, wait times have doubled.

“Since we ramped up the drilling capacity, we started getting these massive delays,” explains Teslyuk. “We’re 7 km [4 miles] behind – 7,000 meters of samples we don’t have data about.”

To circumvent this problem, the company set up its own lab, with the aim of bringing down the wait time from five months to five days. This development means Earth AI is now a fully vertically integrated discovery and exploration company. The company confirmed that it will still use third parties to validate its discoveries before making a final decision on a mine’s economic value.

If Earth AI’s technology continues to deliver at anything close to its claimed success rate, the company could very well become the “shovel kings” of the ongoing technology boom.

Source: Earth AI