How AI Is Changing Mineral Discovery — and Why It Matters

Earth AI, an Austrian startup, secured $20 million in new funding on February 27, 2025. This is the company's sixth funding round since 2017, and the money will help it scale up its approach to finding critical minerals like lithium, cobalt, and rare earths — materials essential to everything from electric vehicles to smartphones.

The announcement matters because Earth AI has already shown results. In 2023, the company made three mineral discoveries — molybdenum, palladium, and lead-silver deposits — while spending just $2.1 million. That's a track record traditional mining companies rarely match.

A Different Approach to Finding Minerals

Earth AI operates differently from traditional mining exploration firms. Most exploration companies hire contractors to drill test holes and collect samples, then send the data to geologists for interpretation. Earth AI maintains its own drilling equipment that runs year-round, feeding continuous data directly into machine learning models.

Think of it this way: traditional exploration is like taking snapshots of the ground at specific times and places. Earth AI's approach is more like a video camera, constantly gathering information that trains algorithms to recognize patterns humans might miss.

The company's current focus is greenfield exploration in Australia — searching for previously undiscovered deposits in areas where geology and regulation favor this data-intensive method. Founder Roman Teslyuk, a geologist with a background in the University of Sydney, started the company after completing Y Combinator's 2019 summer program.

Who Else Is Doing This

Earth AI isn't alone. GeologicAI, a competitor, recently raised $44 million in Series B funding — more than double Earth AI's raise — and acquired another company called Lumo Analytics to broaden its capabilities.

Beyond startups, larger players are moving too. US Strategic Metals signed partnerships with rare earth recycling companies and Saudi Arabia's industrial development center. The activity signals real urgency around critical minerals.

Why the Urgency Now

The push for new mineral discovery and processing stems partly from geopolitics. China has spent decades building control over the entire supply chain — from mining raw materials through processing them into finished components. Electric vehicle maker BYD, for example, owns significant portions of its mineral supply chain, which gives it cost advantages competitors struggle to match.

Western governments and companies increasingly see this as a vulnerability. The European Union has proposed a critical minerals partnership with the US specifically to build alternatives to Chinese supply chains.

How the Technology Works

Earth AI's core idea is to automate what geologists have traditionally done by hand. Instead of relying on periodic surveys and human judgment about where minerals might be, the company feeds continuous drilling data into algorithms trained to recognize geological patterns.

The advantage is processing power and scale. Machine learning can spot correlations in massive datasets faster than human geologists working from snapshots. The company hasn't published detailed information about its algorithms or had its methods peer-reviewed, so the extent of its performance gains compared to traditional exploration remains to be seen.

To put this in historical context, this pattern has played out before in other industries. Semiconductor design underwent a similar shift in the 1990s when software tools replaced manual chip layout. Both mineral exploration and chip design involve analyzing complex spatial patterns, require large upfront investments, and improve over time as more data feeds the system. The parallel suggests AI could genuinely transform mineral discovery — if the technology proves itself at larger scales.

What This Could Mean

If companies like Earth AI and GeologicAI succeed in shortening discovery timelines and raising success rates, it could fundamentally change how the mining industry operates. Traditional exploration is expensive and slow, often taking decades from initial discovery to production. Compressing those timelines would affect everything from EV manufacturing timelines to supply chain stability.

The stakes are high. One critical minerals project is expected to supply 10% of the world's demand for a mineral used in semiconductors, LEDs, and solar panels, illustrating how much rides on successful exploration.

With 11-50 employees and $20 million in new funding, Earth AI has significant room to grow if its model works. The real test, though, is whether early laboratory results translate into stable, profitable mining operations — and whether the company can repeat its early success as it moves into new geological regions and mineral types.

Worth flagging: if AI-assisted exploration delivers on its promise, traditional mining companies that continue relying on older methods risk falling behind. The companies that combine machine learning, proprietary drilling infrastructure, and direct project ownership may build advantages competitors will struggle to match. But it's still early, and the mining industry moves slowly. The next few years will show whether this is a genuine technology shift or another venture-backed overpromise.