In the gold rush era, people were only looking for gold, but ignored that the real "gold mine" was actually the person selling shovels. Nowadays, in a new wave of technology with artificial intelligence and clean energy as the core, while the industry's focus is mostly on the infrastructure and application layers, a start-up company is betting on the bottom "shovel" of this wave - critical mineral raw materials. The company, called Earth AI, is headquartered in the United States and founded by an Australian team. It is positioned as a mineral exploration company that uses artificial intelligence technology to discover key mineral resources more quickly and efficiently.
Currently, the global race around AI and clean energy is driving unprecedented infrastructure expansion. From AI data centers filled with high-performance chips to large-scale photovoltaic power stations, electric vehicle batteries, energy storage systems, etc., these facilities have one thing in common: without exception, they are highly dependent on key minerals such as lithium, copper, nickel, cobalt, graphite, and rare earth elements. Driven by overlapping demand in areas such as consumer electronics, communications and military technology, the United Nations predicts that global trade in critical minerals is expected to triple in 2030 and quadruple in 2040, starting from a scale of approximately US$2.5 trillion in 2023.
However, while the demand curve continues to rise, discoveries of large new deposits are becoming increasingly rare. Multiple studies and industry reports show that despite rising exploration budgets, the number of major mineral discoveries has declined significantly in recent decades. Traditional exploration methods have basically "swept away" relatively easy-to-find ore bodies, forcing companies to pay higher costs and "try their luck" in deeper and more remote areas. However, the success rate has repeatedly reached new lows. In other words, mining companies that rely on traditional methods of prospecting are finding it increasingly difficult to find new economic deposits, despite investing more time, money and manpower.
Earth AI believes that artificial intelligence is expected to change this unfavorable situation. In an ironic twist, the same AI revolution that drove demand for critical minerals is now being used to help find them. Instead of relying mainly on traditional geological exploration methods and empirical intuition, the company chose to use AI models to "gnaw" massive geological data and identify areas that had been ignored but had high mineralization potential. Earth AI's business model combines predictive software with self-developed mobile, low-disturbance drilling technology to find and verify potential mineral sites, and then transfers proven project interests to mining companies. Its AI model uses a combination of decades of historical mining data and satellite remote sensing information to screen and target prospects in untapped “greenfield” areas.
Company founder and CEO Roman Teslyuk noted that while the world’s demand for mineral resources continues to expand, from the energy transition to everyday life, the discovery of new deposits has a reputation for being “cost-intensive” and “time-consuming.” He said Earth AI is reshaping this paradigm through its own self-developed AI systems and drilling technology, and has already made proven discoveries in areas where there has been no previous record of large-scale development.
Judging from the company’s disclosures, the returns from this method are considerable. Earth AI says it has discovered multiple previously unknown mineralized zones in Australia, including ore bodies containing copper, cobalt and gold mineralization, and is targeting them significantly faster than traditional exploration routes. The company claims a project discovery rate of 75%, compared to the industry average of less than 1%. Among the discoveries announced are a huge underground source of nickel and palladium on Australia's east coast, and a deposit also in Australia containing indium, a rare metal critical to AI semiconductors.
In the eyes of industry experts, Earth AI’s solution is particularly critical at a time when supply constraints are no longer just a future risk, but an ongoing reality. Investors seem to agree with this judgment: in early 2025, Earth AI announced the completion of a $20 million Series B round of financing, approximately 14.79 million pounds.
In the past few years, Earth AI's standard process is roughly as follows: the company first uses AI to screen potential mineral sites across Australia; once the target is identified, its geological team goes to the site to collect rock samples and entrusts a third-party laboratory to conduct geochemical analysis; if the test results show economic mineralization prospects, the company will verify it through drilling, obtain the rights and interests of the land, and finally sell the project rights to mining companies interested in development. However, after business expansion, the company found that the bottleneck of this process became more and more obvious-the problem was that the sample testing process was highly dependent on external laboratories.
As the market’s focus on resource discovery heats up, the backlog of orders from third-party laboratories continues to increase, and the sample testing cycle has been lengthened from the original approximately two months to longer. Teslyuk revealed that since the company expanded its drilling capacity, the sample backlog problem has become particularly prominent. He described the team as being "7 kilometers behind", that is, there are about 7,000 meters of samples obtained from drilling that have not yet received analytical data.
In order to get rid of this constraint, Earth AI decided to build its own laboratory, hoping to compress the testing cycle from five months to about five days. Last month, the company announced that its first in-house geochemical analysis laboratory was officially put into operation, allowing it to achieve an internal closed loop in key aspects of the exploration process. This move means that Earth AI has now formed a fully vertically integrated resource discovery and exploration chain from data modeling to target screening, on-site drilling, and sample testing. The company also emphasized that it will still use third-party organizations to independently verify key results before making a final judgment on the economic value of the mining area.
If Earth AI’s technology can continue to maintain a success rate close to its currently claimed success rate in the future, the company may become the “king of shovels” in this wave of technology and energy, playing a key role in the AI-driven mineral “gold rush.”