Any Minecraft enthusiasts in the house?

OGs will remember using X-Ray vision (much to the disappointment of multiplayer server owners) to find diamonds deep underground.

Now, we’re not saying Terra AI is like a video game enhancement… but a greater sense of vision into mining is not a bad way to think about what they’re up to!

Arek and Ethan 🦄

Imagine having to spend billions of dollars on an upfront investment with little to no information about whether it may be worth it. That’s what happens in industries like mining, geothermal energy, and carbon storage. They often make large investment decisions without actually being able to directly see the asset they are exploring. 

To solve this, Terra AI is building platform that helps mining and energy companies better understand what lies beneath the Earth’s surface. 

Terra brings together a company’s existing exploration data, including geological maps, drill cores, geophysics, geochemistry, remote sensing data, and geologists’ own hypotheses, to build detailed models of the subsurface. Besides that, it tests different geological scenarios and identifies where additional data or drilling could be most valuable.

By giving teams a clearer picture of what may exist underground before committing additional capital, Terra’s goal is to reduce unnecessary drilling, lower exploration costs, and shorten the time it takes to move a promising project forward.

Check it out: terraai.com

Today, Terra works directly with customers to apply its technology to specific deposits and projects while it builds toward a self-service platform. 

It serves three main customers: junior exploration companies, major mining companies, and reservoir developers working on projects such as geothermal energy and carbon storage.

For junior exploration companies, Terra can help determine where limited exploration dollars should be spent and improve how a deposit is understood before the company raises additional capital. 

Larger mining companies can use the platform across their portfolios to improve drill targeting and make better development decisions.

For reservoir developers, Terra’s value is less about discovering a new resource and more about understanding how an existing underground system will behave.

Terra AI started with a question that John Mern and Anthony Corso had spent years studying at Stanford: how do you make good decisions when most of the information you need is uncertain?

Both completed their PhDs at Stanford and worked within the Stanford Intelligent Systems Laboratory. The lab has applied ideas across areas like aerospace and autonomous systems where decisions have to be made without having perfect information.

Mern later brought this experience into the natural resources industry. Before founding Terra, he worked at Boeing Phantom Works, Boeing’s advanced research and development arm, on autonomous and unmanned systems. Then, he joined KoBold Metals, where he helped develop AI systems for mineral exploration. That gave him direct exposure to companies spending millions of dollars trying to understand what is underground, but still not knowing for sure until they drill.

Through Stanford, Mern and Corso also connected with Markus Zechner, who had spent years studying uncertainty in underground reservoirs. His work focused on areas like carbon storage and geothermal systems. Together, they saw an opportunity to apply AI and probabilistic modeling to a problem that had historically relied heavily on experience, limited data, and expensive trial and error.

Terra was founded in 2023 to use AI to generate multiple possible geological scenarios and help geologists determine what information to collect next. That approach quickly attracted interest from major investors and industry players. 

Almost every major technology trend happening aboveground depends on resources buried beneath it.

For example, data centers require huge amounts of electricity and copper, while electric vehicles rely on lithium, nickel, copper, and other minerals. The problem is that developing those resources is still extremely uncertain.

A mining company may believe that a region contains copper, lithium, gold, or another valuable resource, but it cannot simply look underground and know exactly where the deposit is or how large it might be. Geologists instead gather information from drill holes, geophysical surveys, geochemistry, and historical records, and then build their best estimate of what the deposit looks like.

That process is expensive and slow. A single drill hole can cost hundreds of thousands of dollars, and even after a large drilling program, companies are still working with only a small sample of the total deposit.

Terra is trying to make that process much more efficient.

Customers provide the platform with the geological data they already have, and Terra’s AI creates a range of possible models for what could exist underground and helps determine which missing piece would give you the most useful information.

That can have a major financial impact. Terra says its technology can reduce drilling meters by 50 to 60 percent, while also improving the precision of resource estimates. For an industry where drilling programs can cost millions of dollars, it can create significant value.

Terra also has an interesting position within the growing AI-for-mining market, because it wants to become more of an “overarching umbrella” for the industry rather than a single point solution. Data collected from drilling companies, sensors, geologists, and other technologies could all feed into Terra’s platform and contribute to a larger subsurface model.

That is what makes the opportunity especially interesting. If Terra only helps companies decide where to drill, it is addressing one part of the exploration process. But if the platform becomes the place where companies combine their geological data, evaluate different scenarios, understand uncertainty, make drilling decisions, and eventually optimize production, Terra could become much more integrated into how natural resource companies operate.

There are still challenges ahead. Mining is a technically complex industry where many decisions are based on decades of geological experience, so convincing teams to rely more heavily on AI will take time. Terra will also need to prove that its technology works across different commodities, geological environments, and stages of development. Competition will likely continue to increase as more startups apply AI to natural resources.

Still, the broader market is moving in Terra’s favor. The International Energy Agency projects that copper supply from announced projects could fall roughly 25% short of demand by 2035, as electrification and next-generation technologies continue to push demand higher. And, with Rio Tinto and BHP Venturesh two of the largest names in global mining, Terra has built strong credibility in a technically demanding industry. 

If Terra can become the platform where more decisions are made, it has the opportunity to become an important player in the future of mining and energy development.