Donald Trump’s quest for critical minerals has taken him from Greenland to Ukraine. He is not alone in wanting more metals, which will be needed for everything from increased electrification to more and bigger artificial-intelligence (AI) data centres and sturdier power grids. According to BCG, a consultancy, about a fifth of the minerals thought to be required by 2035 have yet to be found (see chart 1). Billions of dollars are being invested in trying to find new deposits, but few people know where all the required minerals will come from.
A host of problems with existing projects are only making matters worse. Ore quality—the amount of valuable metal in a sample of rock—is declining, forcing miners to dig deeper; mining permits can take years, if not decades, to finalise (see chart 2); and protectionism and protests are shutting down projects from Chile to Panama and Serbia.
Mining companies are therefore “desperate for new technologies to boost production”, according to Andrew Southam, the boss of KAZ Minerals, a copper miner in central Asia, who spoke at a recent gathering of industry bosses in Riyadh. Fortunately for Mr Southam and his beleaguered colleagues, a suite of new ideas—from better drills and power systems to cleverer uses of data and AI—aim to tackle their most pressing problems.
One set of innovations is focused on mining’s everyday operations such as drilling and processing ore. Miners are borrowing from the oil-and-gas industry, which underwent its own revolution in recent decades with the advent of hydraulic fracturing (fracking). That brought huge gains to the American economy by shrinking consumers’ bills and the country’s reliance on foreign energy.
With this in mind, XtremeX Mining Technology (XMT) has developed a drilling machine adapted from the ones used by big oil companies. Once mines are operational, drilling is the most capital-expensive and time-consuming part for miners, says Govind Friedland, an investor in XMT and a mining entrepreneur. XMT’s rig uses a coil tube—a long, seamless and flexible metal pipe—with a motor driven by alternating current (AC) to drill continuously, more precisely and faster.
The machine could replace the more traditional rotary drilling, a hydraulic-powered process that uses a turning and hammering motion to bore holes and which has been used in the industry for about 60 years. XMT’s drill could help miners go 50% deeper into the ground and use 20% less fuel in the process, because the AC motors are far more efficient than their hydraulic counterparts.
Power is another challenge. Mines require complex energy arrangements in remote areas, often becoming dependent on diesel generators that are dirty, expensive and hard to replenish. The absence of stable supply and grid connectivity also leaves miners vulnerable to intermittent power. Huawei, a Chinese company, has therefore developed the FusionSolar Smart Mine microgrid, a standalone power plant composed of a battery, solar panels and inverters, with a smart energy-management system that allows miners to use as much or as little power they need to run their operations, at a stable frequency and voltage.
A six megawatt-hour (MWh) version of the technology has been installed at an altitude of more than 4,200 metres at an off-grid mine in Argentina. A 90MWh system has been deployed in Mongolia where temperatures can drop to -40°C. The microgrid also uses AI to check on the battery health, predict any faults in the system and release the right amount of energy as needed. All that ends up cutting miners’ operational costs by about a third, says a Huawei executive.
As mines age, companies are also looking for better ways to get more value out of the dirt they dig up. BHP, the world’s largest mining company, and Rio Tinto, the second-largest, are testing new ways of leaching, a chemical process to extract more metal (in their case, copper) from a mineral ore. Because metals are found in different forms on Earth—oxides, sulphides, mixed together or in waste tailings—miners have to use different methods to extract them. Oxides require acids, and sulphides need a bacterial component to dissolve the extra material or oxidise minerals. Acidithiobacillus ferrooxidans, for instance, is used for oxidising ferrous iron and sulphur. Rio Tinto has also developed “Nuton”, its process of infusing crushed rock with proprietary additives and cultivated microorganisms that results in a high-purity copper.
The second group of innovations focuses on better collection of data to make exploring for minerals faster. This means finding higher-grade ores and better assessing and mapping deposits. Daoyun Tech, a Chinese company, makes software that allows companies to collect geological data of a prospective digging site via a drone, survey mining areas and come up with three-dimensional designs and plans. Ivanhoe Electric, a technology-focused exploration company, uses electric currents, which it shoots down to around 1.5km below the ground, to scope out deposits, map subsurface areas and detect sulphide minerals that contain copper, nickel, gold and silver. The precision of the system, which is used by BHP among others, means that less digging is required before deposits are located. It also means that less damage is done to surrounding land.
Freeport-McMoRan, a copper miner, is using sensors on trucks, shovels and machinery to collect data in real time about not just the quality of ore being dug out but the speed of its operations and the performance of its equipment. That information is then fed into AI models that direct Freeport’s machines more precisely and effectively, says Bert Odinet, the firm’s chief innovation officer.
At Freeport’s operations in Indonesia AI models also help predict wet muck spills, where water mixed with ore can rush into mines and endanger equipment and people. The company’s software crunches video and sensor data about water content, rainfall, and the distribution of mud deposits and particle size of the mud in and around the mine, to build its risk-weighted prediction algorithms.
Mining companies aren’t the only ones ratcheting up prospecting efforts through technology. In January, the United States Geological Survey (USGS) and America’s Defence Advanced Research Projects Agency (DARPA) launched a project to develop AI-infused tools to collect and assess the quality of critical-mineral data. The plan includes extracting geospatial data from a vast store of sources, such as about 100,000 old maps and geological surveys; and mapping minerals by georeferencing, a cartography technique used to align latitude and longitude map data with the real world. Only about a tenth of the USGS catalogue has been georeferenced; the rest are in the form of scanned images of maps. (The absence of such domestic data is perhaps why Mr Trump is venturing far and wide in his own search for minerals.)
All this has speeded up the otherwise labour-intensive geological surveying process that can take years and is mostly manual. After collecting, integrating and organising various sets of geoscientific data, the researchers apply machine-learning techniques to extract patterns, predict and forecast where minerals are likely to be buried across America. The goal is “to be able to characterise what we have, know what we have and source what we need domestically”, says Erica Briscoe, a DARPA programme manager.
How quickly mining and drilling companies will invest in all these new technologies, though, is unclear. They are already under pressure from rising labour costs, difficulties in accessing power and long delays in permitting. Then there are the unexpected hitches—according to McKinsey, a consultancy, for mines valued at more than $1bn, costs tend to end up being 80% higher than initial estimates.
To make matters more tricky, it will be years before many of these mining innovations become widely used and budget-friendly. Nevertheless, says Mr Friedland, the industry must start experimenting. “We have to change the way we work.” ■
Curious about the world? To enjoy our mind-expanding science coverage, sign up to Simply Science, our weekly subscriber-only newsletter.