700 Million Tons

Everything you have ever touched was either mined or grown.

The chair you are sitting in. The phone in your pocket. The building around you. The road that brought you here. The wire that carries the signal that lets you read these words. The modern world runs on roughly eighty elements from the periodic table, extracted from the earth’s crust, refined, alloyed, shaped, and assembled into the infrastructure of civilization. The first iPhone contained perhaps fifteen of these elements. The iPhone 15 contains more than eighty. We did not become less dependent on the earth as our technology advanced. We became more dependent, on more elements, from more places, in larger quantities.

The supply chain that delivers these materials is the most important system most people never think about. It is invisible in the way that plumbing is invisible: you notice it only when it fails. And the conditions for failure are now converging from three directions simultaneously.

Three demands, one supply

Three defining trends of this decade, each enormous, each widely celebrated, each treated as essentially inevitable, all require the same critical raw materials.

The first is the green energy transition. Decarbonizing the global economy means replacing hydrocarbons with electrons. This is not a metaphor. It is a materials substitution of extraordinary scale. A typical electric vehicle requires roughly six times more minerals than a gasoline car, copper for the motor windings and wiring harness, lithium and cobalt for the battery, nickel for the cathode, rare earths for the permanent magnets. An onshore wind farm uses nine times more minerals per unit of output than a natural gas plant of comparable capacity. Solar panels require silver, silicon, indium, gallium. The grid infrastructure to connect it all requires aluminum and, above all, copper. To replace fossil fuels with renewables is to replace one set of extracted materials with another, and the replacement set is needed in quantities that dwarf current production.

The second is global rearmament. Europe, after decades of defense austerity, is moving toward spending targets of three to five percent of GDP on defense. Japan is building a military capability it has not possessed since 1945. South Korea, Australia, and the NATO alliance are all increasing procurement. Modern munitions are extraordinarily metals-intensive. A copper-headed artillery shell is not recycled when it detonates, the metal is consumed, scattered, irrecoverable. The munitions expended in Ukraine represent an enormous destruction of refined metals that must be replaced from primary production. Drone warfare, the defining innovation of this conflict, depends on the same minerals that power consumer electronics, plus specialized materials like scandium for next-generation communications. A world that is simultaneously rearming and electrifying is a world that needs more of everything at once.

The third is the AI and data center explosion. By 2026, global data centers will demand roughly one thousand terawatt-hours of electricity per year, approximately equal to Japan’s total electricity consumption. Every server rack in those facilities is a periodic table in miniature: gold for connectors, gallium and indium for semiconductors, tantalum for capacitors, palladium for multilayer ceramic components, tungsten for thermal management, niobium and titanium for structural alloys, rare earths for hard drives and cooling systems, antimony for flame retardants, silver for solder and contacts, barium for insulation. A single AI-powered search query consumes ten to thirty times more energy than a conventional search. The intelligence is abundant. The physical infrastructure to deliver it is not.

Here is the collision. All three trends draw from the same finite pool of critical minerals. You cannot simultaneously green the world economy, arm it for a new era of great-power competition, and build the compute infrastructure for artificial intelligence, not without a dramatic increase in the supply of raw materials that no one has explained how to achieve.

The copper math

Copper makes the physics concrete.

Humanity has mined approximately seven hundred million metric tons of copper in its entire history, ten thousand years of extraction, from the first crude smelting in Anatolia to the modern open-pit operations of Chile and the Democratic Republic of Congo. If you gathered all of it into a single cube and placed it next to the Eiffel Tower, the cube would be substantial but finite. It would have a side length of roughly forty-three meters. That is all the copper the human race has ever produced.

Now consider what comes next. To maintain three percent global GDP growth, not to electrify anything new, not to build additional data centers, not to accelerate the energy transition, just to sustain the existing economic trajectory, the world needs another seven hundred million metric tons of copper in the next eighteen years. The same amount. In a fraction of the time. This is the estimate from S&P Global’s comprehensive study on the future of copper, and it is broadly accepted by industry analysts and mining companies.

The reason copper is irreplaceable is physics, not economics. Copper conducts electricity better than any element in the periodic table except gold and silver, both of which are far too expensive and too scarce to serve as industrial conductors. Every power cable, every grid connection, every EV charging station, every wind turbine nacelle, every data center power distribution unit, copper. The copper wire market alone is a two-hundred-sixty-seven-billion-dollar-per-year industry, and the demand curve is heading straight up.

But the supply side is moving in the opposite direction. Since 1900, the energy required to produce a pound of copper has increased sixteenfold. Water consumption has doubled. The reason is geological: we mine the easy deposits first. The high-grade ores that the industry exploited for most of the twentieth century are largely exhausted. What remains is deeper, lower-grade, more energy-intensive to extract and process. Each successive ton of copper costs more energy, more water, and more capital than the last.

The industry estimates it needs roughly six new tier-one copper mines per year through 2050 to meet projected demand. A tier-one mine is defined by scale, grade, and cost, the kind of asset that can produce over two hundred thousand tons per year at competitive margins. Nobody in the industry can identify where these mines will come from. Discovery rates have been falling for decades. Permitting timelines have been lengthening. Cornell University estimates that copper prices must roughly double from current levels to incentivize the exploration and development pipeline needed to close the gap.

The copper problem is not a copper problem. It is a template. Lithium, cobalt, nickel, rare earths, gallium, germanium, each has its own version of the same arithmetic. The demand is known. The supply is not.

China’s hand on the valve

While the democratic world was building the internet, China was building the supply chain for the physical world.

Over the past thirty to forty years, China has systematically constructed dominance over the extraction, processing, and refining of critical raw materials. This was not accidental. It was strategic industrial policy executed across decades, investment in domestic mining, acquisition of mining assets abroad, construction of processing facilities, development of technical expertise, and cultivation of relationships with resource-rich countries across Africa, South America, and Central Asia.

As of 2025, China has placed export restrictions on gallium, germanium, antimony, tungsten, tellurium, bismuth, indium, and molybdenum. These are names that most people have never heard, but the modern economy cannot function without them. Gallium is essential for semiconductors and LEDs. Germanium is critical for fiber optics and infrared optics. Antimony is used in flame retardants for military applications and in lead-acid batteries. Tungsten provides the hardness in cutting tools and armor-piercing munitions. Indium is indispensable for display technology. Molybdenum strengthens the steel used in pipelines and reactors. Tellurium enables cadmium telluride solar cells. Bismuth serves applications ranging from pharmaceuticals to nuclear shielding.

China controls not just the raw extraction but the processing. Even minerals mined in other countries often flow through Chinese refineries because no comparable processing capacity exists elsewhere. The chokepoint is not at the mine, it is at the refinery, the smelter, the chemical plant that converts raw ore into usable industrial materials.

The strategic implications are stark. China could halt US automobile production within weeks by restricting a handful of materials. This is not trade leverage of the conventional kind, tariffs, quotas, price manipulation. It is structural control over the physical layer of the modern economy. The analogy to OPEC and oil is imprecise but directionally correct, with one critical difference: oil has substitutes (natural gas, nuclear, renewables). For many of these minerals, there are no substitutes at any price.

The United States and Europe have begun to respond. The EU’s Critical Raw Materials Act, bilateral mining agreements with resource-rich nations, executive orders on critical mineral supply chains, these are real but nascent. The challenge is temporal. A mine takes fifteen to twenty years from discovery to first production. Processing facilities require billions in capital and years of construction. Environmental permitting in democratic countries adds additional years. China’s advantage is not a policy choice that can be matched by a countervailing policy choice. It is the accumulated result of four decades of strategic investment. It cannot be replicated on any timeline that matters for the current decade’s challenges.

The mining paradox

Here is the uncomfortable truth at the center of this story: mining itself is part of the problem it is supposed to solve.

Roughly four to five percent of global greenhouse gas emissions come from the simple act of crushing and grinding rock. Not from the trucks or the diesel generators or the transportation of ore to refineries, just from the mechanical process of reducing large rocks to particles small enough to separate the valuable mineral from the worthless gangue. This is one of the most energy-intensive industrial processes on earth, and it scales directly with the volume of rock processed.

The lower the ore grade, the more rock must be moved and crushed to extract the same amount of metal. As grades decline, and they are declining globally, across nearly every commodity, the energy required per unit of output increases. More energy means more emissions. To green the world economy, you need metals. To get those metals, you generate emissions. And as the easy deposits are exhausted, you generate more emissions per ton. It is a thermodynamic trap.

Unless the act of mining itself is transformed.

This is where innovation matters, not the incremental kind, but the kind that changes the fundamental physics of extraction. Pulse-power rock breaking uses electromagnetic pulses to fracture rock from the inside out, rather than crushing it with brute mechanical force. Early results suggest energy reductions of up to eighty percent compared to conventional grinding. Direct lithium extraction pulls lithium from oilfield brines, the produced water that the petroleum industry has been disposing of as waste for decades, bypassing the evaporation ponds that dominate current production in South America. Enhanced geothermal systems tap the heat that exists three miles below the earth’s surface everywhere on the planet. The energy stored in the earth’s crust at accessible depths represents roughly five times current global energy consumption. The crisis is not the absence of energy. It is the cost of drilling to reach it.

These are not theoretical concepts. They are in development, approaching commercialization, backed by major mining and energy companies. The question is not whether they work. It is whether they arrive at scale in time to matter, in time to change the denominator of the equation before the numerator overwhelms it.

The grid nobody built

Even if every ton of metal were mined and refined tomorrow, there is nowhere to send the electricity.

The United States electrical grid is, in places, more than a century old. The power line that caused the Camp Fire in Paradise, California in 2018, the deadliest wildfire in the state’s history, eighty-five people killed, was one hundred and six years old. It was built in 1912, when Woodrow Wilson was running for president and the Titanic had not yet sunk. It was still carrying power in 2018 because nobody had replaced it.

This is not an isolated case. It is the condition of the grid. The American Society of Civil Engineers gives the US energy infrastructure a grade of C-minus. Transmission capacity has not kept pace with demand growth, and the interconnection queue for new generation projects, solar farms, wind farms, battery storage, is measured in years. New generation capacity exists. It simply cannot connect to the grid fast enough to matter.

China, by contrast, is spending eight hundred billion dollars to modernize its transmission infrastructure. The strategic clarity is striking: you cannot electrify an economy without a grid capable of delivering the electrons. The US and Europe invested trillions in broadband, cloud computing, and digital infrastructure, the technologies that generate excitement and venture capital, while systematically underinvesting in the physical grid that powers all of it.

Jensen Huang, the CEO of NVIDIA, made the observation that the cost of artificial intelligence will ultimately converge on the cost of energy. Intelligence is becoming abundant, the models are improving, the algorithms are more efficient, the hardware is scaling. What is not abundant is the energy to run the hardware, and the grid to deliver the energy, and the copper and aluminum that compose the grid. The limiting factor on the AI revolution is not software. It is copper wire.

The constraint

The constraint on our era is not imagination. It is not capital. It is not policy. It is material.

The green energy transition, the AI revolution, and the rearmament of the democratic world are all serious ambitions backed by serious investment and serious political will. They are also all drawing from the same finite pool of physical resources, processed through supply chains that a single country dominates, extracted from an earth whose richest deposits have already been mined.

The math does not add up under current conditions. Something has to give. Either the pace of one or more of these transitions slows, or the efficiency of extraction improves dramatically, or the supply chain itself is restructured, and each of these adjustments operates on a timeline measured in decades, not quarters.

The optimistic case is that mining innovation, lower-energy extraction, geothermal energy, direct processing of unconventional sources, can change the denominator fast enough. That pulse-power fracturing and direct lithium extraction and enhanced geothermal drilling can bend the curve before the demands of decarbonization, defense, and data centers collide in earnest.

The pessimistic case is that geology is a harder constraint than policy. That ore grades will continue to decline, that permitting will remain slow, that China’s supply chain dominance will persist for decades regardless of countervailing legislation, and that the collision between these three demands will be resolved not by engineering but by scarcity, by the market allocating insufficient supply to the highest bidder while the rest waits.

Either way, the constraint is physical. The modern world is built on a layer of extracted materials that most people never think about until someone threatens to cut it off. Three enormous forces are now pulling on that layer simultaneously, and the layer is not infinitely elastic.

Physics, unlike policy, is not negotiable.