The Arithmetic of Net Zero: The Largest Mining Boom in History

This series has set out four dependencies: Europe's on Chinese clean technology, Europe's on Russian materials it could not sanction, America's on a midstream it could not reshore, and the Global South's downstreaming that fed Chinese capital. Treated separately they look like four failures of political will in four capitals. They are better understood as four instances of a single law, and the law is a mismatch of calendars. Demand for transition minerals is set on a political calendar: targets for 2030 and 2035, movable by decree, responsive to elections and summits. Supply is set on a geological and industrial calendar: discovery, permitting, construction, commissioning, measured in decades and immovable by any vote. When a political calendar promises what a geological calendar cannot deliver, the shortfall does not vanish. It is filled by whoever already paid, years ago, to build the capacity. Today that is one country.

The largest mining programme ever attempted

Start by disposing of the word clean. The IEA's Net Zero scenario has clean-energy mineral demand almost tripling by 2030 and quadrupling by 2040, to around 40 million tonnes; on the 2024 scenario, lithium demand rises roughly ninefold and graphite fourfold by 2040, with nickel, cobalt and rare earths doubling and copper up by half. The older 2021 estimates against a 2020 baseline ran higher still, lithium more than fortyfold, before battery chemistry shifted. A typical electric vehicle needs about six times the mineral input of a combustion car; an onshore wind farm about nine times that of a gas plant of equal capacity. The World Bank puts a below-two-degree path at more than three billion tonnes of minerals and metals. Decarbonisation does not retire the mining industry. It conscripts it at a scale without precedent, which is the first half of the mismatch.

The supply calendar cannot be hurried

The second half is that supply runs slow and is slowing further. The IEA puts the average time from discovery to first production at about sixteen and a half years; S&P Global finds it rising toward eighteen for mines begun since 2020, and roughly twenty-nine years in the United States, second-slowest in the world. A mine sanctioned today produces around 2040; anything counted on for 2030 must already be in construction, and most of it is not. Copper is the binding case: after a long discovery drought, deposits found in the past decade hold only about 5 percent of all the copper found since 1990, and grades are falling. The IEA projects a roughly 30 percent copper supply shortfall by 2035, and S&P sees demand near 42 million tonnes by 2040 against supply peaking around 33 million. The metal that wires the entire transition is the one with no scenario for arriving on time.

Demand moves at the speed of a target; supply moves at the speed of a mine. The gap between the two calendars is filled by the only actor that pre-built the capacity.

Recycling answers the next decade, not this one

The circular economy is invoked to close the gap, and for the 2020s it cannot, because the material is not yet available to recycle. The electric-vehicle and battery fleet that would feed a recycling industry is still on the road; end-of-life batteries do not become the largest source of feedstock until around 2035. By 2040, recycled supply is projected to cover only about 5 percent of lithium and copper demand, 7 percent of nickel and 12 percent of cobalt. Recycling is a genuine and important supply story for the 2040s and 2050s. It is simply not available to backfill the 2030 and 2035 targets it is most often invoked to rescue, which is the calendar mismatch in miniature: the fix arrives a decade after the deadline.

The gap defaults to the incumbent

Put the two calendars together and the conclusion is mechanical. If demand multiplies, new supply takes sixteen years, and recycling is a decade late, the shortfall between target and metal can only be met from capacity that already exists. That capacity is overwhelmingly Chinese: the leading refiner for nineteen of twenty strategic minerals, with an average share near 70 percent, roughly 90 percent of rare-earth and anode-grade graphite processing, and over half of lithium and cobalt refining. The single most damning number is not any one share but their trend. The IEA finds the average market share of the top three refined-material suppliers was about 82 percent in 2020, rose to 86 percent by 2024, and is projected to be near 82 percent again in 2035. After a decade and more than a trillion dollars of Western industrial policy, the concentration is essentially flat. The calendars have not moved.

Even the apparent escape hatch leads back to the same place. The shift to lithium-iron-phosphate and sodium-ion chemistries genuinely cuts demand for cobalt and nickel, and it is the reason the 2024 demand multipliers sit far below the 2021 ones. But China makes around 75 percent of battery-grade phosphoric acid and 95 percent of high-purity manganese sulphate, so changing the chemistry swaps one Chinese chokepoint for another. The IEA's own verdict on the whole picture is unusually blunt for the institution: diversification will not materialise through market forces alone. Left to economics, the fastest and cheapest builder wins the gap, and that builder pre-committed its capacity a decade ago.

Where the argument is weakest

The thesis is strongest stated narrowly and weakest stated grandly, and the difference matters. Substitution is real and large: lithium-iron-phosphate went from under a tenth of the EV market to about half, and for nickel, cobalt, graphite and rare earths the IEA now sees supply broadly catching up with demand. The acute, durable gap is concentrated in copper and, into the 2030s, lithium, and in the processing step rather than the mine. Anyone arguing that every mineral is short forever at the old fortyfold multiples is arguing against the data. The defensible claim is narrower and harder to escape: lead times against deadlines, copper, and the processing lock are the parts that do not yield to substitution, and they are enough to carry the conclusion.

What would change this view

The argument earns its keep only if it can be broken, so here is what would break it. We would revise it if the top-three refiner concentration fell below roughly 70 percent by 2035 rather than holding near 82; if discovery-to-production lead times compressed materially through permitting reform; if recycled supply for lithium and rare earths arrived ahead of the mid-2030s schedule; or if Western processing investment sustained double-digit real growth instead of the roughly 2 percent of 2024. Each is measurable, and each is currently pointing the other way. That is why the conclusion holds.

What it reveals

The series resolves into one claim. The dependencies it documented are not, at root, failures of will in Brussels, Washington or Kinshasa, to be fixed by more resolve or a larger subsidy. They are the output of a transition whose demand multiplies on a political calendar while its supply runs on a geological one, with the difference made up by the single actor that pre-built the capacity. The transition does not dissolve dependence on China. On the current trajectory it is the machine that manufactures it. Decarbonisation remains necessary, but it is a resource-security event before it is an environmental one, and a strategy that treats the mine, the refinery and the magnet as someone else's problem is not a plan for agency. It is a schedule for dependence, and the schedule is already written.