Helium

Helium is the strangest entry in this book, because it is the one commodity that is genuinely non-renewable on a human timescale in a physical sense rather than an economic one. It is produced underground by the radioactive decay of uranium and thorium over hundreds of millions of years, accumulates in a small number of natural gas fields, and is so light that once released into the atmosphere it drifts to the top and escapes into space. Every balloon that pops is helium permanently off the planet. Nothing else traded here behaves that way: a barrel of oil burned becomes carbon dioxide that stays in the system, and a tonne of copper is recyclable forever.

It is also, like sulphur, an involuntary byproduct. Nobody drills a well for helium. It occurs at fractions of a percent in certain gas streams and is stripped out during cryogenic processing, which means helium has no supply curve of its own: output is set by how much natural gas somebody else decided to produce, and by whether their plant happens to have a helium unit bolted on. A high helium price does not summon new helium. It can only, over years, justify adding recovery capacity to gas projects that were going ahead anyway.

Ordinary gas fields mostly cannot supply it. Most commercial natural gas carries well under 0.3 percent helium, around the point below which a standalone recovery unit rarely pays, which is why American production clusters in a few unusually rich fields. The largest single source is ExxonMobil's Shute Creek plant at LaBarge, Wyoming, where the gas runs about 0.6 percent helium; Wyoming supplied about 43 percent of US helium from 2000 to 2012, and the plant is often credited with around a fifth of world supply. The older heartland is the Hugoton-Panhandle field across Kansas, Oklahoma and the Texas Panhandle, which fed the federal pipeline, and six new operations started in 2025, three in New Mexico and one each in Colorado, Kansas and Montana. Qatar is the exception: North Field gas holds only about 0.04 percent helium, but liquefaction concentrates it roughly tenfold in the vapour left behind when the methane condenses, and at Ras Laffan's volumes that lean stream becomes a third of the world's supply.

That raises the obvious question of why the American LNG boom has not added a second Qatar, and the answer is the feed gas. US export terminals liquefy pipeline-quality gas that has already been through processing plants hundreds of miles upstream, drawn from basins whose helium content is low and diluted further by mixing in the pipeline grid, so the vapour at the terminal is not rich enough to pay for a recovery unit and is burned as fuel. Qatar's advantage is that a single helium-bearing field feeds its liquefaction trains directly. A helium price high enough to change that calculation at an American terminal has not lasted long enough for anyone to build for it.

Helium is scarce for three reasons that compound. It is genuinely rare on Earth, made only by the slow alpha decay of uranium and thorium and kept only where a gas seal happened to trap it, so the atmosphere holds about five parts per million. Almost nobody looked for it: until about a decade ago nearly every discovery was an accident of drilling for gas, and the first projects targeting helium in its own right, in Tanzania, South Africa, Saskatchewan and Montana, are recent, some in nitrogen-rich gas with no hydrocarbons in it at all. And it is hard to move and harder to keep: it ships as a liquid about four degrees above absolute zero in insulated ISO containers holding roughly 11,000 gallons, which keep it liquid for something like 35 to 45 days before it starts venting, so a cargo held up at a chokepoint is losing product the whole time.

Inventories exist but are small against a third of world supply, and slow to draw. The largest is still the Cliffside field near Amarillo, a natural reservoir that now holds the crude helium Messer bought along with privately owned helium stored there by the industrial gas companies, about 2 billion cubic feet of private gas alone at the last federal count in 2022. Air Products and Linde each store helium in salt caverns at Big Hill in Texas, and a further cavern at Beaumont came online in 2025. Against world demand near 190 million cubic meters a year, about 6.9 billion cubic feet, that is months of cover rather than years, and Cliffside's crude is 60 to 80 percent helium that has to be refined before use, so the rate it can be delivered at matters as much as the volume in the ground. The reserve used to be the buffer of last resort; its sale moved that buffer onto private balance sheets, where it is used to protect the owners' own contracts first.

The uses are unglamorous and almost entirely non-substitutable, which is why shortages bite so hard. The largest is cryogenics, above all cooling the superconducting magnets in MRI scanners, because helium is the only element that stays liquid down to absolute zero. The second is semiconductor and fibre-optic manufacturing, where it is used as an inert purge and coolant, which makes it a quiet input to the entire electronics supply chain. Then come welding shields, leak detection, rocket tank pressurisation, and, far down the list and first to be rationed, party balloons.

The quantities explain who feels a shortage first. A conventional MRI magnet sits in a bath of about 1,500 litres of liquid helium, and since a litre of liquid becomes about 0.75 cubic meters of gas, that one fill is roughly 40,000 cubic feet, enough for some 80,000 party balloons at half a cubic foot each. The scanner does not use it up by scanning: a cryocooler on the magnet re-condenses most of what boils off, so a modern machine may need topping up only every few years. The real loss is a quench, when the magnet stops superconducting, by accident or because someone hits the emergency button, and the whole charge boils off through a vent in the roof within seconds. Newer sealed magnets are designed around that risk and the price: Philips' BlueSeal holds about 7 litres for life, and Siemens Healthineers' DryCool about 0.7 litres. Chipmaking uses helium continuously and as a gas. During plasma etching and deposition it flows between the back of each wafer and the chuck holding it, carrying heat away so the wafer stays at the temperature the process needs, a step repeated hundreds of times per wafer. A hospital with a modern scanner can therefore run for years on the helium already in its magnet, while a fab needs fresh deliveries all the time.

Supply is concentrated to a degree that would be alarming in any other market. The United States, Qatar, Algeria, and Russia account for the overwhelming majority, with Qatar alone around a third and its output attached to the North Field gas complex at Ras Laffan. That is the geographic problem: Qatari helium leaves through the Strait of Hormuz, so a chokepoint crisis is a helium crisis. The 2026 closure demonstrated it, taking roughly a third of world supply out and triggering the fourth named helium shortage this century.

There is no futures contract and no public price. Helium is sold under private long-term contracts negotiated bilaterally between a handful of producers and a handful of industrial gas majors, and the numbers are not published. The nearest thing to a public number is the US Geological Survey's estimated base price for Grade-A helium, about $330 per thousand cubic feet in 2025, roughly $70,000 a tonne, to which producers add surcharges; reported delivered levels ran around 95,000 to 98,000 dollars a tonne through 2026. Both are estimates assembled from private deals rather than a settlement anyone can trade against. The reasons no contract has formed are the familiar ones on this site, stacked: a byproduct with inelastic supply, a tiny number of counterparties, a product that is expensive and lossy to store and ship, and grades and delivery terms that differ from contract to contract.

The result is a market that is repeatedly described as critical and is priced as though it were private. When the reserve that used to absorb shocks has been sold, the supply is a byproduct nobody controls, and a third of it sails through a contested strait, the absence of a price signal stops being a curiosity and becomes the problem.