Three Companies Decide Who Gets an MRI and Who Gets a Chip — and Nobody Elected Them

The MRI machine at a diagnostic centre in Peshawar sits in a cool, humming room. It works perfectly. It needs one thing to keep working: liquid helium, cooled to minus 269 degrees Celsius, circulating through its superconducting magnet.

That helium comes from Qatar. It isn't coming anymore.

---

On March 2, Iranian drones hit Ras Laffan Industrial City, the largest concentration of helium production infrastructure on Earth. QatarEnergy declared force majeure two days later. All exports stopped. A second wave of strikes in mid-March caused what QatarGas called "extensive damage" — repairs will take years, not months.

Thirty percent of global helium supply vanished in a matter of days. Two hundred specialised cryogenic containers sit stranded near the Strait of Hormuz with no viable maritime route to market.

Helium cannot be manufactured. It cannot be synthesised. There is no substitute in the two applications that matter most: cooling MRI magnets and cooling the lithography systems that print semiconductor chips.

The same gas diagnoses your cancer and makes your phone. When there isn't enough of both, someone decides who gets it.

The allocation nobody sees

That someone is not a government. It's not the UN. It's not a public health authority.

Three private companies control the global distribution of helium: Linde PLC, headquartered in Munich, with 24.5% market share. Air Products, based in Allentown, Pennsylvania. Air Liquide, in Paris. Together with a handful of smaller operators — Messer, Iwatani, Taiyo Nippon Sanso — they decide who receives helium when supply is tight.

The hierarchy, confirmed by supply chain intelligence firm Exiger, works like this:

First: semiconductor manufacturers. TSMC, Samsung, Intel. They hold the largest contracts, pay the highest prices, and produce the components that underpin the AI industry, military systems, and consumer electronics. They get helium first.

Second: aerospace and defence. NASA, SpaceX, military rocket programmes. Helium purges fuel systems and pressurises tanks. You can't launch a rocket without it.

Third: hospitals. MRI machines. Cardiac imaging. Tumour staging. Cancer diagnosis.

Fourth: scientific research. CERN. Particle accelerators. Quantum computing labs.

Fifth: industrial users. Welding. Leak detection. These get cut first.

Sixth: party balloons. Eliminated entirely.

Read that list again. Hospitals are third. Behind chips. Behind rockets. A Samsung memory fab and a hospital in Lahore both need helium this month. Samsung's contract is larger, its price tolerance is higher, and its relationship with Linde is longer. The fab gets the helium. The MRI machine does not.

No one votes on this. No government reviews it. It is a commercial decision made in boardrooms in Munich, Pennsylvania, and Paris.

Two industries, one gas, different victims

Samsung and SK Hynix source 65% of their helium from Qatar, according to semiconductor supply chain analysis by Carra Globe. Both companies are now rationing. SK Hynix told Reuters it has "sufficient inventory" — but a South Korean lawmaker, Kim Young-bae of the ruling Democratic Party, warned parliament that the war could disrupt supplies of key chipmaking materials, naming helium specifically.

Samsung declined to comment.

The chip industry's response has been to recycle. Advanced fabs now recapture 90-95% of helium used in certain processes. Linde, Air Liquide, and Air Products all sell on-site recovery infrastructure. But recycling cannot replace what isn't arriving. It stretches the supply; it doesn't create new supply.

Meanwhile, in India, the impact is arriving in hospitals. The Economic Times reported that the helium crunch is "beginning to strain MRI supply chains, pushing up costs and raising the risk of costlier scans and delays in diagnostics." India imports nearly all its helium from Qatar. The Indian Practitioner, a medical trade journal, warned that "hospitals and diagnostic centres may soon face the risk of higher MRI scan prices and potential delays."

The language is careful. "May face." "Potential delays." The reality behind the careful language: if you are waiting for an MRI to determine whether a lump is cancerous, and the machine needs helium that isn't coming, you wait longer. Some people will receive their diagnosis weeks late. Some of those weeks will matter.

The countries nobody mentions

India at least gets coverage. Indian media — Hindi and English — is reporting the MRI crisis as a healthcare emergency.

Africa is not.

Hospitals across sub-Saharan Africa have fewer MRI machines, smaller helium reserves, and weaker contracts with the gas distributors who allocate scarce supply. A diagnostic centre in Lagos or Nairobi is at the very bottom of Linde's priority list. Not because anyone decided to deprioritise African healthcare, but because the allocation follows contract size and price — and African hospitals buy less, pay less, and therefore receive less.

A supply chain analyst at Rare Earth Exchanges described the situation in developing countries as a "diagnostic blackout." Not a shortage. A blackout. Hospitals that have one MRI machine, serving hundreds of thousands of people, losing access to the only gas that keeps it running.

No English-language outlet has reported on this dimension. No African outlet we could find has covered it either. The intersection of helium supply chains and African healthcare is, as of this writing, entirely invisible.

Who profits

When supply is scarce and demand cannot move, whoever holds the gas holds the pricing power.

JPMorgan recently upgraded Linde stock to Overweight, specifically identifying the tightening helium market as a "primary catalyst." Linde's stock sits at $499 a share. The bank's thesis: Linde's US production base (centred on the Shute Creek plant in Wyoming) and its "proven pricing power" make it a clear winner in a shortage.

Analysts at TradingKey wrote that industrial gas suppliers "are positioned to benefit from increased pricing power and potential analyst upgrades." In plain language: the companies that control helium distribution make more money when helium is scarce. Their incentive to resolve the shortage quickly is limited by the fact that the shortage is profitable.

This is not a conspiracy. It is a market structure. Linde, Air Products, and Air Liquide did not bomb Ras Laffan. But they are the mechanism through which the bombing's consequences are distributed — and the distribution follows money, not need.

The Russia problem

There was supposed to be an alternative. Russia's Amur Gas Processing Plant, built to be a major global helium source, was meant to diversify supply away from Qatar. It hasn't worked. Plagued by explosions, technical setbacks, and Western sanctions, Amur is running well below capacity as of early 2026.

But something is shifting. South Korea — home to Samsung and SK Hynix — just purchased 27,000 tonnes of Russian naphtha for the first time in four years. The supply chain pressure is eroding sanctions compliance in real time. If helium from Qatar stays offline, the question becomes whether South Korea, Japan, and eventually the EU begin quietly sourcing Russian helium too.

China, which imports 85% of its helium (split between Russia and Qatar), has announced an emergency push to expand domestic production. TrendForce reported that Chinese output could reach 3 million cubic metres by end of 2026 — but that would cover only about 12% of China's needs. The rest must come from somewhere. Russia is the only option that isn't American.

The helium crisis is quietly becoming a sanctions erosion vector that nobody in Washington is tracking.

What happens next

New helium projects exist — in Saskatchewan, Tanzania, South Africa. All are three to five years from production. They will not help this year. They will not help next year.

The US is now the world's helium chokepoint. ExxonMobil's Shute Creek plant in Wyoming produces 42-47% of American helium. The Midcontinent basin (Texas, Kansas, Oklahoma) adds another 20-25%. Together, American production is the only major supply that didn't transit the Strait of Hormuz.

If the war ends tomorrow and QatarEnergy begins rebuilding Ras Laffan, the facility is years from full capacity. If the war doesn't end, the shortage deepens every month.

Semiconductor helium demand is projected to grow 15-20% annually through the end of the decade, driven by new fabs being built under the CHIPS Act and equivalent programmes worldwide. AI data centres are expected to consume 70% of memory chips produced in 2026. The demand curve is going one direction. The supply curve, since March 2, is going the other.

In Peshawar, the MRI machine still hums. The helium in its magnet will last weeks, perhaps months. When it runs out, the hospital will call its supplier. The supplier will check its allocation list. Samsung will be at the top. The hospital will not.

The machine will go quiet. The patients will wait. Nobody will report it.