On October 27, the Department of Energy added helium to its Critical Material List. Under federal law, that label is reserved for a material that faces a high risk of supply disruption and performs an essential function in energy technology.
Helium qualifies on both grounds. DOE's notice points to semiconductor manufacturing, natural-gas leak testing, quantum computing, and medical and scientific uses. It also connects future demand to the expansion of semiconductor production for AI systems and data centers.
The designation gives helium more policy attention. It does not create another refining plant, fill a storage cavern, shorten a permit review, or guarantee a delivery to a hospital. That gap between recognition and usable supply is the part worth understanding.
Some jobs still require helium
The public association with party balloons hides helium's harder jobs. Its extremely low boiling point makes it useful where equipment must operate at temperatures that other coolants cannot reach. The U.S. Geological Survey says nothing can substitute for helium in cryogenic applications below -429°F.
That includes the superconducting magnets used in MRI scanners and high-field research instruments. Helium also provides controlled, inert environments in parts of semiconductor manufacturing, helps detect leaks in pressurized systems, and supports aerospace and laboratory work. Some of these applications can reduce use or switch gases. The coldest cryogenic work cannot simply choose another product when helium becomes expensive or scarce.
The numbers show how widely the gas is already spread across ordinary systems. USGS estimated that 15% of U.S. helium use in 2025 went to MRI. Controlled atmospheres, fiber optics, and semiconductors accounted for 17%. Analytical, laboratory, scientific, and specialty-gas uses made up another 22%.
The effects are already visible in research. A 2024 National Academies report found that unstable supply and price swings had harmed scientific and technical communities that rely on superconducting magnets. It associated helium access problems with closures of nuclear magnetic resonance laboratories, equipment used in chemistry, biology, materials science, and drug research.
A large reserve is only the first link
The United States is not running out of helium in the ground. USGS estimated 8.49 billion cubic meters of recoverable helium in identified domestic natural-gas reservoirs. In 2025, the country sold or used about 81 million cubic meters, imported 8 million, exported 38 million, and remained a net exporter. Six new U.S. helium operations began producing that year.
Those totals can sound reassuring. They cannot tell a hospital in Colorado, a chip plant in Arizona, or a university laboratory in Massachusetts whether the next shipment will arrive.
Helium often appears in natural-gas deposits at concentrations that must be economically captured. It then has to be separated, refined to the needed purity, stored, loaded into specialized containers, and moved to the customer. USGS counted nine U.S. plants producing crude helium in 2025, five producing Grade-A helium, four purifying helium from other crude sources, and three underground storage locations in Texas.
A constraint at any one of those stages can turn geological abundance into a delivery problem. DOE says limited refining capacity and lengthy permitting still restrict domestic production. Global routes matter too. Marketplace reported in late October that the closure of the Strait of Hormuz had interrupted supply from Qatar, one of the world's largest producers.
Helium consultant Phil Kornbluth called the current episode "helium shortage 5.0," adding, "We don't name them very creatively. It's 1.0, 2.0, 3.0, 4.0, 5.0." The count shows that disruption is a recurring feature of this market. New domestic projects also face financing hurdles before they can add dependable supply.
The AI boom adds pressure to a shared resource
DOE tied the designation partly to semiconductor growth for AI infrastructure. That connection deserves attention without turning AI into the sole explanation for helium risk.
Chipmakers were already competing with hospitals, aerospace companies, welders, laboratories, and other industrial users. More semiconductor production adds demand to a system with limited refining, storage, and transport capacity. Richard Brook of Garrison Ventures put the manufacturing consequence plainly in Marketplace: "If they can't get it, they can't make the parts."
This is one more example of digital expansion creating physical demand somewhere else. A new data center appears on a planning map as land, electricity, cooling, and network capacity. The chips inside it also depend on materials and process gases with their own production geography. The software may scale quickly. The supply chain for a purified gas does not.
That does not mean AI projects should receive priority over medical care or research. It means public officials and major buyers need to know which uses can conserve, recycle, substitute, or wait when supply tightens. USGS says large-volume helium applications in the United States were seldom recycled in 2025, although closed-loop recovery systems were becoming more common. Recovery will not solve every shortage, but wasting a hard-to-replace gas makes less sense as demand grows.
What the designation cannot prove
The new label does not measure how long the current shortage will last. It does not show that a specific hospital faces an MRI interruption, or that every chipmaker is short of helium. The National Academies studied high-field science rather than the entire market. Marketplace's account of Qatar and the Strait of Hormuz describes a fast-moving situation that may change.
The designation also carries no automatic promise of money, faster permits, lower prices, or new capacity. Those outcomes require later decisions by agencies, lawmakers, producers, and buyers. Even a new project takes time to finance, permit, build, and connect to customers.
What the evidence does support is narrower and more useful. Helium supply has failed repeatedly. Some important uses have no practical substitute. Large domestic reserves do not ensure local delivery. Demand from semiconductor expansion is arriving alongside existing medical, scientific, and industrial needs.
Ask for the supply plan
Any hospital, university, manufacturer, or public agency that depends on helium should be able to answer a short set of questions. How much does it use during a normal month and at peak demand? How many suppliers can serve the site? How long would the on-site buffer last? Which uses receive priority during an allocation? Can any helium be recovered, recycled, or replaced?
Those answers will not produce another cubic meter of helium. They will reveal whether an institution understands its exposure before the next disruption arrives.
The practical question for readers is simple: if a local hospital, research center, or major public project depends on helium, who owns its supply and conservation plan, and when was that plan last tested?