Supply Chain Constraints

This week we decided to take stock of the supply chain across five critical American industries: energy, semiconductors, transportation, defense, and materials and mining (a layer that underpins almost everything).

Across these industries, we consider what is constrained, why it matters, what the alternatives are, and what would potentially alleviate it. While researching, we also identified five underlying causes of supply chain chokepoints:

• Engineered: An adversary built a deliberate chokepoint and operates it as a lever
• Eroded: Market forces (such as low margins or volatility) discourage participants, and capacity decays
• Concentrated: One or a few (allied) firms won on capability and scale
• Legislated: We did it to ourselves through regulation
• Certified: The constrained input is human capital, and training cycles are typically measured in years

Major chokepoints in our supply chains also tend to stack several of these causes at once.

To start with, let us consider the underpinning materials layer.

Materials & Mining

Why it matters: Every critical industry relies on it. For example, the neodymium-iron-boron (NdFeB) magnets refined from rare earths turn the motors in electric vehicles (transportation), the fins on guided missiles (defense), and the generators in wind turbines (energy). Castings and forgings made from specialty metals also become essential parts of every industry.

Why it is constrained: China’s dominance has been discussed extensively, and it is very real, but it is downstream. China mines only about 60% of the world’s magnet rare earths, yet it refines 91% and makes 94% of the magnets. Additionally, its grip on separating dysprosium and terbium (the heavy elements that let magnets survive heat) approached 100% until 2025, when Australia’s Lynas began producing the first supply outside China.

The refinement chokehold is a deliberately engineered one by China. For example, in November 2025, Beijing suspended its controls on gallium, germanium, antimony, and graphite anode materials (all of which can be backfilled by the US) while leaving the critical heavy rare earths and magnets restricted.

But the American rare earths industry did not decay solely because of Chinese aggression. Mountain Pass, the country’s only significant rare earth mine, was shut down in 1998 by an EPA action over leaking radioactive wastewater (a legislation chokepoint). Molycorp later reopened the mine in 2010, but went out of business in 2015 after China flooded the market. Additionally, certification is a major issue - only 14 accredited mining-engineering education programs exist in the US, down from 25 in 1982. In 2022, the DoD also highlighted that between “1984 and 2018, the U.S. lost 1,600 foundries [independent commercial plants] and over 4,400 metal casting facilities [all casting locations, including in-house factories].”

The alternatives and potential alleviation: The substitutes are real but imperfect and early. Recycling is the fastest, because magnet scrap already holds the dysprosium and terbium that take decades to mine and separate. Rare-earth-free magnets could sidestep China entirely at a cost in power density. Allied mines and refiners, like Lynas, are growing, but from a small base.

China’s main lever is price; it can flood the market and bankrupt Western entrants. To combat this, in 2025, the DoD secured a deal with MP Materials (a rare earth producer), which set a guaranteed price floor for magnets. This is a great temporary step, but a durable, defensible industrial base for rare earths and magnets will eventually require production that can be price-competitive.

Energy & Power

Why it matters: It is important to note what is not constrained: oil. U.S. crude production hit a record 13.6m barrels per day in 2025, and the country has been a net petroleum exporter for six consecutive years. America is typically not short on oil (though current reserves recently hit its lowest level since 1983 amid growing tensions with Iran), but what we lack is everything around it: the capacity to refine it, the ships to move it, and the turbines and steel to run a grid on anything else.

Why it is constrained: Refining has been eroded. Operable capacity fell by over 250k barrels per calendar day in 2025 to 18.2m. No new major US refinery has been built since 1977 (one is under construction), and distillate stocks sit at multi-year lows. As for the power grid, three Western manufacturers dominate the large gas turbine market, and GE Vernova's roughly 80 GW backlog stretches into 2029. Every large power transformer needs grain-oriented electrical steel (GOES), of which Cleveland-Cliffs is the only U.S. producer.

Every alternative to oil and gas has additional choke points: solar runs on polysilicon that China makes 93.2% of, wind relies on the same rare earth magnets as guided missiles, nuclear fission relies on uranium (primarily supplied by Russia until the 2024 ban), and grid storage of power relies on graphite anodes that China makes over 90% of.

The alternatives and potential alleviation: Currently, it is mostly imports and some workarounds. Refined oil shortfalls are covered by foreign imports and some renewable diesel. Transformer and turbine buyers currently wait in the buying queue or turn to behind-the-meter generation (electricity production at the point of consumption). The clean energy alternatives rely on the same backstops as in the first section: recycling, rare-earth magnets, and allied sources of polysilicon and enriched uranium. For GOES, there is no domestic alternative, only foreign steel.

Durable demand signals for energy would alleviate the constraints of refining turbines and transformers. The Jones Act (from 1920), which requires that all goods transported by water between U.S. ports must be carried on ships that are built in the U.S., owned by U.S. companies, and operated by U.S. citizens, is legislation that means the East Coast imports gasoline from the Netherlands, Belgium, and France rather than from Houston, which is a thousand nautical miles closer. That said, nuclear fuel shows the demand-side fix working: the $2.7bn Department of Energy award in January 2026 is helping to restart domestic enrichment.

Semiconductors & Electronics

Why it matters: Chips sit in everything from cars to interceptors. Beyond the leading-edge fabricators (TSMC) and lithography tools (ASML), the mundane materials and packaging are also a chokepoint.

Why it is constrained: A core example is Ajinomoto, a Japanese food and biotech company that invented monosodium glutamate (MSG), but also Ajinomoto build-up film (ABF), which insulates the packaged substrate beneath a high-performance processor - Ajinomoto commands over 95% of the market. Japan also happens to supply over 90% of extreme ultraviolet photoresists (light-sensitive chemicals used to print tiny chip designs). Unlike Engineered Coercion, these allied firms simply won in the market and built up concentration. Interestingly, though, trailing-edge chip supply, those used by cars, appliances, and most defense electronics, is dominated by China. At the end of 2023, China held about 31% of global legacy chip capacity, heading towards 39% by 2027.

The alternatives and potential alleviation: Novel materials would be the main potential alternatives. Glass substrates, for example, are a potential alternative for ABF. But these take time to develop, commercialize, and scale. Ajinomoto's hold on substrate packaging is also less of an issue because it is an allied Japanese company. As for China’s growing legacy chips dominance, this will require domestic and allied legacy fabs to increase production.

Transportation & Logistics

Why it matters: The machinery that moves goods is just as strategically important as the goods themselves. Without the ability to obtain products and materials in a timely manner, those goods become useless.

Why it is constrained: Roughly 80% of the ship-to-shore cranes at US ports come from one Chinese state-owned manufacturer, ZPMC, and a 2024 congressional investigation found cellular modems on some of them with no documented purpose, hardware the committee called a possible “Trojan horse.” Beyond this engineered risk from China, shipping production is also an issue caused by legislation and market erosion. US yards build well under 1% of the world's commercial tonnage, a collapse we traced in an earlier piece of ours on the US Navy crisis (A System Under Strain).

The alternatives and potential alleviation: Cranes can be purchased from allied makers in Finland and Japan, and a domestic line is now being rebuilt. A $20bn federal maritime initiative, for example, aims to replace Chinese crane equipment and to restart U.S. crane production through PACECO, a Mitsui subsidiary that built American cranes until the late 1980s. Ships can also be sourced from Allied yards in Korea and Japan.

Defense

Why it matters: Defense is an area where the other industries collide. It draws on the same magnets, chips, steel, and fuel, and adds munitions and platforms that have no civilian substitute.

Why it is constrained: Concentration is a major theme: six US solid rocket motor makers in the 1990s have now reduced to two; the country last made its own TNT in 1986; and Radford Army Ammunition Plant is North America's only military-grade nitrocellulose producer. Personnel is also an issue. For example, artillery production is roughly 56k 155mm shells per month against the 100k target. Virginia-class submarines are stuck at near 1.2 boats per year against a requirement of two, which the Navy does not expect to reach until 2032.

The alternatives and potential alleviation: A wave of new entrants is attacking the sub-tier that the primes hollowed out. Anduril and others are building solid rocket motors, Ursa Major and X-Bow on propulsion, and Hadrian on the precision-machined parts that feed every weapons line. The certified people shortage is the one issue money alone cannot rush. A welder, a machinist, or a shopfitter takes years to train, so the only real fix is to start the recruitment and training pipelines now and keep them funded through lulls.

‍

Takeaway: The U.S. is not short of raw materials. Across these industries, the inputs are almost always abundant; it is the downstream layers (conversion, tooling, people) where we are typically constrained. Four of the five causes of a chokepoint are ones we control: a business or market we let erode, a foreign producer we let concentrate, a person we never trained, or a law we implemented ourselves.

These causes also tend to compound; a law makes a business unprofitable, the last producers exit, the certified workforce disperses, and foreign rivals fill the vacuum. We expect the reshoring of the next decade to be judged less by the mines and fabs it announces than by whether the underlying causes were fixed.