America Didn't Lose Its Rare Earth Industry, It Sold It: The CoCo Bond Proposal Explained

The United States did not lose its rare earth industry through carelessness — private American companies sold it, because selling was the economically rational choice, while China bought it for reasons that were never purely economic. That mismatch, argued in a Lawfare article that drew 156 points and 83 comments on Hacker News, is why years of subsidies and policy attention have not restored an alternative supply chain. The article's authors, who work at the intersection of finance and national security, propose a specific financing instrument to fix it: a contingent convertible security (a CoCo bond) that behaves like an ordinary loan in quiet years and converts into producer equity only if Beijing actually weaponizes supply.

The chokepoint in numbers: a $10 billion market that can halt trillions in GDP

China refines more than 90 percent of the world's rare earths and close to 99 percent of heavy rare earths, and it processes nearly all the heavy elements that go into precision-guided weapons and advanced chips. The entire market for rare earth oxides is worth less than $10 billion. That a market that small can grind to a halt trillions of dollars in global gross domestic product is what makes it a chokepoint rather than an ordinary commodity dependency.

Beijing controls that chokepoint by design, not by accident. The article traces the policy lineage to 1964, when Deng Xiaoping inspected an iron ore mine near the Mongolian border holding the world's largest rare earth deposit and ordered engineers to develop steel and rare earths alike. Successive Chinese leaders then deliberately invested in and consolidated control over the industry. Xi Jinping has openly discussed the importance of maintaining global dependency on China's industrial supply chains.

When Beijing tightened export controls in October 2025, Treasury Secretary Scott Bessent characterized it as "pointing a bazooka at the supply chains and industrial base of the entire free world." Those controls have since been suspended only one short reprieve at a time — the latest running until Nov. 10, with a broader trade truce through Jan. 10, 2027 — and the reprieves have repeatedly come only after American concessions. China has also constricted sales even during the truce: in June, with a suspension in force, Beijing placed MP Materials and USA Rare Earth, the two companies the US is counting on to build an alternative supply, on an export control list.

Why the US lost the industry: Mountain Pass and Magnequench were sold, not stolen

Until the late 1980s, the United States was the world's leading producer and refiner of rare earths. The Mountain Pass mine in California leaked roughly 600,000 gallons of radioactive and other waste into the Mojave. As cleanup and regulatory costs mounted and profits fell, the mine shut down in stages starting in the 1990s until it was fully idle by 2002. The work moved to China, where the state subsidized unprofitable operations, tolerated pollution, and sold processed oxides back to the US at rock-bottom prices.

The magnets followed the same path. In 1995, General Motors sold Magnequench, its magnet unit, to a Chinese-backed consortium. By 2006, the last US plant had closed and the machine tools and equipment had been shipped to China.

Economists would call this the correct outcome for a free-market economy: comparative advantage did what it was supposed to do, and a dirty, low-margin business left a wealthy country for a developing one. There was even a stated strategic rationale — drawing China into the trading system would, the wishful thinking went, make it a responsible stakeholder rather than a rival. The article calls this a category error: rare earths were conceived as a market rather than recognized as a geostrategic industry.

Rare earths are a market failure, not a market

A market solves for the lowest cost. A state will pay almost any price to control a chokepoint, because tolerating a money-losing business pales in comparison to the geopolitical leverage it brings. Over time, China has become less market driven, not more, while the US continued to treat as a market what the Chinese Communist Party came to appreciate as a strategic program.

Control over the chokepoint gives Beijing a switch with two settings:

• When China lets the material flow, no private interest will build alternative supply chains at scale, because Chinese state-funded producers can sell at or below cost indefinitely, and no rational private investor will sign up to lose money.
• When China cuts off the flow, incentives reverse overnight, but capacity cannot follow overnight, because standing up the full supply chain and rebuilding lost know-how can take a decade or longer.

The article calls this classic market failure in which the same critical material can become worthless or priceless on command. The precedent is 2010, when Beijing reportedly halted rare earth shipments during a maritime dispute with Japan and Tokyo quickly conceded. The short truces of the past year, each struck around episodic trade disputes, further entrench that market failure by keeping the US and its allies off balance and frustrating any long-term plan to reduce Beijing's leverage.

Government money has to survive politics, and the MP Materials deal blurs two roles

If the market will not rebuild the rare earths supply chain, the government must. Administrations under both parties have tried, and their attempts have run into a predictable morass of misaligned interests. The article lists the conflicts directly: private investors want attractive returns in fat years while having the government carry losses in lean ones; commercial lenders want years of fully contracted offtake to guarantee interest payments while Congress appropriates money one year at a time; markets want flexibility while Congress wants tight oversight and guardrails; and Washington often wants to be the last and lightest money in, which signals to investors that the government is hedging.

The Pentagon's July 2025 deal with MP Materials acknowledges that the market will not finance uneconomic redundancy on its own. It includes $400 million in convertible preferred stock, loans, equity warrants, a 10-year guaranteed price floor, and a 10-year offtake commitment. The article argues the deal was right for its moment — it gets the producer going and sustains it while Beijing floods the market — but that it blurs two different roles. In ordinary years the government is a lender financing capacity the market will not; in a crisis it has insured against a geopolitical event. Those two roles call for two different return profiles. Because the deal does not separate them, the government ends up simultaneously absorbing losses and inviting the charge of nationalization, all without a clear exit strategy.

Up to $100 billion in loan authority is specifically earmarked for critical minerals in the One Big Beautiful Bill Act, so the capital exists. What is missing, the authors argue, is alignment of interests among the parties that would supply and benefit from it, and a political system patient enough to hold a course for the decade or more that rebuilding a supply chain takes.

The proposal: a CoCo bond with two externally observable triggers

The authors propose a contingent convertible security — a CoCo bond — issued by the producer to the government and any private lenders who wish to join, paired with guaranteed offtake at a price floor. Its form changes with two externally observable conditions: the producer's capacity build-out and Beijing's behavior.

In its starting form, and in quiet years, the CoCo bond is a simple loan paying interest. Two triggers change its form:

1. Capacity trigger. The producer must cross a defined capacity threshold by a fixed date. Capacity can be measured as units of output in mining, refining, or manufacturing critical minerals. Management's immediate incentives are tied to the build-out itself.
2. Chokepoint trigger. Either a sustained breach of an agreed price band or an externally observed exercise of the chokepoint — denial of export licenses, a bureaucratic delay beyond a specified period, or another form of "constructive denial" that materially interrupts supply to the United States or allied countries. Critically, no part of this trigger relies on Beijing's confirmation, because Beijing is not incentivized to admit to a cutoff. The trigger must be narrow enough that it cannot be declared for political convenience, but broad enough that an adversary cannot game it simply by changing the form of the restriction.

When the first trigger fires, the lender receives a pro rata share of the producer's free cash flow — in effect, a dividend. The lender paid to build the capacity, so it shares in what that capacity earns. When both triggers fire, the lender also gains the right to convert into common equity. The conversion event does not create new capacity, because the financing has already done that work; it matches the payoff to the risk. The cash participation matters because in a crisis the market may treat a price spike as temporary and event-driven and not value the equity, but the spike still flows through the producer's cash flow, which benefits the lender either way.

Why would a producer agree? Because it gets something the market will not otherwise provide. In effect, the producer buys a put — the price floor that carries it through Beijing's dumping — and pays for it by selling a call — the cash share and the conversion right. It trades some of the windfall in a crisis for the protection of surviving quiet years. The taxpayer is paid for the risk actually borne, when it is borne, rather than for a risk merely feared.

Private co-investors, the Lynas precedent, and existing legal authority

The government does not have to finance the entire CoCo bond. Pre-vetted private co-investors, including foreign co-investors subject to Committee on Foreign Investment in the United States (CFIUS) screening, can offset government money dollar-for-dollar. Ideally, co-investors are the firms that fear a cutoff most, such as defense contractors and automakers. Private co-investors would receive a pro rata right to offtake as a further incentive and could potentially trade the bond with its associated offtake rights with other investors. Japan's JARE consortium did something similar when it financed Australia's Lynas in return for supply. Private holders also give the program a constituency that outlasts any single administration.

The article argues the instrument can start under existing authorities. The Pentagon has leaned on Title III of the Defense Production Act for its recent mineral deals involving equity. The Pentagon's Office of Strategic Capital can lend, but its equity authority remains unclear, and the Senate is currently revising the National Defense Authorization Act for 2027 to clarify those authorities. The authors argue that language should cover instruments like the CoCo bond.

The CoCo bond does not replace stockpiles, price floors, or offtake agreements. Those instruments solve significant but different problems: stockpiles buy time in a crisis, price floors protect producers against predatory pricing, and offtake agreements guarantee demand. The CoCo bond distributes risk and reward to investors in a binary, geopolitical event outside the producer's control, and it lets the government lend as a portfolio — some developers will fail, but the winners, and any crisis, can carry the portfolio as a whole.

Economically, the conversion right works like the equity warrants the Pentagon took in MP Materials. The difference lies in the quiet years, when under this structure the government holds no equity. Outright equity stakes invite the charge of nationalization; cheap financing with no upside smacks of corporate welfare. The authors argue a structure designed to avoid both at the outset — but flexible enough to change form in response to an adversary's actions rather than domestic political developments — has the best chance of surviving enough political cycles to complete a decade-plus build-out.

The Hacker News discussion: separation is the real bottleneck, and the last glut is the stress test

The most substantive pushback in the comments came from @Animats, who argued the binding constraint is separation, not mining, and that the article's proposed vehicle should be tested against history:

The current bottleneck is separation, not mining. There are at least ten startups claiming new separation technologies. They're flaky.

@Animats cited Rare Earth Technologies (about $1 million in investment and 7 employees per PitchBook and LinkedIn, with a site promising a pilot plant in 2025), Northern Minerals in Australia (an AU$0.02 penny stock still completing a Definitive Feasibility Study five years in), and Ucore Rare Metals (lab-scale samples and a demo plant, "more real than the two above" but "moving too slowly"). The comment also raised a point the article omits:

A big problem with this industry, which the original article should have mentioned, is that there was a rare earth glut in 2015, when China flooded the market. Many of the US players went out of business. Mountain Pass mine shut down. The article's proposed investment vehicle should be stress-tested against the historical data from the last glut to see if it holds up.

@leonidasrup challenged the premise of scarcity itself, noting that "rare-earth" is a misnomer: the elements are relatively plentiful in the Earth's crust — cerium is the 25th-most-abundant element at 68 parts per million, more abundant than copper — but are spread thinly as trace impurities, so usable purity requires processing enormous amounts of raw ore at high cost and energy intensity. The same comment flagged thorium, commonly found with rare earths, as the radioactive element that must be separated out.

@maxglute drew a sharper distinction between light and heavy rare earths, arguing that the strategic heavy rare earths used in sensors are geologically rare and economically extracted from limited ionic clay deposits, and that the entire heavy rare earth extraction pipeline is Chinese, built over more than 30 years. In that view, even if the US sorts out light rare earths at Mountain Pass, its strategic defense industrial base is built on heavy rare earths that only China can process at scale.

@BatchJob questioned the framing itself: a shutdown mine is a loss, because it is no longer financially viable to operate, and buying from someone who mines for pennies is the obvious choice. @OgsyedIE made the same point about price signals — energy, labor, pollution tolerance, legal costs, and economies of scale all said to go elsewhere — and noted that people who warned these signals only made sense if globalization lasted forever were dismissed as depressing.

@fmajid offered a simpler alternative to the CoCo bond: "The Feds should just start a rare earth company and contract out its operations. That's how enriched uranium production is handled." @tehjoker made the related political critique that the authors go to great lengths to avoid nationalization, and added that China is "basically in the right here. USA is being aggressive, China is responding."

@throw0101c pointed to a precedent for strategic materials being sold off: the National Helium Reserve, which the US government began selling down in the late 1990s, contributing to current helium shortages that matter for semiconductor lithography. @bubbluppugus generalized the argument beyond minerals — "America didn't lose manufacturing jobs it sold them" — and @dboreham noted the pattern applies to nearly every industry people now worry about being beholden to foreigners. @Georgelemental's only contribution was that the article reads as heavily AI-written.

The bottom line

A market will always sell a chokepoint to whoever runs it cheapest, and by market logic that is the right thing to do. Subsidies alone cannot win it back, because they only affect price, and Beijing prices for geopolitical leverage rather than profit. Neutralizing a chokepoint created through market failure requires capital willing to absorb years of subpar returns or outright losses against an extraordinary geopolitical crisis. The article's design intent is that in quiet years the government earns like a lender, with the price floor protecting against Chinese dumping; when capacity is built, it shares in what it financed; and if Beijing ever reaches for the switch, the taxpayers and private investors who paid to insure against that moment participate in the value their capital created. The risk is asymmetric, the authors argue, so the financing should be too.