Everyone agrees rare earths are the backbone of modern technology. The magnets in your GPU, the guidance systems in a Javelin missile, the superconducting motors in an electric vehicle — they all rely on neodymium, dysprosium, and a handful of other elements most people cannot pronounce. The popular narrative is that the US is finally waking up to this dependency, funding domestic mining projects to reduce its reliance on Chinese processing. Trump-backed miners are opening pits in California and Texas. It sounds like a victory for supply chain security.
But here is the trap. The policy looks right on paper, but the execution has produced a result that is not just neutral, but deeply counterproductive. While headlines celebrate new mines, the ore is being sold to Asian buyers — not American manufacturers. And the reason is chillingly simple: the US has no domestic smelting capacity to speak of, and its defense industry demand is not large enough to absorb the raw output. What we are witnessing is a policy that successfully subsidized mining, and then handed the product to the very processing ecosystem we sought to circumvent.
Let me use an analogy from my days auditing Ethereum bridges. When you see a smart contract that releases funds before validating the final state of the receiving chain, you know the architecture is broken. The US rare earth strategy is exactly that: a withdrawal before settlement. The code says “reduce dependency on China,” but the actual transaction flow says “export raw material to Asia for processing.” The gap between intent and execution is a reentrancy bug in national security policy.
The core of the failure is the bottleneck in separation and refining. Mining is the easy part — you dig rocks out of the ground. The difficult, capital-intensive, and environmentally nasty work is separating the individual rare earth oxides from the ore. China controls roughly 90% of this processing capacity globally. They built it over decades, tolerating the environmental cost while Western nations outsourced their heavy industry. The US now finds itself with shiny new mines and no smelters, like a miner who owns the land but cannot swing a pick.
Consider the data. The US Department of Defense has identified rare earths as critical for the F-35’s radar, laser gyroscopes, and precision guidance systems. Yet the FY2024 defense budget contains no dedicated line item for building a domestic rare earth separation facility. Instead, the administration has funded mine expansions through the Defense Production Act, but those mines are producing ore that can only be commercially processed in Asia. The result is a flow of material that, by default, ends up in Chinese supply chains. We are literally shipping our strategic minerals to the competitor we claim to be securing ourselves against.
Based on my stress-testing work during DeFi Summer, I learned to always model the failure case before the bull case. When we simulated a 40% drop in ETH collateral, we found that liquidation cascades would wipe out 15% of MakerDAO’s total collateral. The parallel here is brutal. If China were to impose export controls on processed rare earths tomorrow — a scenario I assess as highly probable in a Taiwan Strait crisis — the US would face a immediate supply gap for defense electronics. The domestic mines would not help. The ore would be sitting in stockpiles with no way to turn it into usable metal. The gap is not in raw tonnage; it is in the kilns, the solvents, the high-temperature furnaces that only exist in Baotou and Ganzhou.
Chaos is just data that hasn't been stress-tested yet. The current calm in rare earth prices masks a structural vulnerability. EV makers pay $50 a kilogram for neodymium magnets today. In a crisis scenario, that price could triple overnight. More importantly, the Pentagon could not simply substitute aluminum or steel. The physics of magnetic strength is immutable. You cannot build a 1,000-horsepower permanent magnet motor for a destroyer without neodymium. You cannot make a ring laser gyroscope accurate enough for hypersonic navigation without terbium.
Now let me offer a contrarian perspective. Many in the crypto and tech circles assume that supply chain decoupling is inevitable and happening. The data suggests otherwise. US rare earth exports to Asia have increased by 40% since the IRA was passed. The market is functioning as markets do: it seeks the cheapest path to the most efficient processor. That happens to be in China. The policy of “support domestic mining” failed to account for the fact that a mine without a linked smelter is just a hole in the ground that exports its contents. The irony is that the very projects designed to break dependency are, in their current form, actually reinforcing it by feeding raw material into the Asian processing loop.
What would a real solution look like? It would require the Defense Department to commit to multi-year contracts with a domestic separation facility, guaranteeing a price floor for processed oxides. It would require trade policy that imposes a tariff on raw ore exports to incentivize domestic processing. It would require admitting that the market, left to its own devices, will send the ore to where the existing infrastructure is. None of this is being done. The political path of least resistance is to announce a new mine, cut a ribbon, and claim victory. The engineering reality is that the bottleneck remains unsolved.
My time tracing the Luna collapse taught me that the most dangerous fragility is the one everyone assumes is fixed. In 2021, the narrative was that UST was a stablecoin because it was algorithmic and ‘decentralized.’ The reality was that the entire mechanism relied on a single point of failure: the LUNA-UST minting mechanism and the liquidity of Binance. Rare earth policy today has the same flavor. Everyone assumes that domestic mining equals supply security. The data shows the opposite. The gap between the mine and the smelter is the systemic risk that no one is auditing.
The takeaway for macro watchers and crypto investors is this: when the next liquidity cycle turns, and safe-haven assets rally, do not assume that the US has fixed its rare earth problem. The mines are running. The trucks are hauling. But the ore is going to the same place it always went. The only thing that has changed is that we are now paying for the privilege of being a supplier to our own competitor. The question to ask at the next industry conference is not “how much rare earth does the US produce?” It is “what percentage of that production is processed domestically?” The answer, today, is nearly zero. And until that changes, the entire narrative of supply chain decoupling is a fiction.
When the next crisis hits, the liquidity will vanish faster than the headlines can spin. The code doesn't lie. The ledger of global processing capacity is written in Chinese characters. We can either read it now, or pay the price later.