In Q3 2024, US rare earth miners shipped over 15,000 metric tons of ore to Asian buyers. Simultaneously, the Department of Defense quietly flagged a shortage of neodymium magnets for F-35 actuators. This is the market’s way of telling you that the US mining strategy is a failed hedge. When the code bleeds, only the ledger survives. Here, the code is the supply chain; the ledger is the balance of strategic power.
Let me be direct: I’m not a geopolitical analyst. I’m a cryptographer who audited smart contracts during the 2017 boom, migrated $150,000 into Uniswap V2 pools in 2020, and modeled Layer-2 gas costs during the Axie Infinity frenzy. I learned that the weakest link is often hidden in plain sight. The US rare earth policy has a similar flaw: mining is visible, processing is the hidden vulnerability.
Context: The Infrastructure Gap
Rare earth elements (REEs) are the silicon of defense. They power the F-35’s radar, missile guidance systems, laser gyroscopes, and the permanent magnet motors in nuclear submarines. The US holds significant reserves—Mountain Pass in California is the largest operating mine outside China. But here’s the catch: the US has almost zero domestic processing capacity. China controls roughly 90% of global rare earth refining, a monopoly built on decades of environmental cost tolerance and strategic investment.
In 2018, the Trump administration declared rare earths a national security priority, subsidizing mines like MP Materials (which restarted Mountain Pass). The goal was clear: reduce dependency on China. But the policy only addressed upstream mining, not downstream processing. The result is a supply chain paradox: the ore leaves US shores, is processed in Asia (often by Chinese facilities), and returns as expensive oxides—or worse, feeds competitor supply chains.
This is precisely what I saw in the 2017 Symbiont audit. Their smart contract had a reentrancy bug that looked harmless until you traced the state transitions: funds moved out, then back in at a higher cost. The rare earth supply chain has the same pattern. Domestic ore flows out, processed materials flow back in, and the fees are paid to the middleman.
Core: The Order Flow Analysis
Let’s quantify the inefficiency. I’ll use the same risk-adjusted return framework I built for DeFi yield farming. The US mining policy is a capital allocation problem with three layers:
Layer 1: Mining – Capital deployed: ~$2 billion in subsidies and tax breaks since 2020. Output: ~40,000 metric tons of rare earth oxide equivalent per year (MP Materials alone). This is a high-capex, low-margin business. The internal rate of return (IRR) depends entirely on the selling price of concentrate.
Layer 2: Processing – Capital deployed: negligible. No domestic facilities of scale. The US has one pilot separation plant (Mountain Pass’s own upgrade) that covers <2% of demand. The cost to build a full processing facility: $1-2 billion, 5-10 years timeline. China’s processing capacity: >80% global share.
Layer 3: Application – Defense, electronics, EVs. US demand for refined REEs is ~15,000 tons per year, growing at 8% CAGR. But domestic consumption of raw ore is near zero. American manufacturers buy processed oxides, not ore.
The yield on the mining investment is essentially negative. Here’s the math: If the US mine sells concentrate at $5/kg and it costs $3/kg to produce, the gross profit is $2/kg. But that concentrate is exported to Asia, where Chinese processors transform it into oxides worth $50/kg. They capture 90% of the value add. The US taxpayer subsidizes the upstream, while China reaps the downstream profit and strategic leverage.
Apply the concept of impermanent loss from Uniswap. In 2020, I lost 12% during a volatile July spike because I didn’t account for divergence between the two assets. Here, the two assets are “domestic mining output” and “domestic processing capability.” The divergence has widened: mining increased 400% since 2018, processing capability increased <5%. The impermanent loss is strategic autonomy.
During the Axie Infinity gas war, I modeled Layer-2 solutions and found that Optimism’s rollup reduced fees by 90% but introduced a 7-day challenge period. The rare earth processing bottleneck is a similar rollup: you can export the raw data (ore) and get back the final result (oxides), but you’re dependent on the sequencer (China) to finalize. And that sequencer can reorder transactions or censor them at will.
Contrarian: The False Comfort of Domestic Mining
The conventional wisdom is that more domestic mining is a net positive. I disagree. It’s a distraction that creates false security and misallocates capital. The energy spent subsidizing mines could be better spent building processing plants. Instead, the US is subsidizing China’s processing dominance by providing cheap feedstock.
Let me draw from the Celsius collapse. In 2022, I had coded a Python script to monitor Aave and Compound liquidation thresholds. The yield models of Celsius looked sustainable on the surface, but the underlying collateral was undercollateralized. The rare earth policy is undercollateralized in processing. The apparent “liquidity” of domestic ore masks the illiquidity of strategic control.
Moreover, the market is efficient. If the US doesn’t consume raw ore, it will be exported. That’s not a bug; it’s a feature of free markets. But for national security, you must override market signals. The US government could impose export controls on rare earth ore, but it hasn’t. Why? Because the policy is designed to support mining companies, not defense. The signal is clear: the government is not serious about supply chain security.
Compare to the AI-agent trading protocol I designed in 2025. That system executed 10,000 trades per day on Solana, integrating LLM sentiment with deterministic execution. The key was bridging the gap between experimental AI and robust logic. The US rare earth strategy has no bridge. It’s a monolith of mining without the middleware of processing.
Takeaway: The Verdict
The US rare earth strategy needs a fundamental rearchitecting. The current approach is akin to a DeFi protocol with infinite minting but no mechanism to convert the token into a stable asset. You mine the token, but you can’t spend it. The only way to unlock value is to build the processing layer, but that requires 10 years and political will that does not currently exist.
I do not trust whispers; I trust verified hashes. The hash of the US rare earth policy is a hash of failure. Until the government commits to domestic processing, the code bleeds and only the ledger of Chinese dominance survives.
The market has already priced this in. Rare earth mining stocks trade at low multiples, while Chinese processing companies command premium valuations. The arbitrage opportunity? Not in mining. It’s in processing technology, recycling, and rare earth substitutes. But that’s a bet on a decade-long cycle, not a quarterly trade.
In the long run, the US will either build its own processing capacity or face a supply chain catastrophe during the next major conflict. The gas war taught me that speed is a tax. The processing lag is a tax on national security. Pay it now, or pay it later with interest.
The ledger doesn’t lie.