America’s manufacturing Achilles’ heel: McKinsey’s warning on rare earths grows louder

China controls about 70% of global rare earth mining and roughly 90% of refining and processing capacity, creating a major supply-chain chokepoint for defense, EVs, and industrial manufacturing. McKinsey projects up to a 30% global shortfall in magnetic rare earth supply by 2035, while non-China producers may meet less than 20% of demand for dysprosium and terbium. The U.S. is responding with an $8.5 billion Australia pact, new deals in Malaysia and Thailand, and potential reallocation of $2 billion in CHIPS Act funds, but the article says closing the gap will take years.

Analysis

This is less a commodity story than a denial-of-service attack on U.S. industrial policy. The key second-order effect is that the bottleneck is not mining capacity, but qualified separation, metal-making, and magnet finishing; that means even well-funded onshoring programs can miss revenue recognition by years while still consuming capex. The market is likely underestimating how quickly a small supply shock propagates into OEM line stoppages because inventories are thin and qualification cycles are long.

The relative losers are the downstream integrators with the least pricing power and the highest just-in-time exposure: auto OEMs, EV programs, and defense primes that rely on magnets but do not control the input chain. For Ford, the risk is not just a temporary parts delay; it is margin compression from premium sourcing, line inefficiency, and forced design substitutions that can reduce motor performance. MP is not a clean beneficiary at current levels because policy support does not eliminate execution risk, permitting risk, or the probability of Beijing leaning on price to cap Western returns.

The contrarian view is that the headline geopolitical premium may be overbought in the very short term but still under-owned in the medium term. If export controls ease or if stockpiling bridges the next 1-2 quarters, the pure-play scarcity trade can mean-revert sharply; however, the structural gap is a years-long problem, so dips are likely buyable in the supply-chain enablers rather than the miners themselves. Intel is a policy optionality story, not a fundamentals fix; any capital injection helps liquidity, but it does not solve the absence of committed end-demand or the broader U.S. industrial sequencing problem.

Key Decisions for Investors

• Short F vs long a basket of industrials less exposed to rare-earth input risk for the next 3-6 months; thesis is margin compression and production disruption risk at Ford outpaces any policy offset. Use a tight stop if export controls ease materially.
• Buy 6-12 month call spreads on MP only on pullbacks, not strength; treat it as a high-volatility policy proxy with upside if U.S. procurement accelerates, but cap risk because China can pressure pricing. Target a 2:1 payoff.
• Go long defense prime exposure with lower direct magnet sensitivity versus EV/auto OEMs; prefer a relative-value pair long defense / short autos into any additional China export-control headline. The spread should work over 1-2 quarters as defense budgets are slower-moving and more insulated.
• Accumulate INTC only as a policy trade, not a semiconductor beta trade; use limited-risk structures such as January calls or call spreads to capture CHIPS/strategic capital headlines over the next 1-2 quarters. Avoid outright equity until there is evidence of committed end-demand and higher utilization.
• Consider a basket hedge: long U.S. critical-minerals enablers, short China-sensitive end-users, with a 6-12 month horizon. The trade works best if the market keeps pricing this as a temporary headline rather than a persistent supply-chain tax.