The $3?5 Billion Question: What Does Rare Earth Independence Really Cost?

Highlights

• A 2026 working paper estimates reducing U.S. rare earth magnet dependence on China from 85% to 30% would cost $3–5B over a decade—manageable in defense terms, but only if spending is sequenced correctly.
• Diversification costs triple once allied capacity is exhausted: friend-shoring (e.g., Japan, Australia) costs $30–57M per percentage point, while domestic reshoring rises to $96–168M per point—indicating refining is the main bottleneck, not mining or manufacturing.
• Washington’s current approach risks inefficiency by investing heavily in domestic magnet manufacturing (Stage 3) before securing adequate refining capacity (Stage 2), potentially stranding capital and wasting billions on out-of-sequence infrastructure.

---

A March 2026 working paper (opens in a new tab) by Subodh Khanal delivers a rare, quantified answer to a question dominating U.S. industrial policy: what does it actually cost to reduce dependence on China for rare earth magnets? The answer: roughly $3–5 billion over a decade to move from ~85% dependence on China to ~30%—a manageable number in defense terms, but one that hides a far more important truth. Not all dollars spent deliver equal resilience. How Washington sequences that spending may determine whether billions are efficiently deployed—or quietly wasted.

Three Stages, One Bottleneck

Khanal’s model breaks the supply chain into three parts:

• Mining (least constrained)
• Refining (critical bottleneck)
• Magnet manufacturing (high-value, capital-intensive)

The finding is blunt: the problem is not digging material out of the ground—it’s processing and turning it into magnets.

China controls:

• ~60% of mining
• >90% of refining
• ~90% of magnet production

That downstream dominance—not geology—is where strategic vulnerability resides.

The Cost Curve That Changes Everything

The U.S. consumes roughly 2,500 metric tons of NdFeB magnets annually for defense, according to Khanal’s triangulated estimate. From there, the paper models three pathways:

1. Status Quo

• ~85% China dependency
• ~$275M annual cost (unadjusted)
• $530–830M risk-adjusted due to disruption probability

2. Friend-Shoring (85% → ~55%)

• Uses existing allied capacity (Japan, Australia)
• Cost: $30–57M per percentage point reduction
• 10-year cost: ~$0.9–1.7B

3. Reshoring (55% → ~30%)

• Requires building domestic refining + magnet plants
• Cost: $96–168M per percentage point (~3× higher)
• 10-year cost: ~$2.4–4.2B

The Kink: Where Costs Triple

This is the paper’s most important insight.

The cost of diversification is not linear—it jumps sharply once allied capacity runs out.

• Early phase = cheap, fast, leverage existing systems
• Late phase = expensive, slow, build from scratch

Think of it as insurance:

• Friend-shoring = affordable baseline coverage
• Reshoring = costly protection against extreme scenarios

Same goal. Radically different economics.

Washington’s Sequencing Problem

Here’s where policy collides with reality.

The U.S. is investing heavily in:

• domestic magnet manufacturing (Stage 3)

But remains constrained in:

• refining capacity (Stage 2)

That creates a structural mismatch.

You cannot manufacture magnets at scale without sufficient upstream refined material. Khanal’s implication is clear:

Building Stage 3 before fully securing Stage 2 is not just inefficient—it risks stranded capital.

What the Numbers Don’t Fully Capture

The study is transparent about its limits:

• Demand estimates range widely (1,600–4,500 MT)
• Cost data partly inferred from filings and disclosures
• Disruption probabilities require judgment calls

But these uncertainties do not change the central conclusion:

• Reshoring is structurally more expensive than friend-shoring
• The order of investment matters more than the intent

The Strategic Takeaway: Buy the Cheap Resilience First

Khanal doesn’t argue against reshoring. He reframes it.

• Friend-shoring = immediate, cost-effective resilience
• Reshoring = long-term strategic insurance

The mistake is not pursuing both.

The mistake is doing them in the wrong order.

Bottom Line

The United States can reduce its dependence on China for rare earths. The cost—$3–5 billion over a decade—is not prohibitive.

But the structure of that cost is everything. Spend smart, and resilience is achievable.

Spend out of sequence, and the bill rises fast. In rare earths, as in markets:

The first moves are cheap. The last moves are expensive.

Citation: Khanal, S. (2026). The Rare Earth Reshoring Premium: Quantifying the Fiscal Cost of Reducing U.S. Dependence on Chinese NdFeB Magnet Supply Chains. Working Paper.