USA Rare Earth (USAR) is a pre-revenue US/EU rare-earth and magnet integrator — Stillwater (OK magnets), LCM-Cheshire (UK alloys), Lacq (France metallization via Carester JV), Round Top (TX heavy REE). $1.75B cash, debt-free. The April 20 definitive merger with Brazil's Serra Verde (only large-scale producer of all four magnetic REEs outside Asia) plus the Q1 10-Q filed May 14 collectively transformed the story. Stockholder vote on Serra Verde is June 3 — 18 days from this memo.

What the filing says

Execution-progress, not thesis-changing. Bullish deltas:

• CHIPS LOI condition (i) formally satisfied. $500M non-federal raised via $1.5B PIPE (January 2026). First of five conditions off the table in a CEO/CFO-certified filing.
• Round Top PFS pulled forward to Q3 2026 from prior "late 2026/early 2027." Fluor + WSP confirmed as EPCM.
• France C3IV €130M — up to 45% reimbursement on eligible Lacq facility capex (the Carester JV). Bpifrance export credit in parallel. Not in prior 10-K.
• Texas TSIF $14.2M state grant signed May 11 for Round Top.
• Cash $1.749B (vs $360M Dec 31), debt-free post-Barclays repayment.

Bearish friction:

• Zero binding magnet customer contracts. Phase 1a commissioned. Language unchanged from prior: "customer orders expected Q2 2026."
• LCM gross margin 1.9% — under-absorbed fixed costs. 49% of LCM revenue from one unnamed customer.
• CHIPS debt restriction newly explicit: if CHIPS closes, USAR is "precluded from issuing additional debt" except the government senior loan. Forecloses capital structure optionality.
• Burn $57M/quarter ($18.5M opex cash + $38.6M capex), accelerating with Serra Verde and Carester closes.

What the market thinks

USAR $24.39. Market cap $5.4B current; ≈$9.7B fully diluted (≈398M post-deals). Sell-side: 7 Buy / 0 Hold / 0 Sell, mean PT $37.29 (+53% upside). Wedbush/Cantor reiterated $35 on the 10-Q date.

At 10x 2030 EBITDA target ($1.8B) discounted 15% over 4 years, bull-case present value is roughly $26/share. Current price prices close to that. Our forward 12-month probability-weighted EV: +25-30% (Bull 25% / Base 45% / Slow 20% / Bear 10%). Sell-side implied idio premium over MP-beta is roughly +18%; our estimate is closer to +10%. We're behind consensus by ≈8pp.

Options 32d ATM IV 99.7%. Market implies ≈12% probability of -25% move by June 18 (vote-fail proxy). Our prior on vote-fail is 20% — meaningfully more bearish than options imply.

Why the gap exists

Factor decomposition on daily returns (n=76, post-PIPE 90d) gives β_MP = 1.10 (t=11.06, p<0.001). Idiosyncratic variance has DECLINED from 72% (since IPO) → 47% (post-Stillwater) → 33% (post-PIPE, full spec) — well below the 75% Paleologo idio target. REMX (rare-earth ETF) collapses to β≈0 when MP is in the regression — MP IS the relevant sector index.

The bull thesis as written is idio-specific (Serra Verde uniqueness, EU C3IV stacking, CHIPS execution). The empirical return path is currently sector-correlated. Naked long USAR pays ≈117% total vol to capture ≈92% MP vol plus uncompensated noise — Sharpe inferior to owning MP directly for the same non-China RE thesis.

What sell-side coverage underweights: USAR is the only US-listed name with explicit access to the EU CRMA/C3IV subsidy template via the Carester JV at Lacq. The EU framework runs in parallel to US DFC/Orion CMC — 47 EU + 13 non-EU CRMA Strategic Projects designated March 2025, fast-track permitting, multiple French projects (Solvay La Rochelle, MagREEsource Grenoble, LCM-Lacq) accessing C3IV. Coverage focuses on DFC + CHIPS; the additive EU stack is the unrecognized leg.

Risks (ranked by near-term impact)

1. June 3 vote fails (≈20% probability). -25 to -35% drawdown.
2. DFC consent withheld OR Incremental Loan amount not disbursed — both are merger closing conditions. Deal dies. -40 to -50% over 6-9 months.
3. Round Top PFS unfavorable (Q3 2026 target). MP's Litinsky has publicly critiqued heap-leach RE extraction as "phantom feedstocks, technologies unproven at scale."
4. No customer contract by Q3 2026 — erodes the bull thesis quietly.
5. DyTb chemistry shift — humanoid robot magnets may not require heavy REE (Niron iron-nitride, MP GBD IP). Multi-year, not near-term.

Catalysts

• Jun 3 2026 — Serra Verde stockholder vote (80% pass)
• Sep 30 2026 — Round Top PFS target; Carester definitive; first Stillwater contract
• Dec 31 2026 — Serra Verde close (65%); CHIPS → definitive (40%); TMRC merger close (80%)

What would change our mind

• Idio variance rises through 75% post-Serra-Verde close (regression rerun). Confirms catalysts have decoupled USAR from MP. Idio thesis investable on its own merits.
• Binding semiconductor end-user MOU (CHIPS condition ii) — natural first-customers signal, dual-purpose.
• Serra Verde Dy/Tb floor prices disclosed (currently confidential). Enables real Serra Verde NAV.
• MKA (Mkango) Nasdaq listing closes (70% by Dec 31). Cleaner CRMA-Strategic pure-play emerges; capital may rotate.

The trade structure worth considering is MP-hedged USAR (β-neutral pair) — isolates the actual edge claim from sector noise. Naked long at current prices buys consensus that is already largely priced.

Evidence