USA Rare Earth, Inc. (NASDAQ: USAR) — Washington’s Ex-China Magnet Champion, Priced as if the Plant Already Runs

Independent fundamental research. As-of date: 2026-06-20. Price reference: $24.64 (close, 2026-06-18).

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows takes no position and carries no price target; the single opinion in this article is fenced here.

Call: AVOID at ~$24.64 / ~$5.5B market cap. Not-a-short outright. Speculative venture-stage option — size it like a venture bet, not a position. Conviction: LOW (high uncertainty, wide outcome distribution).

USA Rare Earth is a genuinely interesting strategic asset and a genuinely bad stock at this price. The bull narrative — “the leading ex-China, vertically-integrated mine-to-magnet platform, funded by Washington, levered to a structural China-decoupling supercycle” — is real enough that I will not short it: the company sits on ~$1.75B of cash, has a signed $1.6B Department of Commerce/CHIPS funding package (a $277M grant plus a $1.3B 15-year senior loan), is rolling up genuinely scarce ex-China processing know-how (LCM, Carester, the CIX flowsheet), and is acquiring Brazil’s Serra Verde — the only at-scale ex-Asia producer of all four magnetic rare earths, with a U.S.-government-financed, price-floored offtake. Story stocks with sovereign backing and a real geopolitical tailwind can compound improbably and squeeze shorts violently.

But the price is the problem. At ~$5.5B market cap (≈$3.75B enterprise value net of cash) the market is paying a real-company valuation for a company that booked ~$7M of trailing revenue — all of it from one small UK metals subsidiary — and has zero revenue from its flagship magnet plant or its Texas mine. The magnet line is pre-commercial (first customer sales targeted 2H-2026, a 600 mtpa run-rate by year-end vs. a 5,000 mtpa ambition); Round Top is exploration-stage with no S-K 1300 reserves and commercial production targeted for late 2028; the separation flowsheet is still at demonstration scale. To justify today’s EV you must underwrite near-flawless multi-continent execution, a durable two-tier “China vs. ex-China” price premium, binding offtakes that don’t yet exist, and 20–30% mature margins in a business whose global cost-setter is subsidized Chinese capacity — and you must absorb a dilution stack (10.1M earnout shares whose price hurdles are already met, ~127M Serra Verde shares, ~16M+ government warrants, a fresh shelf, and a covenant to raise ≥$600M more equity by end-2027) that is set to push the ~223M share count materially higher. Meanwhile the insiders who know it best are selling: ~$33M of executed director sales plus ~$110M of proposed Form-144 liquidations by SPAC/PE-era backers, against ~$2.2M of token buying, and a CEO who owns zero shares. The factor tape says the same thing — a crowded, 121%-annualized-vol thematic that ran ~6x in 2025, now ~36% off its October high with a fat left tail (−69% one-year max drawdown). Framing: a policy-driven thematic momentum trade and a venture-stage falling knife — not a compounder, and not yet a moat.

What would flip me bullish: the Serra Verde close plus a binding, named, economically-priced offtake for USAR’s own Stillwater magnet output (not a Brazilian SPV’s), at a price that survives a Chinese price war — i.e., proof the two-tier market is real and USAR captures it. What confirms the bear: a slipped Q3-2026 first-oxide milestone or 2H-2026 magnet-sales start, a Serra Verde re-trade/break, or another large dilutive raise — any of which exposes the gap between a ~$5.5B cap and a pre-revenue P&L. An entry that would interest venture-risk capital sits far lower — roughly the $10–14 / ~$2.5–3.5B-cap zone (around the government-warrant strike of $17.17 and below), where you are paying closer to cash-plus-optionality than to a fully-built franchise. Tag: buying the magnet dream before the plant turns on.

📈 Stock Price Action — Five-Year Event Map

USA Rare Earth has only existed as a public company for ~15 months (de-SPAC with Inflection Point Acquisition Corp II; began NASDAQ trading 2025-03-14), so the “five-year” map is necessarily a ~15-month one. The arc is a textbook post-de-SPAC thematic: a violent debut, a washout to ~$5.75, then a policy-driven ~6x melt-up to an all-time high of $38.68 (2025-10-13), and a roll-over to $24.64 today — ~36% off the high, inside a 52-week range of roughly $9.61–$38.68. The pre-merger SPAC trust traded near its ~$10 trust value. Price moves below are FACT (daily price data); attributed drivers are INTERPRETATION.

Cycle narrative. (1) The de-SPAC debuted into a thin, redemption-gutted float — the +72% pop and the −60% washout to $5.75 are both float mechanics, not fundamentals. (2–3) The stock is, from inception, a policy-headline instrument: the spring-2025 China export-control escalation and the July-2025 MP–DoD deal — which established Washington’s willingness to price-floor and equity-fund domestic rare earths — re-rated the entire basket, USAR included. (4–5) Through autumn 2025 it became a crowded retail/thematic momentum trade, peaking at $38.68 on heavy volume as China added heavy REEs (Dy, Tb) to its export-control list. (6) The January-2026 one-two of a $1.6B Commerce funding LOI and a $1.5B PIPE marked the transition from “concept” to “capitalized roll-up,” but also began the dilution that (7) — alongside the announcement of ~$3B+ of share-funded M&A — has driven the stock down ~36% from its high into a volatile, headline-dependent range. This is price history, not a forecast; the opportunity/mispricing judgment lives only in Claude’s Take above.

1. Executive Summary

USA Rare Earth (USAR) is a pre-commercial, development-stage “mine-to-magnet” rare-earth and NdFeB permanent-magnet company that went public via a redemption-gutted de-SPAC in March 2025 and has since been transformed — by a $1.5B PIPE, a $1.6B U.S. Department of Commerce/CHIPS funding package, and three pending acquisitions — into a ~$5.5B-market-cap, ~$1.75B-cash, government-backed roll-up. It is, today, almost entirely aspiration: trailing revenue is ~$7M (100% from one small acquired UK metals subsidiary, Less Common Metals), and there is zero revenue from the flagship Stillwater, Oklahoma magnet plant (pre-commercial; first customer sales targeted 2H-2026) or the Round Top deposit in West Texas (exploration-stage; no reserves; commercial production targeted late 2028).

The central tension: USAR is a genuinely strategic asset — there are very few credible ex-China sintered-NdFeB and heavy-REE-processing capabilities anywhere, and Washington is actively funding this one — wrapped in a genuinely speculative stock. The business has no durable economic moat today; on Greenwald’s taxonomy it has no supply/cost advantage (a ~500 ppm / ~0.06% TREO deposit roughly 100x lower-grade than MP’s Mountain Pass; an unproven flowsheet), no demand captivity (no customers, no qualified products, commodity output), and no scale (China holds it). Its only genuinely scarce assets are ex-China processing/metals know-how (LCM strip-casting, the Carester investment, the proprietary Continuous Ion Exchange process) and conferred government relationships — real, but reversible and not yet monetized. The industry itself is structurally bad for an unsubsidized new entrant (China controls ~85–91% of separation and the bulk of magnet production and sets the marginal price) and artificially attractive only via non-market intervention — price floors, grants, subsidized debt, and a “security premium” some Western customers may pay.

Financially, the story is cash, burn, and dilution — not earnings. GAAP losses (FY2025 net loss $298.5M) are dominated by non-cash SPAC-warrant mark-to-market (~82% of the loss) and are not a usable signal; the real metrics are an operating loss running ~$36M/quarter, capex ~$39M/quarter, ~$57M/quarter total cash use, and a ~$1.75B war chest that removes near-term solvency risk but is dwarfed by a stated ~$4.1B long-term build cost that requires ≥$600M more equity by end-2027 plus the not-yet-fully-committed government debt. Share count has ~2.7x’d in 15 months to ~223M and faces a stacked further-dilution path (10.1M earnout shares already past their price hurdles, ~127M Serra Verde shares, ~16M+ government warrants, an effective shelf). Insiders are net sellers by a wide margin, and the CEO owns no shares.

Valuation can only be framed as embedded expectations: at ~$3.75B EV the market is underwriting near-flawless ramp of plants that are not yet running, a durable two-tier China/ex-China price premium, binding offtakes that don’t yet exist, and 20–30% mature margins — i.e., it is paying roughly half of at-scale peer MP Materials’ enterprise value for a company with ~3% of MP’s revenue and none of MP’s de-risking. The factor tape confirms a crowded, high-volatility thematic that has already rolled over. No recommendation or price target appears below; the single position is in Claude’s Take above.

2. Business Overview

What USAR is today vs. what it plans to be. USA Rare Earth markets itself as an integrated “mine-to-magnet” platform spanning five value-chain links — mining, separation, metals/alloys, magnet manufacturing, and recycling — built to supply U.S.-made NdFeB permanent magnets and heavy-rare-earth materials independent of China. That is the plan. The operating reality as of the report date is far narrower: USAR is a cash-burning, development-stage holding company whose only revenue (~$6M in Q1-2026, ~$1.6M for partial-period FY2025) comes from Less Common Metals (LCM), a small UK rare-earth metals/alloys/strip-casting business acquired in November 2025. There is no revenue from magnets and none from mineral production. The company had ~30 full-time employees at last disclosure, a fact worth holding in mind against a ~$5.5B market cap. [FACT — 10-K FY2025; Q1-2026 10-Q; ROIC profile]

The five links, separated into operating vs. plan:

• (a) Round Top deposit (Hudspeth County, West Texas). A ~1.6-billion-tonne rhyolite laccolith hosting 15 of 17 rare earths — including all heavy REEs (Dy, Tb, Y) — plus gallium, hafnium, zirconium, beryllium, and lithium. It is exploration-stage: USAR explicitly states it has no declared mineral resources under S-K Item 1300 and no reserves, and has not begun extraction. The economic reference point is a legacy 2019 NI 43-101 Preliminary Economic Assessment (the lowest-confidence study tier) carrying a pre-tax NPV(10%) of ~$1.56B, ~70% IRR, ~$350M initial capex, and a ~20-year life — now superseded by a fresh pre-feasibility study (Fluor/WSP, due 2H-2026) and definitive feasibility study (due ~Q1-2027). USAR holds ~2,037 acres plus a purchase option on 5,670 more under a long-term Texas state land lease, and won a $14.2M Texas Semiconductor Innovation Fund grant (May 2026). Commercial production is targeted for late 2028. [FACT — 10-K; 2019 TMRC PEA; PRs]
• (b) Stillwater, Oklahoma magnet plant. A 310,000-sq-ft owned facility — the closest-to-commercial asset. First lab-scale sintered NdFeB magnets were produced in January 2025; “Phase 1a” was commissioned in March 2026; customer-ready production is targeted to start in Q2-2026 with first sales in 2H-2026, ramping to a 600 mtpa run-rate by year-end 2026 and, eventually, toward 1,200 and ~5,000 mtpa (“hundreds of millions of magnets per year”). It remains pre-revenue in magnets as of the report date. [FACT — 10-K; Q1-2026 call]
• © Separation / Continuous Ion Exchange (CIX) + Wheat Ridge, Colorado R&D. A proprietary separation flowsheet; a hydrometallurgical demonstration facility commissioned mid-2026, targeting first separated heavy-rare-earth-oxide production in Q3-2026 (Dy/Tb/Gd/Y oxides; mixed HREE carbonate), fed by Serra Verde, Round Top, and third-party feedstock. Demonstration scale, not commercial. [FACT — 10-K; Q1-2026 call; Benzinga 2026-06-15]
• (d) Less Common Metals (Cheshire, UK). Acquired November 2025 for $197.7M ($103.1M cash + ~6.74M shares). The only operating business; targeting 3,000 mtpa metal/strip-cast capacity by Q4-2026. [FACT — 10-K]
• (e) New roll-ups (post-Jan-2026). Pending acquisition of Brazil’s Serra Verde (Pela Ema mine — an at-scale, producing ex-Asia source of all four magnetic REEs, carrying a 15-year U.S.-government-financed SPV offtake with NdPr/Dy/Tb price floors); a strategic investment in France’s Carester (HREE processing/recycling); and consolidation of 100% of Round Top via an all-stock buyout of JV partner TMRC. See . [FACT — 8-Ks 2026-04-23/05-13/06-05; merger agreement]

Business model and revenue recognition. There is, in substance, no business model yet — there is a capital-formation model (raise equity and government money; acquire and build assets) attached to a future product model. Such recurring revenue as exists is LCM’s lumpy, low-margin metals sales (Q1-2026 gross margin ~1.9%; two customers = 73%+18% of revenue; ~94% Europe/Asia). Verdict: USAR today is ~95% aspiration and ~5% operating — a portfolio of plans and pending deals funded by ~$1.75B of capital-markets and government money, not a business generating returns on capital.

3. Industry Dynamics

Structure. Rare earths are not geologically rare; the bottleneck is separation and refining, where China controls an estimated ~85–91% of global capacity, and magnet manufacturing, where China dominates sintered-NdFeB output. China achieved this through decades of state-subsidized capacity, scale, vertical integration, and a tolerance for the environmental cost of processing — and it actively uses that position as geopolitical leverage. In April 2025, Beijing added seven medium/heavy REEs (including Dy and Tb) to its export-control list, restricting NdFeB magnet materials; Chinese magnet exports fell ~75% over the following two months. That shock is the single catalyst behind the entire Western build-out and behind USAR’s narrative and share price. [FACT — IEA; CFR; China Briefing; S&P Global]

The policy scaffolding. Because Western producers are structurally high-cost against subsidized Chinese capacity, the sector is being kept alive by non-market intervention. The template is the MP Materials–DoD deal (July 2025): a 10-year $110/kg NdPr price floor, a $400M DoD convertible-preferred investment plus warrants, and a commitment to buy 100% of a “10X” magnet facility’s output for 10 years — plus an Apple offtake (>$500M of magnet purchases from 2027, $200M prepayment). This is the benchmark USAR is chasing. USAR’s own government support — the $1.6B DoC/CHIPS package ($277M grant + $1.3B 15-year senior loan) and a $14.2M Texas grant — is funding support (cheap capital), not an MP-style hard product price floor on its own output. The price floors USAR cites attach to the Serra Verde Brazilian SPV offtake it is acquiring, not to Round Top/Stillwater. [FACT — MP IR; 8-K 2026-06-03; 10-K]

Greenwald lens (barriers to entry). For a new entrant, this is a structurally bad industry. The dominant barrier to entry protects the incumbent — China — not the challenger. REE oxides, metals, and magnets are largely commoditized and index-priced, with the marginal cost set by subsidized Chinese capacity. A Western producer therefore has no supply/cost advantage and no demand captivity at market prices; it is, by construction, high-cost. The only thing making it viable is a policy moat — floors, offtake guarantees, grants, subsidized debt, and a “security premium” — which is durable only as long as the policy and the China tension persist. That is a conferred advantage, not an earned one.

Marathon capital-cycle lens. Capital is flooding into “critical minerals / ex-China REE” on the back of the 2025 China shock — MP, Lynas, USAR, Energy Fuels, NioCorp, Vulcan, Ramaco, plus Brazilian/Australian juniors, plus state money from the U.S. (DoD/DoC), France, Japan, and Saudi Arabia. That is precisely the high-returns-attract-capital signal Marathon flags as a top. The crucial nuance: the self-correcting cycle is broken on both sides — China can flood or withhold to discipline rivals (capping Western prices on any recovery), and Western governments are deliberately subsidizing uneconomic capacity for security reasons (divorcing capacity from price signals). The bull case thus literally requires the normal capital cycle not to apply — i.e., a durable two-tier “China vs. ex-China” market. The lithium sector is the cautionary template for exactly this dynamic: Chinese oversupply crushing Western project economics.

Pricing mechanics and the two-tier thesis. The single most important — and least proven — assumption in the entire bull case is that a durable two-tier price market emerges, in which ex-China buyers pay a structural premium over the Chinese domestic price for “secure,” non-Chinese material. Today the global reference prices for NdPr (the light-REE workhorse for magnets) and for Dy/Tb (the heavies that confer high-temperature performance) are still effectively set in China, and Chinese magnet makers enjoy a fully-integrated cost stack — captive mine, captive separation, captive metal, captive magnet — at a scale no Western entrant approaches. A standalone Western producer must clear every link of that chain at sub-scale and at Western environmental/labor/capital cost. The MP–DoD $110/kg NdPr floor exists precisely because the market price is unreliable and frequently below the cost of Western production; the floor is a confession that the economics do not stand alone. For USAR’s own output there is no such floor — so the two-tier premium is not contracted, it is hoped for. If it materializes and holds (because A&D and allied-government buyers genuinely will pay up, and the Jan-1-2027 non-China sourcing rules bite), the economics can work; if China thaws its controls or floods supply to discipline the Western build-out — as it did to bankrupt Molycorp after 2011 — the premium evaporates and the whole sector reverts to structurally loss-making. [Interpretation]

Verdict: structurally bad for an unsubsidized entrant; artificially attractive only for a subsidized, policy-backed one — and only if the policy support and the China-tension premium prove durable and large enough to overcome a structural cost disadvantage. High capital intensity, commodity pricing, a hostile dominant incumbent, and a graveyard of Western failures (Molycorp’s 2015 bankruptcy after the 2011 bubble burst; Lynas surviving only on Japanese strategic financing) make this a low-base-rate venture. The bull case is not that the industry is good — it plainly is not — but that this company will be permanently insulated from the industry’s economics by government support and a security premium. That is a bet on politics persisting, not on a business compounding.

4. Competitive Position

Name the moat — or its absence. USAR has no durable economic moat today. Tested against Greenwald’s three genuine advantage types:

1. Supply/cost advantage — ABSENT. Round Top is a ~500 ppm (~0.06% TREO) very-low-grade rhyolite — roughly 100x lower grade than MP’s Mountain Pass (~6–8% TREO bastnaesite). The thesis depends on cheap whole-ore heap/vat leaching and on the value residing in scarce heavies (Dy/Tb/Y) that China controls and Mountain Pass largely lacks — a real idea, but an unproven, exploration-stage flowsheet with high technical and cost risk. No proven low-cost process exists; against subsidized Chinese capacity USAR is structurally high-cost.
2. Demand captivity / switching costs — ABSENT. No customers, no qualified products, commodity output. The only “captivity” is regulatory (“must buy non-China”), which is a policy artifact, not a customer relationship.
3. Economies of scale + captivity — ABSENT. Sub-scale and pre-revenue; China owns the scale.

What USAR has instead is (i) early positioning in U.S. mine-to-magnet vertical integration — a narrative/option, not yet a moat; (ii) genuinely scarce ex-China processing/metals know-how (the CIX flowsheet, LCM strip-casting, Carester engineering) — the most defensible asset, hard to replicate outside China; and (iii) government/policy relationships (DoC funding, Texas grant, allied-government ties) — real but conferred and reversible.

Deposit economics — why the grade matters. Grade is not a vanity metric; it drives the entire cost structure. At ~6–8% TREO, Mountain Pass moves a tonne of ore and recovers ~60–80 kg of rare-earth oxide; at ~500 ppm (~0.06%) Round Top recovers ~0.6 kg per tonne — roughly two orders of magnitude less per tonne handled. USAR’s counter is that Round Top is a soft rhyolite amenable to whole-ore heap/vat leaching (low strip, low comminution cost, no flotation), and that its value is concentrated in the scarce heavies (Dy, Tb, Y) plus lithium and gallium that Mountain Pass largely lacks and that command far higher per-kg prices than the light REEs. That is a coherent geological argument — but it is unproven at commercial scale, the flowsheet is still at vat-leach/demonstration stage, and a low-grade operation lives or dies on processing cost per kg of recovered, separated, saleable oxide, which no PFS has yet established. Until the 2H-2026 PFS and ~Q1-2027 DFS quantify recovery, reagent consumption, and all-in sustaining cost, Round Top’s heavy-REE endowment is optionality, not a resource — and even the optionality competes against MP (already separating Dy/Tb mid-2026) and against the acquired Serra Verde, which produces the same heavies sooner.

Direct comparison vs. MP Materials. MP is the only at-scale U.S. player and the right yardstick.

MP is roughly 3–4 years ahead on every operating metric — yet USAR trades at ~half MP’s enterprise value on ~3% of MP’s revenue. Other reference points (Lynas, the largest ex-China producer, with real revenue; Energy Fuels/UUUU, which just won a $725M DoD loan; NioCorp/NB) sit in the same policy-driven basket but with varying degrees of actual production.

Verdict: no durable competitive advantage exists today. The closest thing to a real edge is scarce ex-China processing/metals know-how; everything else (“first US mine-to-magnet”) is marketing until proven by commercial output, qualified customers, and competitive cost. Bluntly: USAR is an option on building a moat, not a moat.

5. Growth History and Forward Opportunities

There is no meaningful revenue history — growth is entirely option value on a stack of execution milestones, none yet de-risked by commercial proof. The milestone stack (all management targets, treated as hypothesis):

• Magnets: Stillwater 600 mtpa run-rate (YE-2026) → 1,200 → ~5,000 mtpa ambition; first customer sales 2H-2026.
• Separation: first separated HRE-oxide at the Colorado demo facility in Q3-2026.
• Metals/alloys: LCM 3,000 mtpa (Q4-2026); a planned LCM-France ~3,750 mtpa metals plant (Lacq).
• Mining: Round Top PFS (2H-2026) → DFS (~Q1-2027) → commercial production (late 2028).
• M&A closings: Serra Verde and Carester (targeted Q3-2026); TMRC consolidation.
• 2030 ambition (management, in the 10-K): ~40,000 t/day Round Top feedstock, ~8,000 mtpa MREC/HREE processing, ~10,000 mtpa metals/alloys/strip-cast, ~2,000 mtpa swarf recycling — historically tied to an analyst-cited management projection of ~$2.6B sales / ~$1.2B EBITDA by 2030.

Offtakes — the crux. There are no binding customer offtake agreements with named counterparties and volumes disclosed for USAR’s own magnet/oxide output. Management describes a “demand funnel” and “safety-stock” inquiries — auto OEMs allegedly directing suppliers to hold up to a year of magnet supply, A&D customers facing a January-1-2027 non-China sourcing deadline — but these are narrative, not contracts. The only hard, price-floored offtake in the story belongs to the Serra Verde Brazilian SPV (a U.S.-government-financed vehicle, 15-year, 100% of Phase-1 Nd/Pr/Dy/Tb), which USAR is acquiring, not organically originating.

Verdict: speculative, low-quality, un-de-risked growth — a venture-stage option chain, not a growth business. The single most de-risking near-term item would be the Serra Verde close (it brings actual production plus a government-backed, price-floored offtake); the original Round Top/Stillwater thesis remains years and billions of capex away from cash flow.

6. Financial Quality

For a pre-revenue developer, “financial quality” means cash, burn, dilution, and accounting cleanliness — not margins or returns.

Cash and runway (the central fact). Cash and equivalents were $1,749.6M at 2026-03-31, up from $359.9M at YE-2025 and just $16.8M at YE-2024 — the leap is the January-2026 $1.5B PIPE. Debt is negligible (~$1.3M of finance leases). Quarterly cash use runs ~$57M (operating burn ~$18.5M + capex ~$38.6M); FY2025 saw $49.0M operating burn and $139.6M of investing outflow (including the $103.1M LCM cash component). At the current pace the war chest funds years — but management’s own plan needs ~$4.1B of long-term capex (full Round Top + Stillwater), which it says will require ≥$600M of additional equity by 12/31/2027, a $250M revolver by 12/31/2026, and reliance on the contingent ~$1.58B of government LOIs/agreements. Runway is not the near-term risk; the financing gap to commercial scale is. Both the 10-K and 10-Q state cash is sufficient for at least the next twelve months — there is no going-concern doubt. [FACT — 10-K; 10-Q]

Net loss — distorted by warrant marks. FY2025 GAAP net loss was $298.5M ($297.4M to USAR; EPS $(3.31)), but ~82% of it is a non-cash $244.5M “loss on fair-market-value of financial instruments” — day-one and mark-to-market remarks on SPAC/PIPE warrants and Series A warrants. Quarterly GAAP net income swings wildly with these marks (Q1-25 a +$51.8M gain; Q2/Q3-25 losses of $142.5M/$156.7M; Q1-26 a $(67.0)M loss that included a +$43.6M warrant gain). The underlying operating loss is far smaller and steadier, rising from ~$8.7M to ~$36.5M per quarter as headcount and the plant scale up. GAAP EPS is not a usable signal here; use operating loss and cash burn. Stock-based comp runs ~$4.5–4.9M/quarter. [FACT — 10-K; ROIC, reconciled]

Revenue and margins. FY2025 revenue $1,643K (100% LCM, ~6 weeks); Q1-2026 $5,698K (LCM full quarter) at ~1.9% gross margin, with customer concentration (73%+18%). No magnet or mineral revenue. There are no scale economics to assess — the question is whether they will ever exist at competitive cost.

Balance sheet and share count. Total assets $2,134.8M (2026-03-31): cash $1,749.6M, inventory $28.4M, net PP&E $142.1M, plus $134.8M goodwill + $67.3M intangibles (both from LCM, untested for impairment). Total equity $1,878.4M. Basic common shares went 81.95M (Q1-25) → 148.1M (YE-25) → 223,035,366 (Q1-26 10-Q cover) — ~2.7x in 15 months — plus 1.22M shares of 12% Series A convertible preferred. Current contingent overhang is ~14.9M shares (10.1M earnout + ~2.4M Series A as-converted + ~2.4M Series A warrants); the large investor warrants (~23.8M) were fully exercised during 2025 (raising ~$304M in FY2025, ~$631M cumulatively), so warrant overhang is now small. The 10.1M earnout shares vest at $15/$20 price hurdles — both already exceeded at ~$24.64, so that issuance is near-certain. [FACT — 10-K; 10-Q anti-dilutive table]

The dilution stack, quantified. Per-share value is being eroded from several directions at once, and it is worth assembling the full bridge. Start from ~223.0M common shares (Q1-26 cover) + 1.22M preferred (~2.4M as-converted). Then layer the contingent and pending issuance: ~10.1M earnout shares (price hurdles of $15/$20 already cleared at ~$24.64 — effectively certain); ~3.82M for the all-stock TMRC buy-in; ~127M for Serra Verde; ~16.1M government common shares at $17.17 plus warrants for ~10% of fully-diluted shares (call it ~40M+ on a post-deal base); and an effective S-3 shelf that enables an open-ended ATM. Even excluding the shelf, the path from ~223M toward ~380–420M+ shares is visible in already-announced transactions — i.e., the current share count could rise on the order of ~70–85% before the company sells a single commercial magnet at scale. That is the arithmetic behind the ≥$600M-equity-by-2027 covenant and the reason GAAP “book value per share” and any forward per-share estimate are unstable. A buyer at today’s ~$5.5B market cap is, in effect, pre-funding much of that future issuance. [Fact/Interpretation — share counts from 10-Q and announced deals; the post-deal totals are estimates pending close]

Burn vs. build. The reassuring number — ~$57M/quarter cash use against ~$1.75B — masks the real gap. Current capex (~$39M/quarter) is demonstration- and Phase-1-scale; the commercial build that the equity value is priced on (full Round Top mine + processing + the 5,000 mtpa magnet ramp) is the ~$4.1B figure management itself cites. The $1.75B does not fund that; it funds the next leg and the equity portion of the government milestones. The honest read is that USAR has bought itself runway and credibility, not a fully-financed path to commercial scale — the remaining funding (≥$600M equity + the $1.3B conditional government loan + a $250M revolver) is the bridge that must be crossed, and each crossing is dilutive or covenant-laden.

Verdict on economics with scale: unknowable and unproven. The favorable facts are a fortress cash balance and no going-concern flag; the unfavorable facts are that every dollar of value rests on assets that are not yet operating, the GAAP statements are noise, and the share count is a moving target trending materially higher.

7. Capital Allocation

The entire capital-allocation story is financing and M&A. USAR generates no operating cash; it consumes capital. The track record:

• De-SPAC (March 2025) raised almost nothing. Total cash into the merger was ~$30.9M (PIPE $8.0M + trust $22.8M after heavy redemptions), and after ~$20.8M of forward-purchase prepayments and ~$8.3M of transaction costs, net cash to the operating company was ~$0 — a classic redemption-gutted SPAC outcome. [FACT — 10-K]
• Post-listing raises did the real work: ~$304M of 2025 warrant exercises; a $75M institutional PIPE (May 2025, which triggered the $244M day-one/MTM loss and repriced the Series A warrants from $12.00 to $7.00); the LCM acquisition financing; and the $1.5B PIPE (69.8M shares at ~$21.50, closed Jan-28-2026; net ~$1.45B) that ~doubled the share base and satisfied the non-federal-capital milestone for the government package.
• Government package: the $1.6B DoC/CHIPS agreement ($277M grant + $1.3B 15-year senior secured loan at ~Treasury+150bps) reached definitive documentation June 3, 2026 — a genuine milestone, though loan drawdowns remain conditional on further milestones, and the loan requires the company to issue the government warrants for ~10% of fully-diluted shares, plus the $277M is structured as a common-share purchase (~16.1M shares at $17.17). [FACT — 8-K 2026-06-03; 10-Q]
• Three pending acquisitions stack further dilution: Serra Verde ($300M cash + ~127M shares; headline value floated from ~$2.8B in April to ~$3.64B on the May call as the share-linked consideration moved with the stock; close targeted Q3-2026, subject to a shareholder vote); Carester (France); and TMRC (all-stock, ~3.82M shares, to buy in the 18.7% Round Top JV minority). An S-3 shelf went effective June 5, 2026, enabling further issuance/ATM. [FACT — 8-Ks; PREM14A/PRER14A]

Insider behavior — net bearish. This is decisive for a story stock. Executed open-market sells: Director Michael Blitzer (SPAC-sponsor-affiliated) sold 2,091,849 shares at $15.75 (~$32.9M) in August 2025; Director Carolyn Trabuco sold 13,000 (~$296K) in June 2026 — ~$33.2M total. Genuine open-market buys (code P): only ~101.3K shares / ~$2.2M, all from two directors on January 29, 2026 (alongside the PIPE — optics). Proposed Form 144 liquidations are larger: the Critical Minerals Trust filed to sell 3,877,565 shares (~$105.1M) on June 5, 2026; Bayshore Capital ~$4.8M. CEO Barbara Humpton owns zero shares (she received a fresh ~$11.5M equity grant; FY2025 total comp $11.70M on a $167K salary). The pattern is unambiguous: SPAC/PE-era early money is distributing into post-listing momentum as resale shelves go effective, against trivial token buying and a CEO with no purchased skin in the game. [FACT — Form 4s; Form 144s; DEF 14A]

Incentive alignment. Compensation is overwhelmingly equity (Humpton $11.5M of $11.70M in stock; CFO Steele $4.83M) — standard for a pre-revenue de-SPAC, but a real per-share drag layered on top of financing dilution, and with no disclosed return-on-capital or per-share hurdle (unsurprising given there are no returns to measure).

Verdict: financing has been executed competently and opportunistically — the $1.5B PIPE into strength was well-timed, equity (not debt) is the right instrument for a pre-revenue build, and securing $1.6B of government support is a real achievement — but capital is being consumed, not yet allocated to demonstrably value-creating assets. Returns on every dollar raised are entirely speculative, the dilution path is steep and ongoing, and the people closest to the company are selling.

8. Changes and Headwinds — Last Two Years

USAR has existed publicly for only ~15 months, and the period has been one of near-total transformation. The material changes:

• De-SPAC and listing (March 2025) into a gutted float.
• $75M PIPE (May 2025) — first major post-listing capital; triggered large non-cash warrant remarks.
• LCM acquisition (Nov 2025) — first operating revenue; first goodwill/intangibles.
• $1.6B DoC/CHIPS funding LOI (Jan 26, 2026) → definitive agreements (June 3, 2026).
• $1.5B PIPE (Jan 28, 2026) — the capitalization event that doubled the share base.
• Three M&A deals announced 1H-2026: Serra Verde (Brazil), Carester (France), and 100% TMRC/Round Top consolidation.
• Demonstration milestones: first lab-scale magnets (Jan 2025); first Yttrium metal pour (April 2026); Colorado hydromet demo commissioned, targeting first separated HRE-oxide Q3-2026.
• Management/board churn: Joshua Ballard (CEO from Dec 2024) departed; Barbara Humpton (ex-Siemens USA CEO) is now CEO; CFO William Robert Steele Jr.; CLO David Kronenfeld listed as former; the June 3, 2026 annual meeting saw some directors not re-elected. CEO turnover at a pre-commercial company mid-build is a watch item.

Headwinds. (1) China price/policy risk — Beijing can flood or thaw export controls and collapse the scarcity premium at will. (2) Dilution — a stacked path (earnout, Serra Verde, government warrants, shelf, the ≥$600M-by-2027 equity covenant). (3) Execution — every commercial milestone is still ahead. (4) Governance/political — press reports of congressional conflict-of-interest scrutiny tied to the Commerce Department’s role in the funding; the heavy insider/affiliate selling. (5) Thematic de-rating — the stock is a high-beta basket member already ~36% off its high.

Verdict: the changes have transformed and de-risked the financing (cash, government backing) while adding execution, dilution, and integration risk (three cross-border acquisitions at once). On balance they strengthen the balance sheet and weaken the per-share and focus dimensions of the thesis.

9. Risk Analysis (Risk Matrix)

Net: the risk profile is that of a venture-stage company with a fortress balance sheet but a binary, multi-year, policy-dependent operating thesis — low near-term solvency risk, very high value-realization and dilution risk, and a fat left tail in the stock.

10. Valuation Discussion (Embedded Expectations)

Trailing multiples are meaningless (TTM sales ~$7M; negative EBITDA ~−$84M; GAAP EPS distorted by warrant marks). The only honest frame is embedded expectations and enterprise-value benchmarking.

Live valuation, rebuilt by hand. At $24.64 (close 2026-06-18) on 223,035,366 common shares, market cap ≈ $5.50B. Net of ~$1.75B cash and negligible debt, enterprise value ≈ $3.75B. (Caution: one vendor’s snapshot EV of ~$1.55B uses a stale ~$14.79 price — disregard it; its balance-sheet inputs are current and were used here.) On a fully-diluted basis the picture is worse: add the near-certain 10.1M earnout shares, ~16M+ government warrants ($17.17 strike), ~127M Serra Verde shares, and ~3.82M TMRC shares, and the share count trends well above 223M — so EV per current share understates the true claim on the asset.

What the price embeds. Take management’s ~5,000 mtpa NdFeB ambition (years away; plant pre-commercial). At a $50–90/kg sintered-NdFeB ASP that is ~$250–450M of revenue at full ramp. Against today’s ~$3.75B EV:

So even crediting flawless full-ramp execution years out and healthy China-independent pricing and 20–30% mature margins, today’s EV embeds a rich double-digit forward-sales / ~30–75x forward-EBITDA multiple. The ASP and margin assumptions are themselves generous: Chinese sintered NdFeB has historically sold around $40–70/kg, and the Chinese cost floor is the standing threat to any premium. The current cap discounts near-flawless execution of plants that are not yet running commercially. (ASP and margin figures are ASSUMPTIONS, not company guidance.)

Peer benchmarking. MP Materials trades at ~$8.38B market cap / ~$7.64B EV on real TTM revenue of $254.3M (own mine, ramping magnets, a $110/kg DoD floor, Apple/GM offtakes) — and is still GAAP-unprofitable. USAR, at ~half MP’s EV, has ~3% of MP’s revenue, no operating mine, a pre-commercial magnet line, and no MP-magnitude price floor on its own output. Per dollar of de-risking, USAR is dramatically more expensive than the one comparable that has actually built the thing. Energy Fuels (~$3.82B), NioCorp (~$0.76B), and Lynas round out a re-rated basket within which USAR screens as a premium-priced, earlier-stage name. Own-history valuation percentiles (P/B 7.4x, ~92nd percentile; composite ~77th) is near-useless on ~15 months of history and a negative EPS — read only as “rich versus its own short life.”

A reverse-engineering check. Strip the question to its core: at ~$3.75B EV, and crediting the bull full-ramp revenue of ~$450M (the $90/kg case) at a generous 30% mature EBITDA margin, USAR would earn ~$135M of EBITDA — a ~28x EV/EBITDA multiple on a mature, fully-built, flawlessly-executed business that is still years away and will, by then, sit on a far larger share count. Discount that mature outcome back at even a venture-appropriate rate over the 3–5 years to get there, and the present fair EV shrinks materially below today’s $3.75B unless one assigns high probability to the bull tail and little weight to the bear tail. The market, in other words, is not pricing an expected value across the outcome distribution; it is pricing close to the right tail. That is the precise hallmark of a thematic momentum security — and the precise reason the downside is asymmetric if any of the five “must-be-true” assumptions cracks.

The Serra Verde accretion question. Bulls frame Serra Verde as de-risking — it brings actual production and a price-floored, government-backed offtake. That is true at the asset level. At the per-share level it is far less clear: paying $300M cash + ~127M shares (a ~57% increase on the current count) for a single producing mine is only accretive if the asset’s through-cycle cash generation per dollar of consideration exceeds what the diluted shareholder gives up. With the consideration value floating up with USAR’s own stock (from ~$2.8B to ~$3.64B as the shares rose), the deal is being paid for in an arguably over-valued currency — which can be smart issuer behavior, but means existing holders are funding it. Whether Serra Verde is a bargain or an expensive pivot away from the original Texas thesis is a genuine open question the pending proxy and close will resolve.

Scenario frame (illustrative, NOT targets; ranges, not forecasts). Bear: a milestone slips or China thaws controls; the thematic de-rates and dilution compounds — equity could lose well over half its value back toward a cash-plus-optionality level. Base: Serra Verde closes, magnets begin selling, and the stock trades as a still-speculative but capitalized platform — roughly where it is, with high volatility and dilution drag capping per-share progress. Bull: the two-tier price market proves durable, USAR signs economic offtakes for its own output, and government funding fully converts — a multi-bagger venture outcome if execution lands. The distribution is wide and left-skewed; the market is underwriting the right tail.

11. Variant Perception

Consensus / bull narrative. “The leading ex-China, vertically-integrated mine-to-magnet platform, levered to a structural China-decoupling supercycle, with explicit U.S.-government support and ex-Siemens-USA CEO credibility — a strategic national-security asset.” Sell-side leans constructive (small coverage, “Strong Buy”-skewed, average targets well above spot, though at least one cut on dilution).

Strongest bull case. Real strategic scarcity (very few Western sintered-NdFeB / heavy-REE-processing capabilities exist); an active, well-funded U.S. government counterparty (the MP and UUUU precedents show Washington will floor prices and lend); a genuinely heavy-REE-rich (if low-grade) Round Top resource; a fortress ~$1.75B balance sheet that funds the build without near-term solvency risk; and the Serra Verde acquisition, which would bring actual price-floored, government-backed production. Optionality compounds at each value-chain link.

Strongest bear case. A ~$5.5B cap on ~$7M of trailing revenue; commercial magnet production years away and capital-hungry; Chinese incumbents as the global cost floor able to crush prices at will (the existential risk a funding package does not offset); serial, large dilution; a valuation that embeds flawless execution, premium pricing, and unproven margins; net-bearish insider behavior; and a crowded thematic momentum trade with 121% annualized vol and a −69% one-year max drawdown, already ~36% off its high.

Factor/positioning read (the tape as variant evidence). A quantitative factor model (weak fit on <1yr history — indicative only): beta ~0.62 and rising, positive trailing alpha (a backward-looking residual of the 2025 melt-up), a large negative Liquidity loading and negative LowVol loading (acute downside in any risk-off/liquidity squeeze), a Mining industry tilt, and co-movement with a retail “story-stock” basket. Factor-similar peers are the speculative resource developers (HYMC ~0.98, NioCorp ~0.95, Energy Fuels ~0.94, Standard Lithium, US Antimony, MP ~0.84) — not the cash-flowing miners. The relative-strength peak is ~36% below its high (momentum cresting). The tape says consensus is a crowded, high-volatility thematic LONG that has already rolled over — which can be offsides to the downside if the China-decoupling narrative cools or liquidity tightens, well before fundamentals (years out) can validate it.

The 3–5 assumptions that matter most, and their falsifiers:

1. Execution/ramp to ~5,000 mtpa on time and budget. Falsifier: the Q3-2026 first-oxide or 2H-2026 magnet-sales milestone slips; plant cost/schedule overruns.
2. Premium ex-China ASPs hold (China does not flood). Falsifier: a Chinese NdPr/magnet price war or an export-control thaw collapses the scarcity premium.
3. Binding offtakes at economic prices for USAR’s own output. Falsifier: no major commercial offtake is signed; the company relies on a government floor (or a Brazilian SPV’s) only.
4. Mature EBITDA margins of 20–30%+. Falsifier: unit economics prove uncompetitive vs. Chinese cost without permanent subsidy.
5. Limited further dilution. Falsifier: another large equity raise, the full earnout/Serra Verde/government-warrant issuance, and ATM usage push the share count materially beyond what is modeled.

12. Fact vs. Interpretation Table

13. Open Questions

1. Will USAR’s own output (Stillwater magnets, Round Top oxides) ever secure an MP-style hard product price floor, or only funding support? None exists today.
2. What are the final terms and dilution of the Serra Verde acquisition (value floated $2.8B→$3.64B; ~127M shares + $300M cash), and does it clear the shareholder vote?
3. What is the true fully-diluted share count after the 10.1M earnout (hurdles met), TMRC (~3.82M), government warrants (~16M+ plus ~10% of FD), and any ATM usage?
4. What mature unit cost can the low-grade Round Top flowsheet achieve, and is it competitive without permanent subsidy?
5. Do the non-binding/conditional elements of the government package (the $1.3B loan drawdowns) fully convert to committed, drawn capital — and how does the reported congressional conflict-of-interest scrutiny resolve?
6. Are there any binding, named, economically-priced offtakes behind the “demand funnel,” or is it inquiries only?

14. What Must Be True

Bull case — what must be true: USAR ramps Stillwater toward ~5,000 mtpa roughly on schedule and budget; a durable two-tier “China vs. ex-China” price market emerges and holds; USAR signs binding, economic offtakes for its own output; mature margins reach 20–30%+; the government package fully funds and Serra Verde closes; and incremental dilution stays within a range the per-share value can absorb. Single falsification test: if, by end-2026/early-2027, the Stillwater magnet line is not selling commercially and no binding economic offtake for USAR’s own output exists, the franchise thesis is broken regardless of the cash balance.

Bear case — what must be true: China retains pricing control (floods or thaws to discipline Western rivals); USAR remains structurally high-cost on a low-grade resource and an unproven flowsheet; milestones slip; dilution compounds; and the thematic premium de-rates. Single falsification test: if USAR (a) hits the Q3-2026 first-oxide and 2H-2026 first-magnet-sales milestones, (b) signs a binding, named, economically-priced offtake for its own output that survives a Chinese price move, and © closes the government funding and Serra Verde with dilution at or below plan — the bear case is substantially falsified.

The two cases share one clock: the 2H-2026 → Q1-2027 window, when first oxide, first magnet sales, the Serra Verde close, the Round Top DFS, and government-loan drawdowns all come due. That window is the report’s central falsification event.

15. Source Appendix

See the separate Source Appendix (Appendix B in the combined report) for the full list of primary filings (FY2025 10-K, Q1-2026 10-Q, the 8-K corpus including the $1.5B PIPE and $1.6B DoC package, DEF 14A, Form 4/144, 13D/A, S-3/S-4 and merger proxies), management call transcripts, peer data (MP Materials and the critical-minerals basket), industry sources (China export-control and rare-earth structure), and the market/quantitative data pulls, each with URL and access date.

APPENDIX A — Standard Diligence Questionnaire

USA Rare Earth, Inc. (NASDAQ: USAR) — Standard Diligence Questionnaire (Appendix A)

Supplemental to the research memo. As-of 2026-06-20; price reference $24.64 (2026-06-18). Fact/Interpretation/Assumption labels applied where it matters. For a pre-revenue developer, several earnings-oriented questions are answered with the correct sector analog (cash, burn, dilution, milestones).

General

What thoughtful questions have other investors asked? The recurring debates: (1) Is USAR “the next MP Materials” or “the next Molycorp”? (2) Does any Western rare-earth producer have economic viability without permanent subsidy, or is this entirely a policy trade? (3) How much further dilution is coming (Serra Verde ~127M shares, earnout, government warrants), and what is the true fully-diluted cap? (4) Are the “demand funnel” / “safety stock” customer claims real binding offtakes or narrative? (5) Why is the CEO compensated entirely in equity yet owns no purchased shares, and why are SPAC/PE-era insiders selling? (6) Is the Serra Verde deal an accretive de-risking of the thesis or a stock-funded pivot away from the original Texas story?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? N/A — there are no earnings. The company is pre-revenue (TTM revenue ~$7M, all from the small UK LCM metals unit; net loss distorted by non-cash warrant marks). [Fact] Driven by external environment or internal actions? The equity value is driven overwhelmingly by external forces — China’s export-control policy and U.S. government funding decisions — far more than by internal operating results, which do not yet exist. [Interpretation] How stable are revenues? Not applicable/very unstable: LCM revenue is lumpy with high customer concentration (73%+18%); flagship revenue is zero. Outlook for products/services? Sintered NdFeB magnets and heavy-REE oxides target growing end-markets (EVs, robotics, defense, wind, electronics), but USAR’s participation is unproven. How big will this market be? Global REE/magnet demand is growing structurally (electrification, defense); the addressable ex-China market is policy-created and could be large if the two-tier market is durable — a key open question. [Interpretation]

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More crowded on the supply side (capital flooding into ex-China critical minerals — a Marathon top signal), but the dominant competitor (China) retains structural cost and scale advantages. [Interpretation] How profitable is the business (ROIC/ROE)? Deeply negative; no returns on capital exist — capital is being consumed, not returned. How profitable is the industry? Structurally low/negative for unsubsidized Western entrants (Molycorp bankruptcy 2015; Lynas survived only on Japanese strategic financing); profitable mainly for subsidized Chinese incumbents. Barriers to entry? High capital barriers, but they protect the incumbent (China), not the entrant. Can the business be easily understood? The narrative is simple; the economics (low-grade flowsheet cost, two-tier pricing durability) are genuinely hard to underwrite. Can it be undermined by foreign low-cost labor/capacity? Yes — directly. Subsidized Chinese capacity is the existential competitive threat and the global price-setter. [Fact/Interpretation] Do brands matter? No — commodity output; the only “brand” is “non-China,” a regulatory attribute. Switching costs? None established (no qualified products yet); future magnet qualification could create modest stickiness once customers design USAR magnets in. Moat verdict: none today — an option on building one; the most defensible asset is scarce ex-China processing know-how (CIX, LCM, Carester). [Interpretation]

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The Round Top resource (exploration-stage; no booked reserve value) and the IP/know-how (CIX, LCM strip-casting) carry option value not reflected at fair value. Off-balance-sheet liabilities? The ≥$600M-by-2027 equity-raise commitment and the $250M-revolver requirement are funding obligations tied to the government package; the earnout and warrant liabilities ($181.2M in non-current liabilities) are on-balance-sheet but mark-to-market. [Fact] How conservative is the accounting? Development capex is capitalized to PP&E ($142.1M); R&D expensed; LCM goodwill ($134.8M) + intangibles ($67.3M) untested for impairment; GAAP P&L is whipsawed by non-cash warrant marks (not aggressive, but opaque). No going-concern flag. How CapEx-hungry is the business? Extremely — the stated full build is ~$4.1B of long-term capex; current run-rate ~$39M/quarter. [Fact]

Capital Allocation & Management

How much FCF does the business generate, and how is it used? Negative FCF (~$57M/quarter cash use). There is no FCF to allocate; the company raises and spends. Philosophy: fund a multi-link build with opportunistic equity and government capital. Significant acquisitions recently? Yes — LCM (closed Nov 2025, $197.7M); and three pending: Serra Verde ($300M cash + ~127M shares), Carester, and TMRC (all-stock ~3.82M shares). [Fact] Buying back shares? No — the opposite; serial issuance. Issuing large amounts of stock to insiders? Compensation is overwhelmingly equity (CEO $11.5M of $11.70M FY2025 comp in stock); plus 10.1M earnout shares (price hurdles met) to early holders. Compensation policy: equity-heavy, no return/per-share hurdle (none measurable yet). Motivations of management: build a national-champion platform; note the CEO owns no purchased shares and SPAC/PE-era insiders are net sellers (~$33M executed + ~$110M proposed vs ~$2.2M token buying). [Fact/Interpretation]

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — Delaware C-corp, common stock on NASDAQ (USAR). Not a K-1 issuer. Dividend policy? None; pays no dividend (and should not — it has no cash flow). How profitable is the business? Not — pre-revenue, deeply loss-making. Is net income diverging from cash from operations? Yes, dramatically and benignly — GAAP net loss ($298.5M FY2025) is ~82% non-cash warrant marks, far larger than the ~$49M FY2025 operating cash burn. Use cash burn, not GAAP. [Fact] Valuation: ~$5.5B market cap / ~$3.75B EV net of ~$1.75B cash; trailing multiples meaningless; embedded expectations discount near-flawless full-ramp execution (see the Valuation section).

Risks & Downside

What would cause the stock to decline? A missed Q3-2026 first-oxide or 2H-2026 magnet-sales milestone; a Chinese price war or export-control thaw; a Serra Verde re-trade/break; another large dilutive raise; a thematic/liquidity de-rating; adverse resolution of the reported congressional conflict scrutiny. Risk of catastrophic loss? The ~$1.75B cash balance makes a near-term zero unlikely, but a pre-revenue venture can still impair the majority of its equity value if the operating thesis breaks or dilution compounds — the return distribution is wide and left-skewed (−69% one-year max drawdown already realized). Chance of total loss? Low in the near term (cash-rich), non-trivial over a multi-year horizon if the thesis fails and capital is exhausted. [Interpretation]

Recent News & Events

Has the business environment changed recently? Profoundly: a $1.5B PIPE (Jan 2026), a definitive $1.6B DoC/CHIPS funding package (June 2026), and three M&A deals (Serra Verde/Carester/TMRC) have transformed the company from a three-asset concept into a capitalized, government-backed roll-up. Significant acquisitions? Yes (above). Change in accounting policies? None material flagged beyond ordinary SPAC/warrant accounting. Recent changes — new markets, facilities, management? New facilities (Colorado hydromet demo; LCM-France planned plant); new geographies (Brazil via Serra Verde, France via Carester); CEO change (Ballard → Humpton) and board turnover. The news tape is thin and milestone/promotion-driven, with no adverse company-specific catalyst in the feed — the live risk is thematic de-rating, not a single headline. [Fact/Interpretation]

APPENDIX B — Source Appendix

USA Rare Earth, Inc. (NASDAQ: USAR) — Source Appendix (Appendix B)

Primary sources first. All URLs accessed 2026-06-19/20 unless noted. Company facts reconciled to SEC filings (EDGAR primary); third-party aggregated market data used as cross-checks and labeled.

A. SEC Filings (primary; mirrored to output/USAR/sources/)

• FY2025 Form 10-K (period 2025-12-31, filed 2026-03-30) — net loss $298.5M / EPS $(3.31); $244.5M non-cash warrant FV loss; SG&A $43.1M, R&D $15.9M; cash $359.9M; LCM consideration $197.7M ($103.1M cash + $94.6M equity / ~6.74M shares); 10.1M earnout (hurdles $15/$20); Round Top exploration-stage / no S-K 1300 reserves / late-2028 target; ~$4.1B long-term capex; ≥$600M equity by 12/31/2027 + $250M revolver by 12/31/2026. https://www.sec.gov/Archives/edgar/data/1970622/000197062226000021/usar-20251231.htm
• Q1-2026 Form 10-Q (period 2026-03-31, filed 2026-05-14) — cash $1,749.6M; operating burn $(18.5)M, capex $(38.6)M; $1.5B PIPE (69.8M shares, net $1.45B); govt LOIs/agreements $277M CHIPS + $1.30B loan + ~10% FD warrants & ~16.1M shares @ $17.17; anti-dilutive table 14,911K; cover 223,035,366 common + 1,224,351 Series A preferred; no going concern. https://www.sec.gov/Archives/edgar/data/1970622/000197062226000038/usar-20260331.htm
• FY2024 Form 10-K (filed 2025-03-31) — pre-transformation baseline (net loss $16.4M; cash $16.8M). https://www.sec.gov/Archives/edgar/data/1970622/000197062226… (corpus: sources/10-K/2025-03-31_ea0236222-10k_usarare.htm)
• 8-K 2026-01-26 — $1.5B PIPE securities-purchase agreement (69,767,442 shares @ $21.50) and the $1.6B DoC non-binding LOI. https://www.sec.gov/Archives/edgar/data/1970622/000121390026… (sources/8-K/2026-01-26_ea0274031-8k_usarare.htm)
• 8-K 2026-06-03 — definitive Direct Funding Agreement & Loan Guarantee Agreement with the U.S. Department of Commerce: $277.0M direct funding + $1.3B senior secured debt (15-yr, ~Treasury+150bps) under the CHIPS Act. https://www.sec.gov/Archives/edgar/data/1970622/000121390026064453/ea0293402-8k_usarare.htm
• 8-K 2026-06-05 / 2026-06-15 (and DEFA14A) — Serra Verde merger pro-forma financials & proxy materials (SVRE Holdings / Serra Verde Rare Earths, BVI). https://www.sec.gov/Archives/edgar/data/1970622/000121390026068491/
• 8-K 2026-04-20 / 2026-04-23 / 2026-05-13 — Serra Verde / Carester announcements; preliminary merger proxy (PREM14A); S-4. (sources/8-K/ and sources/PREM14A/2026-05-13_ea0290028-01.htm)
• 8-K 2026-03-05 — TMRC definitive all-stock merger agreement (~3.82M shares; Round Top 100% consolidation). (sources/8-K/2026-03-05_ea0280112-8k425_usarare.htm)
• DEF 14A (filed 2026-04-23) — beneficial ownership (officers & directors 9.1% / 8.9% voting; Gutnick 6.7%; Blitzer 33.6% of Series A preferred); SCT comp Humpton $11.70M (stock $11.5M, salary $167K), Ballard $5.52M, Steele $4.83M. https://www.sec.gov/Archives/edgar/data/1970622/000121390026047155/ea0287130-01.htm
• Schedule 13D/A (filed 2026-06-09) — Mordechai Zev Gutnick 15,580,745 shares / 6.8%. https://www.sec.gov/Archives/edgar/data/1970622/000092189526001587/
• Form 4s (mirrored, sources/4/) — Blitzer sell 2,091,849 @ $15.75 (8/13/2025); Blitzer buy 100,000 @ $21.44 & Trabuco buy 1,300 @ $22.60 (1/29/2026); Trabuco sell 13,000 @ $22.77 (6/8/2026).
• Form 144s (EDGAR) — Critical Minerals Trust 3,877,565 shares ~$105.1M (filed 6/5/2026); Bayshore Capital 176,778 shares ~$4.8M (filed 5/11/2026); Trabuco 13,000 (filed 6/8/2026).
• S-3 / S-3-A (filed 5/13/2026; effective 6/5/2026) — shelf for further issuance/ATM.

B. Management Commentary (hypothesis, not evidence)

• Q1-2026 earnings call (earnings-call transcript) — Stillwater Phase 1a commissioned (Mar 2026), customer-ready production to start Q2-2026, first sales 2H-2026, 600 mtpa target YE-2026; Round Top DFS due ~Q1-2027; LCM 3,000 mtpa by Q4-2026; first Yttrium metal pour (Apr 2026); CFO Steele: Q1 revenue ~$6M (LCM), capex ~$40M, cash ~$1.75B, adj. net loss $24.1M ($0.12/sh); no guidance; Investor Day planned Q3-2026.
• Q3-2025 earnings call (earnings-call transcript).

C. Peer & Industry Sources

• MP Materials investor relations (Q3/Q4-2025) — Mountain Pass NdPr ~720 mt/qtr; Independence magnet plant ramp; $110/kg DoD NdPr price floor; Apple/GM/DoD offtakes; consolidated revenue ~$53.6M Q3-2025. Market cap ~$8.38B, EV ~$7.64B, TTM revenue $254.3M.
• MP Materials–DoD deal (Jul 2025) — MP IR PR 2025-07-10; CGEP Columbia; FAS; Mining.com.
• Energy Fuels (UUUU) $725M DoD OSC conditional loan (Benzinga, 2026-06-18). NioCorp (NB), Lynas as basket references (factor-model data).
• China REE dominance & April-2025 export controls — IEA; CFR (“Leapfrogging China’s Critical Minerals Dominance”); China Briefing; ORF America; S&P Global (“Rare earth supply bottlenecks set to persist in 2026”).
• Round Top deposit — 2019 TMRC NI 43-101 PEA (NPV ~$1.56B, IRR ~70%, ~$350M capex); Junior Mining Network; NS Energy; ScienceDirect (Round Top rhyolite HREE deposit). Texas Semiconductor Innovation Fund $14.2M grant (GlobeNewswire 2026-05-12).
• Western-REE base rate — Molycorp 2015 bankruptcy; Lynas/JOGMEC financing (public record).

D. Market & Quantitative Data (public/third-party; reconciled to filings)

• Daily price history — https://azitrading.com/controls/download-data.php?t=USAR (price arc: post-deSPAC low $5.75, ATH $38.68 on 2025-10-13, $24.64 on 2026-06-18; 52-wk ~$9.61–$38.68; beta ~0.62/alpha; EMAs). Accessed 2026-06-20.
• Valuation percentiles (own-history) (2026-06-18) — P/B 7.41 (~92nd pctile of ~15-mo history), P/S 3.64, composite ~77th; P/E null (negative EPS). Caveat: near-useless on short history.
• News feed — sparse (3 unscored rows): Colorado hydromet demo / Q3-2026 first-oxide (Benzinga 2026-06-15); UUUU $725M DoD loan basket read-through (Benzinga 2026-06-18); Brazil rare-earth macro piece (2026-05-21).
• Fundamental data aggregator — get_company_profile / get_income_statement / get_balance_sheet / get_cash_flow / get_enterprise_value (snapshot EV flagged STALE at ~$14.79; balance-sheet inputs current) and MP comps. Accessed 2026-06-20.
• Factor/risk model (/stock-loadings, /leaderboard, /stock-info, /related-stocks USAR; 2026-06-18) — beta ~0.62, positive alpha, large negative Liquidity & LowVol loadings, Mining tilt; y1 return +94.5% (annualized) / vol 121% / max DD −69.2% / Sharpe 0.76; rs_peak −36.3; factor-similar peers HYMC (0.98), NioCorp (0.95), Energy Fuels (0.94), Standard Lithium, US Antimony, MP (0.84). Weak R² (~0.15) on <1yr history — indicative only.