Macroplane › Investment Theses › ReAlloys Stock ($ALOY): The Ex-China Heavy Rare-Earth Bottleneck Behind the Humanoid Buildout

ReAlloys Stock ($ALOY): The Ex-China Heavy Rare-Earth Bottleneck Behind the Humanoid Buildout

AI supply-chain thesis — mapping bottlenecks, focus companies, and supply-chain exposure for investors.

**Bottleneck theme:** Heavy rare-earth alloying (Dy, Tb, Sm) for NdFeB magnets **Focus:** $ALOY — REALLOYS INC. **Horizon:** 18–36 months for binary execution; staged milestones every 6 months. ReAlloys ($ALOY) is the underappreciated bet on ex-China heavy rare-earth metallization — the missing middle of the magnet supply chain that suddenly matters because humanoid robotics is finally on a real production curve. A single humanoid carries 40–50 NdFeB-wound actuators (every joint, every finger, every wrist), which is roughly **10–20× the rare-earth content of an EV per unit**. Unitree's $608M Shanghai STAR IPO at 67% gross margin on a $27,300 G1 humanoid is the public market's first humanoid pure-play; Tesla Optimus, Figure, and Apptronik are racing it; every one of them needs a non-China magnet path for export-control and defence-customer eligibility. China controls 90%+ of global heavy rare-earth (Dy, Tb, Sm) separation and alloying capacity. Light rare-earths can be mined and refined in the West today — $MP, $LYC.AX, $ILU.AX, $UUUU all sit on that side. The heavy stack is the chokepoint. Beijing's escalating rare-earth export controls have made standing up domestic alloying capacity a national-security imperative, and the same chokepoint applies to the magnets that go into Unitree's BLDC motors, Optimus's hip actuators, and Apptronik's hands. Without the Dy/Tb alloy stage, separated rare-earth oxides do not become magnets, and Western humanoid OEMs cannot ship to defence or strategic customers. ReAlloys (formerly USA Rare Earth's metallization arm) is one of a small handful of Western players with the metallurgical know-how to take separated Dy/Tb/Sm oxides and turn them into the alloys and ingots used in NdFeB sintered magnets. The Pentagon's interest is not theoretical — DPA Title III dollars, IBAS funding, and direct prime-contractor offtake agreements are flowing to qualified domestic alloyers — but the larger demand pool now is robotics. Every Unitree, Optimus, Figure, and Apptronik unit that ships through 2028 needs to know whether its magnets came from a Chinese sintering line that may not survive the next export-control escalation. **The pair trade is structurally cleaner than the single name.** $MP is the upstream miner. $USAR runs the sintered-magnet plant downstream in Oklahoma. ALOY is the metallization layer between them. Long any of them is long the same thesis; long the chain is long the bottleneck. Without ALOY-type metallization between mining and magnets, $MP's REO does not reach an Optimus motor and does not reach an F-35 actuator. **Three reasons this re-rates, not just incrementally appreciates:** 1. **The humanoid TAM dominates the defense TAM.** Even at conservative humanoid unit forecasts (1M units by 2030 at 5–8 kg NdFeB per unit), the magnet demand from humanoids exceeds the entire current F-35 + missile + radar magnet stack by 5–10×. Defense was the catalyst funding the build; robotics is the demand that pays back the capex. 2. **Beijing's export controls re-price every kg of qualified ex-China heavy RE alloy.** The Dy/Tb price spread between Chinese benchmark and ex-China-qualified material is the real revenue line for ALOY — and that spread widens, not narrows, every time export rules tighten. 3. **Verticalized OEMs reset the buying pattern.** Unitree, Tesla, BYD, and the Chinese humanoid cluster will buy domestically when they can. Western OEMs cannot — they need a parallel non-China stack, and ALOY is one of the few names that can sell into it. This is binary execution. Scale a process that has never run at industrial volume outside China, in a politically protected market with subsidised capex and pre-committed offtake. The miss case is that the demand pull is real but ALOY cannot scale fast enough to capture it before a competitor (Quadrant, LCM, Solvay, or a fast-following Korean alloyer) takes the offtake. ## Validation metrics The thesis is **confirmed** if most of these green-line in 12-month windows, **invalidated** if 3+ red-line in 6-month windows. 1. **DPA Title III / IBAS / DOE awards to ALOY** — cumulative ≥$150M by Dec 2027. 2. **Annual heavy-RE alloy throughput** — ≥250 tons/yr by 2027; ≥1,000 tons/yr by 2029. 3. **Named offtake agreements** — ≥3 signed offtakes with magnet OEMs or humanoid platforms by mid-2027. 4. **Customer disclosure naming ALOY** — $MP, $USAR, JL Mag, Quadrant, Vacuumschmelze, or a humanoid OEM citing ALOY by name in earnings or 10-K (rumoured offtakes do not count). 5. **Dy/Tb spot-price spread vs. Chinese benchmark** — widening spread (ex-China > China) confirms pricing power; compression invalidates it. 6. **Humanoid OEM ex-China magnet path announcements** — Tesla, Figure, Apptronik, or 1X explicitly designating an ex-China magnet supplier. 7. **Quarterly cash burn vs. subsidy + pre-payment offset** — net burn ≤$40M/q absent capex catalysts; below this and a bridge financing/dilution event is likely. 8. **Qualification timelines with magnet OEMs** — first commercial-grade qualification complete by Q2 2027 (magnet qualification is typically 12–18 months and gates offtake conversion). 9. **Inventory of separated Dy/Tb feedstock on-site** — ≥6 months of forward production demand. 10. **Pentagon strategic-stockpile commitments** — multi-year DLA-Strategic Materials contract covering ≥30% of nameplate output. **Falsifiability:** if by Dec 2027 ALOY has shipped <100 tons/yr cumulative, has no named-OEM offtake, and the Dy/Tb spread has compressed back to Chinese benchmark, the thesis is wrong regardless of price action. ## Horizon and exit triggers | Window | What to watch | Position posture | | --- | --- | --- | | **0–6 months** | DPA Title III announcement; first humanoid OEM letter of intent; Beijing rare-earth export rule updates | Build into IPO/grant catalysts; size for binary outcome (≤3% of book) | | **6–12 months** | First commercial qualification; offtake count 0→2-3; Unitree post-IPO trading reset prices in the magnet chain | Add on confirmed offtake; cut on missed milestones | | **12–24 months** | First production-volume runs (250+ tons/yr); humanoid OEMs disclose supply chain in 10-K | Trim into news events; binary outcome resolves here | | **24–36 months** | Steady-state ramp at 1,000+ tons/yr; multi-year offtake stack visible in financials | Hold winners; exit losers cleanly | **Exit (positive):** ≥1,000 tons/yr with ≥4 multi-year offtakes by end-2028 → trim 50% into the re-rating, hold the rest. **Exit (negative):** any of (a) DPA Title III award delayed past Q2 2027; (b) first commercial qualification slips past Q4 2027; (c) Dy/Tb spread compresses for two consecutive quarters; (d) a competing Western alloyer (LCM, Quadrant, Solvay) wins the same offtake ALOY is bidding on. This is not investment advice. Do your own research and confirm the supply-chain wiring on Macroplane before sizing a position.

Focus companies in this thesis (3)

  • REALLOYS INC. (ALOY)
  • MP Materials Corp. / DE (MP)
  • USA Rare Earth, Inc. (USAR)

Supply-chain categories covered

  • Rare Earths — Investment-thesis bucket from bottlenecks.app: Rare Earths
  • Permanent Magnets — Production of NdFeB and other rare earth permanent magnets using heavy rare earth alloys.
  • Specialty Metals & Alloys — Tungsten, titanium, cobalt, and other high-performance metals
  • Heavy Rare-Earth Separation — Solvent-extraction and ion-exchange separation of heavy rare-earth elements (dysprosium, terbium, samarium) from mixed concentrates. The narrow upstream chokepoint that determines whether NdFeB sintered magnets keep their coercivity at humanoid joint and EV traction temperatures. Distinct from light-RE separation in that the cascade requires far more stages, China holds ~90% of global capacity, and the metallization step that follows uses different alloy chemistry.
  • Wind Turbines — Wind turbine generators, blades, and gearboxes
  • Actuator Subsystems — Integrated actuator assemblies combining motors, reducers, encoders, and bearings into a single joint module — the layer where Chinese integrators and Western incumbents compete for OEM design wins.
  • EV OEMs — Automakers integrating battery packs into electric vehicles for consumer and fleet markets.
  • Defense Electronics — Radar, EW systems, military communications, and avionics

Thesis milestones & bottleneck markers

  • DPA Title III award — Pentagon separation capacity funding
  • Mountain Pass expansion — MP — MP heavy RE output scale-up
  • Defense prime qualification — First major defense OEM qualifies ALOY heavy RE alloys
  • ALOY revenue inflection — ALOY

Related baskets

  • Rare Earths

More AI supply-chain theses

  • Comfort Systems (FIX) — MEP/HVAC contractor riding the AI data center super-cycle
  • Mission-critical electrical/mechanical construction
  • Celestica Stock (CLS): AI Hardware Manufacturing for Hyperscalers
  • POET Stock Thesis: The Photonic-Integrated CPO Engine Outlier
  • Utility transformers — the actual grid bottleneck (2–4yr lead times)
  • Kulicke & Soffa (KLIC) — Strait of HBM TCB
  • WaveLogic 6 Extreme 1
  • Etch monopoly for HBM TSV + 3D NAND staircase + advanced packaging
  • Boyd $9
  • CoreWeave Stock ($CRWV) Thesis: The Largest Pure-Play AI Cloud at Scale
  • Hybrid-bonding integrated tool jointly with BESI
  • Cisco Acacia

Browse all AI supply-chain theses · Macro trends · Industries · Product categories · Baskets