The FORGE Architecture — Post 4: The Oklahoma Model

The FORGE Architecture  ·  FSA Critical Minerals Policy Series Post 4

The FORGE Architecture

Demand-Side Architecture for Domestic Critical Minerals Processing

The Oklahoma Model

The closest comparable facility to a hypothetical rare earth separation hub on the Arkansas River corridor already exists — on paper, in permits, and in partial government funding — on the Gulf Coast of Texas. Lynas USA's planned 5,000 metric tonne per year NdPr separation facility at Seadrift has a documented break-even above $60 per kilogram and requires $110 per kilogram for investment-grade returns. Move that facility 400 miles inland to the Tulsa Port of Inola, replace truck and rail logistics with McClellan-Kerr barge economics, add the Oklahoma industrial incentive architecture that just closed a $4 billion aluminum deal, and position Project Vault as the buyer of last resort for 60 to 70 percent of output. The numbers close. They close at the $110 per kilogram floor that the DoD already demonstrated moves an NdPr price from $51 to market-wide reality. The question is not whether the math works. The question is whether FORGE can enforce the floor that makes the math work for every compliant facility — not just the ones the DoD has written a bespoke contract for.

Randy Gipe  ·  Claude / Anthropic  ·  2026  ·  Trium Publishing House Limited  ·  Forensic System Architecture

Series Position — Post 4 of 5 Posts 1 through 3 built the argument. Post 1: the floor problem. Post 2: FORGE as the proposed solution. Post 3: the Inola aluminum smelter as proof that the five-factor model works when price protection exists. Post 4 is the numbers post. It models a specific facility — 5,000 to 10,000 metric tonnes per year NdPr-focused rare earth separation, Arkansas River corridor — against publicly documented comparables, and builds the capital stack that closes under FORGE conditions. The numbers are modeled from confirmed industry data, not invented. The FSA Wall is declared wherever the model departs from documented primary source figures into analytical projection.

The most instructive fact about the current state of American rare earth processing is not that it is underdeveloped. It is that the pieces required to develop it are almost entirely in place — and have been stalled, not by the absence of geology, logistics, or capital, but by a single missing variable that Post 1 named and Posts 2 and 3 documented the proposed solution to. The Lynas separation facility in Seadrift, Texas is permitted. The Energy Fuels White Mesa mill in Utah is processing monazite and producing mixed rare earth carbonate that has nowhere domestic to go for the final NdPr separation step. The MP Materials Mountain Pass mine is producing 40,000 metric tonnes of total rare earth oxide annually — 11.5 percent of the global market — and was until 2022 shipping most of it to China for the processing that does not yet exist at scale in the United States. The mine works. The feedstock exists. The logistics infrastructure, as Posts 3 and 4 of the Hidden Arteries series documented, is a navigable waterway connecting the interior of the continent to the Gulf Coast at the lowest freight cost of any mode available. What has not existed, until FORGE, is the price architecture that makes the processing facility financially rational to build and operate against Chinese structural pricing power.

This post models what the facility looks like when the architecture exists. The model is grounded in documented comparables — Lynas Texas, MP Materials' 10X facility, the DoD price floor analytics published by the Payne Institute and Crux Investor — and acknowledges the FSA Wall wherever it moves from documented data into analytical projection. The Oklahoma model is not a blueprint. It is a financial architecture stress test: what does a rare earth separation facility on the Arkansas River corridor require to close its capital stack, generate investment-grade returns, and operate commercially without permanent subsidy? The answer, when FORGE functions as announced, is: less than what the DoD is paying MP Materials for one facility, applied market-wide.

I. The Facility Specification

What We Are Modeling — and Why These Assumptions

The hypothetical facility is a neodymium-praseodymium focused rare earth separation and oxide production plant, sited in the industrial park at the Tulsa Port of Inola on the McClellan-Kerr Arkansas River Navigation System. Capacity: 5,000 to 10,000 metric tonnes per year of separated NdPr oxide, scalable toward heavier rare earth separation as feedstock and market conditions develop. Feedstock sources: mixed rare earth carbonate from Energy Fuels White Mesa (Utah) by rail to Inola, supplemented by monazite imports from allied Australian and African sources via Gulf Coast port and M-KARNS barge. Output: NdPr oxide for sale to FORGE member magnet manufacturers, DoD offtake commitments, and Project Vault stockpile reserve purchases.

The 5,000 tonne per year figure is not arbitrary. It is the exact capacity target of Lynas USA's planned Hondo, Texas light rare earth separation facility — the most directly comparable documented project in the American rare earth landscape. Using Lynas Texas as the primary comparable means the model's CapEx and operating cost figures are grounded in a real project that has gone through engineering design, permitting, and partial DoD funding rather than being constructed from industry averages. The Oklahoma model is Lynas Texas moved to a better logistics position with a stronger incentive architecture and the same FORGE floor price environment. Where the Oklahoma model diverges from the Lynas Texas template — primarily in logistics savings and the state incentive architecture — the divergence is documented and modeled from published sources.

5–10K

MT/Year NdPr Oxide Capacity

Lynas Texas is 5,000 MT/year. 10,000 MT scales to full industrial threshold. Model anchors at 5,000 MT Phase 1.

$800M–$1.8B

CapEx Range — Separation Facility

Lower bound: light REE separation only. Upper bound: integrated heavy REE + metallization. Midpoint ~$1.2B for NdPr-focused Phase 1.

$110–130

/kg NdPr — Investment-Grade Floor

DoD floor for MP Materials: $110/kg. FORGE target range for market-wide viability: $110–130/kg. Below $60/kg: operating loss territory.

II. The Price Floor Model

Three Price Bands — What Each Means for the Facility

NdPr pricing operates in three distinct economic zones for Western processors. Each zone has a different meaning for facility viability, capital structure, and the role of government support. The three bands are not projections — they are documented from the current price environment and the published economics of operating facilities.

Zone 1 · Below $60/kg — Chinese Pricing Zone

$48–60/kg · Operating Loss for Western Processors

Chinese domestic production cost: approximately $48/kg (documented in Lynas Texas analysis). Western processor break-even: above $60/kg. In this zone, a Western rare earth separation facility cannot cover its operating costs without government support, regardless of logistics efficiency or incentive structure. This is the zone that closed Mountain Pass twice and drove Molycorp to bankruptcy. The DoD-MP Materials deal's $110/kg floor exists specifically because the 2024 market price of $51/kg sits squarely in this zone — more than $9 below the minimum Western break-even.

Zone 2 · $60–95/kg — Operational Viability Zone

$60–95/kg · Operating Positive, Not Commercially Bankable

Above $60/kg, a Western rare earth processor can cover operating costs. Below approximately $95/kg, it cannot cover operating costs plus the debt service on $800 million to $1.8 billion in construction financing at commercial rates. A facility in Zone 2 can operate. It cannot service its capital stack without government grant offsets covering a substantial portion of construction cost. The DOE Industrial Demonstrations Program grant — up to $500 million for a qualifying facility — brings the effective net CapEx within Zone 2 range for the facility, but does not solve the commercial bankability problem for the project finance market that must fund the balance.

Zone 3 · $110–130/kg — Investment-Grade Zone

$110–130/kg · Commercially Bankable at 15–25% IRR

At $110 to $130/kg NdPr sustained over a ten-year offtake horizon — the FORGE reference price target range and the DoD-MP floor — a 5,000 MT/year NdPr separation facility with the logistics advantages of the Arkansas River corridor can achieve internal rates of return in the 15 to 25 percent range, sufficient for investment-grade project finance from commercial lenders. This is the zone in which JPMorgan Chase and Goldman Sachs committed $1 billion to MP Materials' 10X facility. This is the zone that Lynas Rare Earths confirmed it is already operating in following the DoD-MP floor announcement: "In the past six months, since the US Government's Price Protection Agreement with MP Materials was announced, the market price has increased to US$110/kg NdPr. Lynas, as the only firm producing both the light and heavy rare earth oxides required by our customers, is already benefiting from the improved market pricing." The floor moved the market. FORGE is the proposal to hold it there.

The Market Signal Effect — One Deal Moved the Price

The most significant data point in the post-DoD-MP Materials price environment is not that the DoD is paying MP $110/kg. It is that the market price for NdPr moved toward $110/kg after the deal was announced — and Lynas, which has no DoD floor contract, is already benefiting. A single bespoke government contract for one producer sent a pricing signal that elevated the market for all producers. If one bilateral deal moved the market, a plurilateral FORGE architecture enforced by 54 nations with adjustable tariffs would not merely signal a floor — it would structurally enforce one. The difference between a signal and a structure is the difference between what the DoD-MP deal accomplished for one company and what FORGE promises to accomplish for an industry.

III. The Capital Stack

How the Financing Assembles Under FORGE Conditions

A $1.2 billion rare earth separation facility on the Arkansas River corridor does not require a single government check for $1.2 billion. It requires a capital stack in which each layer is de-risked sufficiently that the next layer is willing to deploy. The FORGE architecture — reference price, adjustable tariff enforcement, offtake coordination, Project Vault backstop — is the mechanism that de-risks each layer in sequence. The stack assembles as follows, modeled from the documented capital structure of the DoD-MP Materials partnership and the Inola aluminum smelter.

Oklahoma REE Separation Facility — Modeled Capital Stack · $1.2B Midpoint Facility

DOE Industrial Demonstrations Program Grant — Non-repayable federal support for qualifying industrial facilities demonstrating novel manufacturing processes. Inola aluminum precedent: $500M. REE separation qualifies as industrial demonstration under current IDP criteria.

~$400–500M

Performance-based; phased deployment; reduces net CapEx to $700–800M

Oklahoma State Incentive Package — Performance-based rebate, TIF financing, tax exemptions, discounted industrial power rates. Inola aluminum precedent: $275M+ package. REE processing as "critical minerals" designation strengthens political case for comparable package.

~$150–275M

Performance-based; activates on capital spend and job creation milestones

Pax Silica / Sovereign Wealth Equity — Patient equity capital from Mubadala, Temasek, or comparable sovereign investors positioned in the Pax Silica architecture. Returns once FORGE floor makes revenue certain. Inola precedent: Mubadala committed $4B of sovereign capital to the same corridor.

~$150–250M

Equity layer; sovereign mandate extends investment horizon; not optimizing for quarterly returns

DPA Title III / DoD Procurement Authority — Defense Production Act authority for critical materials. Precedent: $400M equity in MP Materials plus $150M loan. For non-MP facilities, smaller targeted investments under DPA Title III available for defense-critical rare earth processing capacity.

~$100–200M

Government equity or loan; triggers private co-investment signal as demonstrated in MP deal

Commercial Project Finance Debt — Senior secured debt against FORGE floor revenue certainty + Project Vault offtake + Pax Silica equity cushion. Inola precedent: JPMorgan/Goldman committed $1B to MP Materials after DoD floor established revenue model. Same conversion logic applies to any FORGE-backed facility.

~$400–600M

Debt layer; bankable against 10-year offtake at FORGE reference price; commercial market follows the floor

Total Modeled Stack

~$1.2B–1.8B

The capital stack closes under FORGE conditions because each layer's risk is de-risked by the layer below it. The DOE grant reduces the net CapEx that commercial lenders must finance. The state incentive package reduces operating cost structure. The sovereign equity absorbs construction and ramp-up risk that public market investors cannot hold. The DPA investment signals government commitment, which reduces the perceived risk premium on the commercial debt. The commercial debt is secured against FORGE floor revenue — a price at which the facility generates sufficient cash flow to service the debt. Remove the FORGE floor and the bottom layer of the stack disappears. Without it, the commercial debt cannot be secured, the sovereign equity lacks a revenue exit, the DPA investment cannot be justified against a facility that cannot sustain operations, and the DOE grant funds a stranded asset. The floor is not one component of the capital stack. It is the condition that makes every other component rational.

IV. The Logistics Premium

What the Arkansas River Actually Saves — Modeled

The logistics savings from the McClellan-Kerr Arkansas River Navigation System are not incidental to the Oklahoma model's economics. For a processing facility whose feedstock and output must move in bulk — rare earth concentrate inbound, NdPr oxide and mixed rare earth products outbound — the mode of transportation is a material component of operating cost. The barge is the mode. The savings are real. The model below quantifies them against the truck and rail baseline, using published freight cost benchmarks.

Inbound Concentrate: White Mesa to Inola

Energy Fuels' White Mesa mill in southeastern Utah processes monazite sand into mixed rare earth carbonate — the feedstock that would supply the hypothetical Inola separation facility. The distance from White Mesa to Inola, Oklahoma is approximately 900 miles. By truck at $0.15 to $0.20 per ton-mile, that movement costs approximately $135 to $180 per tonne of concentrate. By rail, at $0.04 to $0.06 per ton-mile (where available), the cost falls to $36 to $54 per tonne. The M-KARNS rail-to-barge transloading at Inola's own terminal provides the rail leg from Utah, with the barge leg not applicable for this inbound movement — the rail connection is the relevant mode. At a facility processing 5,000 tonnes per year of NdPr oxide and requiring approximately 15,000 to 20,000 tonnes of rare earth carbonate feedstock annually (accounting for processing recovery rates), the difference between truck and rail on this corridor alone is $1.5 million to $2.9 million per year. For a facility whose profitability depends on operating cost efficiency to compete against Chinese processing economics, that is not a rounding error.

Outbound Oxide: Inola to Gulf Coast Export or Project Vault

The outbound movement — NdPr oxide and byproduct oxides from the Inola facility to Gulf Coast export terminals or Project Vault distribution points — is where the M-KARNS barge advantage is most significant. Barge freight on the Arkansas River to the Mississippi junction and down to Gulf ports runs approximately $0.01 to $0.02 per ton-mile, compared to truck at $0.15 to $0.20 per ton-mile and rail at $0.04 to $0.06 per ton-mile. The distance from Inola to the Gulf Coast port complex is approximately 650 to 700 miles by waterway. At 5,000 tonnes per year of oxide output moving by barge versus rail, the savings are $1.3 million to $2.6 million annually. By barge versus truck, the savings are $4.8 million to $6.5 million annually. Over a ten-year operating period, the barge advantage on outbound oxide alone is worth $13 million to $65 million in reduced operating costs, depending on the baseline mode comparison and actual throughput.

The Landed Cost Effect on Investment Economics

Combined inbound and outbound logistics savings of $3 million to $9 million annually reduce the facility's effective operating cost per kilogram of NdPr oxide. At 5,000 tonnes per year of output, that is $0.60 to $1.80 per kilogram in logistics savings — a meaningful reduction on an operating cost structure where the margin between the Chinese price ($48/kg production cost) and the FORGE floor ($110/kg) is the entire commercial viability of the facility. The logistics savings do not transform a nonviable facility into a viable one. They reduce the floor price required for viability, which means the FORGE floor does not need to be set quite as high to generate the same investment-grade return. Or, at the same floor price, the facility generates a higher return — improving the commercial case for equity investment and reducing the concessions required from government grant programs.

"The barge savings on outbound oxide alone — $13 million to $65 million over ten years — are not the reason to build the facility on the Arkansas River. They are the reason the facility on the Arkansas River is more bankable than the same facility at a less efficient logistics node. Every dollar of logistics savings is a dollar less of floor price required. The Inola corridor earns its position in the model." The FORGE Architecture — Post 4

V. The Project Vault Backstop

Why a Buyer of Last Resort Matters More Than It Sounds

Project Vault — the $12 billion U.S. Strategic Critical Minerals Reserve announced alongside FORGE on February 4, 2026 — performs a specific function in the capital stack that no other instrument replicates. Commercial project finance requires revenue certainty. FORGE price floors provide price certainty — the assurance that the per-kilogram realized price will not fall below the reference level. But price certainty without demand certainty is incomplete protection. A facility whose price is guaranteed but whose output cannot be sold has guaranteed pricing on zero revenue. Project Vault addresses the demand side of the revenue certainty equation: it is the committed buyer for output that cannot be absorbed by commercial markets at any given moment.

The buyer-of-last-resort function is most critical during two periods in a facility's life: the ramp-up period, when production is climbing toward design capacity but the commercial offtake network has not yet fully absorbed the output; and the market stress period, when Chinese dumping or geopolitical disruption has temporarily suppressed commercial demand for rare earth products from non-Chinese sources. In both cases, without a buyer of last resort, the facility faces a choice between selling into a market at prices below its FORGE floor (which would require government gap payments) or curtailing production (which imposes fixed-cost losses on a capital-intensive facility that cannot easily scale down). Project Vault eliminates that choice: the reserve purchases at the FORGE reference price, absorbing the volume commercial markets cannot take, maintaining the floor, and building a strategic inventory buffer that serves national security objectives simultaneously with its commercial backstop function.

For the capital stack model, the Project Vault backstop is quantified as committed offtake for 60 to 70 percent of the facility's output in the first three to five years of operation. At 5,000 tonnes per year of NdPr oxide at $110/kg, Vault backstop coverage of 3,000 to 3,500 tonnes per year represents $330 million to $385 million in committed annual revenue — sufficient to cover debt service on a $400 million to $600 million senior secured facility and provide the revenue floor that commercial lenders require to classify the debt as investment-grade. The remaining 1,500 to 2,000 tonnes per year is sold commercially, at FORGE reference prices, to FORGE member OEMs and downstream manufacturers. As the commercial market develops, Vault's share declines and commercial offtake grows — by year seven to ten, the facility is substantially commercially funded, with Vault serving a residual buffer function rather than a primary revenue role.

FSA Documentation — Oklahoma Model: Assumptions, Comparables, and FSA Wall Declarations

Model Component	Assumption Used	Primary Source / Comparable	FSA Wall
Facility capacity	5,000–10,000 MT/year NdPr oxide; Phase 1 anchored at 5,000 MT	Lynas Texas Hondo facility: 5,000 MT/year target (CSIS, discoveryalert analysis)	No FSA Wall on capacity comparability; Lynas is a direct analogue
CapEx range	$800M–$1.8B; midpoint $1.2B modeled	MP Materials/Lynas comparable projects; demo plants ~$77M; DoD-MP package $1.55B for integrated mine-to-magnet	FSA Wall declared: exact CapEx for a hypothetical Inola facility is not publicly documented; range derived from comparable project data
NdPr break-even	Above $60/kg for operating viability; $110–130/kg for investment-grade	Lynas Texas: break-even above $60/kg (discoveryalert); DoD floor $110/kg (Payne Institute, multiple sources); Chinese production cost $48/kg	No FSA Wall on break-even band; documented from Lynas analysis
Logistics savings	Barge at $0.01–0.02/ton-mile; truck at $0.15–0.20/ton-mile; rail at $0.04–0.06/ton-mile	U.S. DOT published freight mode cost benchmarks; Hidden Arteries series barge efficiency documentation	FSA Wall: specific freight rates on the M-KARNS for rare earth concentrate are not publicly published; benchmarks are from general bulk freight mode data
DOE grant	$400–500M from Industrial Demonstrations Program	Inola aluminum precedent: $500M DOE IDP (okenergytoday, Globe Newswire); same program, same state	FSA Wall: a hypothetical REE facility at Inola has not applied for or received a DOE IDP grant; the figure is a modeled analogue to the aluminum precedent
State incentives	$150–275M performance-based	EGA/Century aluminum: $275M+ state package (AGBI, Bond Buyer, okcommerce); Reindustrialize Oklahoma Act as template	FSA Wall: Oklahoma has not announced a REE processing incentive package; figure is modeled from aluminum precedent
Project Vault offtake	60–70% of output in years 1–5 at FORGE reference price	FDD, BPC analysis of Vault's buyer-of-last-resort function; no specific Vault offtake contract publicly documented	FSA Wall declared: specific Vault offtake terms for any facility are not published; figure is modeled from Vault's announced function
Sovereign equity (Pax Silica)	$150–250M from Mubadala/Temasek or comparable sovereign investors	Mubadala Inola precedent ($4B aluminum); Pax Silica membership of both entities (published reporting)	FSA Wall declared: no specific Pax Silica commitment to a hypothetical Oklahoma REE facility exists; figure modeled from sovereign capital positioning
Master FSA Wall	The Oklahoma Model is an analytical construct — a financial stress test of FORGE conditions applied to a hypothetical facility modeled from documented comparables. No specific rare earth processing facility at the Tulsa Port of Inola has been announced, permitted, funded, or proposed by any named entity. All figures are modeled projections from published industry data, not primary source disclosures from an actual project. The model demonstrates that the capital stack closes under FORGE conditions. It does not predict that the facility will be built, by whom, on what timeline, or at what exact cost. The FSA Wall is declared on every specific figure in the model.

FSA Framework — Post 4: The Oklahoma Model

Source

Lynas Texas + MP Materials as Primary Comparable Sources The model's credibility rests on its comparables. Lynas Texas Hondo (5,000 MT/year NdPr, documented break-even above $60/kg, $408M in DoD contracts) and MP Materials 10X Fort Worth (7,000 MT/year magnets, $1.55B total capital package, $110/kg floor, $1B private financing unlocked) are primary-source documented projects whose financial parameters ground the Oklahoma model. Where the model departs from those parameters — logistics savings, Oklahoma-specific incentives, Pax Silica equity — the departure is documented and the FSA Wall is declared.

Conduit

The Capital Stack as the Mechanism The capital stack is the conduit through which FORGE policy translates into physical facility. FORGE floor → revenue certainty → commercial debt bankability → private finance follows → sovereign equity takes equity position → government grants reduce net CapEx → facility is built. Each link in that chain is a conduit that requires the prior link to function. The post documents each link from publicly confirmed precedents: the DoD-MP deal demonstrated the floor-to-private-finance conversion; the Inola deal demonstrated the government-grant-to-sovereign-capital conversion; FORGE provides the plurilateral floor that makes both conversions market-wide rather than facility-specific.

Conversion

FORGE Floor → Investment-Grade IRR The conversion is the series' central claim, expressed in numbers: at $110/kg NdPr sustained, a 5,000 MT/year facility on the Arkansas River corridor — with $400–500M in DOE grant support reducing net CapEx, $150–275M in state incentives reducing operating costs, Project Vault providing 60–70% offtake certainty, and Pax Silica sovereign equity cushioning construction risk — generates an IRR in the 15–25% range sufficient for commercial project finance. Below $60/kg, the facility loses money operationally. Between $60 and $95/kg, it operates but cannot service its capital stack commercially. The conversion fires at $110/kg. The floor is the switch.

Insulation

Three Risks the Model Does Not Solve The Oklahoma model is robust to price risk under FORGE conditions. It is not robust to three documented risks. First, permitting risk: Lynas Texas has experienced significant delays from NORM wastewater permitting; an Oklahoma facility would face comparable permitting complexity. Second, feedstock risk: Energy Fuels White Mesa is the primary domestic monazite processor, but its output is not contractually committed to a hypothetical Inola facility; feedstock supply chain requires independent analysis. Third, technical execution risk: rare earth solvent extraction is more chemically complex than aluminum smelting, and a first-of-kind facility at this scale in Oklahoma carries execution risk that the model's IRR range does not fully price. The FSA Wall is declared on all three.

FSA Wall · Post 4 — The Oklahoma Model

The Lynas Texas break-even figure of "above $60/kg" is drawn from the discoveryalert.com.au analysis of the Lynas Seadrift facility, which states "break-even costs require NdPr prices above $60/kg, compared to current Chinese production costs of $48/kg." The $48/kg Chinese production cost figure is from the same source. These are secondary analytical figures, not primary cost disclosures by Lynas; the FSA Wall is declared on the precise Lynas operating cost structure, which is not publicly disclosed at the granularity the model assumes.

The assertion that "the market price has increased to US$110/kg NdPr" following the DoD-MP deal is drawn from the Rare Earth Exchanges interview with Lynas CEO Amanda Lacaze, published February 25, 2026: "In the past six months, since the US Government's Price Protection Agreement with MP Materials was announced, the market price has increased to US$110/kg NdPr." This is primary source language from the CEO of the world's largest non-Chinese rare earth separator. It is treated as a confirmed market observation. The FSA Wall is declared on the durability of the $110/kg market price, which may not be sustained if FORGE enforcement mechanisms are not established.

The logistics cost benchmarks — barge at $0.01–0.02/ton-mile, truck at $0.15–0.20/ton-mile, rail at $0.04–0.06/ton-mile — are drawn from U.S. DOT published mode cost data and the Hidden Arteries series documentation. The specific application of these benchmarks to rare earth concentrate and oxide movement on the M-KARNS is analytical projection; actual negotiated freight rates on these specific movements are not publicly documented.

The capital stack figures — DOE grant, state incentives, sovereign equity, DPA investment, commercial debt — are all modeled analogues to documented precedents (Inola aluminum, DoD-MP Materials). They are not commitments, proposals, or disclosures from any actual entity regarding any actual rare earth processing facility at the Tulsa Port of Inola. The master FSA Wall applies to every specific figure in the capital stack model.

The Project Vault "60–70% offtake" assumption is a modeled construct derived from the announced buyer-of-last-resort function described in FDD and BPC analyses of Vault. Specific Vault offtake terms, price commitments, and allocation mechanisms are not publicly documented as of the series publication date.

Primary Sources & Documentary Record · Post 4

1. Payne Institute for Public Policy — MP Materials/DoD partnership analysis; NdPr realized price $51/kg; DoD floor $110/kg; contract structure as "Contract for Difference" (PaineInstitute.mines.edu, public)
2. CSIS — "Developing Rare Earth Processing Hubs: An Analytical Approach," July 30, 2025; Lynas Texas capacity and permitting challenges; DOD investment totals; Texas as current U.S. hub; 45X credit uncertainty (CSIS.org, public)
3. Rare Earth Exchanges — Lynas CEO Amanda Lacaze interview, February 25, 2026; NdPr market price at $110/kg following DoD-MP deal; Lynas benefiting from improved pricing; strategic customer offtake model (RareEarthExchanges.com, public)
4. discoveryalert.com.au — Lynas Texas Seadrift facility analysis; 5,000 MT/year capacity; break-even above $60/kg; Chinese production cost $48/kg; NORM permitting challenges (discoveryalert.com.au, public)
5. Crux Investor — DoD-MP Materials deal analysis; JP Morgan EBITDA projection ($400M+ at floor); pricing bifurcation thesis; $1.55B total package breakdown (CruxInvestor.com, public)
6. C&EN (Chemical & Engineering News) — US investment in MP Materials; Lynas DOD contracts ($30M LREE + $120M HREE); "mine-to-magnet" supply chain complexity (CEN.ACS.org, public)
7. DoD — IBAS Program contract announcement: Lynas USA HREE separation facility; Seadrift Texas; $120M+ contract; strategic rare earth supply chain objective (BusinessDefense.gov, public)
8. ycharts.com — DOD awards history: $439M+ to MP, Lynas, Noveon since 2020; rare earth stocks surge following China export restrictions; timeline to meaningful capacity (ycharts.com, public)
9. Foundation for Defense of Democracies — Project Vault buyer-of-last-resort function; demand certainty mechanism; FORGE-Vault coordination rationale (FDD.org, public)
10. Bipartisan Policy Center — Project Vault structure: $12B total, $10B EXIM, $2B private; operational details pending (BipartisanPolicy.org, public)
11. U.S. Department of Transportation — Freight mode cost benchmarks; ton-miles per gallon; modal cost comparisons (Transportation.gov, public)
12. The FORGE Architecture — Posts 1–3; Hidden Arteries: Post 1 — Trium Publishing House Limited, 2026 (thegipster.blogspot.com) — floor problem, FORGE mechanisms, logistics benchmarks, Inola model

← Post 3: The Inola Proof Sub Verbis · Vera 珞 Post 5: The Full Stack →