Thematic: Critical minerals — Phase 2 of the capital cycle, the 2010 base rate, and the price of being wrong on the duration of Chinese export controls

2026-05-29 · long-form · v1

Executive summary

The critical-minerals story the consensus tape is pricing is real, and it is partly mispriced. China controls roughly 60–70% of rare earth mining and over 85% of processing; Beijing's April 2025 controls on heavy rare earths remain active, and the May 20, 2026 Commerce Ministry reaffirmation killed the idea that the post-summit framework was going to relax them quickly T3. Western capacity is being built — Mountain Pass through MP Materials, the Stillwater plant under USA Rare Earth, Lynas's Mt Weld expansion and Seadrift Texas heavy-separation facility, the Australian forced-divestment around Northern Minerals — and the price tape has rewarded the listed producers heavily. NdPr oxide is near $100/kg against a January 2026 print closer to $53, a roughly 90–100% move in five months T3. The Pentagon is now the largest shareholder of MP Materials at an effective 15% stake on a $400 million preferred plus warrants, with a $110/kg NdPr oxide price floor underneath the company T1.

The Friday read is that this is Phase 2 of a Marathon-style capital cycle, not the start of a permanent regime shift in supply economics. Equity is flowing heavily into Western producers — a $1.5 billion PIPE into USA Rare Earth, a separate $1.6 billion Department of Commerce facility, the $400M plus $150M DoD package into MP Materials, a $1 billion bank loan from J.P. Morgan and Goldman Sachs supporting the 10X facility T1; capex is rising sharply; new project announcements are proliferating; sell-side coverage and conference focus have surged. Capital influx of this magnitude historically presages oversupply within a decade and a return of returns toward cost of capital. The 2010 Japan/Senkaku episode — NdPr prices up roughly 10x within a year, then collapsing by 2014–2015 after Western capacity built and China dropped quotas under the WTO ruling — is the textbook base rate, and Molycorp's 2015 bankruptcy is the textbook outcome for the listed-equity bet on the first wave T2.

The asymmetry that matters for AlphaSteve is therefore not "rare earths are a structural winner" — that framing is consensus and is already in price. It is that the duration of Chinese export controls is the load-bearing variable, the price floor changes the failure mode but not the cycle, and the names that survive Phase 4 will not be the ones that ran fastest in Phase 2. Margin of safety on MP Materials lives below the $60 trigger and is sharpened by the DoD floor; on Lynas, on USA Rare Earth, on Northern Minerals, the case is weaker on entry-price grounds given the prints already harvested in 2026 to date. The defense primes and downstream OEMs are second-order beneficiaries with smaller P&L sensitivity. This dossier names the theme, ranks the names, and sets the discipline.

House view reconciliation

The current house view treats critical minerals as a "multi-year structural — dossier-ready" theme with three prior confirmations (China May 20 reaffirmation, MP Materials Q1 print, Australia divestment) plus a fourth from 24/7 Wall St.'s Real Alloys CEO coverage on 2026-05-24 _house-view. The kit Sunday note 2026-05-24 AM flagged the theme as having crossed the dossier threshold and named MP Materials as the lead candidate for a thesis pass with trigger work 2026-05-24-AM. Today's Friday AM extended the same: "MP Materials thesis pass remains top priority this week and is decoupled from the Iran trinary" 2026-05-29-AM.

This long-form does not conflict with the house view; it operationalizes it. It promotes the theme from "dossier-ready" to "dossier (v1) — confirmed multi-year structural, Phase 2 capital-cycle reading with explicit base-rate-grounded variant perception." It does not yet promote any individual name to a full position — the deep-value discipline requires margin-of-safety entry prices, and only MP Materials has a kit trigger ($60) within striking distance of the current tape. Specific house-view edits at the end of the report.

The setup

The shift from "rare earths are a niche policy debate" to "rare earths are a Phase 2 capital cycle" did not happen this month. It happened in two stages. The first stage was Trump's January–March 2025 critical-minerals push and the Executive Order 14241 Presidential Determination on April 7, 2025 that qualified critical minerals for DPA Title III authority and delegated Section 303 powers to the Secretary of Defense T2. The second stage was Beijing's response: the April 2025 controls on seven heavy rare earth elements — dysprosium, terbium, yttrium, scandium, samarium, gadolinium, lutetium — and a broader October 2025 package of controls on processing technology and refining licenses that was suspended for one year under the November 2025 US-China trade truce but expires November 2026 T3.

The May 2026 update is that the controls did not relax meaningfully after the Trump-Xi summit. The Ministry of Commerce statement on 2026-05-20 called the regime "lawful" and pledged cooperation only with "reasonable" US concerns, language consistent with maintaining the post-April-2025 baseline indefinitely T3. The April 29 enforcement draft tightened penalties for unauthorized separation and quota breaches 2026-05-24-AM. Beijing's strategic posture has hardened, not softened, in the months since the summit framework was negotiated.

In the same window, three load-bearing things changed on the Western supply side. MP Materials reported its strongest quarter on record with revenue of $132.9 million (+28% QoQ), 917 metric tons of NdPr oxide production (+63% YoY), and the first $21.1 million of Magnetics Segment revenue from its Independence facility T3. The Australian government, through Treasurer Jim Chalmers, ordered six Chinese-linked shareholders representing roughly 17.58% of Northern Minerals to divest by July 2, 2026 — the second forced-divestment escalation in two years and the cleanest sovereign-level data point of resource nationalism extending into the equity ownership of the ore body itself T3. USA Rare Earth raised $1.5 billion in a January PIPE plus an additional $1.6 billion Department of Commerce facility in April 2026 and commissioned Phase 1a of its Stillwater magnet manufacturing facility in March, with a Q4 2026 run rate of 600 metric tons per year and Phase 1b ramping to 1,200 MTPA by Q1 2027 T1.

The setup is what Marathon called a Phase 2 capital cycle in textbook form. Returns have accelerated (NdPr oxide from roughly $53/kg in January 2026 to roughly $99.61/kg on May 1, a 88% rise; magnet metal NdPr from a base of $126.16/kg on April 1 to $136–$139/kg by late April, +9% MoM extending a YTD gain of approximately 160% T3). Capital is pouring in — billion-dollar equity rounds, sovereign equity stakes, bank lending lines, government loan facilities, foreign-investment review orders forcing rotation of equity capital out of Chinese hands and into Western hands. Sell-side coverage has expanded, conference focus has migrated, mainstream press treats the theme as a foregone conclusion. The reaction in the listed equity has been sharp — MP Materials has roughly doubled off its 2024 lows; USA Rare Earth IPO'd and traded heavily; Lynas's revenue mix has lifted on the prices, not the volumes. The framework's question is not "is there a theme" — there is — but "which phase, what is the base rate, what is the duration."

The analysis

Phase identification — where are we in the cycle

Chancellor's five-phase formulation T2 is the right scaffolding. Phase 1: returns above cost of capital, no one notices. Phase 2: returns attract capital, new entrants emerge, capex rises, equity issuance accelerates, bank lending grows, M&A multiples expand. Phase 3: new capacity online, supply meets demand, margins compress. Phase 4: capital retreats, levered players fail, capacity exits. Phase 5: supply tightens, returns restore.

The critical-minerals cycle does not move uniformly across all elements, and the distinction matters. Light rare earths (lanthanum, cerium, neodymium, praseodymium) cycled hard in 2010–2015 and again in 2020–2022 and are now well into Phase 2 of a third cycle. Heavy rare earths (dysprosium, terbium) are in Phase 1–2 with Western supply largely absent; this is where the April 2025 controls bite hardest and where the Lynas Seadrift Texas facility and the Stillwater Phase 1 magnet line are addressing supply that simply does not exist outside China today. Lithium is in a separate cycle that has rotated into Phase 3 over 2024–2025 after the 2021–2023 capital influx; that piece does not belong in this dossier and is parked.

For light rare earths and the magnet supply chain, the Phase 2 evidence is overwhelming. Capex-to-depreciation ratios across the listed Western producers are running materially above 1.2x — MP Materials is reinvesting more than its trailing operating cash flow in the Independence facility scale-up and the planned 10X facility; USA Rare Earth's Phase 1b and the Cheshire, UK metal-making expansion are both new capital with no offset; Lynas's Mt Weld flotation expansion to a 12,000 tonne per annum NdPr target is incremental capacity coming online over 2026–2028 T2. Equity issuance has been heavy in the cohort — the $1.5 billion USA Rare Earth PIPE in January is the cleanest single signal — and bank-led financing has joined (J.P. Morgan and Goldman Sachs co-led the $1 billion 10X facility loan T3). New project announcements are proliferating: Arafura's Nolans project advancing in Australia, Iluka's Eneabba processing facility, the Texas Seadrift heavy separation plant, the UK Cheshire metal-making expansion, the Stillwater Phase 1a/1b ramp, the Independence and 10X facilities at MP Materials. Sell-side coverage has expanded sharply since the Pentagon-MP Materials announcement on 2025-07-10.

This is what Phase 2 looks like. The framework's discipline is unambiguous: returns are accelerating because capacity has not yet arrived; once capacity arrives — Phase 3 — returns compress and the cohort's equity values follow. The cycle in light rare earths runs 7–10 years on Marathon's mining-clock convention; the 10X facility "first commercial product expected in 2028" timeline T2 and the Stillwater Phase 1b "Q1 2027" milestone T1 anchor when supply pressure begins to relax. The window during which the current price levels sustain is therefore 2026–2027 at the longest, with Phase 3 arriving sometime in 2027–2028 absent further Chinese tightening that pushes the curve out.

The capital flows — what is being built and by whom

The cleanest single way to read a capital cycle is to track the cash-flow direction. Money is moving into Western rare-earth and magnet capacity at a rate that exceeds the entire 2010–2015 cycle's Western private capital aggregate, and the composition is materially different.

At MP Materials, the DoD $400 million convertible preferred plus warrant package combined with a separate $150 million Office of Strategic Capital loan and the J.P. Morgan / Goldman Sachs $1 billion private-credit facility represents roughly $1.55 billion of incremental capital allocated specifically to the 10X facility build and heavy-rare-earth separation at Mountain Pass T1. The $110/kg NdPr oxide price floor is a structural change with no 2010–2015 analog: government is not just a financier but a counterparty with a minimum-price contract for a commodity output. DoD owning roughly 15% of MP Materials means the government has a margin-call risk on its own equity stake and an incentive to keep the price floor live and the operating capacity expanding. Marathon's framework treats government as a Phase 2 source of capital influx, with the modification that government capital is "stickier" than private equity in the Phase 4 retreat — it does not redeem.

At USA Rare Earth, the $1.5 billion PIPE plus the $1.6 billion Department of Commerce facility totals roughly $3.1 billion of equity and quasi-equity capital flowing into a company whose Stillwater Phase 1a magnet output is scaling toward 600 MTPA by end-2026 and Phase 1b to 1,200 MTPA by Q1 2027 T1. The capital intensity here is high — billions of dollars to enable a thousand-tonne-per-year magnet output line that competes against tens of thousands of tonnes of Chinese capacity. Adamas Intelligence's calibration is that the industry needs 42 new magnet production facilities globally by 2030 against only 11 in active development; lead times for new facilities run 4–6 years minimum T2. The capital that has been committed is sufficient for perhaps 4–6 of the missing 31 facilities, depending on geography and capital intensity.

At Lynas, the path is different. The Mt Weld expansion to a 12,000 tonne per annum NdPr target is brownfield rather than greenfield, the Kalgoorlie processing facility produces mixed rare earth carbonate as feedstock for Malaysia, and the Seadrift Texas plant is purpose-built for heavy rare earth separation — the dysprosium and terbium that China's April 2025 controls most directly affect T3. Q3 FY26 gross sales revenue of A$265 million (+115% YoY, +31% QoQ) on volume of 3,131 tonnes (+29% QoQ) at an average selling price of A$84.6/kg (+67% YoY) is the cleanest single income-statement print in the Western rare-earth cohort and is what mid-Phase-2 revenue acceleration looks like in real time T3. Lynas's strategic position — the only large-scale rare-earth refiner outside China currently producing separated REEs at industrial scale — would matter most if the cycle never moved past Phase 2; the Phase 2 framing is precisely that this position becomes less unique as capacity arrives.

At Northern Minerals, the Browns Range project in Western Australia is differentiated by 70% of project value coming from dysprosium and terbium — the heavy REEs most exposed to Chinese controls T3. Northern Minerals is the cleanest pure-play on the heavy-REE side and the Treasurer's May 17 divestment order — six entities representing roughly 17.58% of issued capital ordered to sell by July 2, 2026 T3 — is the cleanest sovereign-level signal of resource nationalism extending into the equity ownership of the asset. The base-rate caveat is that forced rotations of capital out of Chinese hands and into Western hands do not by themselves build refineries.

The pattern is unmistakable. Returns are above cost of capital; capital is flowing in at unprecedented rates; new entrants are forming (USA Rare Earth, Arafura, Iluka, Energy Fuels, Iluka Resources in Australia and others); bank lending is liberal; sovereign equity is participating; M&A activity is emerging. This is Phase 2 of a cycle.

Base rates — what actually happened in 2010, and what does it imply for 2026–2030

The 2010 China–Japan rare earth episode is the single cleanest reference point and is required to honestly frame the 2026 thesis. The sequence: a Chinese fishing boat collided with two Japanese coast guard vessels off the Senkaku Islands on September 7, 2010; the trawler captain was arrested; Chinese customs agents halted REE exports to Japan within weeks; the halt lasted roughly two months; Chinese controls (export quotas) on REEs more broadly tightened through 2010–2011 T2. The price reaction was sharp — global rare earth prices rose roughly 10x in the year following the incident T2.

Three things then happened. First, capital flooded into Western projects — Molycorp's Mountain Pass restart, Lynas's Mt Weld and Kuantan refinery, exploration projects in Canada, Greenland, South Africa, Brazil, Vietnam. Second, by 2013–2014 Chinese supply began to relax (quotas were softened, exports normalized) and the WTO ruled against China's restrictions in 2014, leading China to drop the quotas formally in 2015 T2. Third, prices collapsed back to or below pre-incident levels by 2014–2015, and Molycorp — the listed-equity vehicle that had raised over $1 billion of capital in the post-2010 cycle — filed for Chapter 11 bankruptcy in 2015 T3.

Japan's import dependence on China for REEs fell from approximately 90% pre-2010 to approximately 58–60% by the mid-2020s, mostly through the Lynas/JOGMEC partnership, recycling investment, and substitution research T2. Diversification was real and meaningful but did not become independence. China retained dominance of separation and processing throughout.

The base rates from this single, very-relevant episode tell three things. First, the cycle peak in REE prices following a Chinese supply shock has historically been roughly 4–5 years from the trigger event (peak prices 2011–2012 vs. the September 2010 trigger), with price collapse arriving by year 4–5 (2014–2015). Second, Western capacity that came online in the post-2010 cycle generally did not survive the price collapse intact. Molycorp went bankrupt; Lynas survived only because of JOGMEC's $250 million 2011 capital injection and a multi-year, sub-cost-of-capital period of operating losses through 2015–2018. The base rate for "Western listed-equity rare-earth producers that survived a full cycle with positive cumulative returns to common shareholders": approximately one (Lynas), with caveats. Third, Chinese supply discipline is not durable in the face of WTO action and US/EU economic pressure, historically — China dropped the quotas in 2015 after the WTO ruling.

The 2026 setup differs from 2010 in several specific ways that matter for the base rate. The Pentagon's $400 million equity stake plus the $110/kg NdPr oxide price floor at MP Materials is a structural change with no 2010 analog; government as principal equity holder is different from government as customer or financier T1. The April 2025 Chinese controls focus on heavy rare earths (dysprosium, terbium) and processing technology rather than the broad export quotas of 2010–2014, which makes the WTO route narrower because the controls are framed as national-security export controls rather than trade-restrictive quotas. The downstream demand picture is qualitatively larger: Adamas Intelligence's 2030 magnet REE consumption forecast of $15.65 billion (up 5x from $2.98 billion) is driven by EV traction motors (estimated 200,000–400,000 tonnes of NdFeB demand at 100 million EVs annually by 2030) and offshore wind (roughly 190,000 tonnes of magnets to meet 2030 capacity targets against current ~80,000-tonne annual production capacity) T2. The supply-demand imbalance is more durable than 2010's, which means the Phase 3 transition arrives later — but it does arrive.

The Mauboussin discipline is to use these base rates, not the inside view. The inside view says "this time China has weaponized critical minerals permanently; Western capacity must build." The outside view says "in the one prior episode of comparable size, prices peaked at year 4–5 from the trigger and the Western producers' equity values were destroyed at the Phase 3 transition; the structural change vs. 2010 is the price floor and the demand picture, not the cycle dynamics." Both views can be partially correct simultaneously: the demand picture is real and supports higher trough prices than 2014–2015, but the cycle dynamics still operate.

The variant perception — what the cohort is over-pricing and where the asymmetry lives

The consensus framing of the Western rare-earth cohort treats it as a portfolio of structural winners. The variant perception, which the cycle framework supports, is that this is Phase 2 pricing of names that will face Phase 3 within 24–36 months on the demand side and Phase 4 within 5–7 years on the supply side, and the cohort multiple expansion in the listed names is over-extrapolating both the duration of Chinese controls and the moat depth of the surviving Western producers.

The cleanest specific evidence: MP Materials trades at a market capitalization in the $9–10 billion range at $60 against Q1 2026 annualized EBITDA of approximately $146 million — an EV/EBITDA multiple in the 60–70x range that is pricing the 10X facility ramp through 2028 plus a sustained price premium with no Phase 3 transition embedded in the forward expectations T3. Lynas's Q3 FY26 revenue +115% YoY and +67% YoY pricing is already past the consensus implied trajectory through end-2026; the marginal data point is whether the run rate sustains into a Phase 3 transition or whether the next twelve months see further price acceleration. USA Rare Earth's $1.5 billion PIPE and $1.6 billion Commerce facility are providing the capital to build out capacity that, on Adamas's framework, partially fills a 110,000-tonne 2030 structural deficit — but the equity is being priced as if the company captures a disproportionate share of the rent, which is what the multiple needs to do work but is the part of the base rate that Molycorp's 2015 bankruptcy most directly stress-tests.

The variant view falsification tests are explicit. The variant view holds if (a) China relaxes export controls before November 2026 expiry of the broader suspended package, or under Trump-Xi negotiation pressure, or because the controls themselves are creating retaliatory damage to Chinese auto and industrial customers; (b) Western capacity additions come online materially on the announced timelines (Stillwater 1b in Q1 2027, MP 10X first product in 2028, Lynas Mt Weld at 12,000 tpa, Seadrift Texas operational) and prices begin to compress through 2027; (c) demand-side substitution accelerates faster than expected (Tesla's well-publicized 2023 commitment to a magnet-free motor in next-generation Model 3 / Robotaxi designs is the canonical example; ferrite magnets and reluctance motors are the technology substitutes); (d) the DoD price-floor regime extends to additional producers, increasing capacity supply commitments and dampening pricing power. The variant view is falsified, and the consensus framing wins, if (i) Chinese controls extend through 2028 and the heavy-REE supply position becomes genuinely binding; (ii) capacity additions slip materially; (iii) demand growth runs ahead of even Adamas's projections; (iv) US/EU policy escalates from controls to direct sanctions and Western producers gain pricing power they currently lack.

The most useful framing borrowed from Mauboussin's Expectations Investing is to ask what implied expectations the current price embeds. At a 60x trailing-EBITDA multiple, MP Materials' price implies the 10X facility ramps on schedule, the price floor is structurally relevant rather than a floor against a higher market price, the Independence facility generates meaningful magnet sales by H2 2026 as guided, and there is no Phase 3 compression in NdPr pricing through at least 2030. This is a stack of assumptions where the joint probability is below the consensus framing. The deep-value posture is to wait for either a pullback that absorbs some of this premium or for data that materially tightens one or more of the assumptions.

Variant perception — explicit

The market is pricing the Western rare-earth cohort as if Chinese export controls are permanent, capacity additions arrive on time and at planned cost, demand growth meets or exceeds Adamas's projections, and the listed producers capture the rent of the regime shift. AlphaSteve's variant perception is that capital cycles in this industry have not historically operated that way, the 2010 base rate is unambiguous on the eventual return to lower prices and the survivorship cost in listed equity, the DoD partnership and the price floor have changed the failure mode but not the cycle, and the equity values at current prices are pricing structural moats that the Phase 2 capital influx is in the process of building competitors against.

The variant view extends from the AI-infrastructure capacity dossier's variant view that the cohort multiple expansion is over-extrapolating the duration of supply tightness. The pattern is the same: structural change is real; pace and depth are over-priced; capital influx is the leading indicator that returns will compress within the cycle's natural length. Both variant views can be wrong simultaneously if supply additions are slower and demand growth is faster than the base rate predicts. The discipline is to size to the base rate, not the inside view.

Base rates — explicit

Per the Friday template requirement and the Mauboussin Base Rate Book discipline:

The base rate question for this dossier is "what is the historical frequency of Western capacity buildouts in response to Chinese rare earth export restrictions achieving structural displacement of Chinese supply over a five-year horizon." The relevant reference class is small: 2010–2015 is the one prior episode of comparable scale and intent. The outcome on that one observation: Japan's Chinese dependence dropped from ~90% to ~60% over a decade T2, Western listed-equity producers experienced one survival (Lynas, with JOGMEC's 2011 $250M capital injection) and one bankruptcy (Molycorp, 2015) T3, and Chinese supply discipline broke down within 4–5 years of the trigger under WTO pressure T2.

The base rate calibration is therefore: partial diversification of Western supply over 5–10 years is achievable; full independence is not on the 2010 evidence; and the cycle's compression phase historically arrived at year 4–5. The DoD price floor changes the survivorship math at the bottom but not the cycle dynamics at the top.

The second base-rate question is "what is the base rate for state-backed industrial policies achieving stated capacity targets within stated timelines." Mauboussin's general work and the broader practitioner literature on government industrial policy suggests this rate is poor across sectors — semiconductor industrial policies through the 1990s–2010s, alternative energy buildouts under Solyndra-class programs, agricultural reformulation programs all show high rates of slippage. Where government takes a principal equity stake — as DoD has done in MP Materials — the rate appears to be modestly higher historically (F-35 program, GM 2009 bailout completion vs. Solyndra-class minority-equity disasters) but the data are too sparse to weigh heavily.

The third base-rate question is "what is the base rate for resource nationalism continuing over decade-long horizons." Mixed: the 1973 oil embargo, OPEC's 1979 cycle, copper nationalization in Chile, and various nationalization episodes in the 1960s–1970s show varying durations. A single-supplier-country export control regime has historically not lasted indefinitely under WTO and bilateral diplomatic pressure, with the 2010–2015 China REE episode itself the canonical example of a 4–5 year duration.

The Friday discipline is to weight these base rates against the inside view that "this time is different." The base rate says four out of four prior episodes ended in price collapse and producer-equity destruction within 5–7 years. The inside view says (a) DoD price floor, (b) demand structurally larger, (c) controls focused on heavy REEs where Western supply is genuinely absent. Both can be partially true; the relative weight is what determines position sizing.

Implications for AlphaSteve

The top-down implication is that critical minerals is now a confirmed multi-year structural theme worthy of dedicated dossier tracking, with the variant view that the cohort multiple expansion is Phase 2 of a capital cycle rather than the start of a permanent regime shift. The deep-value posture is to wait for entry prices that absorb the Phase 2 premium and to size positions consistent with the historical base rate, not the inside view. No name in the cohort currently meets the kit's margin-of-safety threshold on the cycle-adjusted framing; MP Materials is the closest to a tradeable trigger, with the DoD price floor changing the failure mode in a way that justifies a more aggressive trigger than 2010-style historicals would imply.

• Portfolio: no position changes from this dossier. Cash remains the deep-value default through the MOU resolution and through the cohort stretch.
• Watchlist: MP Materials (MP) carries with kit trigger $60 / central $85 under the Greenwald-modified doctrine named in 2026-05-29-AM. Lynas (LYC.AX / LYSCF ADR) added to watchlist with provisional trigger work to be completed in the next two weeks. USA Rare Earth (USAR) added to watchlist as a higher-beta entry into the cohort with no provisional trigger — capital structure complexity warrants explicit thesis pass. Northern Minerals (NTU.AX) added to watchlist as a heavy-REE pure-play with provisional trigger work to follow.
• Theses on the workbench: MP Materials full thesis pass remains the top priority for next week. New thesis-builder candidates: Lynas (primary refining, Mt Weld and Seadrift), USA Rare Earth (magnet manufacturing), Northern Minerals (heavy REE).
• Sectors: materials sector view extended — critical minerals sub-sector promoted from observational to multi-year structural with Phase 2 capital-cycle reading. Defense primes (Lockheed Martin, Raytheon/RTX, General Dynamics, Northrop Grumman) reframed as second-order beneficiaries with smaller P&L sensitivity to magnet costs than the listed REE producers but with the lowest cycle sensitivity in the cohort.
• House view updates: Themes section — promote "Critical minerals" from "dossier-ready" to "dossier (v1) — confirmed multi-year structural, Phase 2 capital-cycle reading with explicit base-rate-grounded variant perception"; cross-link to this file. Add provisional sub-position under Business & corporates: "Rare-earth cohort multiple expansion — Phase 2 capital cycle reading." See house-view changes section below.
• Daily-scan adjustments: add screening criteria for rare-earth and magnet-cohort names — track NdPr oxide spot, magnet metal NdPr spot, Brent-WTI-style cross-asset divergences inside the cohort (e.g., listed Western vs. listed Chinese REE producers), and capacity-addition announcements as the Phase 2 leading indicators. Add the November 2026 expiry of the broader Chinese controls suspension as a calendar anchor.

Charts / data

Table 1 — Western rare earth / magnet capacity additions (announced and under construction)

Caption: Western rare earth and magnet capacity additions as of late May 2026, by announced project. Capital influx is heavy; aggregate capacity additions on announced timelines through 2028 represent roughly 4–6 of the 31 "missing" magnet facilities per Adamas's framework. The Phase 2 capital influx is real and the cycle is operating in textbook form.

Table 2 — 2010 base rate timeline

Caption: The 2010 base rate, mapped against the 2026 cycle analog. The DoD price floor and the larger demand picture qualitatively change the trough depth, not the cycle direction. The variant perception is that the cohort multiple is pricing this cycle as if Phase 3 and Phase 4 will not arrive.

Table 3 — NdPr price trajectory recent

Caption: NdPr price acceleration through 2026 H1 is the cleanest indication that the cycle is in Phase 2 — returns above cost of capital and rising sharply. The DoD's $110/kg floor at MP Materials becomes structurally relevant on a price collapse; at current levels, it is below the market price and is providing a Phase 4 backstop rather than current pricing support.

Sources

• T1(https://mpmaterials.com/news/mp-materials-announces-transformational-public-private-partnership-with-the-department-of-defense-to-accelerate-u-s-rare-earth-magnet-independence/)
• T1(https://www.sec.gov/Archives/edgar/data/0001970622/000121390026045339/ea028691001ex99-1.htm)
• T1(https://www.war.gov/News/Releases/Release/Article/4270722/office-of-strategic-capital-announces-first-loan-through-dod-agreement-with-mp/)
• T2
• T2(https://www.morganstanley.com/im/publication/insights/articles/article_bayesandbaserates_ltr.pdf)
• T2
• T2(https://www.csis.org/analysis/chinas-rare-earth-campaign-against-japan)
• T2(https://cepr.org/voxeu/columns/revisiting-china-japan-rare-earths-dispute)
• T2(https://www.weforum.org/stories/2023/10/japan-rare-earth-minerals/)
• T2(https://www.adamasintel.com/unfathomable-rare-earth-demand-growth/)
• T2(https://fas.org/publication/unpacking-dod-and-mp-partnership/)
• T2(https://www.energypolicy.columbia.edu/mp-materials-deal-marks-a-significant-shift-in-us-rare-earths-policy/)
• T3(https://www.usnews.com/news/top-news/articles/2026-05-20/china-says-rare-earth-controls-lawful-will-cooperate-with-us-on-reasonable-concerns)
• T3(https://mccullough.com.au/2026/05/27/australia-orders-china-linked-investors-divest-northern-minerals/)
• T3(https://discoveryalert.com.au/heavy-rare-earths-australia-northern-minerals-china-sale/)
• T3(https://www.gurufocus.com/news/8847298/mp-materials-corp-mp-q1-2026-earnings-call-highlights-record-ndpr-oxide-production-and-strong-revenue-growth)
• T3(https://www.aol.com/articles/mp-materials-mp-q1-2026-233212471.html)
• T3(https://stocksdownunder.com/lynas-rare-earths/)
• T3(https://rare-earth-mining.com/neodymium-price/)
• T3(https://rare-earth-mining.com/rare-earth-market-pricing-analysis-march-2026/)
• T3(https://www.cnbc.com/2025/07/10/pentagon-to-become-largest-shareholder-in-rare-earth-magnet-maker-mp-materials.html)
• T3(https://www.jpmorgan.com/insights/banking/investment-banking/mp-materials)
• T3(https://www.certivo.com/blog-details/china-rare-earth-export-controls-2026-what-new-licensing-rules-mean-for-manufacturers)
• T3(https://discoveryalert.com.au/china-rare-earth-export-controls-critical-minerals-trade-2026/)
• See sources-policy for the citation discipline applied; Wikipedia banned; T1 filings privileged over T3 news for theme-level claims

House view changes this run

1. Themes — promote "Critical minerals — multi-year structural" from "dossier-ready" to "dossier (v1) — confirmed multi-year structural, Phase 2 capital-cycle reading with explicit base-rate-grounded variant perception"; cross-link to this file under 13-Research/long-form/themes/2026-05-29-critical-minerals-capital-cycle-dossier-v1.md. The theme's tracked confirmations stand: China May 20 reaffirmation, MP Materials Q1 print, Australia divestment, 24/7 Wall St. coverage 2026-05-24. The dossier adds the 2010 Senkaku base rate, the Pentagon-MP Materials price-floor structural change, the Adamas 2030 demand and capacity-deficit framing, and the cohort Phase-2 capital-influx synthesis as analytical layers.

2. Add new position under Business & corporates — Rare-earth cohort multiple expansion — Phase 2 capital cycle reading. Position: the listed Western rare-earth and magnet cohort (MP Materials, Lynas, USA Rare Earth, Northern Minerals, Iluka, Arafura, Energy Fuels) is in Phase 2 of a capital cycle, with returns above cost of capital, capital flowing in at unprecedented rates (DoD $400M + $150M into MP; $1.5B PIPE + $1.6B Commerce facility into USA Rare Earth; $1B JPM/GS into MP 10X; sovereign equity stakes; foreign-investment review forcing equity rotation), and capacity additions arriving on a 2026–2028 timeline. The cohort multiple expansion is over-extrapolating the duration of Chinese export controls and the moat depth of the listed Western producers. The DoD price floor at MP Materials ($110/kg NdPr oxide) changes the failure mode but not the cycle dynamics. Confidence medium on the Phase 2 reading; confidence medium on the base-rate-grounded variant perception. Key supporting evidence: 2010 Senkaku base rate; Adamas 2030 demand and capacity-deficit framing; Q1 2026 producer prints; capital-influx data. Material risks to view: Chinese controls extend through 2028; capacity additions slip; demand growth exceeds Adamas; US/EU policy escalates from controls to direct sanctions. Watchlist additions to follow per Implications section.

3. No change to Iran / Strait of Hormuz, US rate path, AI infrastructure capacity, Software / SaaS valuation environment, Equity-market cycle position, USD positioning. This Friday long-form covers a theme that is decoupled from the Iran trinary and operates on a different time horizon than the cohort-pricing-through observations of the past two weeks.

4. Materials sector view — extend with explicit Phase 2 capital-cycle reading on rare earths sub-sector. Other sub-sectors (cement, steel, copper) unchanged from prior view.

5. Calendar anchor added: November 2026 expiry of the broader Chinese export controls suspension package as the cleanest single forward catalyst for theme repricing (in either direction).

6. Backlog: critical-minerals dossier item moves from Tier 2 "queued" to "complete (v1)"; refresh in 8 weeks or earlier on a material capacity-addition slippage / extension event.

Linked

• _house-view — Themes section updated; new Business & corporates position added
• capital-cycle — the Chancellor / Marathon framework operationalized in this dossier; Phase 2 identification
• narrative-cycle — narrative cycle reading: critical minerals at acceleration / early peak narrative phase
• base-rates — the discipline applied through the explicit base rates section
• 09-materials — materials sector file extended
• regulatory-regime — Chinese export controls and Australian foreign investment review tracked
• 2026-05-24-AM — Sunday note where the theme crossed the dossier threshold
• 2026-05-29-AM — Friday AM note confirming MP Materials thesis pass remains top priority and decoupled from Iran trinary
• 2026-05-27-hbm-replaces-cowos-binding-constraint-inversion — Wednesday tech long-form whose variant view on cohort multiple over-extrapolation extends the same pattern to a different sector
• 2026-05-28-ai-memory-cohort-multiple-inflection — Thursday markets long-form whose pattern (cohort multiple expansion over-extrapolating constraint duration) parallels this dossier's reading
• sources-policy — citation discipline applied