Neo Performance Materials

Phase A — Understand the business

Lens 1 · Company Overview

Neo is a rare-earth and rare-metal processing and magnetics company — a mid-stream/downstream converter, not a miner. It buys rare-earth feedstock (carbonate, ore concentrate, scrap) and turns it into separated oxides, metals, alloys, magnetic powders and finished permanent magnets. Three reporting segments:

• Magnequench — bonded and sintered NdFeB magnetic powders and finished magnets. FY2025 revenue $204.6M, adj. EBITDA $28.4M. The crown jewel: Magnequench traces to the magnet business GM/Delco built, sold to Chinese SOEs in 1995; Neo now runs it as the Western-traceable magnet franchise. Shipped +37% more bonded magnets into EV traction motors in 2025.
• Chemicals & Oxides (C&O) — separated rare-earth oxides, emission-control catalysts, specialty chemicals. FY2025 revenue $135.0M, adj. EBITDA $23.4M (+376% YoY). Profitability transformed by exiting legacy Chinese separation assets and ramping a new auto-catalyst line.
• Rare Metals — gallium, hafnium, tantalum, niobium, rhenium; recycling and reclamation. FY2025 revenue $147.7M, adj. EBITDA $43.2M — the highest-margin segment (~29% segment EBITDA margin ).

Global footprint: manufacturing in China, Germany, Canada, Estonia (Silmet separation + new Narva magnet plant), Thailand and the UK; R&D centre in Singapore. HQ Toronto. CEO Rahim Suleman (since Jul 2023). Reporting currency USD, calendar fiscal year.

Contract structure: a mix of tolling/processing spreads (feedstock-in, product-out), specialty long-term supply (e.g. the Energy Fuels carbonate agreement) and increasingly inventory-carry economics in Rare Metals (buy scrap cheap, sell metal into a rising price). Not a take-or-pay utility model — margins move with the processing spread and the metal price, which is the whole investment debate.

Lens 2 · Supply Chain

Neo sits in the middle of the rare-earth chain — its moat and its dependency both live in who supplies it and who buys. Named stakeholders:

Upstream (feedstock in):

• Energy Fuels (UUUU) — ships rare-earth carbonate produced from monazite at its White Mesa Mill, Utah to Neo's Silmet plant, Sillamäe, Estonia; Silmet expects to process ≥840 tonnes/yr TREO minimum. This is the flagship US-ore → European-separation non-China chain.
• Third-party rare-earth concentrate / carbonate producers (historically incl. Chinese and other feedstock); hafnium/tantalum/gallium scrap buyers globally (Neo bought extra hafnium scrap in Q1-26).
• Historically Serra Verde (Brazil) was a natural feedstock partner for the Western chain — but Serra Verde was acquired by USA Rare Earth for ~$2.8B in April 2026, i.e. a competitor now owns that scaled non-China mixed-oxide source. A structural negative for Neo's feedstock optionality (see Lens 13).

Midstream (Neo): Silmet (separation, incl. new heavy-rare-earth line) → Narva (NdFeB magnets) in Estonia; Magnequench powder plants (China/Thailand); metals/recycling in Germany/UK/North America.

Downstream (product out):

• EV / automotive traction-motor OEMs — bonded & sintered NdFeB magnets (Western traceability is the selling point).
• Aerospace & defense — hafnium (superalloy turbine blades), tantalum (capacitors), gallium.
• Emission-control / auto-catalyst customers (C&O).
• Europe's magnet buyers — Narva + VAC (Vacuumschmelze) Hanau are the only two operational rare-earth-magnet plants in Europe.

Chokepoints / single-source risk: (1) Silmet is one of the very few non-China separation plants on earth — a genuine chokepoint asset, but also a single physical node; (2) heavy-rare-earth (Dy/Tb) feedstock is overwhelmingly China-controlled, so Neo's new HRE separation line still needs non-China HRE units to feed it at scale; (3) magnet demand routes through a handful of European auto/industrial buyers. Names present → this lens passes.

Lens 3 · Competitive Advantages (moats)

• Vertical integration + geography = the real moat. Neo is arguably the only company that separates rare earths AND makes finished magnets at commercial scale entirely outside China (Silmet → Narva). Lynas separates but doesn't yet make magnets at scale; MP Materials is scaling magnets in Texas but its separation/feed is US-centric; VAC makes magnets but buys separated oxide. Western traceability (auditable non-China origin) is a contractual moat — OEMs and defense primes will pay a premium and sign long contracts to de-risk, and Chinese competitors "often struggle to meet Western traceability audits".
• Process IP & 30-year operating know-how. Separation and magnet metallurgy are hard to replicate; Neo has run Magnequench and Silmet for decades. Not a green-field permitting story.
• Rare Metals niche dominance — gallium recycling and hafnium reclamation in North America is a small, specialized, high-margin pocket where Neo has scarce capability and pricing power when China restricts gallium exports.
• Bargaining power — asymmetric. Over customers: rising (there is almost no non-China alternative for audited magnets → customers need Neo more than Neo needs any one customer). Over suppliers: weaker in HRE feedstock (China holds the units); stronger in scrap (Neo can wait). Net: the export-control regime is a moat-widener — the more China restricts, the more Neo's non-China integration is worth. Ground: positioning.md/bottlenecks.md are missing on the shelf, so this is web-derived.

Lens 4 · Segments

FY2025 actuals (all ``; mirrored into segments.csv this run):

Q1-2026 vs Q1-2025 (all ``):

Read of the trend — this is the single most important slide in the dossier:

• Rare Metals is the swing factor. It went from 20% of segment EBITDA to the majority — Q1-26 adj EBITDA $23.9M (+176%) on hafnium/gallium/tantalum price surges. Record consolidated EBITDA is overwhelmingly a metals-price event, not a broad-based volume story.
• Magnequench is the durable grower — real +46% revenue, magnet volume up, the structural thesis.
• C&O revenue fell 30% — but that is deliberate portfolio optimization (China separation-asset exit), and EBITDA still rose. Quality-over-scale. Don't misread the top-line drop as demand loss.

The bull sees three engines firing; the bear sees one commodity spike (Rare Metals) doing the heavy lifting and asks what EBITDA looks like when hafnium normalizes.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q1 2026, reported 2026-05-07)

All figures ``:

• Revenue $155.0M vs $121.6M — +27% YoY; beat consensus ~$128M by ~21%.
• Gross profit $50.4M, GM 32.6% (vs 25.3%) — a +730bp gross-margin jump.
• Operating income $26.6M (vs $9.6M) — nearly tripled.
• Adjusted EBITDA $36.2M, 23.4% margin (vs $17.1M, 14.1%) — record quarter, +112% YoY, +930bp margin.
• Adjusted net income $14.9M; adjusted EPS $0.36 basic / $0.34 diluted (vs $0.15) — 83% beat vs ~$0.185 forecast.
• ⚠️ GAAP net LOSS of $(1.6)M (vs $(1.4)M loss) — the company reported a net loss even in a record quarter. The gap between +$14.9M adjusted and −$1.6M GAAP is a ~$16.5M add-back stack (FX, restructuring, non-cash items). Flag for Lens 10.
• Guidance RAISED: FY2026 adj. EBITDA $100–110M, up from $75–80M — a rare mid-year +37% midpoint raise.

Balance-sheet flags (as of 2026-03-31):

• Cash $41.7M; total debt $154.3M (up from $101.8M at Dec-25); net debt $112.5M (from $63.4M).
• Inventory rose to $235M (from $205M at Q4-25) — deliberate hafnium-scrap + feedstock stockpiling into rising prices. This is where the debt went: working-capital-into-a-price-spike, not distress capex.
• FCF turned deeply negative on the TTM view (−CAD $142.7M per stockanalysis ) precisely because of that inventory build — the mirror image of FY2025's +$52.4M FCF.

Market reaction: mixed/"sell-the-news." One transcript headline says the stock dipped on the print, another says it rose ~5% — i.e. a big beat + big raise into a stock already +163% over 52 weeks produced only a muted pop. What's priced in is already a lot.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the shelf (web-only). From call coverage across FY2025 → Q1-2026:

• Tone has shifted from "transformation/portfolio-cleanup" to "harvest/record-results." FY2025 calls emphasized exiting China separation, fixing C&O margins, building Narva. By Q1-26 the language is records, guidance raises, feedstock securing.
• Recurring phrases (2026): "record adjusted EBITDA," "Western/traceable supply chain," "securing attractive feedstock," "strategic inventory," "critical materials strategy."
• Things they stopped saying: the earlier hedging about Chinese separation drag and C&O losses — that problem is now framed as solved.
• Management is focused on: (1) ramping Narva magnets (2,000 → 5,000 tpa); (2) opportunistically stockpiling hafnium/gallium; (3) monetizing the export-control-driven price environment; (4) the Energy Fuels non-China chain. Sentiment: confident, arguably promotional given the stock run — watch for over-anchoring on Rare Metals prices.

Lens 7 · Comps

Non-China rare-earth / magnet peers.

Read: NEO at ~19.5x EV/EBITDA looks cheap versus Lynas at ~50–83x — the recurring "rare-earths winner at a bargain price" thesis. But the comparison is loaded: Lynas's multiple is inflated by a depressed trough-EBITDA denominator, while NEO's ~19.5x sits on peak, hafnium-juiced EBITDA — so NEO's multiple is optically low but on a cyclically high base. On EV/Sales, NEO (2.4x) is dramatically cheaper than Lynas (18.9x), which better reflects that NEO is a lower-margin processor vs a resource owner. Fair read: NEO is the cheapest profitable name in the group, but "cheap" is doing a lot of work if 2026 EBITDA is a peak. Do not fabricate the missing peer multiples — they are not sourced this run.

Lens 8 · Stock-Price Catalysts (last ~5y, moves >5%)

Mostly ``; the pattern is unusually clean for this name.

• The dominant catalyst is Chinese rare-earth policy. April 4, 2025 — China imposed export controls on seven heavy rare earths + magnets; NdPr metal spiked past RMB1M/tonne intraday. This is the axis the whole non-China basket (NEO, MP, Lynas) trades on.
• 2025 price melt-up: NdPr oxide +40%+ YTD in 2025; by 2026, neodymium ~$125/kg, +136% YTD. NEO is +163% over the trailing 52 weeks.
• Company-specific catalysts: Narva magnet-plant opening (Sep 2025, PM of Estonia attended); Silmet HRE line commissioning; the FY2026 guidance raise (May-26); quarterly beats.
• Rhodia patent settlement (€7.1M) — a negative-headline catalyst that proved immaterial (see Lens 10).

What the market actually reacts to for NEO: (1) China export-control headlines (macro/geopolitical — the biggest mover, and exogenous to NEO); (2) rare-earth/rare-metal spot prices (hafnium, gallium, NdPr); (3) guidance changes. It reacts less to volume/operational execution. Translation: NEO is traded as a levered, non-China rare-earth price proxy — beta 1.64. Own it knowing that.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO Rahim Suleman — President & CEO since Jul 7, 2023; CFO 2017–2023. Canadian Chartered Accountant (Univ. Waterloo, M.Acc); ex-CFO of Stackpole International; ex-GE Digital Energy. A finance operator, not a metallurgist or promoter — fitting for a turnaround-and-harvest phase.
• Track record (quantified): on his CEO watch, adj. EBITDA rose $64.4M (FY24) → $75.6M (FY25) → guided $100–110M (FY26); C&O adj. EBITDA +376%; the messy Chinese separation assets exited; Narva built and opened. That is a credible, delivered operational turn — margins up, portfolio cleaned, growth capex landing. Adjusted EPS $0.05 (FY24) → $0.49 (FY25).
• Tenure & skin in the game: Suleman owns ~1.08% directly (~$10.3M); total insider ownership ~0.9–1%. Insider buying noted — the CEO added ~9.6% to his stake. Alignment is modest in % terms but the direction (buying into the run) is a positive tell.
• Ownership DNA — Oaktree. Neo is the reincarnation of Molycorp's downstream assets, bought by Oaktree Capital out of Molycorp's 2015 bankruptcy, IPO'd on the TSX Dec 2017. Board Chair Edgar Lee (since Jun 2025) is ex-Oaktree (ran its $6B Strategic Credit strategy). A distressed-credit lineage → capital discipline and a sell-the-turnaround instinct are in the DNA. Positive for allocation rigor; watch for eventual sponsor/insider exit as a supply overhang.
• Capital allocation: pays a CAD $0.10/qtr dividend (~1.1% yield), buys back stock ($4.0M FY25), and is reinvesting into Narva/HRE — a balanced, shareholder-return-aware posture. The one debate: the Q1-26 debt-funded inventory build is an aggressive bet on metal prices staying high (see Lens 10/13).
• Archetype: professional-manager turnaround operator under a financial sponsor's residual influence — not a visionary founder. Right archetype for this stage (harvest a scarce asset), less obviously the right one for a multi-decade capacity build-out if that's what the moment demands.

Lens 10 · Forensic Red Flags

Web-only (no filings on shelf) — every figure labelled; this lens is necessarily lighter than a filings-grounded one and should be re-run against the SEDAR+ annual/AIF.

1. ⚠️ GAAP loss vs large positive "adjusted" — the headline flag. FY2025 GAAP net LOSS $(10.0)M while adjusted net income +$20.5M; Q4-25 GAAP $(15.6)M loss vs adj +$0.6M; Q1-26 GAAP $(1.6)M loss vs adj +$14.9M. A ~$30M annual GAAP-to-adjusted bridge is material relative to the size of the company. Much is plausibly non-cash (FX on a multi-currency footprint, restructuring from the China exit, intangible amortization), but the "records" narrative is entirely an adjusted-EBITDA/adjusted-EPS story — on a GAAP basis Neo did not make money in FY2025. Verify the add-back composition against the audited statements before underwriting the "profitable" label.
2. ⚠️ Cash flow diverging from earnings — inventory. TTM FCF swung to ≈−CAD $142.7M as inventory rose to $235M and debt rose to $154.3M in a single quarter. Framed by management as opportunistic hafnium-scrap stockpiling into a price spike. If prices hold, this converts to a fat cash gain; if hafnium/gallium reverse, Neo is carrying high-cost inventory funded by debt — the classic commodity-processor working-capital trap. This is the number to watch every quarter.
3. Rising leverage from a net-cash base. Net cash +$13.9M (Dec-24) → net debt $63.4M (Dec-25) → $112.5M (Mar-26). Fast, but off a low base and tied to visible working capital, not losses.
4. Segment concentration in a volatile line. Record consolidated EBITDA leans on Rare Metals commodity prices (hafnium/gallium/tantalum), which are thin, opaque markets — earnings quality is lower than the "record" framing implies.
5. SBC / adjusted-metric reliance — adjusted EPS and adj. EBITDA are the promoted numbers; confirm stock-based comp and one-time add-backs aren't flattering the trend (not separable from web data).

Regulatory findings (required sub-section):

• SEC (EDGAR): No CIK — Neo is a Canadian issuer not required to file with the SEC; no EDGAR LR/AAER search possible. total_sec_findings: 0.
• Litigation (web): Rhodia European patent litigation — SETTLED. Neo agreed to an aggregate €7.1M cash payment to Rhodia (auto-catalyst patents), mutual release; the amount was not materially different from what Neo had already accrued and is not expected to materially affect earnings or limit catalyst sales. A related damages award came in well below Rhodia's original claim. Immaterial.
• Other agencies (FTC/DOJ/FDA/etc.): web search surfaced no material enforcement actions against Neo. No securities class action found.
• Verdict: No material regulatory or securities-enforcement findings — verified via the regulatory-findings file (SEC EDGAR EFTS, no CIK), web search, and public litigation coverage as of 2026-07-06. The one open legal item (Rhodia) is settled and immaterial. Caveat: not grounded in a filings-based Item-3 equivalent — re-verify against the SEDAR+ AIF "Legal Proceedings" section.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (EPS, next three fiscal years)

Bottom-up from FY2025 actuals + FY2026 guidance. Every input labelled; outputs ``. Reporting in USD, adjusted EPS (the number management guides to). Share count ~41.9M.

Anchors: FY2025 adj. EPS $0.49; FY2026 adj. EBITDA guide $100–110M. To bridge EBITDA→adj. EPS: subtract D&A ($25M ), interest ($10M on ~$150M debt ), taxes (~20% ), ÷41.9M shares.

• FY2026 (base): adj. EBITDA $105M (guide midpoint). → pre-tax ≈ $105 − $25 D&A − $10 interest = $70M; after ~20% tax = $56M; ÷41.9M ≈ adj. EPS ~$1.34 ``. (Sense-check: Q1-26 alone did $0.36 adj EPS; four quarters at a Rare-Metals-elevated run-rate landing >$1 is internally consistent — the raise implies a step-change year.)
• Bull FY2026: EBITDA $115M (beat top of guide as hafnium/gallium stay bid) → adj. EPS ~$1.55 ``.
• Bear FY2026: EBITDA $85M (metal prices fade H2) → adj. EPS ~$0.90 ``.
• FY2027 (base): Magnequench/Narva volume grows (5,000 tpa magnet ramp), but Rare Metals prices partially normalize off the 2026 spike. Net EBITDA roughly flat-to-down at ~$95–105M; adj. EPS ~$1.20–1.35 ``.
• Bull: Narva at capacity + prices hold → EBITDA $130M, adj. EPS ~$1.80 ``.
• Bear: hafnium/gallium mean-revert hard → EBITDA $70M, adj. EPS ~$0.70 ``.
• FY2028 (base): the structural non-China magnet story matures — Narva contributes real magnet EBITDA; assume prices settle at an elevated-but-below-2026 plateau. EBITDA ~$110–120M, adj. EPS ~$1.40–1.60 ``.

The whole projection hinges on one variable: how much of the 2026 hafnium/gallium/NdPr price spike is structural (export controls persist) vs cyclical (spike + destock). The base case assumes it partially holds because the export-control regime is unlikely to reverse quickly. No forecast.ts create run — per --watchlist rules and no committed base case yet (the price-durability uncertainty is too wide to log a Brier bet honestly).

Lens 12 · Bull vs Bear

Bull case. Neo is the only vertically-integrated, cash-generative, dividend-paying rare-earth magnet producer entirely outside China — Silmet separation feeding Narva magnets, with a US-ore chain (Energy Fuels) behind it. Every escalation of China's export-control regime (which is now policy, not a one-off) widens the moat and re-prices Neo's Western-traceable output higher. The 2025-26 tape proves it: adj. EBITDA guide doubled to $100–110M, margins up ~930bp, stock +163%. Magnequench magnet volume is structurally growing into EVs/defense; Europe legally must onshore magnets and there are only two plants (Neo's Narva + VAC). Management has a delivered turnaround (China exit, C&O +376%) and is buying feedstock and stock into the trend. At ~19.5x EV/EBITDA / 2.4x sales, it's the cheapest profitable name in a basket where Lynas trades 50–83x. Earnings surprise potential remains high while hafnium/gallium/NdPr stay bid.

Bear case (permanent-impairment risks). (1) This is Molycorp's ghost. Neo is Molycorp's downstream assets — and in the last cycle rare-earth prices collapsed 80–95%, Molycorp burned $1.5B, took on $1.7B of debt, and sold for $20.5M. A processor's margin is a spread; when China re-floods the market (its historical playbook to crush Western entrants), Neo's spread and its debt-funded $235M inventory get crushed together. (2) Record EBITDA is a price event, not a volume moat — Rare Metals (hafnium/gallium) carried Q1-26; those are thin, manipulable markets that mean-revert. (3) Expectations are baked in — a huge beat + a 37% guidance raise produced only a muted stock move; the +163% run already prices a lot of the non-China premium. Pre-mortem (18 months out, thesis broke): China quietly loosens magnet-export quotas to relieve Western OEM pressure → NdPr and hafnium roll over 40% → Neo's high-cost stockpiled inventory writes down, the debt that funded it stays, FY2027 adj. EBITDA prints ~$70M not ~$105M, and a "19.5x on peak" multiple re-rates to "12x on trough" — a ~50% drawdown even with the moat intact. Are multiples too high? Not on 2026 numbers — if 2026 is a durable base. They are too high if 2026 is a peak. Contrarian view the market is missing: the market is treating Neo's record margins as the new run-rate; the more interesting question is whether the magnet business (Magnequench + Narva) can carry the valuation on its own once the commodity spike fades — because that, not hafnium, is the durable moat, and it is currently the smaller profit contributor.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• What structurally breaks the model: Neo doesn't own a mine — it earns a processing/inventory spread. China controls ~90% of separation and the heavy-rare-earth units Neo's new line needs. China can break Neo two ways it has used before: flood the market (crush the spread) or choke HRE feedstock (starve the plant). Neo is structurally short the very inputs its moat depends on.
• Revenue concentration: into a single volatile line (Rare Metals) and a single geopolitical variable (China export policy). If export controls ease even modestly, the entire re-rate thesis softens — and that decision is made in Beijing, not by Neo.
• The moat is thinner than bulls think where it's newest: Narva is at sample/2,000-tpa stage, not proven at 5,000 tpa; the "only non-China magnet plant" claim shares the podium with VAC, and MP Materials + USA Rare Earth (now owning Serra Verde's scaled feed) are racing to build exactly Neo's integrated model with far more capital. Serra Verde — the natural non-China feedstock — is now owned by a competitor (USAR).
• Most dangerous competitor bulls underestimate: not a Chinese major, but MP Materials — bigger balance sheet, US-government backing, comparable Q1-26 EBITDA ($36.6M), and building magnets at scale in Texas. And USA Rare Earth, which just paid $2.8B for Serra Verde to leapfrog into integrated non-China supply.
• Worst capital-allocation move: the debt-funded inventory bet. Buying $30M+ of extra hafnium scrap on leverage into an all-time price high is a trader's bet dressed as strategy — right if you're correct on prices, a balance-sheet hole if you're wrong.
• Accounting/incentives: the GAAP-loss-but-adjusted-profit framing (FY25 GAAP $(10)M) means the promoted "record profitability" is non-GAAP; an Oaktree-legacy board with residual insider stakes has an incentive to keep the narrative and the multiple high into any eventual exit.
• What must hold for today's price: that 2026's hafnium/gallium/NdPr prices are roughly the new normal, that Narva ramps on schedule, and that China does not re-flood. If growth/prices disappoint 20–30%, FY2027 EBITDA drops to ~$70–80M and a stock at ~19.5x on peak EBITDA can lose 40–50%.
• Single scenario that permanently impairs: a repeat of the 2011-2015 rare-earth price collapse (China floods to kill Western capacity) hitting a Neo now carrying $235M inventory + $154M debt. Plausibility: moderate — it's China's proven playbook, and the only thing preventing it is that export controls currently serve China's strategic interest. Geopolitics, not fundamentals, is the swing.

Lens 14 · Management Questions (ordered by information value)

1. Of the FY2026 $100–110M adj. EBITDA guide, how much is volume/spread Neo controls vs spot-price on hafnium/gallium/tantalum you don't — and what does the guide look like at 2024 metal prices? (The single question that reframes the whole valuation.)
2. You raised guidance 37% mid-year. What is the normalized, mid-cycle EBITDA of this portfolio stripping the 2026 rare-metal price spike?
3. The $235M inventory / $154M debt build: at what hafnium/gallium price does that stockpile become a write-down rather than a gain, and what's the hedge?
4. Narva: what is the confirmed path and offtake from sample-stage to 2,000 and then 5,000 tpa, and what magnet EBITDA does that add at 5,000 tpa?
5. FY2025 was a GAAP net loss. Walk through the GAAP-to-adjusted bridge item by item — how much is genuinely non-recurring vs recurring "adjustments"?
6. Heavy-rare-earth feedstock: your new HRE separation line needs Dy/Tb units China controls. Where does that feedstock come from at scale, and is it contracted?
7. MP Materials and USA Rare Earth are building your integrated model with more capital, and USAR now owns Serra Verde. What is durable about Neo's edge in five years?
8. If China eased magnet-export quotas tomorrow, what happens to Neo's realized prices and this year's EBITDA?
9. Capital allocation: with the stock +163%, why buy back stock and fund inventory on debt rather than deleverage — what's the priority stack?
10. Oaktree legacy / insider intentions: is there a residual sponsor position or insider-exit overhang the market should model?
11. Energy Fuels supply chain: what volume is contracted vs aspirational, and how much of Silmet's feed does it actually cover?
12. Which segment do you expect to carry the valuation in 2028 — magnets or rare metals — and why?
13. What is your through-cycle capex plan for Western capacity, and can it be self-funded if EBITDA halves?
14. Customer contracts: what share of magnet/oxide revenue is under multi-year take-or-pay vs spot, and how sticky are the Western-traceability premiums?
15. What single operational risk (a plant, a permit, a feedstock line) keeps you up at night that isn't priced by a market focused only on China headlines?

---