Vale

Phase A — Understand the business

Lens 1 · Company Overview

Vale is the world's largest producer of iron ore and iron-ore pellets and a top-3 nickel and increasingly-material copper producer. Two reported segments ``:

• Iron Ore Solutions — extraction of iron ore + pellets/briquettes + the integrated logistics spine (railways, maritime terminals, ports). Three Brazilian systems: Northern (Carajás / S11D — the crown jewel, ~99–100% process recovery, 66%+ Fe), Southeastern, Southern. FY2025 revenue $30,130M (78.5% of total) ``.
• Vale Base Metals (VBM) — nickel (Sudbury/Thompson/Voisey's Bay in Canada, Sorowako in Indonesia, Onça Puma in Brazil) + copper (Salobo/Sossego in Carajás; by-product copper from Canadian nickel). FY2025 revenue $8,273M (21.5%) ``.

Business model = price-taker on globally-traded commodities, competing on cost and grade. Iron ore priced off the 62% Fe CFR China index (avg $102.4/dmt in 2025, −6.5% YoY) ``; nickel/copper off LME/SHFE. There is no pricing power — the entire equity story is being the lowest-cost, highest-grade tonne on the seaborne cost curve and shipping volume.

Customer concentration is geographic, not contractual: "No customer individually represented 10% or more of our net operating revenue in 2025" , but **China alone was 50.5% of revenue ($19,405M)** — the single most important fact about this company. Contracts are largely index-linked spot/provisional pricing, not take-or-pay; revenue is a pure function of volume × a China-set price.

Lens 2 · Supply Chain

Vale is unusually vertically integrated for a miner — it owns much of its own chain, which is the moat (Lens 3). Named links ``:

• Upstream inputs: diesel/heavy fuel, explosives, grinding media, and — critically — maritime freight (spot Capesize rates swing COGS materially: +$783M cost impact from higher iron-ore spot freight in 2025) and Brazilian grid + self-generated hydro power (energy is the dominant nickel-processing cost at Sorowako/PTVI).
• Vale (the transformation): mine → beneficiation → owned railways (Carajás Railroad, Vitória-Minas) → owned ports (Ponta da Madeira / Pier IV can berth 420,000 DWT Valemax vessels at 16,000 t/hr) → its own Valemax fleet + long-term COAs.
• Midstream: 8 pelletizing plants in Brazil + 2 in Oman + 2 briquette plants; nickel refining in Clydach (Wales, UK) and Japan; the Copper Cliff Nickel Refinery + Port Colborne (Ontario).
• Downstream customers: Chinese mills (Baowu et al.) = 50.5% of revenue; Japan 6.3%, Germany 4.8%, rest of Asia/Europe. Base-metals offtake now partly committed to Manara Minerals (Saudi) under the VBM partnership.
• Chokepoints / single-source: the S11D conveyor-belt truckless system and the Carajás Railroad are single points of failure for ~170Mt/yr of the Northern System ``; Brazilian tailings-dam licensing is the binding regulatory chokepoint on volume (Lens 10). Names or it didn't happen: Baowu (buyer), Mitsui (JV partner in VLI logistics + 6.7% shareholder), BHP (Samarco 50/50 JV partner and co-defendant), Glencore (Sudbury JV), Sumitomo (VBM minority in Indonesia), Ero Copper (Brazil earn-in), PT Antam (Hu'u JV).

Lens 3 · Competitive Advantages (moats)

This is a genuine Tier-1 cost-and-grade moat, and it is the whole thesis:

• Lowest-quartile cost. Iron ore C1 cash cost $21.3/t in 2025 (−2% YoY, 2nd straight year of cuts); all-in $54.2/t (−3%) . Against a ~$100/t index that is a **~$46/t all-in margin** and a >40% EBITDA margin through the cycle. 2026 guidance C1 $20–21.5/t . Only Rio Tinto's Pilbara sits in the same cost tier.
• Grade. Carajás/S11D ore is 65%+ Fe — a premium product as the market decarbonises (DRI/green-steel feedstock), which partially offsets the Simandou threat (Lens 13) since Vale competes on the high-grade end rather than being undercut on cost.
• Irreplaceable integrated logistics. The captive rail-and-port network from Carajás to Valemax berths is non-replicable — "unique, non-replicable assets" ``. A new entrant cannot buy this; Simandou had to build a 600km railway and a new port to enter at all.
• Bargaining power: weak over customers (China sets the price), strong over the cost base (owns the chain, self-supplies logistics/power). The moat protects the margin, not the price.

Lens 4 · Segments (revenue · EBITDA · geography)

Revenue by segment, FY2025 ``:

Adjusted EBITDA by segment, FY2025 `` — the more revealing cut:

The trend that matters: iron-ore EBITDA is decelerating (pellets −35% on collapsing pellet premia; fines flat), while base metals doubled (+131%) — copper EBITDA up 81% and nickel off the floor (+526% from a near-zero base). Vale is quietly transitioning from a pure iron-ore play into an iron-ore-plus-copper major. Geography ``: China 50.5% ($19,405M), Japan 6.3%, rest-of-Asia 10.1% → Asia 66.9%; Europe 14.9% (Germany 4.8%); Americas 13.0% (Brazil 8.2%, US 2.9%). The China dependence is structural and rising in importance even as it shrinks in growth.

---

Phase B — Measure performance

Lens 5 · Earnings Result (FY2025 print)

The headline is ugly and the underlying is fine — the gap is the analysis. FY2025 consolidated income statement ``:

Read: EBITDA rose 4.2% (higher iron ore/copper/nickel volumes +$326M, weaker average BRL +$374M) ``, but net income fell 67% — entirely below the EBITDA line, from two non-cash/one-off hits:

1. $4,599M impairment (vs +$301M gain in 2024) — predominantly nickel/base-metals assets, reflecting the years-long nickel price rout.
2. $2,670M tax charge (vs $721M) — a write-off of deferred tax assets on tax-loss carryforwards in foreign subsidiaries (again, the nickel business) . Financial results actually *improved* (a +$1,616M derivative mark-to-market gain as the BRL appreciated 11.1% into year-end). **Balance-sheet flags:** inventories jumped to **$5,937M from $4,605M** — a ~$1.3B build worth watching (production ran ahead of sales into a soft Q4 iron-ore tape). Market reaction: the market saw straight through the reported number — the stock is +47% over 2025 and +18% in the last month ``, rewarding the EBITDA/volume/copper story and ignoring the impairment. Classic "kitchen-sink the nickel book, own the cash flow."

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the shelf (transcripts/ empty) — from web/IR ``. Management's message has shifted decisively over ~4 quarters:

• Then (2023–24): defensive — Samarco/Brumadinho remediation, "de-risking," succession noise, nickel pain.
• Now (4Q25 → 1Q26): offensive and disciplined. CEO Gustavo Pimenta hammers three phrases: capital discipline ("efficiency gains, not project cancellations" — 2025 capex cut to $5.4–5.7B), copper growth at low capital intensity ("double copper in Carajás at 20% capital intensity, that's where my money should be"), and cost leadership (C1 to a new low). What they've stopped saying: the crisis-management language of the disaster years — Samarco is now framed as a capped, provisioned, closed chapter. Tone is confident, almost mid-cap-growth in flavour, unusual for an iron-ore incumbent.

Lens 7 · Comps

Global diversified-major and iron-ore peers. Multiples ``; sources conflict on Vale's EV/EBITDA (5.8x vs ~3.5x cited) — flagged, not silently reconciled.

Read: Vale trades at roughly the same EV/EBITDA as BHP/Rio (~5.5–5.9x) but at a lower P/E and a ~1/3-the-scale market cap, carrying a persistent "Brazil discount" for governance/political risk + the residual dam-liability tail. The discount is narrowing as the Samarco settlement caps the tail and copper re-rates the growth profile. Sum-of-the-parts angle : mark VBM at the Manara price (10% = $2.5B → $26B EV; Vale owns 90% = ~$23.4B attributable). Total EV ≈ mkt cap ~$70B + net debt ~$10.6B = ~$80.6B; strip VBM → **iron ore implied EV ≈ ~$57B on ~$13.8B EBITDA = ~4.1x** . On that math the world's best iron-ore franchise is priced at a discount to its own base-metals arm — the crux of the bull case.

Lens 8 · Stock-Price Catalysts (what actually moves VALE)

Pattern over ~5yr ``, the >5% movers:

• The 62% Fe iron-ore price / China demand — the dominant driver by far; 50.5% of revenue and the highest-beta input. Every China-stimulus headline and every port-inventory print moves the stock.
• Disaster events & settlements — Brumadinho (Jan 2019) was the defining drawdown; the Oct-2024 Samarco Definitive Settlement (R$170B) removed a major tail and helped the 2025 re-rating.
• Copper — the new catalyst — Morgan Stanley's upgrade to Overweight ("attractive copper growth exposure… linked to electrification, not steel") is emblematic of why the stock ran from ~$9 to ~$16.65 ``.
• Governance/political — the messy 2024 CEO succession + Lula-government friction is a recurring Brazil-specific swing factor. The tell: this stock reacts to the iron-ore tape and to copper optionality, and shrugs off reported EPS (it rose through a 67% net-income drop). Own the commodities and the SOTP, not the P&L.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO: Gustavo Pimenta (since Oct 2024; CFO Nov-2021→2024). Finance-bred, chosen over 14 external candidates to end a "messy succession" ``. Archetype = professional manager / capital allocator, not founder — appropriate for a mature, cash-return incumbent. Early track record is coherent: cost cuts (C1 to record lows), capex discipline, and a clean pivot to copper-led growth.
• Skin in the game: low personal ownership (typical of a state-adjacent Brazilian major); no controlling shareholder since 2020. Governance is diffuse — PREVI 8.0%, BlackRock 7.0%, Mitsui 6.7%, plus 12 golden shares held by the Brazilian government with veto power ``. This is the double-edged sword: no controller means no controller-abuse, but the state exerts influence via golden shares, permitting, and political pressure (PREVI-linked directors are prominent on the board).
• Capital-allocation history: iron-ore ROIC has been strong through the cycle; the nickel M&A legacy is the black mark — the impairments and DTA write-offs of 2025 trace to over-paid, over-built nickel assets from prior regimes. The current signal is good: $5.9B dividends/IOC declared in 2025, no buyback executed (120M-share authorization left unused — prudent, given the tape), and a copper program pitched explicitly on return ("20% capital intensity").
• Red flags: a board member resigned during the succession citing "leaks, manipulations, political influences and conflicts" ``; Lula publicly criticised Vale over dividends-vs-victims and asset sales. Political-interference risk is real and idiosyncratic to Vale.

Lens 10 · Forensic Red Flags

Acting forensically on FY2025 ``:

• Impairments & DTA write-offs are the honest read of a bad nickel book, not a manipulation — they depress earnings, the conservative direction. If anything the 2025 print is under-stated vs cash generation.
• Reported FCF is muddied by reparation cash. Net operating cash flow $8,801M − capex $6,006M = ~$2,795M reported FCF . But that is *after* ~$3.55B of dam-reparation cash (Brumadinho $874M in-OCF + Samarco $2,298M in-investing + de-characterisation $378M). **Underlying pre-reparation FCF ≈ ~$6.3B** . Bulls quote the underlying; the reparation cash is real and runs for years — don't fully add it back.
• Dividend > earnings. $5,923M declared vs $2,352M net income to shareholders = the payout is funded from cash flow/EBITDA, not accounting earnings. Sustainable while iron ore holds; thin cover if the tape breaks (the policy is a minimum 30% of adjusted EBITDA — it flexes down automatically) ``.
• Inventory build (+$1.3B) — modest yellow flag; production outran sales into soft 4Q pricing.
• Provisions are the balance-sheet story: Samarco/Fundão provision $2.6B (Vale's share of the shortfall guarantee), Brumadinho + de-characterisation multi-year, aggregate reparation spend $5,960M expected after 2026 on top of ~$2.4B in 2026 ``. These are disclosed, court-ratified, and capped — but large.

Regulatory findings (required):

• SEC: No Litigation Releases or AAERs naming Vale, 2021-07-10→2026-07-10, via EDGAR EFTS ``. (Historically Vale settled US securities-fraud charges over Brumadinho dam-safety disclosures in 2023; outside the current search window and not in the file.)
• Item 3 / Legal Proceedings (10-K equivalent, 20-F) : the material litigation is the dam disasters. **Fundão/Samarco (Mariana 2015, 19 deaths):** Definitive Settlement **R$170B (~$30B)**, ratified by Brazil's STF Nov-2024, paid over 20 years; R$73.1B disbursed 2015→2025; Fundação Renova liquidated Nov-2025 with Samarco now primary obligor and Vale/BHP each guaranteeing up to 50% of any Samarco shortfall (Vale provision **$2.6B**). **The live tail is the UK Claim** — ~610,000 claimants in London; in **November 2025 the English court found BHP liable**, with the **quantum trial set for April 2027–March 2028**; Vale is exposed via a July-2024 liability-sharing agreement with BHP and booked a **$449M top-up provision in 2025** on the increased likelihood of loss . Brumadinho (2019, 270 deaths): Vale's own dam; R$37.7B 2021 global settlement, ongoing indemnification.
• Non-SEC: No new material FTC/DOJ-type enforcement found `` beyond the Brazil/UK dam proceedings above.
• Net: No accounting-fraud findings; the risk is liability magnitude (the UK quantum trial is an open, potentially multi-£-billion tail), not accounting integrity.

---

Phase D — Project & stress-test

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

Bottom-up from FY2025 actuals + 2026 guidance + Q1-2026 run-rate. Base shares ~4.27B ``.

Inputs : 2026 guidance — iron ore 335–345Mt, copper 350–380kt, nickel 175–200kt, capex $5.4–5.7B, C1 $20–21.5/t. Q1-2026 already printed iron ore 69.7Mt (+3%), copper 102.3kt (+12.5% @ $13,143/t realised), nickel 49.3kt (+12.3%) . Iron-ore index ~$100/t and structurally soft (Simandou + China). 2026 EBITDA consensus ~$17.5B ``.

Base path: FY2026 EPS ~$2.00, FY2027 ~$2.00 (consensus $1.99; iron-ore drag offset by copper ramp), FY2028 ~$2.10 `` as copper volumes build toward the 900kt target. The swing factor is not operations — it's the iron-ore price, i.e. China. Per --watchlist rules, no forecast.ts create logged in this unattended sweep; base call = FY2026 non-GAAP EPS ≥ $2.00 for a future human-gated forecast.

Lens 12 · Bull vs Bear

Bull. The best iron-ore cost/grade franchise on earth (C1 $21/t, >40% margins, ~$46/t cushion) is being handed a second growth engine in copper — doubling to ~700–900kt at "20% capital intensity" into an electrification supercycle — while trading at ~7–8x P/E, ~5.9% yield, and a SOTP that prices the iron-ore business below its own base-metals arm (~4x EBITDA). The Samarco tail is now capped and provisioned; leverage is trivial (net debt ~$10.6B, ~0.7x EBITDA ``); and the VBM IPO (readiness targeted mid-2026) is a live value-crystallisation catalyst. Buy the option on copper, get paid ~6% to wait, own the cheapest quality tonne in the world.

Bear. 78% of revenue and 89% of segment EBITDA is iron ore, and 50.5% of all revenue is China — right as (a) Chinese crude steel fell 4.4% to 960.8Mt in 2025 with exports at a record 119Mt masking genuinely weak domestic demand , and (b) **Simandou** floods **16Mt in 2026 ramping to ~90–120Mtpy of high-grade Guinean ore straight into Vale's premium niche**, pivoting the seaborne market into structural surplus (World Bank −3% price in 2026; record 160Mt Chinese port stocks) . The stock has already run ~80% off its low to sit exactly on the $17.32 consensus PT — the deep-value gap is closed. The dam liabilities are capped, not closed: the UK quantum trial (2027–28) is an open, potentially multi-£-billion tail. And Brazil-political/governance risk (golden shares, Lula friction) never fully clears.

Pre-mortem (18 months out, thesis broke): China stimulus faded, iron ore settled at $80–85 as Simandou ramped ahead of schedule, iron-ore EBITDA fell ~30%, the dividend flexed down (it's 30%-of-EBITDA), copper couldn't offset a $4–5B iron-ore EBITDA hole, and the stock round-tripped to ~$11 — with the UK quantum ruling landing as an added overhang.

Multiples too high? No — they're fair-to-cheap. The issue is not the multiple, it's the E: a cyclical peak-ish iron-ore price against a structural supply shock. Contrarian view the market is missing: consensus is anchored on "cheap iron-ore major + copper call option = Buy," but the honest framing is a high-quality cyclical trading at fair value at a rolling-over point in its dominant cycle, where the copper story is real but too small (21% of revenue) to offset an iron-ore de-rate for another 2–3 years.

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

• Where it breaks: the iron-ore price. Half the revenue is one country's property-and-infrastructure cycle, and that cycle is in structural, not cyclical, decline — China's steel intensity has peaked. Vale cannot grow its way out via volume (adding tonnes into a surplus depresses its own price).
• The competitor bulls underrate: Simandou. It is not marginal high-cost tonnage that Vale's low cost sees off — it is new, low-cost, high-grade (65%+) supply aimed at exactly Vale's premium Carajás segment, backed by Rio Tinto + Chinese capital that wants supply diversification away from Australia/Brazil. It compresses the premium that justifies Vale's grade moat.
• Concentration shift: if China's mills keep exporting finished steel to survive (119Mt, record), any trade-war escalation that closes those export markets collapses Chinese steel output and Vale's volumes simultaneously.
• Worst capital allocation: the nickel empire — billions in impairments and DTA write-offs in 2025 alone are the receipt for a decade of value destruction; the copper program is being sold on the same "growth" logic that built that nickel book.
• What must hold for the price: iron ore ≥ ~$95, copper strong, and no adverse UK quantum surprise. A 20–30% iron-ore-driven EBITDA miss (→ ~$12–13B) takes EPS to ~$1.3–1.4 and, on an unchanged ~8x, the stock to ~$11–12 — a ~30% drawdown from $16.65.
• The single scenario that permanently impairs: an adverse UK Claim quantum ruling (2027–28) at the top of the ~£36B claim range, landing while iron ore is in a Simandou-driven trough — liability + cash-flow squeeze at once.

Lens 14 · Management Questions (ordered by information value)

1. With Simandou ramping into your premium grade segment, what iron-ore price deck are you underwriting new capex against, and at what price does the dividend flex below its 30%-of-EBITDA floor?
2. VBM is marked at $26B EV by Manara but sits inside a ~$70B parent — what is the specific trigger, timeline, and structure (IPO vs. further private stake sales) for crystallising that value, and does mid-2026 readiness still hold?
3. What is your true through-cycle break-even iron-ore price including sustaining capex and annual reparation cash, and how does that compare to Rio's Pilbara + Simandou?
4. Quantify Vale's maximum exposure under the BHP liability-sharing agreement if the UK Claim quantum lands at the upper end — and what is not provisioned today?
5. Copper to 900kt at "20% capital intensity" — show the project-by-project IRR and the assumed copper price; why should we trust this after the nickel impairments?
6. Is the nickel book now fully sized to reality, or is there further impairment/DTA risk if nickel stays below $18k/t?
7. How much of the 2025 inventory build reverses in 2026, and what does it say about your read of near-term Chinese demand?
8. What is your explicit capital-return framework ranking dividends vs. buybacks vs. the copper build vs. debt — and why no buyback in 2025 with the stock near its lows?
9. How do the 12 golden shares and current political environment constrain asset sales, dividends, or the VBM IPO in practice?
10. What is the realistic annual reparation cash-outflow schedule for Brumadinho + Fundão for 2026–2030, and when does it materially step down?
11. Where are you on the DRI/green-steel megatrend — can high-grade + briquettes structurally lift realised premia enough to offset Simandou?
12. What would it take to reduce China revenue dependence below 45%, and is that even a goal?
13. How exposed is COGS to Capesize freight and the BRL, and how much is hedged into 2026–27?
14. What is your M&A appetite in copper, and would you do a large deal after the nickel experience?
15. What single leading indicator do you watch to know the iron-ore cycle is turning?

---