Two filings landed on the same day. On July 21, 2026, Vale — the world's second-largest iron ore producer and the heaviest weight in Brazil's benchmark equity index — released its second-quarter production and sales report: the best Q2 in iron ore since 2018, with records in copper and nickel on top of it. Hours later came a second notice, this one carrying Official Letter CVM No. 196/2026, in which CVM, Brazil's securities regulator, demanded explanations about the shareholder meeting that decides — today — who chairs the company's board. Over the previous five months the shares had fallen 21%. That contrast is the subject of this piece, and it explains why the mine stopped being the thing that sets the price of VALE3.
Buy, hold or walk away? The answer first
The question is blunt, so the answer goes up front. The full VALE3 analysis we published today lands on a score of 6.5/10 and a HOLD verdict, with an estimated fair value of R$ 77.00 (range: R$ 60 to R$ 95). Against the R$ 72.24 reference price, that is 6.6% of headroom, plus an expected dividend of roughly 5.7% over the next twelve months. Total expected return in the base case: something around 12.3%.
One number settles the argument: the Selic — Brazil's benchmark interest rate — sits at 14.25%. In the base case, VALE3 pays an investor less than cash parked in Brazilian government bonds, and it pays that less while carrying iron ore price risk, currency risk and boardroom risk. The company isn't in trouble. The issue is that today's price already pays for an outcome that still has to show up. Hence the verdict: hold the shares, collect the dividend, don't rush to buy.
If you already own it, there is no operational reason to sell — the business is delivering. If you don't own it, the asymmetry is lukewarm. A genuine margin of safety only appears below R$ 65, where the expected dividend yield clears 6.4%. Above R$ 85, the stock starts demanding that everything go right at once.
First things first: VALE4 and VALE5 no longer exist
This is the single most searched question about the company, and it deserves a clean answer. In December 2017, as part of its move to Novo Mercado — the top corporate-governance tier of B3, Brazil's stock exchange, which requires a single class of voting shares — Vale converted every preferred share into common stock. Since then there has been one ticker and one class: VALE3.
Anyone holding VALE5 lost nothing and had to do nothing: the position became VALE3 automatically. There is no "VALE4", no "VALE6", no other code trading today. Every analysis of Vale is an analysis of VALE3. For investors outside Brazil, the same company trades in New York as the ADR under the ticker VALE — an American Depositary Receipt is a certificate issued by a US bank representing shares of a foreign company, which is what lets you buy it in dollars on the NYSE.
One small exception is routinely reported badly. The Brazilian federal government still holds 12 special-class preferred shares — the so-called golden share. It grants veto power over exactly three matters: relocating the headquarters, changing the corporate purpose and changing the company's name. It carries no meaningful economic rights. The golden share does not elect directors, does not set the dividend and does not steer strategy. Keeping that straight matters, because it is a completely separate thing from the ownership fight going on right now — which comes later in this piece.
What Q2 2026 showed: the operation is at its best in years
What gives those figures meaning is the margin per tonne. Vale sold fines at US$ 95.0/t against an all-in cost of US$ 55.4/t — that measure bundles the cost of getting ore out of the ground, freight to China, royalties and sustaining the operations. Roughly US$ 39.6 per tonne is left over. Multiply that by something close to 320 million tonnes a year and you have, in practice, where all of the company's cash comes from. No other business in Brazil produces absolute margin at that scale.
The point is that none of it is under threat. The projects delivered — Capanema and VGR1 are ramping up, Serra Sul +20 is 86% complete and starts in the second half — production guidance was maintained at 335-345 Mt, and every 2025 target was met or beaten. If the stock is down 21% since February, the problem lives outside the mine.
The three ceilings holding the price down
Three forces are pressing on VALE3 simultaneously, and not one of them is solved by digging more ore. They are independent of each other, which means the stock can be hurt by one while the other two stay exactly where they are.
Ceiling 1 — the ore market flipped into surplus
Iron ore is hovering around US$ 100 per tonne, with market consensus near US$ 95/t for 2026 and a downward bias. Demand is shrinking on the side that matters: China's steel output fell 3% in the first half of 2026, a second consecutive year running below one billion tonnes. Meanwhile inventories at Chinese ports sit at roughly 160 Mt, the highest in more than three years. Full warehouses make for buyers in no hurry.
One technical detail with a very practical consequence: in 2026 the reference index moved from 62% to 61% iron content. The benchmark that prices the entire market was downgraded — a signal that the average ore available worldwide has gotten poorer, and that the yardstick adjusted to that reality.
Simandou attacks precisely where Vale is strongest. Carajás' trump card is the quality premium: ore at roughly 65% iron with little silica, for which steelmakers pay above the index because it takes less energy and emits less carbon to turn into steel. It is the one segment where Vale holds a structural edge over the Australians. The Simandou mine in Guinea produces ore at exactly that ~65% Fe grade — the first shipment left in November 2025, 2026 estimates run from 16 to 20 Mt, and the ambition is 120 Mt a year in the 2030s. This is not volume that crushes the average price; it is volume that erodes the premium. The near-term relief is that the ramp-up looks slower than advertised — RBC stretched its curve from 30 to 48 months. The structural direction, though, only points one way.
Ceiling 2 — the currency eats the margin with nobody making a mistake
Vale earns in dollars and pays a large share of its costs in reais. With the real at R$ 5.08 (the central bank's PTAX reference rate on 07/21/2026), every real of wages, power, diesel and services inside Brazil costs more dollars. The C1 cash cost — what it takes to pull the ore out and load it onto the ship, excluding freight, royalties and depreciation — came in at US$ 23.6/t in Q1 2026, up 12% in a year. The company itself flagged in the release that the year should close at the top of guidance: C1 of US$ 20-21.5/t and all-in of US$ 52-56/t.
The official sensitivity is the best tool an investor has here: every R$ 0.10 move in the exchange rate shifts C1 by US$ 0.25/t. A real appreciating from R$ 5.08 to R$ 4.80 would add about US$ 0.70/t to the cost — across 320 Mt a year, roughly US$ 220 million evaporating without anyone at Vale making a single bad call.
| Producer | C1 cost (guidance) | Currency exposure |
|---|---|---|
| Fortescue (Australia) | US$ 17.5-18.5/t | Australian dollar, no Brazil risk |
| BHP (Australia) | US$ 18.0-19.5/t | Australian dollar, no Brazil risk |
| Vale (Brazil) | US$ 20.0-21.5/t | Costs in reais, revenue in dollars — trending to the top of the range |
The reading is uncomfortable and has to be said out loud: Vale is the most expensive of the group on C1, and the gap does not come from inefficiency — it comes from the currency. It is also a risk that runs against what most Brazilian investors are rooting for: a strong real is good for the country and bad for Vale.
Ceiling 3 — governance turned into a public fight
A clean timeline helps here, because the headlines made a mess of it:
| Date | What happened |
|---|---|
| 06/11/2026 | Previ — the pension fund of Banco do Brasil employees, which owns 7.02% of the capital — calls a shareholder meeting to remove Daniel Stieler as board chairman and nominates José Maurício Pereira Coelho for the seat. |
| June 2026 | The board recommends rejecting the removal and puts forward Ieda Gomes Yell as an alternative to Previ's nominee. |
| 07/06/2026 | Stieler resigns — the central item on the agenda is left without an object. |
| 07/14/2026 | The board elects Wilfred Bruijn interim chairman, with Bruijn himself and Sousa Oliveira abstaining. |
| 07/20/2026 | CVM issues Official Letter No. 196/2026 after a query from two minority shareholders (Geração L. Par and Banco Clássico) challenging the validity of that election: the agenda items were conditional on a removal that the resignation had rendered moot. |
| 07/22/2026 | The meeting chooses between Manuel Lino Silva de Sousa Oliveira (backed by Previ) and Marcelo Gasparino (currently vice-chairman of the board). |
The right frame is not political, it's structural. Vale is a corporation in the literal sense: a company with no controlling shareholder and a 96.01% free float — the slice of shares trading freely, outside the hands of any controller. Its largest holder owns 7.02%. On paper it is one of the best governance structures on the Brazilian market: Novo Mercado listing, a single share class, full tag-along rights, a permanent fiscal council elected by minority holders, 8 independent directors out of 13, and PwC as auditor since 2019.
The mechanism, no accusations attached: it is precisely that dispersed ownership that lets a 7% holder set the company's agenda. When nobody owns the place, whoever brings organization and resolve gets to call the meeting. That is neither illegal nor irregular — it is the design working as designed. The side effect is that the virtue doubles as an entry point: pulverized capital isn't immune to politics, it just changes the address the pressure arrives at. For an investor, what shifts isn't strategy (both candidates are career directors who back current management) — it's the multiple a foreign buyer is willing to pay for a company picking its leadership under a regulator's letter.
The P/E trap: why 20x is the wrong number
Open any quote site and you will see VALE3 trading at about 20 times earnings and conclude the stock is expensive. That conclusion is wrong, and the reason is an accounting one.
Trailing twelve-month profit still carries the US$ 3.8 billion loss booked in the fourth quarter of 2025. It came from two things that do not repeat: a US$ 3.5 billion impairment on the Canadian nickel assets, plus an additional Samarco provision.
Impairment, in one sentence: it is the moment a company admits on its books that an asset it once bought is worth less today than the balance sheet says — no cash leaves the till, but the loss shows up in the quarter's results. In Vale's case, it was the acknowledgment that the Canadian nickel mines inherited from the 2006 Inco acquisition are worth less than the books claimed.
| Gauge | Accounting basis | Recurring basis |
|---|---|---|
| 2025 profit | US$ 2.4 bn (R$ 13.8 bn) | US$ 7.8 bn (pro forma) |
| Price / earnings | ~20x | ~8x |
| EV / EBITDA | — | ~5.1x |
EV/EBITDA measures what the whole business is worth (market capitalization plus net debt) against the cash the operation throws off in a year, before interest, taxes and depreciation. At 5.1x, Vale sits below the 5.5-5.9x the market pays for BHP and Rio Tinto. In other words: VALE3 is not expensive. But it isn't scraping the bottom either — in 2024 and 2025 the shares traded below R$ 55 on comparable multiples. Cheap on a multiple is not the same as favorable asymmetry when risk-free money pays 14.25%.
How much it pays in dividends — and why less than the rear-view mirror suggests
Vale is a formula-driven payer, not a generous one. Its Shareholder Remuneration Policy mandates a minimum of 30% of adjusted EBITDA less sustaining capital expenditure, split into two payments: one in September (calculated on the first half) and one in March (on the second half), with the option of JCP in December, deducted from the March installment. JCP — interest on capital, a Brazilian tax-efficient form of shareholder payout — lets the company pay less corporate tax because the amount is deducted from its own taxable profit, while the recipient has 15% withheld at source.
| Year | Payout per share | Context |
|---|---|---|
| 2021 | R$ 14.65 | Peak ore prices, with extraordinary payouts |
| 2022 | R$ 7.58 | End of the supercycle |
| 2023 | R$ 6.08 | Normalized prices |
| 2024 | R$ 5.35 | Worst year of the recent cycle |
| 2025 | R$ 7.62 | Recovery |
| 12 months to Jul 2026 | R$ 5.4772 | 7.58% yield on R$ 72.24 |
Look at the swing: from R$ 14.65 to R$ 5.35 in three years. A cyclical company's dividend is not predictable income. Our expectation for the next twelve months is R$ 3.80 to R$ 4.60 per share, centered on R$ 4.15 — a dividend yield of roughly 5.7%. That is less than the 7.58% in the rear-view mirror, and the gap isn't pessimism: it is taking extraordinary payouts out of the math. There is no comfortable room for them with net debt expanded to US$ 17.8 bn, R$ 6.2 bn of Mariana payments due in 2026 alone, and capex guided at US$ 5.4-5.7 bn.
Two details that hit the wallet. First: since January 1, 2026, dividends above R$ 50,000 per month paid by the same company to the same individual are subject to a 10% withholding tax in Brazil — and companies naturally respond by shifting distributions toward JCP. Vale already did exactly that in the March 2026 cycle (R$ 1.56 of JCP against R$ 0.76 of dividend). Second: buying a stock just to "catch the dividend" creates no return. On the ex-dividend date — the first session on which a buyer no longer receives the payout — the price opens discounted by roughly the amount distributed. You collect with one hand what you lose with the other; what decides the return is the thesis and the entry price.
The liability nobody adds to the math
There is a line in Vale's cash flow that almost never makes it into a multiple-based analysis: the disbursement for dam-disaster reparations. It is large, it is contracted, and it runs through 2043.
| Year | Vale's contribution — Mariana |
|---|---|
| 2026 | R$ 6.2 bn |
| 2027 | R$ 3.1 bn |
| 2028 | R$ 2.1 bn |
| 2029 | R$ 3.5 bn |
| 2030 | R$ 3.2 bn |
Add to that, in 2026 alone, US$ 0.9 billion for Brumadinho and US$ 0.7 billion for dam decharacterization. The year's total comes to roughly US$ 2.7 billion. For scale: recurring free cash flow in 2025 was US$ 4.76 billion. In other words, more than half of all the free cash the operation generated is already committed to reparations before any decision about dividends, capex or buybacks. That number is why nobody should expect an extraordinary dividend in 2026 — and it shows up in no P/E ratio anywhere.
One risk remains open. The English courts have already recognized BHP's liability in the Mariana case and rejected its appeal. Since Samarco is a 50/50 joint venture between Vale and BHP, the damages-quantification phase in London has a path to spilling onto the Brazilian side of the bill.
Three scenarios, with observable triggers
| Scenario | Target price | Probability | Triggers to watch |
|---|---|---|---|
| Bull — ore holds and the real weakens | R$ 95.00 | ~25% | Ore above US$ 105/t; the real in the R$ 5.50-5.80 band; copper and nickel firm alongside a Vale Base Metals IPO; the shareholder dispute settled without litigation; pro forma EBITDA near US$ 18 bn with a full September dividend |
| Base — sideways, collecting the dividend | R$ 77.00 | ~50% | Ore between US$ 90 and US$ 100/t; guidance of 335-345 Mt met with costs at the top of the range; EBITDA of US$ 15.5-16.5 bn and a minimum-policy dividend; noisy governance without rupture; Simandou shipping 16-25 Mt |
| Bear — the surplus materializes | R$ 55.00 | ~25% | Ore below US$ 85/t with Simandou accelerating toward 40 Mt+; the real at R$ 4.80-5.00 pushing C1 above US$ 22/t; the board dispute heading to court with management changes; politically driven capex or a rewrite of the dividend policy; new dam-related provisions |
In the base case the stock oscillates between R$ 68 and R$ 85 and virtually all of the return comes from the dividend — the scenario in which an investor ties with fixed income and discovers they took on volatility for free. In the bear case there is a double penalty: earnings fall and the multiple compresses along with them, the classic behavior of a cyclical in a bad cycle.
The verdict, and what to put on the calendar
Vale is one of the best-run operating companies on the Brazilian market going through one of the worst institutional moments in its recent history. The operation is not the problem: 84.3 Mt produced, records in copper and nickel, guidance maintained, costs high because of the currency rather than inefficiency. The problem is that the price of VALE3 is not made of production — it is made of the ore index, the exchange rate and institutional confidence, and all three are working against it at the same time. The 21% drop since February isn't the market punishing the mine; it's the market repricing everything the mine doesn't control.
Verdict: HOLD — score 6.5/10. Estimated fair value of R$ 77.00 (range R$ 60 to R$ 95), with ~6.6% of headroom plus an expected dividend of ~5.7%, against a risk-free Selic paying 14.25%.
It makes sense for investors who want dollar and commodity exposure inside a Brazilian portfolio, those chasing one of the largest cash flows on B3 while accepting a cyclical payout, and anyone already holding at a lower cost — the operation has given no reason to sell. It does not make sense for anyone who needs predictable income (the dividend already halved from one year to the next), anyone with no tolerance for institutional risk, and anyone who mistakes "giant company" for "safe investment" — VALE3 went from R$ 118 in 2021 to R$ 49 in 2025.
The objective trigger for adding to a position is a price below R$ 65, not the headline out of the shareholder meeting.
What to watch, with dates: the Q2 2026 financial results come out on 07/30/2026 — and the number that matters is not production, which we already know was excellent. It's the C1 cash cost: does it confirm the top of the US$ 20-21.5/t guidance, or blow through it? Next comes the announcement of the September dividend installment (historically between July and August) and how much of it arrives as JCP. Then, month by month: the 61% Fe ore index in Qingdao (below US$ 85/t and the bear case starts materializing), inventories at Chinese ports (a sustained drop below 145 Mt would signal real demand), and Simandou's monthly shipments.
This piece covers the central points. The full dossier — year-by-year fundamentals, risks, governance, management, scenarios and four valuation methods with the assumptions laid out — is in the complete VALE3 analysis, and there is also a PDF report. Nothing here is investment advice or an official document of Vale S.A.: it is a reading of public filings (CVM, company releases, the Reference Form) and market data through July 22, 2026. The decision to invest is always yours.
