Why the Iron Ore Bears Are Wrong About Vale

Three things have created the consensus. First, Chinese steel output rolled over in 2024 and has not recovered. Second, Vale's share price has de-rated from the $20 handle in 2022 to the $14-17 range through 2025-2026. Third, the Brazilian tax and environmental litigation overhang has not gone away, and every quarter seems to bring another Samarco or Renova Foundation headline. Each of these is real. None of them are new.

What has changed is that the cash flow has stabilised at a level the bears said could not persist. Free cash flow in 2025 of $3.06 billion compares to $2.88 billion in 2024. Net income of $2.47 billion is down year-on-year, but from a base that already assumes normalised iron ore. The company is not grinding toward distress; it is pinned to a margin that reflects the cost curve position of the Carajás complex. That cost curve position is the moat the bears are ignoring.

Start with the China demand argument. Yes, Chinese crude steel output has declined from the 2020 peak. Yes, property completions are weak. But iron ore demand is not just Chinese property. It is also Chinese manufacturing, Indian urbanisation, ASEAN infrastructure, and, increasingly, the energy transition capex cycle. BHP's own supply deficit modelling for the copper market uses iron ore as the anchor demand proxy. The demand picture is not what it was in 2019, but it is also not the collapse the VALE multiple implies.

Now the pricing argument. Iron ore fines benchmark prices have ranged between $90 and $130 per tonne through 2025. Vale's all-in sustaining cost at Carajás is in the low $30s per tonne. That is $60-100 per tonne of cash margin at spot prices. On roughly 320 million tonnes of annual production, the operating cash flow that throws off is consistent with the mid-single-digit billions Vale has been printing, quarter after quarter, regardless of where spot sits within its range.

The third bear argument, balance sheet risk, has already been largely retired. Vale finished 2025 with $7.4 billion of cash against $19.4 billion of debt. Net debt to EBITDA sits at roughly 1.3x, which is investment-grade territory. The company has been returning capital through buybacks and the dividend channel throughout the down-cycle. If you are genuinely worried about Vale's solvency at current commodity prices, you are either using the wrong cost curve data or you have a view on iron ore going to $50. Both are possible. Neither is consensus.

The fourth bear argument, legal and environmental liabilities, is the only one that has real asymmetry. Brumadinho and Mariana settlements have dragged on, and every new ruling has potential to move the stock by 5-10%. But at a $74 billion market cap, the residual exposure is already partly priced. And Vale's current management has made balance sheet discipline and community restitution the core of the equity story, not a footnote.

Vale's trailing P/E of 31.7x is a data artefact because the 2025 earnings base is depressed. The forward P/E of 7.8x is the correct lens. A price-to-sales of 0.35x puts Vale in the cheapest decile of global mining. Enterprise value at roughly $90 billion is less than 4x run-rate EBITDA, which historically has been a floor for large-cap diversified miners. Go back to 2015-2016 and check: the bottom for Vale's EV/EBITDA was around 3.5x, and that was during an actual liquidity crisis and a pending Samarco settlement.

The dividend policy matters here too. Vale has returned capital aggressively through both regular distributions and special payouts since 2020. The payout track record means any meaningful iron ore rebound gets pushed straight through to shareholders. A 20% move in average realised prices on stable volumes produces something like $4-5 billion of incremental EBITDA, most of which flows to dividends and buybacks under current policy.

This is why the bear framing is wrong. Bears are modelling the earnings decline, which is real. They are not modelling the capital return on the normalisation, which is the return driver.