Vale's Iron Ore Cycle Has Already Peaked, and the Dividend Will Be the Tell

Iron ore at $98 per tonne (the current 62% Fe spot reference) sits roughly 35% below the 2021 peak of $230 and approximately 8% above the 75th percentile cost of marginal global supply. The marginal cost analysis is what matters. When iron ore prices sit only modestly above the marginal cost curve, the price is sustained by capacity constraints. When new supply arrives, the price falls toward the marginal cost.

New supply is arriving. Simandou Phase One in Guinea is targeting first ore in late 2026, with full capacity of 60 million tonnes by 2030. Roy Hill, Onslow, and Iron Bridge in Western Australia are collectively adding 80 million tonnes annually over the next 18 months. Vale's own production is recovering toward 320 million tonnes per year, up from the 300 million tonne cap that was in place post-Brumadinho. Aggregate, the market is absorbing approximately 150 million tonnes of incremental supply over a 36-month window against a Chinese steel demand profile that is structurally declining.

China is the demand-side concern that the bull narrative most often dismisses. Chinese steel production peaked in 2020 at 1.06 billion tonnes and has declined every year since to roughly 880 million tonnes in 2025. The decline is structural, driven by the rebalancing away from infrastructure-heavy GDP composition. The bull case requires this trend to reverse. There is no data supporting a reversal.

Vale's operating margin of 28% in 2025 reflects iron ore at $98. At $80, which is closer to the implied marginal cost given supply additions, operating margin compresses to approximately 18%. Operating income would fall to approximately $7 billion against the $11 billion print in 2025. Free cash flow against $6 billion of sustaining capex would compress to roughly $1 billion. The dividend, at $4.4 billion per annum, would be uncovered.

Vale paid out approximately $4.5 billion in dividends in 2025 against $3.1 billion of free cash flow. The shortfall was funded by balance sheet cash. The company carries approximately $14 billion of net debt against $11 billion of trailing operating income. Net debt to EBITDA stands at roughly 0.9x, which on the surface looks comfortable.

The figure is misleading. EBITDA is sensitive to iron ore price. At $80 per tonne iron ore, sustaining EBITDA would compress to approximately $9 billion, pushing net debt to EBITDA closer to 1.6x. That is the level at which Brazilian sovereign-rating-constrained miners typically begin to limit dividends. Vale's covenant restrictions are not tight, but the rating agency methodology is. A push above 1.5x leverage with a negative cycle outlook usually triggers a credit watch, which historically has preceded a dividend reset.

The Brumadinho-related obligations are the often-ignored variable. Vale committed to a $7 billion settlement framework in 2024, with annual payments of approximately $1 billion through 2029. Those payments are above and beyond operating capex and are non-discretionary. They reduce the free cash flow available for distribution by roughly 25% annually. Built into a normalising cycle assumption, the dividend is not affordable at the current rate.

The Risk Desk's base case is a 30-40% dividend cut announced within 12 months. The bear case is a 50% cut. The probability-weighted outcome for the share price under either scenario is materially below current levels.