淡水河谷

Vale's second quarter adjusted EBITDA of USD 4.1 billion increased 19% from a year ago and 4% from the prior quarter. Higher iron ore, copper, and nickel prices and volumes more than offset increased unit costs in the key iron ore business. It declared a BRL 2.03 (about USD 0.41) per share dividend.

Why it matters: Like competitors, it is suffering from higher energy and freight costs due to the Iran war and foreign currency headwinds. Guidance for iron ore unit cash costs is raised to between USD 22.50 and USD 23.50 per metric ton, from USD 20.00 to USD 21.50. Seasonally higher second-half volumes likely partially offset these headwinds, but we now estimate USD 23.00 for 2026, from USD 21.50, and at the guidance midpoint. Unit cash costs were around USD 24 for the second quarter and first half, up from USD 22.20 a year ago. But higher byproduct prices led by gold, solid production, and cost out see improved unit cost guidance for the base metals business. Forecast EPS and DPS for 2026 to 2028 fall modestly overall, but our longer-term estimates are little changed. We think higher energy and freight costs are temporary.

The bottom line: Shares trade close to no-moat Vale's unchanged fair value of USD 15, with the changes to our estimates immaterial to our valuation.

Between the lines: Our Capital Allocation Rating is Poor given potential for overreach from the Brazilian government. It was instrumental in the removal of previous CEO Eduardo Bartolomeo and also likely influenced the recent change in Vale's Chair. We considered upgrading to Standard given it has shed noncore assets in recent years while increasing cash returns to shareholders, and improving relations with stakeholders while making amends to those affected by the tailings dam disasters at Feijao and Samarco. While we don't expect materially value-destructive actions from increasing government influence under the current administration, on balance we think a Poor rating is appropriate.

Iron ore accounted for around 70% of second quarter EBITDA. With 80 million metric tons sold in the second quarter, up 3% on a year ago, and around 150 million for the half, it is on track to meet our unchanged forecast for around 325 million metric tons in 2026, a 3% increase on 2025.

Similar to the first quarter of 2026, production volumes in the second quarter were the highest since the Feijao dam failure at Brumadinho in early 2019 saw it take around one-quarter or 100 million metric tons of capacity offline. We think further increases in volumes are likely, and estimate around 360 million by 2030. This will likely help reduce unit cash costs over this period.

Base metals had another strong quarter, with EBITDA of USD 1.3 billion, up 80% on a year ago due to the surging copper price (up 57%) and, to a lesser extent, stronger nickel prices (up 14%), along with higher sales volumes of each metal. Together with higher byproduct credits, stronger volumes led to dramatic improvements in unit costs from last year.

Our fair value estimate for no-moat Vale is USD 15 per share.

We assume iron ore averages about USD 100 per metric ton from 2026 to 2028 based on the futures curve. Based on our estimate of the long-run marginal cost of production, our assumed midcycle price is around USD 75 per metric ton from 2030. Solid demand from China, which accounts for around 75% of the seaborne iron ore trade, is supportive of near-term prices. However, longer-term we expect demand from China to moderate as steel production peaks and starts to decline as its economy moves away from one reliant on fixed-asset investment to a more consumption-based economy. China’s falling population, along with rising scrap-based production, also contribute to reduced demand for iron ore in our view. We also think additional supply is likely, led by Simandou and Vale. Hence, we expect a long-term price substantially below the current spot of around USD 100 per metric ton.

Cuing off the futures curve, our assumed average copper price from 2026 to 2028 is about USD 6.00 per pound. Based on our estimate of the long-run marginal cost of production, we assume a midcycle price of about USD 3.80 per pound from 2030.

In addition, we assume nickel prices of about USD 7.70 per pound from 2027, based on spot.

We assume a 9.2% cost of capital in our five-year explicit forecast and use an enterprise value/EBITDA multiple of 6 to value cash flows generated beyond year five.

As a commodity producer, Vale is a price taker and needs low-cost mines with long lives and a low installed capital base to support the longer-term excess returns needed to justify an economic moat. We forecast midcycle returns on invested capital similar to its weighted average cost of capital of about 9%. Vale’s iron ore solutions segment accounts for about 80% of midcycle EBITDA, but ROIC is weighed down by its base metals business, which generates low returns on its invested capital base.

Our forecast is based on assumed midcycle prices from 2030 of roughly USD 75 per metric ton for iron ore (which we note is materially less than the average price of around USD 100 per metric ton over the past decade) and USD 3.80 per pound for copper, based on our estimates of the marginal costs of production. We also assume nickel prices of about USD 7.80 per pound from 2027 based on spot prices. As its midcycle ROIC is similar to its WACC, we don’t assign a moat to Vale.

In calculating ROIC, we have added back to invested capital about USD 39 billion in asset and intangible write-downs taken over the past decade on the basis that these amounts relate to assets developed or acquired in the ordinary course of business and so should be included when calculating ROIC. Some of the more material amounts include USD 17 billion in relation to nickel, USD 11 billion in relation to coal, and USD 6 billion in relation to fertilizers.

Looking at each of Vale’s segments in turn:

Iron ore solutions (no moat): Vale’s ore has an average iron content of around 63%-64%, higher than the 61%-62% produced by BHP and Rio Tinto and the 57%-58% produced by Fortescue. As such, it generates a premium compared with the 61% benchmark price. However, Vale’s iron ore assets have cash costs that place it around the middle of the seaborne iron ore cost curve, above its lower-cost competitors, including BHP and Rio, but better than Fortescue once the latter’s lower-quality ore (lower iron content) is taken into account. Anglo American and other smaller competitors also sit to the right of Vale on the industry cost curve. Vale’s competitive disadvantage versus its lower-cost Australian competitors is due to higher freight costs incurred to ship its ore to China, which accounts for around 75% of annual seaborne iron ore demand. With substantial resources, we think new mines will be periodically developed to continue to feed and utilize the installed infrastructure base of Vale’s three iron ore systems. The remainder of this segment is relatively immaterial. We estimate that the iron ore solutions segment will produce around 80% of midcycle EBITDA. However, the iron ore operations’ middling position on the industry cost curve means we think it is unlikely to generate returns above Vale’s cost of capital.

Base metals (no moat): The segment has an inflated capital base due to the purchase of Canada’s Inco in 2006 and the subsequent development of high-cost mines in Brazil and New Caledonia. Both the nickel and copper operations lack a cost advantage, with both located in the upper half of the relevant industry cost curves. We forecast nickel sales to rise to around 220,000 metric tons in 2030, up from about 170,000 metric tons in 2025, driven by increased production at its Canadian mines, led by Sudbury. We forecast copper sales to increase to about 420,000 metric tons in 2030, up from around 370,000 metric tons in 2025, driven by increased production at Salobo in Brazil and at its Canadian nickel mines, which also produce copper. This will likely lead to cash operating costs for both nickel and copper falling, but the company’s operations are likely to remain middling at best on the respective industry cost curves. We forecast this segment will produce around 20% of midcycle EBITDA. However, as we forecast low-single-digit returns on invested capital from midcycle, we don’t deem the base metals segment as moatworthy either.

Exploration projects (no moat): These are an immaterial part of Vale and way too early in their potential development to assign any of them a moat.

Vale's lower unit costs relative to producers in India and China, as well as high-cost producers such as Anglo American, allow it to remain profitable even in depressed price environments.

Vale's high-quality ore with an iron content of around 63%-64% could attract significant premiums relative to the 61% benchmark.

Vale could benefit from efforts to decarbonize steelmaking. Steelmakers prefer higher-grade inputs, such as Vale’s iron ore, as they require less metallurgical coal and result in lower carbon emissions per metric ton of steel produced.

We believe China's real estate- and infrastructure-driven economic growth has likely peaked, and prices for materials such as iron ore will decline.

The Brazilian government occasionally reviews mining royalties, which could allow it to extract further value from Vale's iron ore assets.

Vale's investments in nonferrous business lines, such as coal, nickel, and copper, have been procyclical and expensive, destroying value. Returns have been poor and have weighed down overall group returns.