Chapter 1

The Price-Taker

Alcoa Corporation makes a metal it cannot price. It mines bauxite, refines that ore into alumina, and smelts alumina into primary aluminum — the raw ingot that fabricators elsewhere turn into car panels, cans, and window frames. What it does not do is set what any of that is worth. Aluminum trades on the London Metal Exchange and is priced daily; alumina is sold against a published index; and the extra a US buyer pays to take metal on the ground in Ohio rather than in Rotterdam is a regional premium fixed by trade policy, not by Alcoa [1] [2]. The company is, in the plainest sense, a price-taker: its revenue is three external numbers multiplied by the tonnes it ships, and its profit is those numbers minus a cost base it can grind down but never escape.

That purity is a deliberate inheritance. On November 1, 2016, the old Alcoa split in two — the value-added downstream business (rolled sheet, aerospace parts) left as Arconic, and the upstream mining-and-smelting half kept the Alcoa name [3]. What remains is a vertically integrated commodity producer across 25 sites in eight countries, run as exactly two reportable segments — Alumina (bauxite and refining) and Aluminum (smelting, casting, and most of the power assets) — with no downstream cushion to absorb a swing in market prices [4]. In 2025 it turned $12.8 billion of sales into $1,940 million of Segment Adjusted EBITDA, split $882 million from Alumina and $1,058 million from Aluminum [5] [6].

This report reads Alcoa as what its own filings describe: a pure-play upstream aluminum price-taker whose entire result is set by two external commodity curves and a US Section 232-inflated Midwest premium it does not control, sitting 46% below its high at a low forward multiple that embeds a doubling of premium-driven EBITDA. It is run by a credible operator that keeps its cost and balance-sheet promises but has funded serial acquisitions with roughly 46% two-year dilution while quietly shelving the shareholder returns it announced at the last peak. This chapter builds the machine — the three prices, the thin spread, and the two-speed cycle. The chapters that follow test the cost edge that spread depends on (Cost Is the Only Moat), the ledger of how the company was paid for (The Operator's Word), and what today's price is quietly assuming (What the Price Assumes).

Three prices set the top line

For a price-taker, revenue is almost the whole story. A tonne of Alcoa's primary aluminum sells for three stacked components, and the 10-K names them in order: "(i) the published LME aluminum price for commodity grade P1020 aluminum, (ii) the published regional premium applicable to the delivery locale, and (iii) a negotiated product premium that accounts for factors such as shape and alloy" [7]. The base is the LME — the global exchange price, which Alcoa realizes on a roughly 15-day lag and which averaged $2,614 per tonne in 2025 [8]. On top sits the regional premium — in the United States, the Midwest premium — which compensates for physically delivering metal into a particular market. Last is a small product premium for a specific shape or alloy. Stacked together, those components put Alcoa's realized aluminum price at $3,376 per tonne in 2025 — the $2,614 LME base plus $762 of premiums layered above it [9] [10].

Source: FY2025 Annual Report (Form 10-K), MD&A — realized price ($3,376 in 2025, $2,841 in 2024) and LME 15-day-lag average ($2,614, $2,409); premiums are the residual [11] [12].

The premium layer nearly doubled in a single year — from $432 to $762 per tonne — and that jump, not a move in the underlying metal, is where trade policy enters the income statement. It is taken up under "why now" below. The alumina leg works the same way with one input: alumina is priced off the Alumina Price Index (API), a weighted average Alcoa calculates from three published spot indices, and it realized $415 per tonne in 2025 [13] [14].

None of this is Alcoa's peculiarity; it is the structure of the whole upstream industry. Century Aluminum, a US pure-play smelter, describes the identical three-component price to its own shareholders and states the position bluntly: "Our operating results depend on the market for primary aluminum which can be volatile and subject to many factors beyond our control" [15]. Producers here do not compete on price — they all sell into the same exchange. They compete on cost. That is why the spread, not the price, is the figure that matters most.

A thin spread, heavily levered

Alcoa keeps very little of each sales dollar as gross margin, and that thinness is the source of its violence. In 2025, cost of goods sold was $10,658 million against $12,831 million of sales — 83% of revenue, leaving a gross margin near 17%; selling and administrative expense was a rounding error at roughly 2% of sales [16]. When cost sits that close to price, a few-percent move in either one swings operating profit by hundreds of millions. The economics resolve to a per-tonne spread — realized price minus operating cost — that the segment note discloses directly for each leg.

Source: derived from FY2025 Annual Report (Form 10-K) segment notes — Aluminum $3,376/$2,600 (2025) and $2,841/$2,410 (2024); Alumina $415/$317 and $472/$309 [17] [18].

The aluminum bars carry the point: the spread widened from $431 to $776 a tonne in one year — a $345-per-tonne gain on roughly 2.3 million tonnes of production. That is the operating leverage, and it explains why physical output barely matters to the result. Aluminum production was 2,319 thousand tonnes in 2025 versus 2,215 the year before, and alumina production actually fell, to 9,640 thousand tonnes from 10,034 — small moves [19] [20]. Yet total Segment Adjusted EBITDA has swung from $2,280 million in 2022 down to $734 million in the 2023 trough and back to $1,940 million in 2025 — a three-fold range on nearly flat tonnes [21] [22]. The same leverage runs straight through to cash: cash provided from operations nearly doubled to $1,185 million in 2025 from $622 million a year earlier, so that free cash flow — operating cash flow less capital expenditure — swung from negative $440 million in the 2023 trough to positive $567 million in 2025 [23].

Source: FY2023 Annual Report (Form 10-K) segment reconciliation (FY2021–FY2023) and FY2025 Annual Report (Form 10-K) segment information (FY2024–FY2025) [24] [25].

The cost side of the spread is not a single number but a short list of commodity inputs, and Alcoa discloses their weights. In alumina refining, the cash cost is roughly a quarter bauxite, a third conversion (labor and plant), and the balance energy and caustic soda. In aluminum smelting, the alumina intermediate is the single largest line at about a third, followed by power and carbon.

Source: Alcoa investor overview, March 2026, composition of 4Q25 production cash costs [26].

Two features of that list carry through the rest of the report. First, energy — natural gas and other power in refining, electric power in smelting — is a dominant, differentiating cost, and where a producer buys its power decides whether it survives a downcycle; that is the cost edge examined in the next chapter. Second, alumina is both a product Alcoa sells and, at roughly a third of smelting cost, the largest input it feeds to itself — which is why the two segments cannot be read in isolation.

Two segments, two cycles

Because alumina is the aluminum segment's biggest raw material, a fall in the alumina price does two opposite things to Alcoa at once: it guts the Alumina segment's revenue and, in the same breath, cheapens the Aluminum segment's cost base. In 2025 both happened, and the two segments moved in opposite directions inside a single year.

Source: FY2025 Annual Report (Form 10-K), segment notes — Alumina and Aluminum Segment Adjusted EBITDA, FY2024 and FY2025 [27] [28].

The Alumina segment's EBITDA fell $526 million as the API dropped and alumina slid into a global surplus, driven by refinery expansions in China and Indonesia after prices had touched an all-time high in the fourth quarter of 2024 [29] [30]. The Aluminum segment's EBITDA rose $401 million over the same span, lifted by a 9% higher aluminum price and helped by the very same cheaper alumina flowing into its costs, while metal itself stayed in deficit on historically low inventories [31] [32]. A refining glut and a metal shortage coexisted. The practical consequence: "the aluminum cycle" is not one number for this company — the vertical chain both amplifies consolidated swings and partly hedges them from the inside, and a single-segment read is misleading in either direction.

The surplus left a mark on the balance sheet as well as the income statement. In the fourth quarter of 2025, Alcoa wrote the Alumina reporting unit's goodwill down to zero, "a charge of $144 in Impairment of goodwill," driven by declining alumina prices and rising Australian mine costs [33]. The durability of Alcoa's cost position on that leg — where the write-down points — is the subject of the next chapter.

Why now

Alcoa is in front of investors today for three dated reasons, all of which flow from the machine above.

The first is policy. In 2025 the US moved its Section 232 tariff on Canadian aluminum from an exempt 10% to 25% on March 12 and to 50% on June 4 [34]. That tariff cuts both ways for Alcoa. It lifts the Midwest premium — up 211% year over year — which is the layer that took Alcoa's realized aluminum price from $2,841 to $3,376 a tonne, so a US-sited producer benefits [35]. But Alcoa also ships large volumes of Canadian-made metal into the US and pays the tariff on it: total Section 232 tariff costs were $571 million in 2025, which management says the elevated premium currently covers [36] [37]. The premium is a benefit and the tariff a cost, and the net is roughly a wash only while both hold — a two-sided dependence, not a one-way windfall.

The second is a second policy layer taking effect now. The EU's Carbon Border Adjustment Mechanism — a levy on carbon-intensive imports that has applied to aluminum in a transitional phase since October 2023 — moves into fuller force, repricing metal by its carbon content in Alcoa's European markets [38]. How much that is worth, and to whom, belongs with the cost-and-carbon discussion that follows.

The third is the tape. Alcoa trades around an $11.8 billion market value, roughly 46% below its three-year high — a deeper drawdown than the aluminum group as a whole — even as consensus looks for double-digit revenue growth ahead. What that price is underwriting, and why a headline year of record statutory earnings should not be taken at face value, is the closing chapter's work (What the Price Assumes).

For now, the machine is the point. Three prices Alcoa does not set — LME, API, and a tariff-driven Midwest premium — multiplied by tonnes it can barely move, minus a cost base built of energy, bauxite, alumina, caustic, and carbon, produce a per-tonne spread thin enough that a modest price move is a swing of hundreds of millions, and a chain integrated enough that its two halves can pull against each other in the same year. Everything that follows — the cost edge that defends the spread, the capital that was spent reshaping the company, and the price the market now puts on it — is built on that structure.