AANYSEThe short version
Alcoa Corporation
Alcoa is a US-listed, vertically integrated aluminum producer — bauxite, alumina and metal — selling into globally priced markets it cannot control. These fourteen screens test whether its 46% drawdown belongs to the contrarian dislocation pattern this framework hunts.
From $22.57 in April 2025 the shares ran 271% to an $83.79 peak on 2 June 2026, then fell 46% to $45.26 by 31 July — a full round trip in sixteen months.
Mkt cap $11.8BNet debt $842.0MEV $12.7BP/E FY27E 7.6×ND/EBITDA FY27E 0.3×
$45.26
Share price, 31 Jul 2026
$11.8B
Market cap
4.5%
Adj. FCF yield (FY25)
−46%
Since the June peak
SwipeScroll▾
Snapshot
Alcoa Corporation in numbers
Price
$45.26as of 2026-07-31
Mkt cap
$11.8B
Net debt
$842.0M
EV
$12.7B
12m perf
+51.0%
3m ADV
$316.3M
| Year to Dec (USD) | 2023 | 2024 | 2025 | 2026E | 2027E | 2028E |
|---|---|---|---|---|---|---|
| Sales | 10.6B | 11.9B | 12.8B | 14.8B | 15.2B | 14.4B |
| EBITDA | – | – | – | 3.0B | 3.1B | 3.0B |
| EBIT | – | – | – | 2.3B | 2.7B | 2.8B |
| EBIT margin | – | – | – | 15.8% | 17.9% | 19.4% |
| EPS | −3.65 | 0.26 | 4.37 | 6.65 | 5.93 | 7.04 |
| EV/EBITDA | – | – | – | 4.2× | 4.0× | 4.2× |
| EV/EBIT | – | – | – | 5.4× | 4.6× | 4.5× |
| P/E | n/a | 174.1× | 10.4× | 6.8× | 7.6× | 6.4× |
| FCF yield | −3.7% | 0.4% | 4.8% | 7.4% | 13.3% | 16.1% |
| Gearing | 20.4% | 27.3% | – | – | – | – |
Consensus: S&P Capital IQ (CapIQ) · as of 2026-08-03Derived from run data; ratios use the latest price.
IThe business
The business
Alcoa mines bauxite, refines alumina and smelts aluminum, setting none of the prices
FY2025 revenue by product
Aluminum$8.5B65%
Alumina$3.7B28%
Bauxite$0.7B5%
Energy$0.2B2%
FY2025 revenue was $12.8B from 25 sites in eight countries.
- Two steps upstream. Bauxite ore becomes alumina powder at five refineries, then aluminum ingot at eleven smelters; fabricators, not Alcoa, turn metal into cans, car parts and wire.
- A price-taker. Aluminum sells at the London Metal Exchange quote plus premiums, alumina against a published index. Alcoa owns no downstream fabrication and sets no prices.
- Small in a global market. 3.1% of world primary aluminum and 6.4% of world alumina — the largest producer outside China, but China is 59% of the market.
The profit pool
Two segments that keep trading places, and one is now losing money
Segment adjusted EBITDA
- The halves invert. Alumina segment EBITDA swung from $1,408M in FY2024 to $882M in FY2025 to negative $136M in 1H2026, while Aluminum climbed the other way to $1,767M.
- Half the assets earn nothing. Alumina carries $5.7B of assets against Aluminum's $6.2B; management wrote its goodwill to zero in Q4 2025.
- Why alumina broke. A global supply surplus from Chinese and Indonesian refinery expansion pushed realized alumina below cash cost — $329/t against $352/t in 1H2026.
IIThe record
The cash-flow record
A decade of lumpy, unpredictable free cash flow
Reported free cash flow
- Unpredictable, not just volatile. Reported FCF ran from −$715M to $819M to −$440M across FY2016–FY2025; the rolling five-year average drifts between $100M and $349M, a 0.45 coefficient of variation.
- The framework wants stable. Volatile year to year is fine; unpredictable is not. The two negative years came from price and restructuring, not a healthy underwriting cycle.
- Adjusted can't be rebuilt cleanly. Stock-based comp is missing from the data feed for every year, so the framework's adjusted-FCF series is not computable and the reported series stands in.
IIIThe story now
The fit
Does not fit the framework (P1 not met)
The pillar ledger
| Pillar | Result | Deciding number |
|---|---|---|
| Year-10 durability (P1) | Not met | 3.1% Al share; alumina below cost |
| Yield vs the bar (P3c) | Not met | 4.5% vs 8%; 355 bps short |
| Forward path (P3d) | Not met | ~15% chance of the 10% bar |
| Buyback engine (P4b) | Not met | No buyback since 2022; shares +47% |
| Diagnosis (P5) | Met | 66% temporary; price cut > value cut |
- The gate decides it. The framework's one pure gate asks for very high conviction that year-10 revenue and cash flow beat today's. On a 3.1%-share price-taker whose larger segment sells below cash cost, that conviction is absent — and nothing offsets a gate.
- The counter-fact. The diagnosis pillar leans temporary: the blind trial put the impairment at 66% temporary, and the price fell more than intrinsic value did. But P1 sits upstream, so that cannot rescue the case.
- Held to low confidence. A name-mask probe left a 0.21 probability gap, above the 0.20 line, so a prior-driven-risk flag stands — even though both model families agreed on the gate.
The dislocation
A 46% fall in eight weeks, but the tape shows a slide, not a panic
−46%
$83.79 (2 Jun) to $45.26 (31 Jul)
1.23×
Volume vs the 2× capitulation line
~19.5%
Of the fall from dated events
- Two dated triggers, small effect. The 30 June South32 deal and the 16 July Q2 print together explain about 19.5% of the drop; the rest was a June retreat in the aluminum price management itself calls sentiment.
- No capitulation. Peak 20-day volume ran 1.23× normal against the framework's 2× line; no session hit the 4.8–6.7× band of this stock's biggest volume days.
- The seller is unknown. Short-interest, insider and ownership records all predate the June fall, so forced-versus-informed selling cannot be evidenced either way.
Year-10 durability
No pricing moat, and a physical base that has shrunk for a decade
Production volume
- Leadership without the market. Alcoa leads alumina outside China, but that is 41% of the world; China added 13.4M tonnes of aluminum capacity in a decade and sets the price.
- The base is shrinking. Alumina production fell 27.3% and smelting capacity 22.2% over ten years; priced revenue rose because metal prices did, not volume.
- The one real barrier. Western greenfield smelters cost $7,500–$9,100/t against $1,150 in China, so new Western supply is uneconomic — a genuine entry barrier, not a monopoly's pricing power.
Self-help
The buyback flywheel is off while the share count climbs on deals
Shares outstanding
- +47% shares in two years. Weighted shares went from 178M in FY2023 to 261M in FY2025 as acquisitions were settled in stock, with about 17M more committed to South32.
- No buyback since 2022. Zero shares repurchased in FY2023, FY2024, FY2025 or 1H2026, though a $500M authorization sits unused. The flywheel the dislocation should unlock is not turning.
- Allocation points the other way. Debt paydown ranked first on every call since 2023, then a $3.1B cash acquisition — the framework's own falsifier of a pivot away from repurchases.
Damage math
The price cut runs about $2.9B ahead of the value cut
Base-anchored damage
Value actually lost$2B41%
Price beyond value$2.9B59%
- A real gap. On the most defensible pairing, price destroyed about $4.9B of market value against roughly $2.0B of intrinsic value — a $2.9B mispricing.
- 66% temporary. A blind trial of two cited briefs, judged three ways in different reading orders, put the impairment at 66% temporary, spread 0.11.
- Half is spike give-back. Much of the drop reverses a two-week geopolitical price spike; the line that actually broke is Alumina, and part of that is structural.
The clock
Dated catalysts through 2027, but none of them sets the metal price
The catalyst calendar
| When | Event | What it tests |
|---|---|---|
| 15 Oct 2026 | Q3 2026 results | Sequential fall from 2Q26's $2.12 EPS |
| Oct–Nov 2026 | South32 shareholder vote | 5% ticking fee on $3.1B begins |
| End 2026 | WA mining approval | Lifts the bauxite-supply overhang |
| 1H 2027 | South32 deal closes | ~17M shares issued; leverage ≤2.0x |
- Four drawdowns, three recoveries. On its nine-and-three-quarter-year record the stock has had four episodes below −45%; an 18-month recovery to the prior high happened about 22% of the time, median forward return 0.0%.
- No repricing calendar. Unlike an insurer whose premiums reset on a schedule, Alcoa's recovery waits on an undated aluminum price; the catalysts above test the thesis but none sets the metal.
- The near-term print fades. Consensus has quarterly EBITDA easing from 2Q26's $901M toward ~$693M by 2Q27 — the reset price, not the spike, is the run rate the Street models.
IVThe price
Yield versus the bar
Cheap on the headline, short of the bar on the framework's own math
Adjusted FCF yield versus the 8% bar
3-yr avg adj. yield
0.2%
FY2025 adj. yield
4.5%
FY2027 consensus, adjusted
7.2%
FY2027 consensus, face
13.2%
- Below the fortress bar. Adjusted FCF yield is 4.45% on FY2025, 355 bps under the 8% bar a net-cash balance sheet demands. A strong balance sheet raises the bar, not lowers it.
- Consensus clears it, until you adjust. Street FCF implies 13.25% for FY2027, but SBC and acquisition deductions cut it to 7.18%, and the $3.1B South32 deal reinstates a $620M-a-year drag.
- About one chance in seven. The jury put the odds of clearing the 10% bar in any year to FY2029 at roughly 15% (p=0.19) — the widest disagreement on the board.
What it would take
Recovering the lost $10B leans on an aluminum price no event sets
+$2.4B
EBITDA to recover value lost since June
+21–24%
Required LME aluminum price move
$4,752/t
Q2 realized price it must clear
- The math of the round trip. Recovering the ~$10.1B of value lost since 2 June, at the 4.2× EV/EBITDA that fall implies, needs about $2.4B more EBITDA.
- That is a metal-price call. It equates to a $1,008–$1,154/t move in the LME aluminum price, 21–24% above the $4,752/t Alcoa realized in Q2, with no dated corporate event to set it.
- No clean price target. The framework's re-rating math is unavailable because the adjusted-FCF series is not computable, so no price-at-bar can be struck.
What you pay
Cheap on sales, unremarkable on cash
Years of FCF to retire the float
Consensus FY2028 FCF
yrs6.2
Reported FY2025 FCF
yrs20.8
Adjusted FY2025 FCF
yrs22.5
- Cheap on the surface. The $11.8B market cap is 0.92× FY2025 sales and below the $19.6B gross cost of the plant — the beaten-commodity look the framework hunts for.
- Thin on cash. At FY2025 free cash flow it takes about 21 years to retire the float; even on the consensus FY2028 ramp, six. The absurdity check runs the other way.
- Options are facts, not a plan. Listed calls run to January 2028 at roughly 59% implied volatility; the report records this as context, never a recommendation.
What to watch
A cheap, well-financed leader in a real dislocation — but no moat, no capitulation and the buyback engine off
- 01Share count inflects upward
- 02Capital allocation pivots to debt paydown over repurchases
- 03Alumina remains EBITDA-negative for four consecutive quarters
- 04The industry repricing cycle fails to materialize
This distills a fixed fit test built tab by tab; the full report carries every number and its source.
Compiled from the full report · 2026-08-03 · For information, not investment advice.