Constellium SE
Rolls and shapes aluminium to order: coil for drink cans, sheet for car bodies, thick plate for aircraft wings, and crash beams for cars.
The scores
Growth2/4 →
Shipments fell 1% in H1 2026 and rose 4% in 2025; the revenue and profit growth is metal price, not demand.
Business quality2/4 →
Aerospace qualification is a real barrier, but 61% of revenue is can and auto sheet, where packaging volumes fell while its largest rival burned.
Financial position3/4 ↑
Net debt of $1.76B with nothing maturing before 2028 and an undrawn revolver; not trivial, and normalised leverage of 2.6 times exceeds its own target.
Execution record3/4 →
Muscle Shoals and the flood recovered, guidance raised twice, buybacks bought cheaply; against that, a corrected prior-period disclosure in the metric it is judged on.
Per share capture4/4 →
Basic count down 4.3%, buybacks cash-funded, no dividend, no issuance and no convertible; stock awards recapture about half of it.
Potential2/4 →
The prize grows about 1.3% a year, no customer volume or term is disclosed anywhere, and the upside turns on a tariff.
Risk profile2/4 →
The Section 232 tariff is externally controlled, sits inside the horizon, can change in one proclamation, and it decides the valuation.
Valuation3/4 →
5.1 times guided profit and 7.5 times a normalised one: the price reflects the business without the windfall, neither discounting nor extrapolating it.
Evidence quality3/4 →
Central facts are filed and corroborated by Boeing, Novelis, Century Aluminum and a trade body; the normalisation rests on arithmetic and one management claim.
The Verdict
We rate Constellium neutral. It rolls aluminium into coil for drink cans, sheet for cars and plate for aircraft. The business is mixed: profit more than doubled this year while it shipped less metal, thanks to a US tariff, cheap metal costs, an accounting timing gain and a rival's mill knocked out by fire. Aircraft metal is the one lasting improvement. The price already assumes most of that windfall fades, so it is fair, not cheap. The view hinges on how long the tariff lasts: if it holds while aircraft work keeps growing, today's price proves cheap. The earnings are a tariff, not a franchise.
The long casePlane makers are building more jets after years of shortage, and Constellium's aircraft metal rides that; most of its sales are cans and cars, which this recovery does not reach.
The evidence under this - 8 claims
- The normalisation uses the FY2025 packaging margin, which was itself lifted by the tariff in its second half, and a through-cycle aerospace margin management may be sandbagging. If instead the current environment persists, the same arithmetic gives about $970M and 5.3 times, which is not a demanding price at all - the range between the two is the whole argument.
- It also excludes the growth investments management says arrive in 2027 and 2028 at returns above 15%, which are real capital already largely spent and not in either figure.
- 2026-09-26editedRe-struck at the September 25, 2026 close on the refreshed pack: enterprise value $5,133M, so 7.5 times the normalised $680M (was 8.0) and 5.3 times the current-environment $970M (was 5.6). The normalisation itself is unchanged.
- 2026-09-26editedRe-struck on the research pack refreshed at the 2026-09-26 review (S14) at the September 25 close: the price fell 8.3% with no quarter reported, so the multiples moved and the six-year medians did not.
This is the third upgrade of the year on the same driver, and the company beat consensus by 129%, 122% and 97% in the three prints before Q2, so its own guidance has been a poor predictor of its own results. The new range still implies a fourth quarter well below the third.
- 2026-10-28editedQ3 2026 release raised the guide to $1,030M-$1,060M; the Q3 test of this claim resolved in favour of a higher range
The company
Constellium does not sell aluminium; it sells aluminium shaped, alloyed and treated to a customer's specification. Its biggest business rolls coil for drink cans and sheet for car bodies: 61% of first-half sales.
The other main business, a quarter of sales but 42% of segment profit, makes thick plate for aircraft wings and defence, plus crash beams and battery cases for cars. That metal has to pass the customer's own lengthy approval, and for aircraft the regulator's strength tests too.
The evidence under this - 4 claims
This is the company describing its own moat, which is the least independent witness available on the question. The independent version is narrower: MMPDS qualification is a real, dated barrier in aerospace structure, but it says nothing about beverage can stock, which is a specification-driven commodity where Constellium's own packaging volumes fell 3% in H1 2026 while a rival was out of the market.
How it makes money
It passes the cost of the raw metal on to customers and hedges the rest, so in principle it earns a fee for every tonne it rolls, not a bet on the aluminium price.
In practice two things leak through: a timing gain or loss when prices move between buying metal and selling it, and the gap between the price of the scrap it melts down and the price of the metal it sells.
The evidence under this - 4 claims
He also said the company's own recycling work - scrap selection, melt-loss reduction, throughput - is inside that same bucket, so the market share of it is not separable from the self-help share either. That cuts both ways: it is a reason the figure is genuinely hard to state, not only a reason to withhold it.
The same executive said in the previous answer that incremental recycling benefits would taper off in the second half, and the company's own guidance implies a materially lower second half. A locked spread and a lower guided profit are consistent only if something else falls away, and management did not say what.
Where the money comes from, stream by stream
| Stream | Share of sales (H1 2026) | How it earns it |
|---|---|---|
| Packaging rolled products | 42.9% | Can body and lid sheet rolled for beverage can makers; the metal is passed through and the conversion margin kept. |
| Automotive rolled products | 16.5% | Aluminium sheet for car bodies and closures, sold to carmakers under qualified, multi-year supply programmes. |
| Aerospace rolled products | 13.0% | Certified plate and sheet for aircraft and space structures, qualified part by part with each airframer. |
| Transportation, industry, defence and other rolled products | 10.4% | Plate and sheet for trucks, rail, defence and industrial uses, largely from the Aerospace & Transportation mills. |
| Automotive extruded products | 10.1% | Crash-management systems and structural extrusions made to carmakers' designs. |
| Other extruded products | 6.0% | Industrial extrusions sold to general manufacturing customers. |
| Specialty and other thin-rolled products | 1.1% | Thin foil-like specialty rolled products in small volumes. |
Who runs it
Ingrid Joerg became chief executive in January 2026 after a decade running the aerospace and then the can and car sheet businesses; Jack Guo has run the finances since 2023. Insiders own only about 1% of the company, and the biggest holder, a French state investor, sold a large block in May. The largest part of the yearly bonus rides on segment profit, which includes the tariff windfall, and the share awards on returns against other materials stocks.
The evidence under this - 4 claims
The 5% holders' figures are as of their own filings, not March 2026: BlackRock's dates from September 2024, and Bpifrance sold 4.19 million shares in May 2026 (C49), so its 9.2% is already out of date.
- Segment Adjusted EBITDA excludes metal price lag but keeps the tariff-driven premium and scrap-spread gains the thesis questions (C08, C09), so the bonus metric rewards the windfall as fully as the operations.
- The free cash flow metric also excludes factoring effects, so it is not the free cash flow the release headlines.
Why it wins, and who could beat it
Why customers come back
A buyer cannot change supplier without approving the new metal all over again, and the customer, not Constellium, bears that cost. For an aircraft, the alloy must earn strength values the US aviation regulator accepts.
That protected half is the one that grew when demand returned: aircraft and transport tonnage rose 20% in the first half as aerospace and defence orders recovered.
The evidence under this - 3 claims
This is the company describing its own moat, which is the least independent witness available on the question. The independent version is narrower: MMPDS qualification is a real, dated barrier in aerospace structure, but it says nothing about beverage can stock, which is a specification-driven commodity where Constellium's own packaging volumes fell 3% in H1 2026 while a rival was out of the market.
The company says the same segment's transportation, industry and defence line also benefited from the North American automotive rolled shortage, so part of this tonnage is the same temporary substitution that lifted Packaging & Automotive Rolled Products, not aerospace recovery.
The threat
One threat per business. Novelis's Oswego mill in New York, shut after two fires in late 2025, restarted in June 2026, and Constellium expects the car-sheet volumes it gained from the shortage to normalise by year-end.
In aircraft, carbon fibre has taken the wings of the newest wide-body jets, though not Boeing's 737, whose build rate is rising from 42 to 47 a month.
The evidence under this - 3 claims
- The chief executive also says Constellium's own U.S. automotive capacity is limited and expected to be fully booked for several years, and that platform nominations run for the life of the contract - so the volume it holds under contract does not move when a rival restarts, only the spot substitution does.
- On the call management said the automotive benefit from the Oswego outage is tapering off in Q4 and that US can-sheet imports surged to 350 thousand tonnes by August because of it, leaving a digestible inventory overhang.
- 2026-10-28editedQ3 call: the shortage benefit tapers in Q4 and can-sheet imports surged
The rivals: all private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | How far along | Priced against it |
|---|---|---|---|
| Novelis Inc. (Oswego, New York rolling mill)Priced against it: Listed abroad | On sale | Listed abroad | |
| Not listed | Carbon-fibre composite primary structure (Boeing 787, Airbus A350 wings and fuselage)Priced against it: Not listed | On sale | Not listed |
How big it could get
How big this could get
Take its biggest market. American drink cans used about 1.38 million tonnes of aluminium in 2023, roughly 106 billion cans, and that volume has grown about 1.3% a year for a decade.
Constellium's can-and-car-sheet business shipped 1,086 thousand tonnes in 2025 across two continents: about four-fifths the size of the entire US can market. Growth there comes slowly.
The evidence under this - 2 claims
The results
The numbers
Sales grew 26% over the twelve months to Sep 2026, faster than the 15% in the year to Dec 2025.
It keeps 5.7% of sales as profit, up from 1.9% in the year to Dec 2023.
Yearly sales growth has ranged from −8.3% to +15% over the last 3 years, with sales falling in 2 of them.
Shareholders' money earned 53% a year.
It turned 3.2% of sales into free cash, $319M over the last twelve months.
It owes $1.66B more than the cash it holds, about 2.6 years of the cash its operations bring in.
The share count fell 5.4% in 3.8 years, and pay in shares costs 0.3% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-10-28. Fiscal years end in December.
| Dec 2022 | Dec 2023 | Dec 2024 | Dec 2025 | 12 mo to Sep 2026 | |
|---|---|---|---|---|---|
| Size and growth | |||||
| Sales | $8.53B | $7.83B | $7.33B | $8.45B | $10.0B |
| Sales growth | - | −8.3% | −6.3% | +15% | +26% |
| Profit | |||||
| Net margin | 3.6% | 1.9% | 0.8% | 3.2% | 5.7% |
| What growth costs, per sales dollar | |||||
| Research and development | 0.5% | 0.7% | 0.7% | 0.6% | 0.5% |
| Selling, general and admin | 3.3% | 4.1% | 4.3% | 3.9% | 3.6% |
| Pay in shares | 0.2% | 0.3% | 0.3% | 0.2% | 0.3% |
| Capital spending | 3.4% | 4.7% | 5.6% | 3.9% | 3.2% |
| Cash | |||||
| Cash from operations | $365M | $432M | $301M | $489M | $638M |
| Free cash flow | $76.0M | $66.0M | −$112M | $159M | $319M |
| Balance sheet | |||||
| Cash and short-term investments | - | $223M | $141M | $120M | $158M |
| Debt | - | $1.93B | $1.92B | $1.94B | $1.82B |
| Net cash | - | −$1.71B | −$1.78B | −$1.82B | −$1.66B |
| Returns and shares | |||||
| Shares (diluted) | 147M | 148M | 148M | 142M | 139M |
What growth costs
Growth does not make this business much more profitable on its own. Office, research and plant spending rise roughly in step with the metal shipped, so the recent jump in profit came from metal prices and scrap, not from cheaper growth.
The evidence under this - 2 claims
Revenue is mostly pass-through metal (C01), so a falling cost per sales dollar since 2024 reflects higher aluminium prices in the denominator rather than efficiency; per tonne the overhead rose slightly.
The last report (29 Jul)
Quarter ended June 2026, filed figures.
Sales rose 22% to $2.64B on 1% fewer tonnes. Profit before interest, tax and depreciation and before the timing effect was a record $280M for a third quarter, against $196M a year earlier.
Can-and-car-sheet profit was $608 a tonne against $298, with the rival's mill back since June. Free cash flow was $128M, so $223M over nine months. The forecast for 2026 went up to $1.03B to $1.06B, which still implies a fourth quarter well below the third.
On the call management tied the lower fourth quarter to seasonality, an outage and the car-sheet shortage fading, and expects stronger aerospace volumes and prices next year.
What it did to the thesisIt passed the test the thesis set: profit per tonne stayed above $550, cash arrived and the forecast rose. But the segment's gain still leans on the car-sheet shortage and cheap metal costs, so the view stays Neutral until the fourth quarter shows it lasting.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $1.04 against $0.86 expected; $1.04 as filed, with no adjustments.
The evidence under this - 7 claims
The release credits the Packaging & Automotive Rolled Products gain to favourable price and mix and favourable metal costs, and says automotive rolled volumes again benefited from North American supply shortages - the same two drivers the thesis worries are temporary. The segment's per-tonne figure slipped from $621 in Q2 to $608 and Q3's metal price lag was a $4M negative, so the Q3 profit contains no timing gain.
This is the third upgrade of the year on the same driver, and the company beat consensus by 129%, 122% and 97% in the three prints before Q2, so its own guidance has been a poor predictor of its own results. The new range still implies a fourth quarter well below the third.
- 2026-10-28editedQ3 2026 release raised the guide to $1,030M-$1,060M; the Q3 test of this claim resolved in favour of a higher range
- The chief executive attributed the first-half to second-half shape in July to ordinary seasonality - lower summer and December demand, European maintenance outages scheduled into the second half. H2 2025 Segment Adjusted EBITDA was $408M, so the new range is 12% to 20% above last year's second half rather than a decline against a comparable period. The Q4 gap to Q3 is wider than that seasonality alone would make, which the release does not explain and the call was not available to ask.
- On the call the chief executive explained the fall from $280M: normal seasonality, a recycling outage moved to the fourth quarter, the automotive benefit from the rival mill fading, new tariffs on cars and parts, and $55M of full-year headquarters cost phasing into Q4.
- 2026-10-28editedRe-struck on the raised guide and the Q3 actual: H2 now $458M-$488M, Q4 $178M-$208M
- 2026-10-28editedQ3 call explained the implied Q4 drop (seasonality, outage, fading automotive shortage, tariffs, headquarters spending)
- 2026-10-28editedRe-sourced on the Q3 2026 release: net debt $1,658M, leverage 1.6 times; $100M of the 2028 notes redeemed in July and the asset-based facility enlarged to $700M on 22 October
- The chief executive also says Constellium's own U.S. automotive capacity is limited and expected to be fully booked for several years, and that platform nominations run for the life of the contract - so the volume it holds under contract does not move when a rival restarts, only the spot substitution does.
- On the call management said the automotive benefit from the Oswego outage is tapering off in Q4 and that US can-sheet imports surged to 350 thousand tonnes by August because of it, leaving a digestible inventory overhang.
- 2026-10-28editedQ3 call: the shortage benefit tapers in Q4 and can-sheet imports surged
Asked for the size of the benefit from these projects, management gave only the 15% return hurdle and the earlier capital figure, and no EBITDA number.
What management did with the money
It bought back 8.9 million shares in 2025 at an average of $12.92 and 1.8 million this year at $26.67; it pays no dividend. Net debt fell only from $1.82B to $1.76B, so the lower borrowing ratio it reports comes almost entirely from the higher profit.
Insiders in 2026Q2: 2 purchases and 11 sales, from their Form 4 filings.
The evidence under this - 3 claims
- 2026-10-28editedRe-sourced on the Q3 2026 release: net debt $1,658M, leverage 1.6 times; $100M of the 2028 notes redeemed in July and the asset-based facility enlarged to $700M on 22 October
Cash generation is real even if the ratio moved for arithmetic reasons: the company redeemed $100M of its 2028 notes in July 2026 and bought back $48M of stock in the half, both from cash. A denominator-driven ratio is not the same as an unimproved balance sheet.
The price, and where we disagree
In short. Fair, not cheap. The price already assumes most of this year's unusually high profit fades. It would be cheap only if the tariff boost and the cost savings last.
The shares are up 66.6% in a year but 8.3% lower than when this thesis was written. They fell 5.0% the day after the record quarter and raised forecast, and sit 32.7% below their high. The market is already asking how long this lasts.
The September-quarter results arrived on October 28 and passed the first test: can-and-car-sheet profit held at $608 a tonne with the rival mill back, and the forecast rose. The next answer is the fourth quarter, which the new forecast puts at $178M to $208M, in February.
The evidence under this - 6 claims
- 2026-09-26editedRe-struck on the research pack refreshed at the 2026-09-26 review (S14) at the September 25 close: the price fell 8.3% with no quarter reported, so the multiples moved and the six-year medians did not.
- The chief executive also says Constellium's own U.S. automotive capacity is limited and expected to be fully booked for several years, and that platform nominations run for the life of the contract - so the volume it holds under contract does not move when a rival restarts, only the spot substitution does.
- On the call management said the automotive benefit from the Oswego outage is tapering off in Q4 and that US can-sheet imports surged to 350 thousand tonnes by August because of it, leaving a digestible inventory overhang.
- 2026-10-28editedQ3 call: the shortage benefit tapers in Q4 and can-sheet imports surged
The release credits the Packaging & Automotive Rolled Products gain to favourable price and mix and favourable metal costs, and says automotive rolled volumes again benefited from North American supply shortages - the same two drivers the thesis worries are temporary. The segment's per-tonne figure slipped from $621 in Q2 to $608 and Q3's metal price lag was a $4M negative, so the Q3 profit contains no timing gain.
This is the third upgrade of the year on the same driver, and the company beat consensus by 129%, 122% and 97% in the three prints before Q2, so its own guidance has been a poor predictor of its own results. The new range still implies a fourth quarter well below the third.
- 2026-10-28editedQ3 2026 release raised the guide to $1,030M-$1,060M; the Q3 test of this claim resolved in favour of a higher range
How we read the multiple
Which profit you believe decides the price. Counting its debt, the business costs about $5.13B, which is:
Even management's 2028 target of $900M is below the roughly $1B it forecasts for this year, part of it a one-off gain from timing on metal prices.
- About 4.8 times last year's cash profit, 5.1 times this year's forecast: cheap for almost any business.
- About 7.5 times a normal year's, built from today's tonnes at the company's own long-run margins.
- On that normal year, its debt would sit above its own target.
Against its own history. Cheaper than its own recent years for each dollar of profit, but that profit includes the timing gain, which is exactly what flatters the comparison.
The evidence under this - 5 claims
- The normalisation uses the FY2025 packaging margin, which was itself lifted by the tariff in its second half, and a through-cycle aerospace margin management may be sandbagging. If instead the current environment persists, the same arithmetic gives about $970M and 5.3 times, which is not a demanding price at all - the range between the two is the whole argument.
- It also excludes the growth investments management says arrive in 2027 and 2028 at returns above 15%, which are real capital already largely spent and not in either figure.
- 2026-09-26editedRe-struck at the September 25, 2026 close on the refreshed pack: enterprise value $5,133M, so 7.5 times the normalised $680M (was 8.0) and 5.3 times the current-environment $970M (was 5.6). The normalisation itself is unchanged.
This is the third upgrade of the year on the same driver, and the company beat consensus by 129%, 122% and 97% in the three prints before Q2, so its own guidance has been a poor predictor of its own results. The new range still implies a fourth quarter well below the third.
- 2026-10-28editedQ3 2026 release raised the guide to $1,030M-$1,060M; the Q3 test of this claim resolved in favour of a higher range
The company says it expects to reach the 2028 target two years early and describes 2028 as a milestone rather than a destination, and it has named growth investments - the Airware cast house, the Muscle Shoals and Ravenswood casting complexes, the Nostitz and Fayssac recycling centres - that land in 2027 and 2028 and are not in the 2026 number.
- 2026-09-26editedRe-struck on the research pack refreshed at the 2026-09-26 review (S14) at the September 25 close: the price fell 8.3% with no quarter reported, so the multiples moved and the six-year medians did not.
- Metal price lag is not always a gain: it was negative $92M in 2023. Stripping it out of a good year without adding it back in a bad one would flatter the through-cycle figure, so the honest reading is that the line is volatile in both directions and that trailing earnings today sit at the favourable end of it.
- Trailing net income also carries a $60M unrealised derivative loss running the other way, which the company excludes from its own adjusted measure; leaving it in makes this estimate conservative rather than harsh.
- 2026-09-26editedRe-struck at the September 25, 2026 close on the refreshed pack: market value $3,373M, so about 10.9 times the lag-free earnings against 6.2 reported.
Priced against its rivals
Its rivals have not yet been checked against their own filings, so there is no group to price it against; the price is judged against its own history instead.
Where we disagree with the market
A cheap cyclical whose earnings have stepped up for good: record profit, a raised forecast, and 14 of 17 analysts positive.
The earnings are a tariff, not a franchise. The market half-knows it, so the question is how long the tariff lasts.
Few are reading the company's own forecast arithmetic: $980M to $1.02B for the year, against $572M already earned by June, means a second half 22% to 29% lower.
We are wrong if the tariff holds while aircraft tonnage keeps compounding. Then the windfall becomes the base and today's price proves cheap.
What analysts say, and the evidence
The evidence under this - 6 claims
- 2026-09-26editedRe-struck on the research pack refreshed at the 2026-09-26 review (S14) at the September 25 close: the price fell 8.3% with no quarter reported, so the multiples moved and the six-year medians did not.
This is the third upgrade of the year on the same driver, and the company beat consensus by 129%, 122% and 97% in the three prints before Q2, so its own guidance has been a poor predictor of its own results. The new range still implies a fourth quarter well below the third.
- 2026-10-28editedQ3 2026 release raised the guide to $1,030M-$1,060M; the Q3 test of this claim resolved in favour of a higher range
- The chief executive attributed the first-half to second-half shape in July to ordinary seasonality - lower summer and December demand, European maintenance outages scheduled into the second half. H2 2025 Segment Adjusted EBITDA was $408M, so the new range is 12% to 20% above last year's second half rather than a decline against a comparable period. The Q4 gap to Q3 is wider than that seasonality alone would make, which the release does not explain and the call was not available to ask.
- On the call the chief executive explained the fall from $280M: normal seasonality, a recycling outage moved to the fourth quarter, the automotive benefit from the rival mill fading, new tariffs on cars and parts, and $55M of full-year headquarters cost phasing into Q4.
- 2026-10-28editedRe-struck on the raised guide and the Q3 actual: H2 now $458M-$488M, Q4 $178M-$208M
- 2026-10-28editedQ3 call explained the implied Q4 drop (seasonality, outage, fading automotive shortage, tariffs, headquarters spending)
The company says the same segment's transportation, industry and defence line also benefited from the North American automotive rolled shortage, so part of this tonnage is the same temporary substitution that lifted Packaging & Automotive Rolled Products, not aerospace recovery.
What could go wrong, and how we would know
The risks
- A tariff can be reversed. The extra price Americans pay for aluminium rose 183% to $2,405 a tonne on a 50% tariff nobody can forecast.
- The rival mill is back. Novelis restarted Oswego in June, and management expects the car-sheet gains from the shortage to normalise by year-end.
- Profit not yet cash. Of the first half's $798M profit, only $95M reached free cash flow; stock and customer bills absorbed $457M.
- Two customers, a fifth of sales. Two unnamed customers were 21.2% of 2025 revenue, and the ten largest about 56%.
- Past accounting errors. It corrected errors in its timing-gain figures for 2024 and 2025, serious enough to trigger a review of executive pay.
The evidence under this - 8 claims
- The chief executive also says Constellium's own U.S. automotive capacity is limited and expected to be fully booked for several years, and that platform nominations run for the life of the contract - so the volume it holds under contract does not move when a rival restarts, only the spot substitution does.
- On the call management said the automotive benefit from the Oswego outage is tapering off in Q4 and that US can-sheet imports surged to 350 thousand tonnes by August because of it, leaving a digestible inventory overhang.
- 2026-10-28editedQ3 call: the shortage benefit tapers in Q4 and can-sheet imports surged
The company assessed the errors as immaterial to every previously issued statement, and says they did not change consolidated cost of sales or the consolidated income statement at all - only the split within cost of sales, and therefore the non-cash lag line and one segment's adjusted profit. The auditor issued a clean opinion and management concluded internal control was effective.
What we still do not know
- How much of the cost windfall is the scrap market and how much the company's own recycling. The finance chief declined to say.
- Whether stock and customer bills stop swallowing cash in the second half, which the $300M cash forecast needs.
- Whether the European Union taxes scrap exports from 2027, which management says would help its European plants.
The evidence under this - 3 claims
He also said the company's own recycling work - scrap selection, melt-loss reduction, throughput - is inside that same bucket, so the market share of it is not separable from the self-help share either. That cuts both ways: it is a reason the figure is genuinely hard to state, not only a reason to withhold it.
- The guided cash costs are annual and need not split evenly: capital expenditure and cash taxes are both commonly second-half weighted, and if more of the $330M fell in H1 the H2 requirement is easier than this arithmetic makes it.
- Working capital can also release without prices falling if volumes fall, or if the company factors more receivables - it sold $1,990M in H1 against a $713M limit, so the programme has room to run harder.
She also says the 30% the industry association asked for is probably not realistic, so the company's own expectation of the size is below what it lobbied for, and the measure requires member-state votes after any decision - two steps at which it can be delayed or diluted.
What would change our view
- Neutral→AttractiveThe per-tonne economics hold and the cash follows
It needs all of:
- ○Profit per tonne in cans and car sheet stays above $550 through 2027, showing the earnings do not depend on metal prices.checked at the next results, due by 17 Feb
- ○Free cash flow reaches the $300M forecast for 2026, and stays above $250M once metal prices stop rising.checked at the next results, due by 17 Feb
- Neutral→UnattractiveThe premium or the volumes give way
It needs any of:
- ○The extra price Americans pay over the world aluminium price falls below $1,000 a tonne.checked at the next results, due by 17 Feb
- ○Aircraft tonnage stops growing while Boeing builds 737s faster, or 2026 profit before the timing gain lands outside $1.03B to $1.06B.checked at the next results, due by 17 Feb
What is coming
- 28 October 2026 Q3 2026 results: Whether the Packaging & Automotive cost bucket stays positive with Novelis back, and whether second-half profit falls as far as the guidance implies.
Every dated event ahead
- 28 October 2026 Q3 2026 results: Whether the Packaging & Automotive cost bucket stays positive with Novelis back, and whether second-half profit falls as far as the guidance implies.
- 31 December 2026 Share repurchase programme expires: Whether the board renews buybacks at today's price, having bought most of it at about half the current price; renewal or lapse says what management thinks the shares are worth.
- January 2027 EU aluminium scrap export measure: Whether Europe taxes scrap exports as management expects, which would support European scrap spreads from 2027; no measure leaves that spread to the market.
- February 2027 FY2026 results and 2027 guidance: Whether 2027 guidance keeps the tariff and scrap gains or falls back toward the $900M 2028 target, which dates the windfall.
Since we wrote this
- 28 OctEarningsQ3 profit rose 43%; forecast raised; view heldCan-and-car-sheet profit held at $608 a tonne with the rival mill back, but the gain still leans on shortages and cheap metal.
- 28 OctEarningsCall explained the Q4 drop; view holds at NeutralManagement tied the lower fourth quarter to seasonality, an outage and the fading car-sheet shortage, and gave no figure for 2027 project gains.
- 5 OctThesis reviewThe long case now sits under the verdictAfter the five-name pilot, the owner asked for the long case on every thesis.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
- 26 SepPrice movePrice figures moved to the September 25 close; view heldThe business now costs 5.1 times this year's forecast profit and 7.5 times a normal year's. Which profit is real is still the question, and the October results are the first clean answer.
- 7 SepThesis writtenNeutral, medium convictionWritten five weeks after a record half-year in which profit more than doubled on 1% less metal.
Sources 19
Sources - 19
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-02-25the pass-through pricing model, the Competition section (classes, no names), customer concentration, the segment note, the three-year Segment Adjusted EBITDA and metal price lag series, and the revision of prior-period metal price lagbacksC01C23C32C33C43C47C51C52
- S02primarycompany filing2026-07-29the segment tables, the LME and regional premium averages, the cash flow statement and working-capital components, the balance sheet, the debt note, the share counts and the scrap-spread definitionbacksC02C03C04C05C06C08C09C12C18C19C20C22C24C26C27C28C41C42C46C52
- S03primarycompany release2026-07-29shipments in tonnes, Segment Adjusted EBITDA per tonne by segment, free cash flow, leverage, the raised FY2026 guidance and the 2028 targetsbacksC02C03C04C05C08C09C11C12C13C15C17C18C22C24C28C42
- S04primarytranscript2026-07-29Q2 2026 earnings call transcript (question section and prepared remarks)FMP earning-call-transcript, CSTM 2026Q2the CEO's through-cycle A&T margin guidance, the CFO's refusal to size the scrap-spread benefit and his pointer to the P&ARP cost bucket, the H2 scrap lock, the Oswego normalisation comment and the EU scrap export tax expectation. The analysts' questions are a separate origin from management's answers
You should think that the EUR 2,000 per ton are an exceptional quarter on the A&T performance. ... Our guidance for the longer term through cycle is EUR 1,300 per ton, and we feel that's a fair value that we are guiding.
backsC10C14C22C39C40 - S05primarycompany release2026-02-18FY2025 shipments, segment margins per tonne, free cash flow, the original FY2026 guidance of $780-820M, the 2028 targets and the CEO's attribution of the price rise to the U.S. tariff announcementsbacksC11C13C15C17C24C28C41
- S06secondaryvendor data2026-09-07Research pack, frozen at content/CSTM/research-pack.jsoncontent/CSTM/research-pack.jsonprice, market value, enterprise value, the trailing multiples and their six-year medians, returns, the 20-day range, the analyst split, FINRA short interest and the measured news reactionsbacksC38
- S07primarycompany filing2026-08-06Century Aluminum Form 10-Q, quarter ended June 30, 2026 - Section 232 Aluminum Tariffs Century Aluminum Companya second filer's account of the tariff steps and dates, and its own statement that the Midwest premium rose afterwards with a material positive effect on resultsbacksC07
- S08primarycompany filing2026-05Novelis Inc. Q4 FY2026 results (Form 8-K) Novelis Inc.the scale of the rival that was out of the market (3,557 kt shipped in FY2026), the Oswego fire losses and management's expectation of restarting and normalising shipmentsbacksC21
- S09primarycompany release2026-06-10the dated confirmation from the rival itself that the mill Constellium's automotive windfall depended on is back online and ramping supply
aerospace, automotive, beverage can and specialties industries throughout North America, Europe, Asia and South America
backsC21 - S10primaryindustry data2024-12The Aluminum Association, 2024 Aluminum Can KPI Report The Aluminum Associationthe size and growth of the U.S. beverage can pond, measured by a trade body rather than by Constellium, as an annual series from 2012 to 2023
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 / Pounds of Cans Shipped by U.S. Producers (MMlbs) 2644 2581.3 2555.7 2528.7 2529.2 2645.7 2622.6 2689.5 2838.6 2808.9 2840.9 3035.7
backsC36 - S11primaryindustry data2026-09-07MMPDS (Metallic Materials Properties Development and Standardization), Battelle - definition and FAA acceptance Battelle / MMPDSthe qualification mechanism behind the aerospace moat, stated by the standards body rather than by the supplier
the 'A' and 'B' basis values published for materials in the MMPDS have been determined by the FAA to satisfy the material strength probability levels required by Title 14 of the Code of Federal Regulations
backsC34 - S12primarycompany filing2026-07The Boeing Company Form 10-Q, quarter ended June 30, 2026 - 737 Program The Boeing Companya customer's own statement of the narrowbody build rate and its direction, which is the rate behind machine.currentbacksC35
- S13primarycompany release2026-07-29the four-year free cash flow track record with its deferred-purchase-price footnote, the segment Adjusted EBITDA bridges, and the components behind the 2026 free cash flow guidance - capital expenditure, cash interest, cash taxes and working capitalbacksC44C45
- S14secondaryvendor data2026-09-26Research pack, refreshed at the 2026-09-26 review and frozen at content/CSTM/research-pack.json (market tier at the 2026-09-25 close)content/CSTM/research-pack.jsonprice, market value, enterprise value, the trailing multiples (the vendor's trailing twelve months running through the June 2026 quarter) and their six-year medians, returns, the 252-day high, and FINRA short interest at the 15 September 2026 settlementbacksC29C37
- S15primarycompany filing2026-04-10beneficial ownership as of March 17, 2026, the CEO succession, the 2025 annual bonus metrics, goals and payouts, the long-term award design and the 2022-2025 PSU outcome, and the Bpifrance board nomination rightbacksC47C48C50
- S16primarycompany filing2026-05-28Bpifrance Participations Schedule 13D/A (Amendment No. 6) on Constellium SE, event of May 26, 2026 Bpifrance Participations / Caisse des Dépôtsthe largest holder's block sale of 4,190,000 shares at $33.89, its remaining 6.2% stake and its 60-day lock-upbacksC49
- S17primarycompany release2026-10-28Q3 and nine-month shipments, revenue, Segment Adjusted EBITDA per tonne by segment, free cash flow, net debt, leverage and the raised FY2026 guidancebacksC15C24C53
- S18primarycompany release2026-10-28the Q3 2026 segment profit bridges (volume, price and mix, costs, exchange), net debt and leverage by quarter, and the raised 2026 guidancebacksC24C53
- S19primarytranscript2026-10-28Q3 2026 earnings call transcript (FMP), held 2026-10-28data/cache/transcripts/CSTM-2026Q3.jsonManagement's answers on the fourth-quarter guide, 2027 aerospace pricing, scrap availability, can-sheet imports and the Vision 2028 savings
We also see that automotive benefits from the Oswego outage are tapering off in Q4 today.
backsC16C22C54