Crown Holdings, Inc. (NYSE: CCK) — The Company That Got Rich by Not Building, and Is Building Again

Independent equity research Report date: 2026-07-26 · Price as-of: $117.88 (2026-07-24 close) · Shares out: 108,766,371 (2026-06-30) · Market cap: ~$12.8B · EV: ~$18.7B Coverage: Initiation · Sector: Materials · Containers & Packaging (Metal Cans & Transit Packaging)

This article takes no position and contains no price target. The single, deliberate exception is the Claude's Take block immediately below, which is clearly labeled as the author’s own subjective view. Sections 1–15 that follow it are recommendation-free by design.

⚡ Claude’s Take

Claude’s own independent subjective opinion. General information only — not investment advice, and not a recommendation to buy or sell any security. The analytical body (Sections 1–15) below carries no position and no price target.

Verdict: HOLD here / accumulate on weakness in the $95–$105 zone (~7.5–8.0x EV/EBITDA, ~11.5–12.5x adjusted EPS). Not a short. Conviction: medium.

Tag: “Discipline was the whole thesis. Management just put it on a timer.”

Crown is the best-run business in aluminium beverage cans, and the numbers say so without ambiguity. It earns a 14.0% return on invested capital and 25.8% on tangible capital, against Ball’s 9.1%; it runs a 13.2% operating margin against Ball’s 10.6%; it carries 2.5x net leverage against Ball’s sub-investment-grade credit; and it trades at 8.8x EV/EBITDA versus Ball at ~11.0x. On the substrate’s own terms, this is the higher-quality asset at the lower price, and that gap is not a rounding error — it is 2.2 turns. If you must own a can maker, this is the one.

But the thing that produced the last four years of shareholder return is not what the story says it is. Crown’s free cash flow went from negative $36M in 2022 to $1,117M in 2025. Roughly half of that came from a 51% cut in capital expenditure — $839M down to $413M — not from operations. That was the correct decision, brilliantly executed into an industry that had just over-built for a hard-seltzer boom that never arrived; Crown then bought back 17% of its shares with the proceeds, funded from actual free cash flow, which is precisely what Ball failed to do. Credit where it is due. The problem is that this is a capital-cycle trade being narrated as a quality story, and the cycle has turned: capex is guided back up 33% to ~$550M, with new capacity in Brazil, Greece, Spain and a greenfield entry into Northern India, at a moment when management itself puts industry utilisation at “mid to high nineties” and 110% at seasonal peak. High returns are attracting capital again, including Crown’s own. Meanwhile the crown jewel is already softening: Americas Beverage segment income fell to $475M in H1-2026 from $504M, with the CEO conceding matching last year is “a bridge too far,” because non-metal costs are indexed to PPI and PPI has under-recovered actual inflation. And the headline “adjusted EPS +16%” rests on segment income up 5.3% — about 38% of reported EPS growth is share count, on a buyback the CFO has already guided down to ~$200M in H2 from $517M in H1.

The framing is a quality-improving cyclical in the late-favourable phase of its capital cycle, priced fairly. That is not a criticism — it is a description of a symmetric distribution. The factor evidence supports it rather than the momentum read the tape implies: despite a +17% quarter and a price 4.6% off its all-time high, Crown carries no momentum loading whatsoever across all four nested models, loading instead on Value (+0.32), Dividend Yield (+0.20), low Beta (−0.20) and negative Growth. This is not a crowded trade and it is not a falling knife; it is a value name whose factor identity hasn’t caught up with its price. The five-year record is the sobering counterweight: +4.2% annualised with a −48.3% drawdown and a 0.075 Sharpe. At a 49.9th-percentile own-history P/E and a 7.0% FCF yield (6.2% after the minority dividends that rank ahead of you), you are paid adequately to wait and not remotely paid to chase.

Triggers. Flips bullish: evidence at the 2027 contract renewals that non-metal costs are being repriced on a mechanism that actually tracks Crown’s cost base rather than PPI — that single change would convert a negotiated spread into something closer to pricing power and put Americas segment income back on a growth path. Flips bearish: capex guidance drifting through ~$700M without contracted volume behind it, or Transit Packaging taking its first goodwill impairment — either would confirm that the discipline was a phase, not a policy.

📈 Stock Price Action — Five-Year Event Map

Crown has completed a full round trip and is now marginally through the far side. The stock closed as high as $123.57 (2022-03-25), collapsed to $63.91 (2022-11-02) — a 48% decline in seven months — and has since recovered to $117.88, leaving it −4.6% off its five-year high with a 52-week range of $89.36–$117.99. It trades above its 21-day ($111.63), 50-day ($107.17) and 200-day ($102.43) EMAs, with the 50 above the 200. The five-year annualised return is nonetheless only +4.2%, with a −48.3% maximum drawdown and a Sharpe of 0.075 — the stock has taken half a decade to get back to where it started. (Source: AZI 5-year price CSV and FactorsToday leaderboard, pulled 2026-07-26.)

Cycle narrative. (1) The 2021–early-2022 advance was a genuine demand shock extrapolated too far: on-premise closures and the hard-seltzer craze drove canned-beverage volumes up, and the whole industry — Crown included — committed capital accordingly, taking Crown’s capex to $816M in 2021 and $839M in 2022 against a $391M base in 2019. (2–3) It was substantially a fad. Demand cracked, customers destocked, European energy costs spiked, and Crown spent 2022 generating negative $36M of free cash flow while carrying peak capex — the stock more than halved to $63.91. (4) The November-2022 reversal was the capitulation low; the market began paying for the announced retrenchment rather than the realised results. (5) The February-2024 −19.7% drawdown was a genuine fundamental disappointment on FY2024 guidance, and marks the last time the stock was materially cheap. (6–7) The 2024–25 recovery is the capex cut becoming visible in reported cash flow: OCF minus capex went $660M (2023) → $789M (2024) → $1,117M (2025), and buybacks scaled from $12M to $505M alongside it. (8) The 2026 leg is operational — global volumes +5% in each of Q1 and Q2, a Q2 adjusted EPS beat at $2.49, and a full-year guidance raise from $7.90–$8.30 to $8.30–$8.50 on 2026-07-20. Price moves are Fact; attributed drivers are Interpretation.

1. Executive Summary

Crown Holdings is a $12.4B-revenue global metal packaging converter, founded in 1892, operating 179 plants across 39 countries with ~23,000 employees. 73% of consolidated net sales come from the global beverage can business, and 61% of sales originate outside the United States. It reports four segments — Americas Beverage, European Beverage, Asia Pacific and Transit Packaging — plus an “Other” grouping containing the North American tinplate (food can, aerosol, closures) and beverage can equipment operations.

The business is genuinely good, and better than its closest peer by a wide margin. FY2025 delivered net sales of $12,365M, a 22.03% gross margin and 13.23% operating margin — both the highest in an eleven-year series — with EBITDA of $2,092M. Return on invested capital, computed from the June-2026 balance sheet, is 14.0%; stripping the $4,014M of goodwill and acquired intangibles out of invested capital, the underlying converting franchise earns 25.8% on tangible capital. Against a WACC of roughly 7.5–8%, this is a business creating real economic value. The comparison that frames it: Ball Corporation, the world’s largest can maker and Crown’s most direct competitor, has never cleared ~10.4% ROIC in eight years and ran 9.1% in FY2025, on a sub-investment-grade credit — and trades at ~11.0x EBITDA against Crown’s 8.8x.

The moat is real, local, and bilateral. Cans are cheap, bulky and freight-uneconomic, so plants sit beside customer fillers; this is a textbook Greenwald economies-of-scale-plus-captivity advantage, and it holds — the big three (Ball, Crown, Ardagh) have held ~66% of the global market with minimal share drift. But the customers are themselves oligopolists, and full contractual aluminium pass-through makes Crown’s cost base transparent to the people negotiating against it. The 2026 numbers show the limit precisely: Americas Beverage segment income fell to $475M in H1-2026 from $504M even as segment sales rose 18.5% and North American volumes grew 5%, because — in the CEO’s own words — “PPI [is] not enough to fully capture all the cost increases.” Crown passes metal through completely and carries the basis risk on everything else. That is a negotiated conversion spread, not pricing power.

What actually produced the last four years of returns was capital discipline, not operations. Free cash flow went from negative $36M in 2022 to $1,117M in 2025. Approximately half of that swing is a 51% reduction in capital expenditure, from $839M to $413M — capex fell from 6.5% of sales to 3.3%. The rest is margin recovery as the post-2022 industry retrenchment restored pricing. Crown then converted the cash into a 17% reduction in share count since 2021, funded from genuine free cash flow rather than, as at Ball, from divestiture proceeds and incremental debt. The segment capex disclosure reveals how deliberate the allocation was: Americas Beverage ran at 1.06x depreciation (bare maintenance), Asia Pacific at 0.39x and Transit Packaging at 0.70x (harvest), with essentially all growth capital directed to European Beverage at 3.15x.

The central risk is that the cycle is turning back. FY2026 capex is guided to ~$550M, up 33%, funding capacity additions in Brazil, Greece and Spain plus a greenfield entry into Northern India (~$250M for a two-line plant, with volume commitments covering “70% or a little bit higher”). This is happening when management puts North American utilisation at “mid to high nineties” adjusted and 110% at seasonal peak — i.e., at the top. The additions are contract-anchored, which distinguishes them from the 2021 error, but the direction is unambiguous, and Marathon’s asset-growth observation is that the marginal dollar committed at peak utilisation earns least.

Capital allocation carries one large, unresolved failure. Signode Industrial Group, acquired in 2018 for $3,937M, is now the Transit Packaging segment. It generated FY2025 segment income of $258M — roughly $195M after tax, a ~5.0% return on the price paid, below cost of capital — and it is shrinking: sales down 10.2% and segment income down 22.1% over two years, with H1-2026 down again. It carries $1,494M, or 47%, of group goodwill, entirely unimpaired. It is being managed as a harvest asset. This single transaction is the reason consolidated ROIC sits ~12 points below what the underlying franchise earns.

Valuation is fair, not cheap. At $117.88, market cap $12,821M plus net debt $5,406M plus minorities $499M gives an EV of ~$18,726M — 8.8x TTM adjusted EBITDA of $2,134M, 14.0x the $8.40 midpoint of FY2026 adjusted EPS guidance, and a 7.0% adjusted FCF yield (6.2% after the ~$110M of dividends owed to noncontrolling interests, which rank ahead of Crown shareholders). Against its own decade, the P/E sits at the 49.9th percentile — squarely mid-range. The price embeds roughly 1.3–1.5% of perpetual real FCF growth, comfortably inside management’s own 1–3% volume range. It does not require an inflection; it requires durability — that 2025–26 is a plateau rather than a cycle peak.

Verdict framing for the committee. Three linchpins. (1) The franchise genuinely earns excess returns on tangible capital and is materially better than its nearest listed comparable, at a lower multiple. (2) The shareholder outcome of the last four years is a capital-cycle trade — a capex cut plus a real, self-funded buyback — and both of those engines are now decelerating by management’s own guidance. (3) The largest and most profitable segment is already declining because the non-metal pass-through mechanism does not work, and 2027 laps a World Cup. The bull case is not that Crown re-rates; it is that Crown compounds mid-single-digit EPS while paying you a 6–7% cash yield. That is a defensible proposition at this price and an attractive one meaningfully below it.

2. Business Overview

Crown Holdings, Inc. is a Pennsylvania corporation founded in 1892 — originally as the maker of the crown cork bottle cap — and headquartered at 14025 Riveredge Drive, Tampa, Florida, having relocated from Yardley, Pennsylvania. At December 31, 2025 it operated 179 plants in 39 countries with approximately 23,000 employees, generating consolidated net sales of $12,365M, of which 61% originated outside the United States.

2.1 What the company actually does

Crown converts metal — principally aluminium coil, secondarily tinplate steel — into rigid containers. Approximately 73% of consolidated net sales derive from the global beverage can business: two-piece aluminium cans and the ends that seal them, produced on high-speed lines at rates measured in thousands of cans per minute, sold under multi-year supply agreements to beverage fillers. The remaining ~27% spans North American food cans, aerosol cans and closures; glass bottles and metal crowns in Mexico and Brazil; beverage can-making equipment and tooling; and the Transit Packaging business acquired with Signode in 2018.

The economics are conversion economics. Crown does not take a position in aluminium: metal cost is passed through to customers under contractual provisions, supplemented by derivative hedging. In Q2-2026, of the $519M year-over-year increase in net sales, $395M was pass-through of higher material costs and $32M was favourable currency — meaning genuine activity growth accounted for less than a fifth of the reported revenue increase. This is the single most important structural fact about reading Crown’s financials: revenue is a poor proxy for activity, and revenue-based margins are systematically misleading. Segment income in dollars, and volume in units, are the numbers that carry information.

2.2 Segment structure and economics

The FY2025 segment disclosure (10-K Note Z) is the clearest window into where the money is made:

Americas Beverage is the franchise. At $5,615M of sales and $1,030M of segment income it is 45% of group revenue and 57% of reportable segment income, at an 18.3% margin — the highest of any segment. It spans the US, Brazil, Canada, Colombia and Mexico, and includes the Mexican glass business (two factories, five furnaces), which management describes as performing strongly because “as the economy tightens, especially in the lower income economies, glass does better.”

European Beverage covers Europe, the Middle East and North Africa. It is the recovery story: segment income has risen from $199M (2023) to $276M (2024) to $334M (2025) — up 68% in two years — as the European energy shock unwound and volumes recovered. It is also the only segment receiving material growth capital, at 3.15x depreciation, funding new lines in Greece (first commercialised July 2026) and Spain.

Asia Pacific is small ($1,096M) but structurally interesting: sales have fallen from $1,297M in 2023 while segment income rose from $154M to $183M, taking the margin from 11.9% to 16.7%. This is portfolio pruning — the Myanmar beverage can plant was sold in February 2026 — combined with genuine growth in Vietnam, Cambodia, Thailand and Indonesia. Volumes grew double digits in H1-2026. It receives $17M of capex against $44M of depreciation.

Transit Packaging is the Signode business: strapping, industrial film, protective packaging (airbags, edge protectors, honeycomb), and the automation equipment that applies and removes consumables, sold into metals, construction, agriculture, corrugated and general industry. It is the only segment in secular decline (the capital-allocation section below).

“Other” — the North American tinplate businesses and the beverage can equipment operations — deserves more attention than its label suggests. It contributed the upside surprise in Q2-2026: Donahue attributed “at least 2 thirds of the beat in other” to the can-making equipment business, noting Crown supplies “many can companies around the world, including some of our direct competitors here in The United States.” That is a genuinely differentiated asset — Crown sells the machines its competitors make cans on — but it is lumpy and recognised on a build-and-ship basis, which makes it a poor foundation for extrapolating a quarter.

2.3 Revenue quality and contract structure

Beverage can revenue is recurring in the meaningful sense: multi-year supply agreements with fillers whose plants Crown’s plants are physically co-located with. Switching requires the customer to re-source from a plant further away and absorb the freight penalty. Volumes are contracted; metal is passed through in full; non-metal costs are recovered through index-based mechanisms, predominantly PPI-linked.

Transit Packaging is materially lower quality on this axis, and the 10-K says so explicitly: “the Company’s relationship with several of its customers, particularly in Transit Packaging, is noncontractual, and as a result its customers may unilaterally reduce their purchases of its products.”

Verdict: A well-structured, genuinely recurring core business with real revenue visibility, attached to a lower-quality industrial consumables segment with no contractual protection. Roughly three-quarters of the company is a good business; the rest dilutes it.

3. Industry Dynamics

3.1 Structure

Global aluminium beverage can manufacture is a three-firm oligopoly: Ball, Crown and Ardagh Metal Packaging together account for roughly 66% of the global market. Regionally the concentration is higher still — per Ball Corporation Form 10-K disclosures, that company alone holds ~36% of a ~139-billion-unit North American market, ~39% of a ~97B-unit EMEA market and ~46% of a ~43B-unit South American market. Share has drifted less than two points in decades. On Greenwald’s market-share-stability test — the most reliable single indicator of a genuine barrier to entry — the industry passes cleanly.

The barrier’s mechanism is physical, not legal or technological. A beverage can is a low-value, high-volume, air-filled object. Shipping empty cans any distance destroys the economics, so plants are built beside customer filling lines, often on multi-year anchor contracts. A greenfield entrant must therefore find a filler willing to switch, build adjacent to it, and win volume against an incumbent whose plant is already paid down and whose marginal cost is lower. This is Greenwald’s economies of scale plus customer captivity operating at regional rather than national scope — and it is why, despite a genuinely commoditised physical product, the industry has not fragmented.

3.2 Where the capital cycle sits — the load-bearing analysis

This industry is a near-perfect Marathon case study, and locating it correctly is the most important structural judgment in this memo.

Phase 1 — over-investment (2020–2022). COVID shifted beverage consumption from on-premise to at-home, and the hard-seltzer craze added a genuine but temporary demand layer. The industry declared itself sold out for years and committed capital accordingly. Crown’s capex went $391M (2019) → $554M (2020) → $816M (2021) → $839M (2022) — from 4.1% to 6.5% of sales. Ball did the same.

Phase 2 — the bust (2022–2023). Seltzer collapsed, customers destocked, European energy costs spiked. Crown generated negative $36M of free cash flow in 2022, its stock fell 48% from $123.57 to $63.91, and both majors retrenched — Ball closing plants in St. Paul, Kent WA and Phoenix; Crown cutting capex 51% to $403M by 2024.

Phase 3 — the payoff (2024–2026, now). Capacity stopped growing while volumes recovered. The result is the current tightness. Trade press for 2026 reports canmakers “tight on capacity” with North American supply “expected to be particularly constrained”; Crown reported its highest-ever shipments month in March 2026. Donahue’s own capacity arithmetic on the Q2-2026 call is the best available datum: on rated equipment speed the industry runs “perhaps 92%-93%”, but “if you adjust that for changeovers for sizes, label changes, maintenance, you know, we have got to be in the mid to high nineties, which is… real terms from April to August, it is 110% utilization.”

Crown’s margin recovery — gross margin 17.77% (2022) → 22.03% (2025), operating margin 9.92% → 13.23% — is very substantially the return on this industry-wide discipline. The company is being paid, right now, for capacity that nobody built.

Phase 4 — the turn (beginning). This is the forward-looking point that matters. Crown’s FY2026 capex guidance is ~$550M, a 33% increase, funding: a line expansion at Ponta Grossa, Brazil; two new lines in Greece (the first commercialised July 2026); Spain; and a greenfield entry into Northern India, at roughly $250M for a two-line plant. Donahue was explicit that more may follow: “There may come a time when we need to consider more capacity. And when we believe we can do that in a responsible way we will take a look at that much closer.”

Two things are true simultaneously. The additions are contract-anchored — greenfield projects carry volume commitments for “the large majority of the volume… somewhere like India, you might get commitments for 70% or a little bit higher” — which is a materially more disciplined posture than 2021’s build-and-they-will-come. And the direction is still up, at peak utilisation, across four countries at once, by the industry’s most disciplined operator. When the most disciplined participant starts adding, the constraint that generated the returns is being relaxed.

3.3 Demand: mature, with a favourable substrate mix shift

Long-run beverage can volume growth is low-single-digit. Donahue characterises it as “largely a mature market, obviously, with some modicum of growth”, with a normal range of “zero to 2 or zero to 1 versus 2 to 3”; trade press puts the canmakers’ stated long-term range at 1–3%.

The offsetting positive is genuine substrate substitution into cans. Growth in energy drinks, flavoured alcohols, flavoured teas and sparkling alcohols is pulling volume out of coffee cups, plastic bottles, glass and draft. Donahue: “where we see growth is the continuing growth in energy drinks which largely come in cans and offset many products that are consumed in other substrates… So all these are positive substrate moves for the can industry.” Aluminium’s infinite recyclability — cans are the world’s most recycled beverage package — makes this a regulatory tailwind rather than a headwind, which is unusual in packaging and is the single most durable structural positive in the sector.

A caution on the 2026 numbers: North American market growth of ~3.5% in Q2-2026 is flattered by World Cup demand, worth on Donahue’s estimate “a few hundred million cans” in the quarter. That is a direct and quantified 2027 comparison headwind, and management flagged it unprompted.

3.4 Competitive intensity and regulation

Intensity is moderate and, importantly, bilateral. The three majors do not compete on price destructively — the capital intensity and regional plant economics discourage it — but the customer side is at least as concentrated as the supply side. Ball discloses that AB InBev, Coca-Cola Bottlers’ Sales & Services and Red Bull together represent ~40% of its net sales; Crown does not disclose comparable figures, but its risk factors acknowledge that customer consolidation “has been accompanied by pressure from customers for lower prices.”

Regulatory exposure is modest but real: bisphenol-A, a starting material for internal can coatings, is banned in baby bottles in the EU and Canada, has had its tolerable daily intake recommendation lowered by EFSA, and is under consideration by the US EPA. A forced global coatings reformulation would be a capital and qualification cost, not an existential threat, but it is the one regulatory item with genuine cost potential.

Verdict: a structurally adequate industry, not a structurally good one. The positives — consolidated supply, genuine regional entry barriers, full commodity pass-through, a secular substrate tailwind, and a recyclability story that aligns regulation with the product — are real and durable. The negatives are equally structural: countervailing customer power of the same magnitude as supplier power, a cost base made transparent to the buyer by the pass-through mechanism itself, non-metal recovery mechanisms that demonstrably under-recover, terminal volume growth of 1–3%, and a demonstrated industry tendency to convert every demand spike into an over-build. This industry produces adequate returns for its best operator and mediocre returns for everyone else — which is precisely the pattern the ROIC data shows.

4. Competitive Position

4.1 Naming the moat

In Greenwald’s taxonomy Crown holds a regional economies-of-scale advantage reinforced by customer captivity — advantage type three, the only one he regards as genuinely durable. The mechanism has three components:

1. Freight-driven regional monopoly. Empty cans are uneconomic to ship far. Each plant serves a geographic radius, and within that radius the incumbent’s marginal cost is unbeatable by anyone who must build first.
2. Co-location and integration. Crown’s plants sit beside customer filling lines, in some cases with direct feed. Switching means re-sourcing from further away and absorbing freight.
3. Scale in a capital-intensive, high-speed process. A modern two-line plant costs ~$250M (CFO Clothier). Capacity is lumpy, and sub-scale entrants cannot match line speeds or yields.

The financial proof is required before the claim can stand. It is present: Crown earns 25.8% on tangible invested capital and 14.0% on total invested capital, against a WACC of ~7.5–8%. A business with no advantage in a commoditised, capital-intensive process does not earn 26% on tangible capital for years. The moat is real and it shows up where it must — in the returns.

4.2 Where the moat stops — the 2026 natural experiment

The limitation is that the advantage is bilateral and the surplus is negotiated, and H1-2026 provides an unusually clean natural experiment for exactly how much of the surplus Crown keeps.

Consider Americas Beverage in H1-2026 against H1-2025:

Volumes up 5%, revenue up 18.5%, profit down 5.8%. The revenue increase is mostly aluminium pass-through and therefore economically meaningless, but the profit decline is not. In a period of tight capacity, rising volumes and a market growing 3–3.5%, Crown’s most advantaged segment made less money than the year before.

The reason, from the CEO directly: “North America, we had a number of cost increases this year that we knew we would not fully recover in our pass through models, i.e., PPI not enough to fully capture all the cost increases. As well as the negative mix associated with lower sales in Brazil.”

This is the mechanism laid bare. Aluminium — the largest input — is passed through contractually and completely, so Crown neither gains nor loses on it. Everything else — labour, energy, freight, coatings, maintenance — is recovered through index formulas tied to PPI, and Crown carries the basis risk between a published index and its own actual cost base. When actual inflation runs ahead of the index, the shortfall lands on Crown, and the moat provides no protection because the contract, not the moat, sets the price. Management knew this going in (“that we knew we would not fully recover”) — meaning it was conceded at the negotiating table, which is itself the evidence of where the bargaining power sits.

CFO Clothier declined to quantify the non-metal exposure when asked directly: “hesitant to wanna answer that question only because we do not like to give away too much of our cost model or pricing model.” That is a reasonable commercial position and an unhelpful analytical one; it is recorded as an Open Question.

4.3 Head-to-head versus Ball

The most informative comparison available, drawn from both companies’ primary filings:

On seven of nine dimensions Crown is the better business, and it trades at a 2.2-turn discount. The two dimensions where Ball wins are scale (Ball is #1 globally) and disclosure quality (Ball names its customers and their share; Crown does not).

The scale point deserves a moment, because it cuts against intuition. Ball is larger and earns less. That is strong evidence that in this industry scale beyond regional sufficiency does not create additional rent — once you are the incumbent plant next to the filler, being larger globally does not improve your negotiating position against a customer who is also global. It may worsen it, by making you the obvious counterparty for the largest, most sophisticated, most price-aggressive buyers. Crown’s superior returns appear to come from a better mix — a higher-margin Americas franchise, the equipment business, Mexican glass, North American tinplate — rather than from any scale advantage.

4.4 Switching costs and network effects — pressure-tested

Switching costs: real but bounded. A customer switching supplier faces freight penalties, qualification work and line-integration costs. But these are transaction costs, not lock-in: they are meaningful at the margin of a contract renewal and irrelevant to the price negotiated within it. Crown’s contracts come up for renewal; Donahue: “there is wins and losses every year. We have got some wins. We have got some losses. But in total, we are gonna be flat to up.” That is the language of a competitive re-bid, not of a captive customer base.

Network effects: none. There is no mechanism by which an additional Crown customer makes Crown more valuable to existing customers. Any claim to the contrary should be rejected.

A genuine and underrated differentiator: the equipment business. Crown makes and sells beverage-can-making machinery to its own competitors — “We supply many can companies around the world, including some of our direct competitors here in The United States.” This gives Crown a real-time read on industry capacity additions before they show up in anyone’s shipments, and it monetises competitors’ expansion. It is small and lumpy, but it is a structural informational edge no peer has.

Verdict: a durable but narrow advantage that produces genuine excess returns on tangible capital, materially better executed than at its nearest competitor, and demonstrably insufficient to protect the spread against non-metal inflation. Crown has a moat and keeps some of the rent — more than Ball does. It does not have pricing power in the sense of being able to raise price at will, and H1-2026 proves it in the most direct way available: record volumes, tight capacity, and falling segment profit in the best segment.

5. Growth History and Forward Opportunities

5.1 The historical record — and why revenue is the wrong lens

(a) FY2021 includes a ~$1,510M non-operating charge on the divestiture of the European Tinplate business. Operating margin that year was a normal 11.72%; the loss was a portfolio action, not an operating failure.

Revenue has gone essentially nowhere in four years — $12,943M in 2022 to $12,365M in 2025, a 4.5% decline — while EBITDA rose 20% and operating margin expanded 331 basis points. This is the pass-through effect working in reverse: aluminium prices fell from the 2022 spike, deflating revenue without touching profit. Anyone reading Crown’s growth off the revenue line will reach the wrong conclusion in both directions.

The honest growth measures are volume and segment income. On those: global beverage can volumes grew 5% in Q1-2026 and 5% in Q2-2026; reportable segment income has compounded from $1,560M (2023) to $1,728M (2024) to $1,805M (2025), a 7.6% two-year CAGR.

5.2 Growth by segment — where it is and is not coming from

The composition has shifted decisively. Americas drove 2023–25 and has now stalled; Europe is the current growth engine, up 68% in two years and still growing; Asia is growing profit on falling revenue (mix improvement plus portfolio pruning); Transit is in outright decline.

This matters for forecasting because Europe is a recovery — it is regaining margin lost to the 2022 energy shock (10.3% margin in 2023 → 14.4% in 2025), not breaking new ground. Recovery growth is finite by construction. Once European margins normalise at mid-teens, the group loses its principal growth contributor, and the burden returns to a stalled Americas.

5.3 Organic versus acquired

Growth over the past four years has been entirely organic and, more accurately, entirely self-help. Crown made no material acquisitions: FY2024 and FY2025 both show zero cash paid for acquisitions; FY2023 shows $126M. Donahue on current appetite: “certainly not contemplating any large M and A. Frankly, not contemplating any M and A.”

The margin expansion came from three sources: industry pricing discipline post-2022; genuine operating cost-out and footprint rationalisation (restructuring charges of $22M, $49M and $23M in 2025/24/23, plus asset sales and impairments of $53M/$19M/$72M); and portfolio pruning (European Tinplate 2021, Myanmar 2026).

5.4 Forward opportunities

1. India — the only genuinely new market. A greenfield two-line plant in Northern India at ~$250M, with contracted volume commitments of “70% or a little bit higher.” This is Crown’s entry into one of the world’s fastest-growing beverage markets and is the single most interesting forward item in the file. It is also small: two lines against a 179-plant network, and the site is not yet finally negotiated (“we have not disclosed that yet because we are still negotiating land cost”).

2. European capacity additions. Greece (line one commercialised July 2026, line two late in the year) and Spain. These serve demonstrated demand — European volumes +7% in Q2-2026 with “growth noted across almost all countries” — and Donahue describes Greece as “much needed capacity to our European system.” This is the highest-confidence growth capital in the plan.

3. Brazil (Ponta Grossa). Explicitly not a demand-driven addition: “We need more size capability in the Southeast, and that is the reason for the addition even though the market for us has been soft this year. it is just regional size expansion.” Adding capacity into a market where your own volumes fell high-single-digits is defensible on mix grounds and uncomfortable on capital-cycle grounds.

4. North American food can — quietly improved. Now “well balanced between human and pet food”, with ~40% pet food — a structurally more stable end market. Donahue: “Cats do not know if it is August or January.” Volumes fell 3% in Q2-2026 against a 9% prior-year comp. This is a genuine, underappreciated quality improvement in a business the market largely ignores.

5. Continued share count reduction. Mechanically the largest single contributor to per-share growth (the capital-allocation section below), and management expects “another 900 million to a billion dollars of free cash flow next year as well.”

5.5 The 2027 problem

Three identified headwinds converge in 2027. The World Cup contributed “a few hundred million cans” to Q2-2026 alone and does not repeat. Brazil is forecasting a flat full year after being down high-single-digits in H1, which management is explicitly discounting: “we are being a little cautious on what we think our Brazilian team is going to deliver.” And the Middle East crisis is costing Asia Pacific an incremental “$4 million to $5 million” per quarter that either subsides or gets built into 2027 pricing.

Donahue’s own 2027 framing was notably hedged: North American volumes “probably” up, with the caveat that if growth “return[s] to more historical levels of growth, be it zero to 2 or zero to 1 versus 2 to 3, then as business moves around and we pick and choose which business we want based on profitability, We could be flatter or up.”

Verdict: high-quality but decelerating growth, increasingly dependent on share count rather than operations. The volume growth is real (5% globally, twice in a row, in a 1–3% market), the margin expansion is real and earned, and the pruning has been intelligent. But the Americas engine has stalled, Europe’s contribution is a finite recovery, Transit is shrinking, and 2027 laps a quantified one-off. Underlying operating growth from here looks like low-to-mid single digits, with the buyback supplying the difference — which is exactly what H1-2026 already shows (adjusted net income +8.6%, adjusted EPS +13.9%).

6. Financial Quality

6.1 Margin structure and its trajectory

Crown’s margin expansion since 2022 is the most impressive thing in the financials and needs to be understood mechanically rather than admired.

Gross margin 17.77% → 22.03% and operating margin 9.92% → 13.23% between FY2022 and FY2025. Three drivers, in order of contribution:

1. Aluminium price normalisation flattering the ratio. Revenue fell from $12,943M to $12,365M largely on lower metal pass-through while profit rose — mathematically inflating margin without any operational improvement. This is presentational, not economic.
2. Industry pricing discipline post-2022. Real and economic.
3. Cost-out and footprint rationalisation. Real and economic, evidenced by the restructuring charges taken.

The clean way to see the genuine improvement is EBITDA per unit of activity: EBITDA rose from $1,744M (2022) to $2,092M (2025), +20%, while volumes over that span grew far less. That is a real improvement. But at least part of the margin percentage expansion is arithmetic, and a return of aluminium prices to 2022 levels would compress reported margins without changing the economics.

6.2 The cash flow inflection — the single most important table in this memo

(a) implied from H1 OCF of $659M and the ≥$900M adjusted FCF guide. (g) = company guidance. Capex and OCF 2019–2025 from SEC EDGAR XBRL (us-gaap:PaymentsToAcquirePropertyPlantAndEquipment and NetCashProvidedByUsedInOperatingActivities), 10-K/FY basis.

Free cash flow swung $1,153M between 2022 and 2025. Operating cash flow contributed $727M of that; capital expenditure contributed $426M — 37% of the entire swing came from spending less. Capex fell from 6.5% of sales to 3.3%.

The critical diagnostic question is whether that cut is sustainable efficiency or deferred investment. The segment table answers it with unusual precision:

• Americas Beverage at 1.06x depreciation — this is maintenance-only spend on the segment generating 57% of profit. Sustainable in the near term; not indefinitely, and not while adding size capability in Brazil.
• European Beverage at 3.15x depreciation — all the growth capital, aimed at demonstrated demand.
• Asia Pacific at 0.39x depreciation — spending less than a quarter of what the assets are consuming, in the fastest-growing region, at a 16.7% margin. This is harvesting, and it is the most questionable line in the disclosure.
• Transit Packaging at 0.70x depreciation — harvesting a declining asset. Rational given the returns; an admission nonetheless.

So: the capex cut is real, deliberate, and partly a decision to under-invest in the best-growing and highest-margin non-core segments. Group capex at 1.34x depreciation is not starvation. But the mix tells you management has been funding the buyback partly by declining to reinvest in Asia. The 2026 step-up to $550M is, in that light, less a change of policy than a deferred bill arriving.

6.3 Returns on capital — the decomposition that matters

Computed from the 2026-06-30 balance sheet and TTM P&L:

ROIC.ai’s independently computed FY2025 figure is 12.65%, up from 7.25% in 2020 and 10.61% in 2022 — a different convention, the same trajectory and the same conclusion. Return on common equity was 17.2% in FY2025.

The ~12-point gap between total and tangible ROIC is the acquisition history. The operating franchise earns roughly 26% on the capital actually deployed in plant and working capital; goodwill and intangibles of $4,014M — of which $1,494M sits in Transit Packaging alone — drag the consolidated figure to 14%. Both numbers are true and both must be quoted: quoting only 14% understates the quality of the business Crown runs; quoting only 25.8% understates what shareholders actually paid for.

6.4 Balance sheet — stronger than it was, weaker than the headline

Three observations.

First, deleveraging is genuine. Net leverage at 2.5x is at management’s stated long-term target and improved from Q1. Total debt fell from $6,062M a year ago in gross terms while the company bought back $517M of stock in six months.

Second, tangible common equity is negative at approximately −$1,147M. Goodwill and intangibles ($4,014M) exceed total equity ($3,366M). This is normal for a serial acquirer and is not itself alarming for a business with contracted cash flows — but it means P/B is analytically meaningless for Crown and any percentile rank based on it should be discarded (see the valuation section).

Third, and least discussed: the $927M supplier finance program. Crown runs supplier finance arrangements under which banks pay its suppliers early and Crown pays the bank on original invoice terms of up to 150 days. The confirmed balance was $927M at both 2025-12-31 and 2024-12-31, with $3,304M of additions and $3,324M of settlements during 2025. It sits inside accounts payable and is excluded from the company’s net debt and its 2.5x leverage ratio.

Economically this is a debt-like obligation. Treating it as such takes net leverage from 2.5x to approximately 3.0x ($5,406M + $927M) ÷ $2,134M TTM adjusted EBITDA. Two mitigating points: the balance has been flat for two years, so it is not currently manufacturing cash-flow growth; and Crown discloses it properly, which many users of these programs do not. But it is a $927M call on liquidity that the headline leverage ratio omits, and if the banks withdrew, working capital would absorb it.

6.5 Earnings quality

Cash conversion is good. FY2025 operating cash flow of $1,530M against net income including minorities of $883M is 1.73x. Operating cash flow exceeded net income in every year 2019–2025 except 2022 (0.94x, the working-capital build at peak aluminium). No accrual red flag.

The GAAP-to-adjusted gap is moderate and honestly composed. H1-2026 GAAP diluted EPS $3.78 versus adjusted $4.34 — a $0.56 gap, 15% of GAAP EPS. The components are: intangibles amortization $0.66, restructuring and other $0.04, debt extinguishment $0.02, less tax $0.15 and minorities $0.01. Amortization of acquired intangibles is the overwhelming majority — a genuine non-cash charge, but one that represents real capital spent on acquisitions, and adding it back is the standard convention that flatters every serial acquirer. It is scheduled to decline: $140M (2026), $137M, $137M, $124M, $95M (2030). Restructuring charges have fallen sharply ($4M in H1-2026 versus $45M in H1-2025), which is a genuine sign of a settled footprint.

One-time items to normalise: FY2021’s ~$1,510M European Tinplate divestiture charge (any multi-year EPS or ROE series is meaningless without it); FY2025 restructuring and other of $83M including $53M of asset sales and impairments; and the FY2025 asbestos provision of $11M.

Asbestos. Crown Cork carries a legacy liability from a 1963 acquisition of an insulation business. At 2025-12-31: 59,900 claims outstanding (up from 57,500 at end-2022), 1,300 new claims in 2025, and an accrual of $177M. Cash paid was $19M in 2025, $15M in 2024, $17M in 2023 — roughly 1.7% of free cash flow, immaterial to the annual numbers. There has been no insurance coverage since 1998. The accrual assigns zero value to the 18,000 post-1964 claims and to claims in the 23 states plus Texas and Pennsylvania that have enacted successor-liability caps. That is a defensible legal position that has held for two decades, and the 10-K discloses the dependency explicitly: an adverse constitutional ruling on those statutes “could have a material impact on the Company.” It is a genuine tail risk carried at a fraction of gross exposure, not a running cost.

Verdict: economics do improve with scale, and materially so, but the recent improvement is more capital-cycle and capital-allocation than operating leverage. The franchise converts profit to cash reliably, earns a real excess return on tangible capital, and has repaired its balance sheet. The qualifications are that margin percentages are flattered by falling metal prices, the capex cut that produced the FCF was partly a decision to under-invest in Asia, and headline leverage understates the obligation by roughly half a turn.

7. Capital Allocation

Capital allocation at Crown is a study in contrasts: one very large historical error that has never been written down, and a subsequent four-year record of genuinely disciplined execution.

7.1 The error: Signode / Transit Packaging

In 2018 Crown paid $3,937M net cash for Signode Industrial Group, creating the Transit Packaging segment. Eight years on:

FY2025 segment income of $258M, taxed at the 24.4% adjusted rate, is roughly $195M after tax against a $3,937M purchase price — a ~5.0% return, against a WACC of 7.5–8%. The business has declined in revenue and profit for three consecutive years, its margin has fallen 210bp, H1-2026 is down again ($121M versus $132M), and it carries 47% of the group’s $3,155M of goodwill with zero accumulated impairment (group-wide accumulated impairments total $113M, none in Transit).

This is the clearest capital-allocation failure in the file, and it is a large one — a $3.9B outlay, roughly 30% of today’s market capitalisation, earning below its cost of capital in a business with no identifiable moat and non-contractual customer relationships. It is the single reason consolidated ROIC (14.0%) sits ~12 points below the tangible-capital return of the underlying franchise.

Management’s revealed judgment is visible in the capital allocation, not the commentary: Transit receives 0.70x depreciation. Crown is neither fixing it nor selling it; it is harvesting it. That is probably the correct decision now — the alternative, selling into a “tepid global industrial production environment”, would crystallise the loss at the bottom — but it is a tacit acknowledgement.

The offsetting good portfolio decision: the 2021 divestiture of ~80% of European Tinplate, which cost a ~$1,510M non-operating charge but brought in $2,255M of proceeds and exited a structurally inferior business. Painful optically, correct strategically.

7.2 The buyback — and the test that actually matters

(a) H1 FCF = H1 OCF $659M − H1 capex $203M. Crown’s cash generation is heavily H2-weighted, so the H1 ratio is not annualisable.

Weighted-average diluted share count: 131.3M (2021) → 121.4M (2022) → 119.7M (2023) → 119.4M (2024) → 115.7M (2025), with shares outstanding at 108,766,371 on 2026-06-30. That is a −17% reduction since 2021, and management states repurchases equalled “almost 7% of outstanding Company shares over previous twelve months.”

The critical test is the funding source, and Crown passes it. In FY2025 Crown returned $625M of buyback plus common dividends against $1,117M of free cash flow, while reducing leverage. This is the sharpest single contrast with Ball: per Ball Corporation public filings, Ball returned $1,932M in FY2024 against negative free cash flow, financing its EPS algorithm with BAE divestiture proceeds and incremental debt. Same mechanical effect on EPS; materially different quality of decision. Crown is retiring stock at 8–9x EBITDA on its own business with cash it actually earned.

The honest qualification is how much of reported growth this represents. In H1-2026:

Roughly 5.3 percentage points of the 13.9% — about 38% of reported EPS growth — is share count, and in Q2 specifically the “+16% adjusted EPS” headline rests on segment income up 5.3%. This is not a criticism of the buyback, which is well-funded and value-accretive at these multiples. It is a caution against reading the EPS line as an operating result.

And the pace is decelerating by guidance. H1-2026 repurchases were $517M; Clothier guided “close to… $200 million worth of stock back in the second half” — a 61% step-down. The remaining authorization was $1,296M of a $2,000M program at 2025-12-31.

7.3 Dividend

FY2025 dividends paid were $120M ($1.04/share, an 11.7% payout ratio). H1-2026 dividends of $77M imply a materially higher per-share rate — Donahue: “we took a big step at the beginning of this year to bring the dividend up to a level that more appropriately reflects our confidence in our future cash flow generation capabilities.” The trailing yield is ~1.03%. A “refreshed dividend policy” is flagged for board discussion at year-end 2026.

A payout ratio in the low teens on a mature, low-growth, cash-generative business is conservative to a fault. Donahue’s framing of the problem is unusually candid: “the fortunate problem we have is we have a lot of cash.”

7.4 Incentive design and alignment — good design, soft calibration

From the DEF 14A filed 2026-03-23:

Annual bonus vests on economic profit and modified operating cash flow (MOCF). Long-term equity is ~two-thirds performance-based, vesting on relative total shareholder return and return on invested capital. CEO stock ownership guideline 6x salary (3x for other NEOs); hedging and pledging prohibited; 50% after-tax share retention required for two years post-vest.

The design is close to best-in-class for a capital-intensive converter. Economic profit charges management for the cost of capital in the annual plan; ROIC and relative TSR govern the long-term plan. This is precisely the structure that should prevent another Signode, and its presence is consistent with — and plausibly causal of — the post-2022 capex discipline. Credit it.

The calibration is soft. Corporate-level NEOs including the CEO were paid 100% above target (i.e., 200% of target) on the annual bonus in both 2024 and 2025, and 52.5% above target in 2023. Three consecutive years of large over-target payouts on cash-flow and economic-profit metrics — in a period when the cash flow was substantially produced by cutting capital expenditure, an input the MOCF metric rewards directly — suggests thresholds set conservatively. There is a real design tension here worth naming: a bonus metric built on operating cash flow pays management for cutting capex, and Crown cut capex 51%.

CEO Timothy Donahue’s total compensation was $17,501,363 (2025), up from $13,618,150 (2024) and $11,928,223 (2023) — a 47% increase in two years, comprising salary $1.45M, stock awards $8.63M, non-equity incentive $4.21M, pension value change $3.16M and other $0.06M.

7.5 Insider behaviour

Every Form 4 filed since 2024-01-01 was retrieved and parsed — 171 filings. The non-derivative transaction code histogram: A (grants) 121 · F (tax withholding) 47 · S (open-market sales) 28 · D 15 · G 3.

There were zero code-P open-market purchases. Open-market sales totalled $32.45M, dominated by CEO Donahue (222,500 shares sold against 332,062 granted), followed by Djalma Novaes Jr, EVP/COO (49,488), Gerard Gifford, EVP/CAO (38,595) and director James H. Miller (10,006).

The selling itself is weak evidence — routine diversification for executives whose stock has nearly doubled. The stronger signal is the absence of buying. This 30-month window spans a stock that traded in the $70s in early 2024 and $77 in April 2025; not one officer or director committed personal capital at those levels. That is consistent either with management regarding the shares as fairly valued throughout, or simply with an equity-heavy package that makes purchases redundant. It is not a bullish tell.

Verdict: management has allocated capital well for four years and badly once, expensively. The post-2022 record — capex discipline, self-funded buyback, deleveraging, portfolio pruning, no M&A, and an incentive plan that charges for capital — is genuinely good and materially better than the closest peer. The $3.9B Signode acquisition remains an unimpaired sub-cost-of-capital asset in secular decline, and the capex step-up now underway is the first real test of whether the discipline is a policy or a phase.

8. Changes and Headwinds — Last Two Years

8.1 Timeline of material events

8.2 The headwinds, honestly stated

1. Americas Beverage has stopped growing. H1-2026 segment income $475M versus $504M. The CEO expects the full year to fall short of 2025’s $1,030M: “it might be a bridge too far for this year to equal last year in segment income in the Americas… If we do not get to a billion of segment income, we will get real close to that number.” The segment producing 57% of profit is guided down.

2. The non-metal pass-through does not work. Structural, conceded, and unquantified (the competitive-position section). This is the most important headwind because it is a mechanism, not a cycle.

3. Brazil. Volumes down high-single-digits in H1-2026 on customer mix — Crown services “the other 2 big beer companies there and not the larger 1”, and the largest player promoted aggressively into the World Cup. Management is explicitly discounting its own team’s flat-for-the-year forecast.

4. Transit Packaging. Three consecutive years of declining revenue and profit in a “tepid global industrial production environment.”

5. The Middle East crisis. An incremental “$4 million to $5 million” quarterly cost in Asia Pacific, plus a flagged H2 headwind.

6. The 2027 World Cup comp. A quantified “few hundred million cans” of Q2-2026 volume that does not repeat.

7. Capex normalising upward. $413M → ~$550M guided, with four expansion projects live.

8. Buyback deceleration. $517M in H1 to ~$200M guided in H2.

8.3 What strengthened

The balance sheet (2.5x, at target); the European franchise (segment income +68% in two years); the North American food can business (~40% pet food, structurally more stable); portfolio focus (European Tinplate 2021, Myanmar 2026, no M&A); and the incentive structure (economic profit and ROIC metrics).

Verdict: the changes of the last two years have strengthened the balance sheet and the per-share economics while the operating headwinds have accumulated. Crown today is a financially better company than in 2024 — less levered, more cash-generative, fewer shares — facing a demonstrably harder operating year ahead in its most important segment. Both statements are true, and the tension between them is the investment case.

9. Risk Analysis

The three that matter. Risk 1 is the most important because it is a mechanism, already visible in reported results, and management has conceded it. Risk 2 is the most important forward risk, because it is the thing that ended the last cycle and Crown is now participating in it. Risk 4 is the largest unquantifiable, and it is unquantifiable because Crown chooses not to disclose it.

Risk 8 deserves more attention than it typically receives: a $927M facility, terminable on 30 days’ notice by either party, excluded from a leverage ratio that is the company’s stated principal measure of liquidity.

10. Valuation Discussion

No price target and no recommendation appears in this section. The purpose is to establish what the current price embeds.

10.1 Building the enterprise value correctly

Market cap independently confirmed at $12,821,379,072 by FactorsToday. Third-party aggregator enterprise values struck at period-end closes were not used.

Noncontrolling interests are included deliberately: Crown consolidates majority-owned subsidiaries whose earnings appear in EBITDA but whose economics belong to minority partners. Guided 2026 noncontrolling interest expense is ~$150M and dividends to minorities ~$110M — real cash claims ranking ahead of Crown shareholders.

10.2 Multiples

10.3 Own-history percentiles — and why the composite must be decomposed

AZI’s valuation index (2026-07-24, n_components = 3) returns:

This is the mirror image of the case where all three metrics agree and confirm each other. Here they disagree sharply, and each disagrees for a diagnosable reason:

• P/S at the 83rd percentile is an artifact in both numerator and denominator. Sales per share is depressed because aluminium prices normalised down from 2022 (deflating revenue without touching profit) and because share count fell 17%. Neither movement carries valuation information.
• P/B at the 28.7th percentile is an artifact in the opposite direction, and is the least meaningful of the three. Book equity was rebuilt from the FY2021 loss year, while goodwill and intangibles of $4,014M exceed total equity of $3,366M — tangible common equity is approximately −$1,147M. A price-to-book ratio on a negative-tangible-book converter is not a valuation metric.
• P/E at the 49.9th percentile is the usable signal. GAAP EPS is not badly distorted (the adjusted-to-GAAP gap is 15%, overwhelmingly intangible amortization), and the denominator does not collapse the way it does for a deep cyclical.

Conclusion: fairly valued against its own decade. Not cheap, not rich, dead centre.

10.4 Cross-sectional comparison

Ball figures derived from Ball Corporation FY2025 Form 10-K and public market data as of 2026-07-13.

Crown is the higher-return, higher-margin, better-capitalised business at a 2.2-turn EBITDA discount to its nearest listed competitor. FactorsToday’s factor-similarity screen independently confirms the comp set: ATR (AptarGroup) 0.904, PPG 0.897, BALL 0.896 are the three closest listed analogues.

The necessary caveat: relative cheapness against a peer that on these metrics screens expensive is bounded comfort. It establishes that if one owns exposure to this substrate, Crown is the vehicle. It does not establish that the substrate is worth owning.

10.5 Embedded expectations — what must be true for $117.88

At $117.88 the market capitalisation is $12,821M against guided adjusted free cash flow of at least $900M — or ~$790M after the ~$110M of dividends owed to minority partners, which is the figure that actually accrues to Crown shareholders.

A simple perpetuity test: capitalising ~$900M of sustainable free cash flow at an 8% cost of equity with zero real growth supports ~$11,250M of equity value. The $12,821M market capitalisation therefore embeds roughly 1.3–1.5% of perpetual real free cash flow growth.

That is a modest hurdle. It sits comfortably inside management’s own long-term volume range of 1–3%, and below the mid-point. The price is not demanding an inflection, a re-rating, or a margin breakthrough.

What the price does assume is durability. Specifically, it assumes that the 2025–26 level of margin and cash generation is a plateau rather than a peak — that a 13.2% operating margin achieved at mid-to-high-90s industry utilisation, with capex at 3.3% of sales and a 51% cut still working through the asset base, is a sustainable steady state rather than the top of a capital cycle. That is the single assumption the entire valuation rests on, and the capital-cycle discussion explains why it is genuinely contestable.

10.6 Scenarios

Expectation bands, not price targets.

Bear. Non-metal under-recovery persists and Americas segment income falls below $950M; Transit continues to erode; 2027 laps the World Cup and Brazil disappoints; capex normalises to ~$600M and the buyback settles at ~$400M/year. Adjusted EPS flat to modestly down from $8.40 through 2028; FCF ~$800M. A business growing nothing, on a de-rated 11–12x, is worth materially less than today.

Base. Global volumes +2–3%; Europe and Asia continue to offset a flat-to-slightly-down Americas; buyback ~$700M/year; capex settles $550–600M. Adjusted EPS compounds mid-to-high single digits to roughly $9.00–$9.50 by 2028, with about 60% of the growth from share count. FCF holds $900M–$1,000M — consistent with Donahue’s own expectation of “another 900 million to a billion dollars of free cash flow next year as well.” A 13–15x multiple on that band brackets today’s price. Today’s price is approximately the base case.

Bull. India and the European additions land on contract; Transit stabilises; the 2027 renewals reprice non-metal recovery onto a mechanism that tracks Crown’s actual costs; Americas returns to growth. Adjusted EPS above $10 by 2028, FCF above $1.1B, and a modest re-rate toward 10x EBITDA on a higher EBITDA base.

The distribution is unusually symmetric for a cyclical trading near its high. The downside is cushioned by contracted volumes, full metal pass-through, 2.5x leverage and a real buyback; the upside is capped by a mature market, a transparent cost base and a multiple that has historically not exceeded the low teens on EBITDA. This is the signature of a fairly-priced quality cyclical rather than a mispricing in either direction.

11. Variant Perception

11.1 Consensus

The consensus view of Crown in mid-2026 is: a well-run can maker executing a successful margin and cash-flow turnaround, raising guidance, buying back stock aggressively, cheaper than Ball, with a strong balance sheet and a favourable secular substrate story. Every element of that is defensible on the evidence, and it is why the stock is up 84% from its 2022 low and sits 4.6% from an all-time high.

11.2 The strongest bull case

Crown is the highest-quality asset in a consolidated oligopoly, earning 25.8% on tangible capital, trading at 8.8x EBITDA and a 7% free cash flow yield, with a management team that has demonstrated genuine capital discipline (capex cut 51%, no M&A, 17% of shares retired from operating cash flow) and an incentive plan built on economic profit and ROIC. Volumes are growing 5% — twice the market rate — in a market with a real secular substrate tailwind from energy drinks and flavoured alcohols. The industry is capacity-constrained, and Crown is adding contract-anchored capacity into demonstrated demand, including entry into India. Meanwhile the direct competitor trades 2.2 turns higher on half the return. At 14x guided earnings you are paying a market-average multiple for an above-average business with a below-average beta of 0.59.

11.3 The strongest bear case

Everything that produced the re-rating is reversing simultaneously. The free cash flow that funded the buyback came 37% from cutting capital expenditure, and capex is now guided up 33%. The buyback that supplied ~38% of EPS growth is guided down 61% in H2. The Americas segment that produces 57% of profit is already declining, and the reason is structural — non-metal costs indexed to a PPI formula that under-recovers, which no volume growth can fix. Transit Packaging, a $3.9B acquisition, earns 5% and has shrunk for three straight years with $1,494M of unimpaired goodwill on it. 2027 laps a World Cup that management itself sized at hundreds of millions of cans. And the industry is doing what this industry always does at peak utilisation — building. The last time it did that, the stock fell 48%. You are paying a full own-history multiple, 4.6% from an all-time high, for the top of a capital cycle, with insiders having bought precisely zero shares in thirty months.

11.4 The 3–5 assumptions that actually matter

1. Is the 2025–26 margin level a plateau or a peak? Everything rests here. Plateau → the base case works and the stock is fairly priced. Peak → the multiple and the earnings compress together.
2. Can non-metal cost recovery be repriced at the 2027 renewals? This is the difference between a negotiated spread and genuine pricing power, and it is the single highest-value unknown.
3. Does the capex step-up stop at ~$550M? Contract-anchored expansion at $550M is discipline. Drift toward $700M+ without commitments is the 2021 error repeating.
4. Does Americas stabilise in 2027, or is 2026 the start of a trend? 57% of profit turns on this.
5. Is Transit Packaging worth its carrying value? $1,494M of goodwill on a declining, sub-cost-of-capital asset. Non-cash if it goes, but it would confirm the capital-allocation read.

11.5 Where consensus may be offsides — the factor evidence

The positioning data is the most interesting input here, and it says something genuinely counter-intuitive.

Despite a +17% actual quarter (FactorsToday reports m3 at +87.2% annualised; de-annualised, +17.0%), a price 4.6% off its all-time high, and trading above its 21-, 50- and 200-day EMAs, Crown carries no momentum loading at all. Momentum does not appear as an active factor in any of the four nested ElasticNet models — and because the model is L1-sparse, absence means the coefficient was driven to zero, not that it is missing.

What Crown does load on (reading within the Base + Sector + Industry model, R² 0.355): Market +0.724, Materials +0.360, Value +0.321, DividendYield +0.204, OilPrice −0.197, BetaFactor −0.197, Growth −0.108, Quality +0.097. Beta is 0.593. Idiosyncratic volatility is 20.7% annualised on a model R² of 0.346 — meaning roughly two-thirds of Crown’s return variance is company-specific.

The interpretation: the market still classifies Crown as a defensive, low-beta, value/dividend materials name, and its factor identity has not caught up with its tape. Three consequences follow.

• This is not a crowded momentum trade. There is no momentum cohort positioned in it to unwind. The main mechanical de-rating risk that afflicts stocks 4.6% from their highs does not apply here.
• It is also not a falling knife. Value/low-beta names near their highs with positive earnings revisions are not distressed.
• Therefore any de-rating must come from company-specific news — which, given 20.7% idiosyncratic volatility and two-thirds stock-specific variance, is exactly where the risk sits. The events that move Crown are Crown’s own: the Americas segment print, the capex guide, a Transit impairment.

The sobering counterweight consensus underweights: Crown’s five-year annualised return is +4.2%, with a −48.3% maximum drawdown and a Sharpe of 0.075. The 10-K’s own comparative table has $100 invested at end-2020 growing to $108 by end-2025, against $196 for the S&P 500 and $100 for the Dow Jones US Containers & Packaging index. The trailing-year numbers are strong because they measure the recovery leg of a round trip that took four and a half years to complete. The variant perception is not that Crown is a bad business — it plainly is not — but that a stock which has spent half a decade going nowhere, and whose factor identity is still “cheap defensive materials,” is being valued as though the last two years’ capital-cycle windfall is the new run-rate.

12. Fact vs. Interpretation

13. Open Questions

1. What is Crown’s customer concentration? Crown discloses no named-customer percentages, unlike Ball (AB InBev 15%, CCBSS 14%, Red Bull 11%). Without it, the single largest risk to the thesis cannot be sized. Priority: highest.
2. What proportion of the cost base is recovered via PPI-indexed formulas, and what was the 2026 shortfall in dollars? CFO Clothier expressly declined to answer: “we do not like to give away too much of our cost model or pricing model.” This is the mechanism driving the Americas decline.
3. When do the major North American contracts come up for renewal, and is the non-metal mechanism on the table? The bull case turns on this.
4. What are the numeric ROIC vesting targets in the performance-share plan? Metric identity is confirmed; calibration is not — and given three consecutive years of large over-target annual bonus payouts, calibration is the open issue.
5. What is Transit Packaging’s goodwill headroom? $1,494M carried against three years of declining revenue and profit. Headroom percentages are not disclosed.
6. Why is Asia Pacific being funded at 0.39x depreciation while delivering a 16.7% margin and double-digit volume growth? This is the least-explained allocation in the file.
7. What is total 2026–2028 committed growth capex across Brazil, Greece, Spain and India, and how much of the India volume beyond the ~70% commitment is speculative?
8. What is segment-level invested capital? Crown states total assets by segment are “not provided to the CODM” (Note Z), so segment ROIC cannot be computed. Capex/depreciation is the only available proxy.
9. What are the terms and counterparties of the $927M supplier finance program, and how concentrated is it among banks?
10. What is the CEO succession plan? Donahue holds Chairman, President, CEO and CODM roles simultaneously.

14. What Must Be True

14.1 For the bull case

14.2 For the bear case

The single most informative future datapoint for either side is the FY2027 outcome for Americas Beverage segment income together with the FY2027 capex guide. Those two numbers, taken together, resolve B1/B2/B3 and R1/R2 simultaneously — whether the conversion spread can be defended, and whether the discipline that produced the re-rating survives contact with a tight market. They will be known with the Q4-2026 print in February 2027.

15. Source Appendix

Full appendix accompanies this memo as Appendix B in the combined report.

Primary — SEC filings (550-filing census covering 2021-07-01 to 2026-07-24)

• Form 10-K FY2025, filed 2026-02-27 — Items 1, 1A, 3, 5, 7; Notes F, G, H, L, P, Q, U, Z.
• Forms 10-K FY2021–FY2024, filed 2022-02-28, 2023-02-27, 2024-02-27, 2025-03-03.
• Form 8-K filed 2026-07-20, Exhibit 99 — Q2 2026 earnings release (segment tables, condensed balance sheet and cash flow, adjusted EBITDA and net leverage reconciliation, FY2026 guidance).
• DEF 14A filed 2026-03-23 — Compensation Discussion & Analysis, Summary Compensation Table, ownership guidelines.
• 171 Forms 4 and 4/A filed 2024-01-01 to 2026-07-24, primary XML, parsed in full.
• SEC EDGAR XBRL companyconcept API, CIK 0001219601 — PaymentsToAcquirePropertyPlantAndEquipment, NetCashProvidedByUsedInOperatingActivities, RevenueFromContractWithCustomerExcludingAssessedTax.

Primary — management commentary

• Crown Holdings Q2 2026 earnings conference call, 2026-07-21. Speakers: Timothy J. Donahue (Chairman, President & CEO), Kevin C. Clothier (SVP & CFO).

Third-party quantitative (aggregated data — cross-checks, not primary; reconciled to filings)

• ROIC.ai MCP: company profile, income statement, balance sheet, cash flow, profitability ratios, earnings call transcript. Pulled 2026-07-26.
• AZI price history CSV (azitrading.com/controls/download-data.php?t=CCK), 11,684 rows, pulled 2026-07-26.
• AZI valuation_index own-history percentile ranks, 2026-07-24 basis.
• FactorsToday: /stock-loadings/CCK, /leaderboard/CCK, /stock-info/CCK, /stock-specific-vol/CCK, /related-stocks/CCK. Pulled 2026-07-26.

Industry and trade press

• Packaging Dive / Food Dive, “Canmakers say capacity is tight as beverage industry gears up for big summer,” 2026.
• Vinetur, “Can Makers Brace for a Tight Summer,” 2026-05-07.

Peer company filings

• Ball Corporation (NYSE: BALL) FY2025 Form 10-K and public market data as of 2026-07-13 — source of the peer ROIC, margin, leverage, EV/EBITDA and regional share figures used in the peer-comparison and valuation sections.

Analytical frameworks

• Greenwald & Kahn, Competition Demystified — barriers to entry taxonomy, market-share-stability test (the industry and competitive-position sections).
• Marathon Asset Management / Chancellor, Capital Returns — capital-cycle and asset-growth analysis (the industry, growth and capital-allocation sections).

Data not available / gaps recorded

• ROIC.ai get_company_news returned an empty array for CCK; the the recent-events section event timeline was built from the 61-filing 8-K census, earnings releases, transcript and trade press.
• ROIC.ai cash-flow object exposes no discrete capex field for CCK; all capex figures come from EDGAR XBRL and the 10-K.
• FactorsToday /factor-returns/historic returned an empty array; the market-wide factor-regime overlay is unavailable.
• Crown discloses neither named-customer concentration nor segment total assets.

All figures reconciled to primary filings except where explicitly labelled as third-party aggregated data. Price as-of 2026-07-24 close. This article is general information, not investment advice.

APPENDIX A — Standard Diligence Questionnaire

Crown Holdings, Inc. (NYSE: CCK) · Report date 2026-07-26 · Price as-of $117.88

A supplemental diligence questionnaire accompanying the analysis above. Answers are labelled Fact / Interpretation / Assumption where the distinction matters.

General

What thoughtful questions have other investors asked about this company?

The Q2-2026 call (2026-07-21) is a good proxy for the live sell-side debate, and the questions cluster into four themes:

1. Can Americas earnings stop declining? George Staphos (Bank of America) opened with exactly this: guidance had been for Americas earnings down, the quarter came in “relatively flat which was better than expected” — could the full year be flat? Donahue’s answer was no: “it might be a bridge too far for this year to equal last year in segment income in the Americas.” This is the central question, and management answered it in the negative.

2. Is capacity being added responsibly? Philip Ng (Jefferies) pressed on North American slack capacity, drawing out Donahue’s utilisation arithmetic (92–93% on rated speed, “mid to high nineties” adjusted, 110% April-to-August). Jeffrey Zekauskas (JPMorgan) drilled into the India greenfield economics — cost (~$250M), contract structure, and volume commitment level (~70%). Investors are, correctly, watching the capital cycle.

3. Why doesn’t volume growth flow through to earnings? Edlain Rodriguez (Mizuho) asked the sharpest question of the call: “5% volume growth, double digit volume growth, but the earnings growth is that much lower than that?” This produced the PPI under-recovery admission — the most analytically valuable disclosure of the quarter.

4. What happens to the cash? Christopher Parkinson (Wolfe) and Katan Mamtora’s colleague (BMO) both pushed on capital allocation and the dividend. Donahue: “the fortunate problem we have is we have a lot of cash”, with a refreshed dividend policy flagged for year-end board discussion and “not contemplating any M and A.”

A fifth question that is not being asked, and should be: what is Transit Packaging worth, and when does the $1,494M of goodwill get tested? No analyst raised Signode on the call. (Interpretation.)

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

Closer to a high than a low. (Interpretation, well-supported.) FY2025 gross margin (22.03%) and operating margin (13.23%) are the highest in an eleven-year series. Industry utilisation is at “mid to high nineties” adjusted and 110% at seasonal peak on management’s own estimate. Capex sits at 3.3% of sales against a 4.1–6.6% historical range. Free cash flow of $1,117M is a record. Every one of those is a top-of-cycle reading.

The qualification is that this is a shallow cycle for the beverage portion. Beverage can volumes are defensive — people drink in recessions — and 73% of sales sit there. The genuinely cyclical exposure is Transit Packaging (industrial production) at ~16% of sales, and that segment is already at a low, not a high.

Driven by the external environment or internal actions?

Both, roughly evenly, and separating them is the key analytical task. (Interpretation.)

• External: the industry-wide post-2022 capacity discipline that restored pricing. Crown did not cause this; it benefited from it.
• External, presentational: aluminium price normalisation, which inflated margin percentages without economic content.
• Internal, real: a 51% capex cut, footprint rationalisation (restructuring charges of $22M/$49M/$23M in 2025/24/23), portfolio pruning (European Tinplate 2021, Myanmar 2026), and a 17% share count reduction funded from operating cash flow.

How stable are revenues?

Revenue is stable in units and misleading in dollars. (Fact.) Aluminium pass-through means revenue moves with metal prices independent of activity: of the $519M year-over-year Q2-2026 revenue increase, $395M was material-cost pass-through and $32M was FX. Net sales fell 4.5% from 2022 to 2025 while EBITDA rose 20%. Volume and segment income are the stable, informative series.

Outlook for products/services?

Favourable at the substrate level. Growth in energy drinks, flavoured alcohols, flavoured teas and sparkling alcohols is pulling volume into cans from coffee cups, plastic bottles, glass and draft. Aluminium’s infinite recyclability aligns regulation with the product — unusual in packaging. (Fact, from the transcript and 10-K.)

How big will this market be — growing, shrinking, domestic or international?

Global and mature. North America is roughly 139 billion units, EMEA ~97B, South America ~43B (per Ball Corporation Form 10-K disclosures). Long-run volume growth is 1–3%; Donahue’s own range is “zero to 2 or zero to 1 versus 2 to 3.” For Crown specifically, 61% of sales are non-US, and the growth mix is shifting toward Europe (segment income +68% in two years), Asia (double-digit volumes) and now India.

2026 is running above trend — ~3.5% North American market growth — flattered by World Cup demand worth “a few hundred million cans” in Q2 alone. That reverses in 2027. (Fact, management-stated.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive?

Roughly stable, with a modest tilt toward more competitive at the margin. (Interpretation.) The three-firm structure (Ball, Crown, Ardagh at ~66% globally) is not changing and share drift is minimal. But capacity is being added — by Crown in four countries, plus new West Coast capacity Donahue referenced from smaller players, plus “creep” as sub-scale operators debottleneck. Donahue: “there is some new capacity coming on the West Coast… there is been other capacity that is come on, and some of these smaller companies that have brought capacity on will get better, and they will they will create more capacity from their own creep.”

How profitable is the business (ROIC, ROE)?

(Interpretation, computed from primary filings.)

• ROIC 14.0% — NOPAT $1,229M (TTM EBIT $1,626M × (1 − 24.4%)) over invested capital $8,772M (equity $3,366M + debt $6,062M − cash $656M), at 2026-06-30.
• ROIC on tangible capital 25.8% — removing $4,014M of goodwill and intangibles.
• ROE 17.2% (FY2025), ROIC 12.65% on ROIC.ai’s convention, up from 7.25% in 2020.

Against a WACC of ~7.5–8%, Crown creates real economic value. The 12-point gap between total and tangible ROIC is the acquisition history — overwhelmingly Signode.

How profitable is the industry — how many competitors, what barriers to entry?

Three global majors plus regional players. Industry profitability is mediocre and Crown is the exception, not the rule — Ball, the world’s largest, has never cleared ~10.4% ROIC in eight years and ran 9.1% in FY2025. (Fact, from Ball Corporation Form 10-K disclosures.)

Barriers are real and physical: empty cans are freight-uneconomic, so plants co-locate with customer fillers, giving the incumbent a regional cost advantage no entrant can match without first winning the contract. A two-line plant costs ~$250M. This is Greenwald’s economies-of-scale-plus-captivity, and the industry passes his market-share-stability test cleanly.

Can the business be easily understood?

Yes — unusually so. Crown converts metal into cans on contracts, passes metal through, and earns a conversion spread. The complications are three and all are knowable: the pass-through mechanics (which make revenue misleading), the segment structure, and the noncontrolling interests. There is no technological obsolescence risk and no product cycle.

Can it be undermined by foreign low-cost labour?

No. This is one of the genuinely attractive features. Freight economics make cans a local product — you cannot ship empty cans from a low-wage country economically. Labour is a modest share of conversion cost, and the process is highly automated. (Interpretation, well-supported by the industry’s physical economics.)

Do brands matter?

No, at the Crown level; enormously, at the customer level. Consumers do not choose a can by its maker. Crown’s customers’ brands are what create the demand, and that is precisely the source of the countervailing power problem — Crown’s revenue depends on brands it does not own, sold by counterparties large enough to negotiate the surplus away.

What is the nature of competition?

Competition for contracts, not for daily price. Contracts are multi-year with volume commitments; the competition occurs at renewal. Donahue: “there is wins and losses every year. We have got some wins. We have got some losses. But in total, we are gonna be flat to up.” Within a contract, the competition is against the index formula, not against a rival.

Customers’ switching costs?

Real but bounded. Switching means re-sourcing from a plant further away, absorbing freight, and requalifying. That is meaningful at renewal and irrelevant within a contract. It is a transaction cost, not lock-in. (Interpretation.) The 2026 evidence that switching costs do not confer pricing power: Crown grew North American volume 5% against a 3–3.5% market while its Americas segment income fell 5.8%.

Financial Condition & Balance Sheet

Assets not fully recognised on the balance sheet?

Two of substance. (Interpretation.)

1. The beverage can equipment and tooling business. Crown designs and sells can-making machinery to competitors worldwide — “including some of our direct competitors here in The United States.” It sits inside “Other,” carries no separate valuation, and provides a real informational edge on industry capacity additions before they appear in shipments.
2. Fully-depreciated productive plant. Gross PP&E is $9,148M against net $5,380M — $3,961M of accumulated depreciation. In a business where a paid-down plant beside a customer filler is the moat itself, book value materially understates competitive value.

Off-balance-sheet liabilities?

Yes, and one is significant. (Fact.)

• Supplier finance program obligations of $927M (2025-12-31 and 2024-12-31), sitting inside accounts payable with terms up to 150 days, terminable by either party on 30 days’ notice. Excluded from the company’s net debt and its 2.5x leverage ratio. Treating it as debt takes leverage to ~3.0x.
• Asbestos — $177M accrued against 59,900 outstanding claims, with zero value assigned to 18,000 post-1964 claims and to claims in states with successor-liability caps. Gross exposure is unquantified by design.
• Operating leases and pension obligations ($289–314M pension liability) are conventional and disclosed.

How conservative is the accounting?

Mixed, leaning conservative on operations and aggressive on one legacy item. (Interpretation.)

Conservative: operating cash flow exceeded net income in every year 2019–2025 except 2022 (1.73x in FY2025); no revenue-recognition complexity; restructuring charges taken rather than deferred; clean five-year filing record with no NT filings, no Item 4.02 non-reliance, no material weakness, no restatement; the supplier finance program is properly disclosed even though it need not be flattering.

Less conservative: Transit Packaging’s $1,494M of goodwill has never been impaired despite three consecutive years of declining revenue and segment income and a 210bp margin decline. Group accumulated impairments total $113M, none in Transit. That is the one line where the accounting is doing work the operating results do not support.

The GAAP-to-adjusted gap is moderate and honest: H1-2026 $3.78 GAAP versus $4.34 adjusted, 15% of GAAP EPS, overwhelmingly intangible amortization ($0.66 of the $0.56 net gap), which is scheduled to decline from $148M to $95M by 2030.

How CapEx-hungry is the business?

Historically very; currently, deliberately, not. (Fact.) Capex ran 4.1–6.6% of sales from 2019 to 2023, peaking at $839M (6.5%) in 2022. It was cut to $403M (3.4%) in 2024 and $413M (3.3%) in 2025, and is guided back to ~$550M (~4.3%) in 2026.

Maintenance capex is roughly depreciation, ~$308M, or 2.5% of sales. The segment split shows where the discipline actually fell: Americas Beverage 1.06x depreciation, European Beverage 3.15x, Asia Pacific 0.39x, Transit 0.70x. The group is not starving itself, but Asia and Transit have been harvested to fund the buyback.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy?

(Fact.) FY2025 free cash flow (OCF $1,530M − capex $413M) was $1,117M. FY2026 guidance is ≥$900M after ~$550M capex. Management expects “another 900 million to a billion dollars of free cash flow next year as well.”

FY2025 uses: $505M buyback, $120M common dividends, ~$137M dividends to minority partners, balance to debt reduction. Returns were funded from free cash flow with room to spare, while leverage fell — the single most important quality distinction versus Ball, which returned $1,932M in FY2024 against negative FCF.

The stated philosophy, from Clothier: “a disciplined, balanced capital allocation framework” — invest in customer-anchored growth, maintain 2.5x leverage, return the rest opportunistically. Donahue’s framing of priorities: “the first thing to do is to service our customers and be present when our customers need us. And from there, all manners of success should follow.”

Significant acquisitions recently?

No — and that is a deliberate, stated position. Zero cash paid for acquisitions in FY2024 and FY2025; $126M in FY2023. Donahue: “certainly not contemplating any large M and A. Frankly, not contemplating any M and A.”

The relevant acquisition is historical: Signode Industrial Group, $3,937M in 2018, now Transit Packaging. FY2025 segment income $258M ≈ $195M after tax = ~5.0% on the price paid, below cost of capital, with revenue down 10.2% and income down 22.1% over two years. It carries 47% of group goodwill, unimpaired, and is being run at 0.70x depreciation — harvested rather than fixed or sold.

Buying back shares?

Aggressively, and well. (Fact.) $12M (2023) → $217M (2024) → $505M (2025) → $517M in H1-2026 alone ($305M in Q2). Weighted-average diluted shares fell from 131.3M (2021) to 115.7M (2025); shares outstanding were 108,766,371 at 2026-06-30 — −17% since 2021, with “almost 7%” retired over the trailing twelve months.

Authorization: $2,000M (July 2024, through 2027), $1,296M remaining at 2025-12-31. The pace is guided down sharply — Clothier expects “close to… $200 million” in H2-2026 versus $517M in H1.

Two caveats. Repurchases at 8–9x EBITDA on a 14% ROIC business are value-accretive — this is a good use of cash. But ~38% of H1-2026’s reported 13.9% adjusted EPS growth is share count (adjusted net income grew 8.6%), so the EPS line materially overstates operating progress.

Issuing large amounts of new shares to insiders?

No. Stock-based compensation was $48M in FY2025 on $12,365M of revenue (0.4%) and against $505M of repurchases — buybacks exceeded SBC more than tenfold. Share count is falling decisively. The Form 4 corpus shows 121 grant events (code A) since 2024-01-01 totalling $3.26M in value — modest.

Compensation policy of directors/management?

(Fact, DEF 14A filed 2026-03-23.)

• Annual bonus: economic profit and modified operating cash flow (MOCF).
• Long-term equity: ~two-thirds performance-based, vesting on relative TSR and ROIC.
• Ownership guidelines: CEO 6x salary, other NEOs 3x; all in compliance.
• Hedging and pledging prohibited; 50% after-tax share retention for two years post-vest.
• Clawback policies in place, including the NYSE-mandated recovery policy effective 2023-10-02.
• CEO total compensation: $17,501,363 (2025), $13,618,150 (2024), $11,928,223 (2023).

Design versus calibration are separate verdicts. (Interpretation.) The design is close to best-in-class for a capital-intensive converter — economic profit and ROIC both charge for the cost of capital, which is exactly the structure that should prevent another Signode. The calibration is soft: corporate NEOs including the CEO were paid 100% above target (200% of target) in both 2024 and 2025, and 52.5% above target in 2023.

There is also a design tension worth naming: an MOCF-based bonus metric pays management for cutting capital expenditure, and Crown cut capex 51%. The outcome was correct, but the incentive was pointing the same way regardless of whether it was.

Motivations of management?

Equity-weighted and reasonably aligned on paper: ~two-thirds of LTI performance-based on ROIC and relative TSR, 6x CEO ownership requirement, no hedging or pledging. (Fact.)

The behavioural evidence is less encouraging. Across 171 Form 4 filings since 2024-01-01 there were zero code-P open-market purchases, against $32.45M of open-market sales led by CEO Donahue (222,500 shares). The window spans the stock in the $70s in early 2024 and $77 in April 2025. Not one insider bought at those levels. (Fact; the inference that this is a mild negative is Interpretation.)

Governance note: Donahue holds Chairman, President, CEO and CODM roles simultaneously — a concentration of authority with no disclosed succession plan reviewed here.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer?

No. Crown Holdings is a Pennsylvania corporation, NYSE-listed common stock (CUSIP 228368106, ISIN US2283681060), issuing a standard Form 1099-DIV. No K-1, no ADR structure, no MLP complexity. It also has secondary listings (GETTEX:CWN, LSE:0I4X) which are not the primary trading venue.

Dividend policy?

(Fact.) Crown paid no dividend at all until 2021. FY2025 dividends paid $120M ($1.04/share) — an 11.7% payout ratio, trailing yield ~1.03%. H1-2026 dividends of $77M imply a materially higher per-share rate following what Donahue called “a big step at the beginning of this year to bring the dividend up to a level that more appropriately reflects our confidence in our future cash flow generation.”

A “refreshed dividend policy” is flagged for board discussion at year-end 2026. A low-teens payout on a mature, cash-generative business is conservative; the company’s revealed preference has been buybacks over dividends by roughly 4:1.

How profitable is the business?

Covered above: ROIC 14.0% (25.8% tangible), ROE 17.2%, operating margin 13.23%, EBITDA margin 16.92%, net margin 5.97% (FY2025). Margins are at eleven-year highs.

Is net income diverging from cash from operations?

No — the divergence runs the right way. (Fact.) FY2025 operating cash flow of $1,530M against net income including minorities of $883M is 1.73x. Operating cash flow exceeded net income in every year 2019–2025 except FY2022 (0.94x), which is fully explained by a $561M working-capital build at peak aluminium prices. There is no accrual-quality red flag.

Risks & Downside

What factors would cause the stock to decline?

In rough order of probability × impact (Interpretation):

1. A second consecutive year of Americas Beverage decline — 57% of segment income, already guided down for 2026.
2. Capex guidance drifting materially above ~$600M — confirming the capital cycle has turned and the discipline was a phase.
3. A 2027 miss on the combined World Cup, Brazil and buyback-deceleration comps.
4. Evidence that non-metal cost recovery cannot be repriced at the 2027 renewals.
5. A Transit Packaging goodwill impairment — non-cash, but it would confirm the capital misallocation and force a reassessment of management’s acquisition judgment.
6. An aluminium spike re-running the FY2022 working-capital drain.

Note the positioning nuance: because Crown carries no momentum factor loading despite trading 4.6% off its all-time high, there is no crowded momentum cohort to unwind. With ~two-thirds of return variance stock-specific and 20.7% idiosyncratic volatility, a de-rating would have to be caused by Crown’s own news, not by a factor rotation.

Risk of a catastrophic loss?

Low, but not zero, and it is identifiable. (Interpretation.) The scenario is not operational — contracted volumes, full metal pass-through and a defensive end market make a demand collapse implausible. Two tail risks exist:

1. Asbestos. 59,900 outstanding claims against a $177M accrual that assigns zero value to 18,000 post-1964 claims and to claims in ~25 states with successor-liability caps. The 10-K states plainly that an adverse ruling on the constitutionality of those statutes “could have a material impact on the Company.” Crown Cork has had no insurance coverage since 1998. Current cash cost is only $19M/year, so this is a legal-outcome tail, not a running cost.
2. Supplier finance withdrawal. A $927M facility terminable on 30 days’ notice, excluded from the stated leverage ratio. Withdrawal would be an immediate working-capital call of roughly 10 months of free cash flow.

Chance of a total loss?

Very low. Net leverage is 2.5x (3.0x including supplier finance) on an investment-grade credit; the business generates ~$900M–$1.1B of annual free cash flow; volumes are contracted; the largest input cost is contractually passed through; and the assets are irreplaceable regional plants beside customer fillers with real liquidation value. The relevant risk is a multi-year de-rating and flat earnings — which is precisely what happened between 2021 and 2025, when the stock delivered +4.2% annualised with a −48.3% drawdown — not permanent capital impairment.

Recent News & Events

Has the business environment changed recently?

Yes, in two directions at once. (Fact and Interpretation.)

Improving: global beverage can volumes grew 5% in both Q1 and Q2 2026 — roughly twice the market rate — with Asia up double digits, Europe +7% and North America +5%. Industry capacity is tight (“mid to high nineties” adjusted utilisation, 110% at seasonal peak). Crown raised FY2026 adjusted EPS guidance on 2026-07-20 from $7.90–$8.30 to $8.30–$8.50, with adjusted FCF of at least $900M.

Deteriorating: Americas Beverage segment income fell to $475M in H1-2026 from $504M, and management guides the full year below 2025’s $1,030M. Latin American volumes are down (Brazil high-single-digits, on customer mix — Crown services the smaller two brewers, not the largest). The Middle East crisis costs Asia Pacific an incremental “$4 million to $5 million” per quarter. Transit Packaging remains in a “tepid global industrial production environment.” And capex is guided up 33%.

Significant acquisitions?

None. Zero acquisition spend in FY2024 and FY2025, and Donahue has stated none is contemplated. The only recent portfolio action is a divestiture: the Myanmar beverage can plant, sold February 2026, explicitly not expected to be material.

Change in accounting policies?

None identified. The five-year filing corpus contains no Item 4.02 non-reliance 8-K, no NT 10-K or NT 10-Q, no restatement, and no reported material weakness. Segment reporting was refreshed to the current CODM-based presentation consistent with ASU 2023-07, which added the segment expense and capex disclosures used throughout this analysis — a genuine improvement in transparency.

Recent changes — new markets, facilities, management?

(Fact.)

• New market: greenfield entry into Northern India — Crown’s first — at ~$250M for a two-line plant with volume commitments of “70% or a little bit higher.” Site under final land-cost negotiation, not yet disclosed.
• New facilities: first Greece line commercialised July 2026, second due late 2026; Spain due late 2026; Ponta Grossa, Brazil line addition (for size capability in the Southeast, explicitly not demand-driven).
• Divested: Myanmar beverage can plant (February 2026).
• Management: Djalma Novaes Jr. promoted from President – Americas Division to Executive Vice President and Chief Operating Officer effective 2025-07-01. Timothy Donahue continues as Chairman, President and CEO; Kevin Clothier as SVP and CFO.
• Capital structure: dividend raised materially in Q1-2026; $517M of stock repurchased in H1-2026, ~7% of shares over the trailing year.

Prepared as a supplemental appendix to the Crown Holdings research memo dated 2026-07-26. All figures reconciled to primary SEC filings, the 2026-07-20 earnings release, the 2026-07-21 earnings call, and the DEF 14A filed 2026-03-23, except where labelled as third-party aggregated data.

APPENDIX B — Source Appendix

Crown Holdings, Inc. (NYSE: CCK) · Report date 2026-07-26

Every non-obvious factual claim in the memo and Appendix A traces to a source below. Primary sources are listed first. Third-party aggregated data is labelled as such and was used for cross-checking, never as the authority where a filing exists.

1. Primary — SEC filings

Full SEC filing corpus enumerated for the period 2021-07-01 to 2026-07-24 (550 filings). Company CIK 0001219601.

Form census, 2021-07-01 to 2026-07-24: 334 Form 4 · 61 8-K · 41 Form 144 · 20 SC · 15 10-Q · 13 Form 3 · 11 11-K · 5 10-K · 5 DEF 14A · 5 SD · 5 Form 4/A · 2 S-8 · 2 S-4 · 2 424B3 · 1 PRE 14A · 1 DEFA14A · 1 S-4/A · plus registration and effectiveness notices.

Notable absences (verified, and material to the earnings-quality conclusion): no NT 10-K, no NT 10-Q, no 8-K Item 4.02 (non-reliance), no adverse ICFR opinion, no restatement.

SEC EDGAR XBRL companyconcept API (CIK 0001219601, 10-K/FY basis), pulled 2026-07-26 — the authority for the capital-expenditure and operating-cash-flow series that underpins the cash-flow section:

• us-gaap:PaymentsToAcquirePropertyPlantAndEquipment — 2018 $462M · 2019 $391M · 2020 $554M · 2021 $816M · 2022 $839M · 2023 $793M · 2024 $403M · 2025 $413M
• us-gaap:NetCashProvidedByUsedInOperatingActivities — 2019 $1,163M · 2020 $1,315M · 2021 $905M · 2022 $803M · 2023 $1,453M · 2024 $1,192M · 2025 $1,530M
• us-gaap:RevenueFromContractWithCustomerExcludingAssessedTax — 2021 $11,394M · 2022 $12,943M · 2023 $12,010M · 2024 $11,801M · 2025 $12,365M

2. Primary — management commentary

Verbatim quotations relied upon:

• On Americas: “it might be a bridge too far for this year to equal last year in segment income in the Americas… If we do not get to a billion of segment income, we will get real close to that number in the Americas.” — Donahue
• On the pass-through mechanism: “North America, we had a number of cost increases this year that we knew we would not fully recover in our pass through models, i.e., PPI not enough to fully capture all the cost increases.” — Donahue
• On utilisation: “if you took the rated speed of the equipment that is out there, you would posit that perhaps the industry is running at 92%-93%. But if you adjust that for changeovers for sizes, label changes, maintenance… we have got to be in the mid to high nineties, which is… real terms from April to August, it is 110% utilization.” — Donahue
• On future capacity: “There may come a time when we need to consider more capacity. And when we believe we can do that in a responsible way we will take a look at that much closer.” — Donahue
• On India economics: “new plant largely cost around $250 million to build. You know, you are putting 2 high speed lines in a plant”; “somewhere like India, you might get commitments for 70% or a little bit higher.” — Clothier
• On M&A: “certainly not contemplating any large M and A. Frankly, not contemplating any M and A.” — Donahue
• On H2 buyback: “we should be able to buy close to, you know, $200 million worth of stock back in the second half.” — Clothier
• On 2027 FCF: “another 900 million to a billion dollars of free cash flow next year as well.” — Donahue
• On the World Cup comp: “if the World Cup was worth a few hundred million cans in Q2, if we wanna just throw a dart and try to pick a number you have gotta try to overcome that.” — Donahue
• On declining to quantify non-metal exposure: “hesitant to wanna answer that question only because we do not like to give away too much of our cost model or pricing model.” — Clothier
• On the equipment business: “the large majority of the beat in other maybe at least 2 thirds of the beat in other, had to do with our can-making equipment business… We supply many can companies around the world, including some of our direct competitors here in The United States.” — Donahue
• On the dividend: “we took a big step at the beginning of this year to bring the dividend up to a level that more appropriately reflects our confidence in our future cash flow generation capabilities.” — Donahue
• On cash: “the fortunate problem we have is we have a lot of cash.” — Donahue
• On Brazil: “our Brazilian team being down high single digits in the first half is projecting that they are going to be flat for the year. So some of our second half caution is just putting some caution against our own Brazilian forecast.” — Donahue
• On food can mix: “About 40% of the business now is pet food so that is a very stable business.” — Donahue

Analysts on the call (used in Appendix A “questions other investors have asked”): George Staphos (Bank of America), Philip Ng (Jefferies), Christopher Parkinson (Wolfe Research), Michael Roxland (Truist), Hillary (Deutsche Bank), Arun Viswanathan (RBC Capital Markets), Joshua Spector (UBS), Jeffrey Zekauskas (JPMorgan), Edlain Rodriguez (Mizuho), Katan Mamtora (BMO Capital Markets).

3. Third-party quantitative sources

Aggregated data. Used as cross-checks and for market/factor data unavailable in filings. Where any figure conflicted with a filing, the filing governed.

4. Industry and trade press

• Packaging Dive / Food Dive, “Canmakers say capacity is tight as beverage industry gears up for big summer,” 2026 — North American can supply “expected to be particularly constrained”; Crown’s highest-ever shipments month in March 2026; long-term volume range 1–3%.
• Vinetur, “Can Makers Brace for a Tight Summer,” 2026-05-07 — industry capacity tightness corroboration.
• Verified Market Research / Mordor Intelligence, aluminium can market structure — Ball, Crown and Ardagh ≈ 66% of the global aluminium beverage can market; Ball ~36% of North America.

5. Peer company filings

• Ball Corporation (NYSE: BALL) — FY2025 Form 10-K and public market data as of 2026-07-13. Source of the peer comparison: Ball FY2025 ROIC 9.1%, operating margin 10.6%, EV ~$22.1B / ~11.0x EBITDA, BB+/Ba1 credit rating, top-three customers ~40% of net sales, FY2024 shareholder returns of $1,932M against negative free cash flow, and regional market-size figures (North America ~139B units, EMEA ~97B, South America ~43B).

No position in CCK or any company mentioned is stated or implied anywhere in this article.

6. Analytical frameworks

• Greenwald & Kahn, Competition Demystified — barriers-to-entry taxonomy (the moat section: Crown classified as regional economies-of-scale plus customer captivity); market-share-stability test (the industry-structure section: industry passes); the requirement that a moat claim tie to a financial outcome (the moat section: satisfied by 25.8% return on tangible capital).
• Marathon Asset Management / Chancellor, Capital Returns — capital-cycle analysis (the capital-cycle discussion: the 2020–22 over-build → 2022–23 bust → 2024–26 payoff → 2026 turn); the asset-growth anomaly applied to the capex step-up (the capital-cycle discussion, the capital-allocation section); supply-side focus on capacity rather than demand forecasting.

7. Gaps and unavailable data — recorded, not estimated

Source appendix prepared 2026-07-26. Price as-of 2026-07-24 close ($117.88). Balance-sheet data as-of 2026-06-30.