Ball Corporation (NYSE: BALL) — A Buyback Algorithm Priced as a Value Stock

An independent equity research note Report date: 2026-07-14 · Price as-of: $61.30 (2026-07-13 close) · Shares out: 266.2M (2026-05-01) · Market cap: ~$16.3B · EV: ~$22.1B Coverage: Initiation · Sector: Materials · Metal & Glass Containers (Aluminum 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

The author’s own independent opinion, offered as general information and not investment advice. The analytical body (Sections 1–15) below carries no position and no price target.

Verdict: AVOID here / HOLD if owned. Accumulate only in the mid-$40s to low-$50s (~8.5–9.5x EV/EBITDA on comparable EBITDA, ~$45–52/share). Conviction: medium-high.

Tag: “The can is winning. The converter isn’t.”

Ball is the world’s #1 aluminum beverage can maker — #1 in all three of its regions, with share that has barely moved in decades — and it is still not a good business. Over eight years its return on invested capital has run 7.2–10.4%, never once clearing ~10.4%, even at the 2021 hard-seltzer peak when the industry claimed it was sold out for three years. That is a business earning roughly its cost of capital. The moat is real but it is local — cans are cheap and bulky, so freight economics hand each plant a regional near-monopoly — and the surplus that moat generates is negotiated straight back out by customers who are themselves giants: AB InBev, Coca-Cola’s bottler co-op, and Red Bull are ~40% of sales. Ball has a moat and no rent.

What the market is mispricing is not the business — it’s the arithmetic. The stock screens cheap: 20.9th percentile against its own ten-year multiple range. That percentile is a trap. The upper half of that distribution is the 2020–21 seltzer bubble (52x P/E, 19x EV/EBITDA) that the 2022–23 bust proved was a fiction; measuring today against a discredited peak manufactures cheapness. Struck properly at $61.30, Ball trades at ~11.0x EBITDA versus Crown Holdings at ~9.1x — a premium to a direct competitor earning 12.6% ROIC against Ball’s 9.1%. Ball is the lower-quality asset at the higher price. Worse, the “10%+ EPS growth algorithm” that re-rated the stock is not operating growth: FY25 comparable net earnings rose 0.8% while comparable EPS rose 12.6%. The entire difference is share count — and the share count was shrunk with one-time BAE divestiture proceeds and incremental debt, not with free cash flow. In FY24 Ball returned $1,932M against negative FCF. That engine is now out of fuel: the 2026 buyback drops to ~$600M, and shares outstanding actually rose in Q1’26. The framing is contrarian value, and the tape is neutral-to-supportive (above both EMAs, no crowded momentum) — but the value isn’t there. A reverse-DCF at an 8% WACC cannot reach $61.30 even with infinite ROIC; the price requires a ~6.4% discount rate, which a BB+/Ba1 high-yield credit does not deserve. Ball is priced for a franchise it has never been.

Triggers. Flips bullish: evidence that the 2027+ contract renewals are being repriced at structurally better terms — i.e., ROIC breaking durably above ~12% — proving pricing power finally shifted toward the converter. Flips more bearish: NA volumes turning negative while the Midwest premium stays above $1/lb, squeezing the pass-through lag and forcing leverage above 3.5x on a sub-IG balance sheet.

📈 Stock Price Action — Five-Year Event Map

Ball has round-tripped a full cycle and is still a third below its high. The stock closed as high as $91.78 (2021-09-02) at the hard-seltzer peak, collapsed to $41.22 (2023-10-23) as that demand proved a fad, and has clawed back to $61.30 — -33.2% off its five-year high, with a 52-week range of $44.35–$67.84 (intraday basis) and the stock now -8.9% off its 52-week high. It sits above both its 50-day EMA ($59.12) and 200-day EMA ($57.26), with the 50 above the 200 — an uptrend structure, but one operating inside a five-year range, not a breakout. (Source: AZI 5-year price CSV, pull 2026-07-13/14.)

Cycle narrative. (1) The 2020–21 boom was a genuine demand shock — canned-beverage consumption spiked with on-premise closures and the hard-seltzer craze — and the market extrapolated it into a secular substrate-shift story, taking Ball to a ~52x P/E. (2–3) It was a fad. Seltzer collapsed, customers destocked, and the 2022-08-04 print delivered a -18.6% day, the sharpest five-year move, when volume guidance broke. (4) The 2023 bounce was mechanical, not fundamental: the aluminum destock released working capital and drove FY23 OCF to $1,863M. (5) The October-2023 low at $41.22 marked capitulation, with Ball closing plants (St. Paul, Kent WA, Phoenix) into the overbuild it had helped create. (6) The BAE close on 2024-02-16 was the pivotal capital event — ~$4.5B of after-tax cash that funded ~$1.7B of 2024 buyback and ~$2B of debt paydown, mechanically lifting EPS. (7–8) Since, the stock has recovered on the buyback-driven EPS algorithm and a tightening capacity backdrop, with the Q4’25 print (+9.0%) re-rating it on a guided 10%+ EPS growth. The five-year annualized return remains -4.4% with a -55% max drawdown (FactorsToday leaderboard) — a lost half-decade. Price moves are Fact; attributed drivers are Interpretation.

1. Executive Summary

Ball Corporation is the world’s largest producer of aluminum beverage cans, holding the #1 position in each of its three reportable regions: North & Central America (~36% of a ~139-billion-unit market), EMEA (~39% of ~97B units), and South America (~46% of ~43B units). Following the sale of its aerospace business to BAE Systems for $5.6B — announced August 2023, closed February 16, 2024 — it is a pure-play packaging converter. FY2025 net sales were $13,161M, with reportable segment comparable operating earnings of $1,594M and GAAP diluted EPS of $3.30.

The business has a real but narrow moat that produces stability, not returns. Cans are cheap, bulky, and freight-uneconomic, so plants co-locate at customer fillers, handing each incumbent a regional near-monopoly. This is a genuine Greenwald economies-of-scale-plus-captivity advantage, and it shows: Ball’s share has drifted less than two points in decades. But the advantage is bilateral. Ball’s customers are as concentrated as Ball is — AB InBev (15% of net sales), Coca-Cola Bottlers’ Sales & Services (14%), and Red Bull (11%) together are ~40% — and they negotiate the surplus away. The financial proof is unambiguous: ROIC has run 7.2–10.4% across 2018–2025, never clearing ~10.4% even at the 2021 peak, against a WACC of roughly 7.5–8%. Ball passes the market-share-stability test and fails the franchise-return test. It is a good business, not a wonderful one.

The reported growth is a buyback artifact. Ball guides to “10%+ comparable diluted EPS growth” and delivered +12.6% in FY2025 ($3.17 → $3.57). But comparable net earnings over the same period rose from $977M to $985M — 0.8%. Every point of the EPS growth beyond that came from retiring stock: share count fell from 349M (2016) to 265.4M (2025), with the shrink heavily back-loaded at -8.3% in each of 2024 and 2025. That repurchase was not funded from operations. In FY2024 Ball returned $1,932M to shareholders against negative free cash flow (OCF of just $115M, depressed by $766M of aerospace cash taxes, against $345M+ of capex); in FY2025 it returned $1,541M against $788M of FCF while total debt rose $1,319M. The per-share growth that re-rated the stock was purchased with divestiture proceeds and borrowings, not earned. That engine has now stopped: the 2026 plan cuts buyback to ~$600M, and shares outstanding actually rose to 266.2M by 2026-05-01 as option issuance outran repurchase.

The stock is not cheap, despite screening that way. AZI’s valuation index puts Ball at the 20.9th percentile of its own ten-year range — but the upper half of that range is the 2020–21 seltzer bubble (52x P/E, ~19x EV/EBITDA) that the subsequent bust proved false; measuring against a discredited peak manufactures cheapness. Struck correctly at $61.30 (market cap $16.3B + FY25 year-end net debt $5.8B = EV ~$22.1B), Ball trades at ~11.0x FY25 EBITDA — versus Crown Holdings at ~9.1x with a 12.6% ROIC and 12.9% operating margin, against Ball’s 9.1% and 10.6%. Ball is the lower-quality asset at the higher price. A reverse-DCF is blunt: at an 8% WACC and 2.5% perpetual growth, no ROIC solves for $61.30 — at infinite ROIC the value is $59.11. The price requires a ~6.4% discount rate. That premise is directly contradicted by Ball’s credit: BB+ (S&P) / Ba1 (Moody’s) — sub-investment-grade.

Balance sheet quality is thinner than the headline. Total equity of $5,421M is ~99% Rexam-deal intangibles: strip $4,379M of goodwill and $982M of other intangibles and tangible common equity is ~$60M (~$0.23/share) — and it was negative in 2022 (-$1.9B) and 2023 (-$1.6B). P/B is meaningless here and should be ignored. GAAP earnings are separately distorted: FY2024’s headline $13.00 diluted EPS contained a $3,584M aerospace disposal gain; continuing-operations EPS was $1.37. Ongoing Rexam intangible amortization of ~$135M/year is a permanent GAAP drag.

Verdict framing for the committee: Ball is a structurally fair-to-decent industry position — a consolidated regional oligopoly currently sitting in the favorable half of the capital cycle, with utilization back to the mid-90s% and supply discipline restored — attached to a return profile that has never justified a franchise multiple. The three linchpins are (1) ROIC ≈ WACC, (2) top-3 customers at ~40% of sales, and (3) an EPS algorithm dependent on a buyback that is now stepping down by more than half. The bull case does not rest on an operating claim; it rests on a discount-rate assertion that the company’s own high-yield rating undermines.

2. Business Overview

Ball Corporation, founded in 1880 and headquartered in Westminster, Colorado, employs roughly 16,000 people and is the world’s largest manufacturer of aluminum beverage containers. Its history is one of serial reinvention — from glass jars, to aerospace, to metal packaging — and the most recent turn completed on February 16, 2024, when it closed the sale of its aerospace segment to BAE Systems for $5.6B, leaving a pure-play packaging converter for the first time in decades.

Segments and revenue mix (FY2025, 10-K filed 2026-02-19). Ball reports three segments plus “Other”:

Geographically, the US is $6,163M, Brazil $1,494M, and other markets $5,504M. South America is the highest-margin region at 15.1% — a function of a more concentrated market (Ball at ~46% share) and less-developed customer countervailing power — while North America and EMEA both sit at ~12.3–12.4%.

How it makes money — and what that structure implies. Ball converts aluminum coil into cans, ends, and closures, and sells them under multi-year supply contracts to beverage fillers. The economics are conversion economics: Ball earns a spread on the conversion, not on the metal. Aluminum cost is passed through to customers via contractual provisions in most sales contracts, supplemented by derivative hedging. This is the single most important structural fact about the business, and it cuts both ways. It insulates Ball from commodity direction — a genuine risk reducer, and the reason the equity beta is only 0.68 — but it also means Ball has no exposure to the upside of its principal input and, more fundamentally, that the customer knows precisely what Ball’s cost base is. A pass-through contract is a transparent contract, and transparency is the enemy of pricing power. Ball is compensated for a spread that its counterparty can compute.

Revenue durability. Revenue is recurring in the practical sense — cans are a consumable, purchased continuously by long-tenured customers under multi-year agreements, and the 10-K describes “high customer retention.” But this is volume recurrence, not contractual lock-in of price. The distinction matters enormously and is developed in the competitive-position section below: the customer must come back, but on renewal it comes back with leverage. Ball’s Q1-2026 commentary that it is “sold out for 2026, over 90% for 2027, and over 50% through the decade” should be read as management framing, not verified fact — Ball discloses no backlog, no contracted volume, and no utilization figures, and the phrase “sold out” appears zero times in the 10-K. The claim is structurally unfalsifiable from outside, and an identical claim circulated in October 2020, immediately before the worst demand bust in the industry’s modern history.

Recent portfolio moves. The cups business was deconsolidated on 2025-03-21. On 2025-08-27 Ball deconsolidated its Saudi operation (selling 41% of a 51% stake, retaining 10%). In January 2026 it acquired Benepack — 80% of ORG’s European beverage-can business (Belgium and Hungary) for $218M/€184M total, of which only $95M was cash, the balance assumed debt. Bolt-ons remain small: $159M of cash for acquisitions in FY2025, $75M in Q1’26.

Verdict. A focused, simple, well-run pure-play converter with a genuinely leading position in a consumable product — but a business model whose defining feature, cost pass-through, structurally caps the value it can capture. Understandability is high; the economics are transparent to the customer as well as to us. This is a toll-taker whose toll is set by the people driving through.

3. Industry Dynamics

Structure: a consolidated regional oligopoly. The aluminum beverage can industry is, on paper, exactly the structure investors are taught to want. In North America, five firms make “substantially all” of a ~139-billion-unit market. In EMEA, four firms serve ~97B units. In South America, four firms serve ~43B units. Globally the industry is Ball, Crown Holdings (CCK), Ardagh Metal Packaging (AMBP), and CANPACK. Concentration this tight normally produces pricing discipline and excess returns.

It doesn’t here, and understanding why is the analytical heart of this report.

Why the oligopoly doesn’t pay. Three forces neutralize it:

1. Countervailing customer power. The industry’s concentration is mirrored — and exceeded — on the other side of the table. Ball’s top three customers are ~40% of sales. AB InBev, Coca-Cola’s bottling co-op, and Red Bull each buy in volumes that make them irreplaceable to any single converter, and each is sophisticated enough to run a competitive tender on a product with published input costs. When a concentrated seller meets a more concentrated buyer, the surplus flows to the buyer.
2. Product commoditization. A 12-ounce aluminum can is a specification, not a product. Ball’s cans are not meaningfully differentiated from Crown’s. The only differentiators — printing quality, specialty formats, delivery reliability — are real but marginal, worth basis points of price, not multiples of return.
3. A structural tendency to over-build. This is the Marathon capital-cycle point and it is the industry’s defining pathology, examined below.

The capital cycle — the industry’s recurring self-harm. The 2020–23 episode is a textbook Marathon cycle and the single most instructive thing in this report’s evidence base. In 2020–21, COVID and the hard-seltzer craze produced a genuine demand shock; the industry declared itself “sold out 24–36 months” (Credit Suisse, October 2020), and Ball projected a 10-billion-can 2020 deficit. High returns duly attracted capital: the industry added roughly 30% to North American capacity, Ball among the most aggressive (Glendale AZ, Pittston PA, and others). Then seltzer collapsed. Customers destocked. By 2022–23 Ball was closing the plants it had just built or expanded (St. Paul, Kent WA, Phoenix) and taking writedowns — $420M of business-consolidation charges in FY2024 alone. Ball’s ROIC fell from 10.4% (2021) to 7.3% (2022) and 7.2% (2023).

The lesson is not that management erred — every competitor did the same thing, which is precisely Marathon’s point. The lesson is that this industry converts demand booms into capacity, not into profit. Any bull case resting on “the can is winning” must explain why the next demand upswing will not be capitalized away exactly as the last one was.

Where we are now: the favorable half of the cycle. Credit where due — the current setup is genuinely better. North American operating rates are back to the mid-90s%, Europe outperformed in 2025, and supply discipline has returned (BofA, inside.beer, Packaging Dive, 2026). Ardagh Metal Packaging, the weakest player, is financially distressed at ~7.1x net debt/EBITDA and in no position to add capacity. This is the phase of the capital cycle in which returns improve — and it is why Ball’s ROIC recovered to 9.1% in 2025 and why volumes and pricing are firming. An investor buying today is buying into a favorable cycle position. The question is what they are paying for it, and whether a mid-90s% utilization rate that still only produces a 9.1% ROIC tells you something permanent about the industry.

It does. That is the finding. Ball’s peak ROIC, at the top of the greatest demand boom in the industry’s history with the industry claiming it was sold out for three years, was 10.4%. If a business cannot clear 11% at the top of its cycle, the cycle is not the problem.

Demand: real substrate shift, oversold as a growth story. The secular case is legitimate but modest. Aluminum’s global recycling rate is ~75% (per the 10-K), Ball’s cans average 74% recycled content, and cans benefit from recyclability, the “billboard effect” of full-surface printing, and the growth of energy drinks (Red Bull, Monster, Celsius) which are can-native. Third-party sizing puts the global aluminum can market at ~$42B in 2025 growing ~4.7% CAGR to ~$57B by 2034 (Coherent Market Insights) — though this vendor sizing carries no published methodology and should be treated as an assumption, not evidence; the 10-K’s own unit figures are the reliable anchor. Ball’s own volume algorithm is +1–3% NA, +3–5% EMEA, +4–6% SA. Actual Q1-2026 global volume growth was +1%. That is the honest number: this is a GDP-plus-a-bit consumable, not a growth market.

Regulation and tariffs. Section 232 aluminum tariffs doubled from 25% to 50% effective 2025-06-04, and the Midwest Premium exceeded $1/lb for the first time in late January 2026. Ball’s CFO expects “some direct tariff cost in 2026,” with metal cost largely passed through. The pass-through protects margins but creates two second-order risks: a timing lag between cost incurrence and recovery, and — more strategically — higher can prices blunt the substrate-shift advantage versus glass and PET. The tariff that protects domestic converters also raises the price of the product they are trying to get shelf space for.

Verdict: a FAIR industry, tilting good near-term — structurally better than fragmented, structurally worse than differentiated. The consolidated regional oligopoly and the current favorable capital-cycle position are real positives, and the near-term earns the “tilting good.” But a commodity product sold to more-concentrated customers under cost-pass-through contracts, in a market with a demonstrated reflex to convert every boom into overcapacity, is not a structurally attractive industry. It is a defensible one. The disconfirming evidence I weighed most heavily is the current mid-90s% utilization with restored discipline — a genuinely good setup — and it still only yields a 9.1% ROIC. That is the industry telling you what it is.

4. Competitive Position

The moat is real. Name it precisely: economies of scale plus customer captivity, operating locally.

In the Greenwald taxonomy, Ball’s advantage is not a brand, not a network effect, not a patent. It is the classic “think local” case, and it is one of the cleaner real-world examples of the mechanism. An empty aluminum can is mostly air: it is cheap per unit and bulky per unit, which makes freight cost a punishing fraction of delivered value. Shipping empty cans any meaningful distance destroys the economics. Consequently, can plants are built adjacent to — sometimes physically connected to — customer filling lines. The relevant market is therefore not “global cans” but “cans within economic truck distance of this filler.” Within that radius, the incumbent with the co-located plant has a structural cost advantage no distant competitor can match, and the customer faces real switching friction: qualifying a new supplier means new logistics, new specifications, and supply-chain risk on a line that cannot stop.

The evidence that this is a genuine barrier, not a story. Apply the share-stability test — Greenwald’s most reliable diagnostic. Ball has been #1 in North America, EMEA, and South America for decades, and share has drifted less than two points. Market share that stable, for that long, in a market with four or five participants, is close to dispositive: it means new entrants cannot profitably attack and existing rivals cannot profitably take share. That is a barrier to entry. Ball is not “a company in a crowded market with weak differentiation.” It is an entrenched incumbent behind a real wall.

And yet the moat produces no rent. This is the central paradox of the investment.

Run the second Greenwald test — the ROIC test. A franchise generates returns durably above the cost of capital, typically 15–25%. Ball’s ROIC, 2018 through 2025:

Against a WACC of roughly 7.5–8%, this is a business earning approximately its cost of capital, with a cyclical peak of 10.4%. Gross margin has been flat at ~19–21% for the entire period. Ball passes the share-stability test and fails the return test — a combination that is uncommon and diagnostic. It means the barrier exists but the surplus it creates is being transferred, not retained.

Where does the surplus go? To the customers. This is the bilateral bargaining problem, and the numbers make it concrete. AB InBev is 15% of net sales (16% in 2024, 15% in 2023), Coca-Cola Bottlers’ Sales & Services 14% (13%, 13%), and Red Bull 11% (9%, 8%) — ~40% between three counterparties, and note Red Bull’s rising share of the mix. Consider the negotiation from AB InBev’s side: it knows Ball’s aluminum cost precisely, because the contract passes it through. It knows the conversion cost is a well-understood industrial process. It buys enough volume to fill a plant. It can credibly threaten to shift incremental volume to Crown or CANPACK at the next renewal, or to backward-integrate. Ball’s local monopoly gives it the power to refuse a bad price on this line today — but not the power to extract a good one over a multi-year contract, because Ball’s plant is a sunk, immobile asset that is worthless without that customer’s volume. The co-location that creates Ball’s advantage over rivals simultaneously creates the customer’s advantage over Ball. The moat and the trap are the same wall.

Head-to-head versus Crown Holdings — the comparison that should trouble a Ball bull. Crown is the closest direct analogue: same product, same customers, same structure.

Ball is the scale leader — larger, #1 everywhere — and earns 350bp less ROIC and 230bp lower operating margin than Crown, while trading at a ~1.9-turn premium. If scale were the source of durable advantage in this industry, the largest player should out-earn the second-largest. It does not. That is powerful evidence that scale in cans stops paying above a fairly low threshold — you need enough plants to be co-located with your customers, and beyond that additional global scale buys little, because the competitive unit is the plant, not the company. Ball’s global #1 position is a fact; it is not obviously an asset.

Pressure-testing the counter-arguments. Three deserve a fair hearing:

• “Ball’s scale gives it procurement advantage on aluminum.” Weak. Aluminum is an exchange-traded commodity with transparent pricing, and the cost passes through regardless. Procurement scale in a passed-through commodity accrues largely to the customer.
• “Ball’s innovation — specialty cans, sleek/slim formats — differentiates it.” Partially true and worth something; specialty formats carry better margin. But these are rapidly copied, and if they were durably differentiating, Ball’s margins would exceed Crown’s. They don’t.
• “The 2022–23 trough was cyclical; normalized ROIC is higher.” Fair — and I credit it. Take the 2021 peak, 10.4%, as the honest cyclical ceiling. A business whose peak return in its best year in a generation is 10.4% is not a franchise. Normalizing helps Ball’s near-term earnings; it does not rescue the return profile.

Verdict: a DURABLE but NARROW and LOW-RETURN advantage — a real moat with modest rent. The barrier is genuine, geographically bounded, and has held for decades; Ball is entrenched and will not be displaced. But the advantage manifests as stability — of share, of volume, of a margin floor — and not as level of returns. In plain terms: the moat protects Ball from competitors but not from customers. Applying the discipline stated at the outset — if a moat claim cannot be tied to a financial outcome that would deteriorate without it, it is not a moat — Ball’s moat passes, but the outcome it protects is survival at ~9% ROIC, not superior economics. It is worth owning at a price that assumes exactly that and nothing more. It does not justify a compounder multiple.

5. Growth History and Forward Opportunities

The historical record: volume growth is real, small, and cyclical. Ball’s net sales went $12,062M (2023) → $11,795M (2024) → $13,161M (2025). The FY24 decline and FY25 rebound (+11.6%) are not a growth trend; they reflect the destock/restock cycle, aluminum price pass-through, FX, and portfolio changes. Decomposing FY2025’s growth from the MD&A gives the honest picture:

Two observations. First, roughly half of FY25’s segment growth is price/mix, which is substantially aluminum pass-through — i.e., revenue that arrives with its own cost attached and does not drop to earnings. Revenue growth in a pass-through converter is a low-information number, and this is why. Second, the genuine volume contribution (~$678M on a $11.8B base, ~5.7%) reflects a restocking rebound off a depressed base, not a run-rate.

The run-rate is visible in Q1-2026: global volume +1%. Against a company algorithm of +1–3% NA, +3–5% EMEA, +4–6% SA — which would blend to roughly +2–4% — actual delivered volume is at or below the bottom of the range. Q1’26 sales of $3,603M (+16% vs $3,097M) again overstate the reality: reportable comparable operating earnings rose only +7% ($406M vs $378M), and the sales/earnings gap is pass-through and the Benepack consolidation.

Long-run per-share growth has come from share count, not from the business. This is the finding that reframes the whole story, and it is worth stating precisely because Ball’s own guidance language obscures it. Ball’s stated long-term financial goal is “comparable diluted EPS growth >10%/yr.” In FY2025 it delivered +12.6% ($3.17 → $3.57). But:

Comparable net earnings: $977M (FY24) → $985M (FY25) = +0.8%. Comparable diluted EPS: $3.17 → $3.57 = +12.6%. The entire difference is a ~11% reduction in share count.

The earnings of the business were flat. The earnings per share grew 12.6% because there were 11% fewer shares. (One caveat, stated for honesty: FY24’s $3.17 includes aerospace through 2024-02-16, so the like-for-like rate is not perfectly clean — but the direction and the overwhelming magnitude of the share-count contribution are not in question.)

Share count: 349M (2016) → 315.7M (2023) → 289.4M (2024) → 265.4M (2025) — and the shrink is heavily back-loaded: roughly -1.2% to -1.6%/year through 2023, then -8.3% in each of 2024 and 2025, contributing ~+9%/year to EPS. Those two years are when the stock re-rated. They are also precisely the two years Ball had $4.5B of BAE cash to spend.

Forward opportunities — assessed honestly.

• Substrate shift (glass/PET → aluminum). Real, slow, and partially offset. Aluminum’s recyclability and the billboard effect are durable advantages, and energy drinks — the fastest-growing beverage category — are can-native. But 50% Section 232 tariffs and a >$1/lb Midwest Premium raise the delivered price of a can relative to PET, blunting exactly this edge. Net: a modest tailwind, worth perhaps a point of volume, not a growth story.
• Emerging-market volume (South America). The best genuine opportunity. SA carries the highest margin (15.1%) and the fastest algorithm (+4–6%), and Ball holds ~46% share. But it is only 16.4% of sales, so even excellent SA performance moves the consolidated needle slowly — and it brings Brazilian FX and macro volatility.
• Specialty/aerosol/PHC. The “Other” segment is $730M (5.5% of sales), with PHC at 1.5B units (21% of a 6.8B market). Higher-margin, but too small to matter to the thesis.
• Bolt-on M&A. Benepack (Jan 2026, $218M for 80% of ORG’s European can business) and Florida Can (Q1 2025) are sensible, small, capability-and-footprint additions. At $159M/year of acquisition cash, this is maintenance, not a growth engine.
• The 2027+ repricing. The single most important forward variable, and the one we cannot see. Ball says it is >90% contracted for 2027 and >50% through the decade. “Contracted” is not “priced well.” If the tight capacity backdrop is allowing Ball to renew at structurally better conversion spreads, ROIC breaks higher and the bull case is right. If renewals merely lock in current spreads with concentrated customers who know Ball’s cost base, the algorithm is unchanged. Ball discloses nothing that lets an outsider distinguish these. This is the genuine blind spot in the analysis.

Verdict: LOW-QUALITY growth — real at the volume line, but low-magnitude, cyclical, and historically manufactured at the per-share line. Underlying volume growth of ~1% (Q1’26 actual) on a GDP-linked consumable, with roughly half of headline revenue growth being pass-through, is not a growth business. The per-share growth that investors actually experienced was bought, not earned — and it was bought with a non-repeatable asset sale. The disconfirming evidence I weighed: the capacity backdrop is genuinely tight and could deliver better-than-algorithm pricing on renewal, which would be true operating growth. That possibility is live, unverifiable from disclosure, and is the strongest argument the bulls have. It is also, notably, an argument about something Ball refuses to quantify.

6. Financial Quality

Quality of earnings: three material distortions, all of which flatter the headline.

First, the aerospace gain. FY2024 GAAP diluted EPS of $13.00 is meaningless as an earnings figure: it contains a $3,584M discontinued-operations gain on the BAE disposal (gain on disposal $4,634M pre-tax). Continuing-operations diluted EPS was $1.37. The wedge — $13.00 headline versus $1.37 of actual continuing earnings — is the largest QoE distortion in this file, and any screen or model keying on FY24 GAAP EPS is wrong by roughly 10x. FY2025 is clean by comparison: GAAP diluted EPS $3.30 (continuing $3.30; disc-ops nil), net income $912M.

Second, Rexam amortization. Ball carries ~$135M/year of amortization of acquired intangibles from the 2016 Rexam deal — $135M in FY25, and roughly $135–139M/year ongoing. This is a genuine GAAP charge against a real (if sunk) cash outlay. Ball’s “comparable” earnings add it back. Both lenses are defensible; the honest treatment is to acknowledge that comparable EPS of $3.57 excludes a charge for an $8.4B acquisition that Ball actually paid for.

Third, “record adjusted FCF.” Ball reported “record $956M adjusted free cash flow” for FY2025. The reconciliation: $788M of actual FCF (OCF $1,262M − capex $474M) plus a $168M aerospace cash-tax add-back. That $168M was a real cash outflow. The adjustment is disclosed and arguably defensible as non-recurring, but “record” is doing work that the underlying $788M does not support.

The segment-to-EBT bridge (FY2025) — where $1,594M becomes $1,128M:

Note the business-consolidation line: a +$41M credit in 2025 against a -$420M charge in 2024 and -$133M in 2023. A swing of $461M year-over-year from a line item that is, by name, non-recurring — and which was a benefit in the year Ball reported strong results. This is not aggressive accounting, but it is a reminder that the “comparable” framing has been excluding large charges for years and then booking a credit when reversals land.

Margins: flat, and that is the tell. Gross margin has run ~19–21% for eight years. Blended comparable segment operating margin is ~12.1% (NCA 12.3%, EMEA 12.4%, SA 15.1%). GAAP operating margin is 10.6%. These margins do not expand with scale. Ball has grown revenue from ~$8B pre-Rexam to $13.2B and the margin structure is unchanged. That is the direct financial refutation of a scale-advantage thesis: if scale conferred economics, eight years and a transformational acquisition would have shown it.

Returns. ROIC 9.1% (2025), 7.2–10.4% across 2018–25, versus WACC ~7.5–8%. ROE requires a warning. ROIC.ai reports ROE of 6.87% and book value per share of $49.70 — both are garbled and both must be discarded. Real total equity is $5,421M on 265.4M shares = $20.43/share (AZI independently corroborates at $20.94), and ROE on $912M of continuing net income is 16.8%. But that 16.8% is itself an artifact, not a quality signal: it is high only because the denominator is tiny relative to the asset base — the equity has been repurchased away. A high ROE produced by shrinking equity with borrowed money is a leverage statistic, not a profitability statistic. ROIC (9.1%) is the honest measure here, and it is the one to use.

Balance sheet: thin, intangible, and sub-investment-grade.

Ball’s equity is ~99% Rexam-deal intangibles. Tangible common equity is essentially zero — and was negative in 2022 (-$1.9B) and 2023 (-$1.6B). P/B is meaningless for Ball and should not be used, which also means the AZI valuation index’s P/B percentile (20.3rd) is uninformative and its composite is contaminated by it.

Debt: total $7,012M (2025) against $5,693M (2024) — up $1,319M — with cash of $1,212M for net debt of ~$5.8B, or ~2.7x comparable EBITDA. By Q1-2026 total debt had risen to $7.86B with cash of $749M: net debt ~$7.1B, a seasonal peak on aluminum inventory build (~3.3x). The ladder is well-termed and cheap — blended coupon ~4.5%, revolvers undrawn until June 2030 — with maturities spread across €600M 1.5% 2027, $1,000M 6.0% 2029, $1,300M 2.875% 2030, a $1,500M Term A at 4.97% 2030, $850M 3.125% 2031, €850M 4.25% 2032, and $750M 5.5% 2033.

But the rating is BB+ (S&P) / Ba1 (Moody’s), both stable — sub-investment-grade. This is a correction to the working assumption of BBB and it matters twice over. Operationally, it modestly raises refinancing cost and narrows access in stress. Analytically, it is decisive for the valuation, because the bull case set out in the valuation section depends on Ball deserving a ~6.4% WACC. A high-yield credit does not earn a 6.4% cost of capital. Ball’s low equity beta (0.68) reflects the demand defensiveness of selling cans for beer and Coke; it does not reflect a low-risk balance sheet, and conflating the two is the error at the center of the bull case.

Cash flow: volatile, working-capital-driven, and weaker than it looks.

FY2023’s $1,863M was inflated by the aluminum destock releasing working capital. FY2024’s $115M was crushed by $766M of aerospace cash taxes (total aerospace cash tax $830M: $766M in 2024, $168M in 2025, with a $104M refund in January 2026), producing negative FCF. FY2025’s $1,262M absorbed a -$453M inventory build (aluminum) offset by +$716M of accounts payable — meaning a meaningful slice of FY25 OCF is a payables stretch, which is a one-time benefit that reverses. D&A ran $622M/$620M/$686M (a basis mismatch with continuing ops, reconciling via disc-ops D&A of $0M/$9M/$81M).

Capital intensity: high, and now rising to meet D&A. Capex was $474M in FY25 against $622M of D&A — Ball has been under-investing relative to depreciation, which flatters recent FCF. The 2026 guide takes capex to $600–650M, i.e., approximately D&A. That is the correct, disciplined level for a business at ~1% volume growth — and it is a ~$150M/year headwind to FCF versus 2025’s run-rate. The 2026 guide of “>$900M adjusted FCF” should be read against this and against the $956M “record” that itself included a $168M add-back.

Verdict: economics do NOT improve with scale — and this is the report’s most important financial finding. Eight years, an $8.4B transformational acquisition, and $5B of revenue growth have produced no margin expansion and no ROIC improvement. Gross margin is flat at ~19–21%; ROIC has never cleared 10.4%. Earnings quality is fair but requires three separate normalizations, all of which cut the same way. The balance sheet is adequately termed but tangibly empty and rated below investment grade, and the reported ROE of 16.8% is a leverage artifact of having repurchased the equity away rather than evidence of profitability. Ball is a competently run, financially stable, low-return converter. Nothing in the financials supports a franchise multiple.

7. Capital Allocation

Capital allocation is where a mediocre-return business either compounds or leaks, and Ball’s record here is genuinely mixed — competent and shareholder-oriented in intent, but built on a foundation of one large, expensive mistake and one large, non-repeatable windfall.

The Rexam acquisition (2016, $8.4B) — the deal that defines the balance sheet. Rexam made Ball the global #1 and created the scale position that the entire bull case rests on. Judged on strategy, it worked: Ball got the #1 slot in NA, EMEA, and SA. Judged on economics, it is harder to defend. The deal loaded ~$4.4B of goodwill and ~$5.4B of intangibles onto the balance sheet against $5.4B of equity, and left ~$135M/year of amortization that will drag GAAP earnings for years. Ten years on, ROIC has never exceeded 10.4%. Ball bought the industry’s leading position and did not earn a franchise return from it — which is the clearest possible evidence for the conclusion drawn above that scale in cans stops paying. The price paid capitalized the benefits to the seller. Today’s near-zero tangible equity is the fossil record of that transaction.

The aerospace divestiture (announced Aug 2023, closed 2024-02-16, $5.6B) — the best decision management made. Selling aerospace to BAE at $5.6B (~$4.67B pre-tax gain, ~$4.5B after-tax cash) was, in my judgment, straightforwardly good capital allocation: it monetized a business at a full price, in a strategically unrelated segment, into a strong buyer’s market for defense assets. Roughly $2B went to net-debt paydown and ~$2B to buyback. The criticism is not the sale; it is what the proceeds were used to obscure.

The buyback: the right tool, applied with the wrong funding, at the wrong time.

The 2024–25 program was enormous — $3.0B of repurchase across two years, against combined FCF of ~$558M. Ball’s own FY24 10-K language is candid: the buyback was “completed using cash on hand, OCF, proceeds from the sale of businesses and available borrowings.” In plain terms, Ball bought back ~$3B of stock with divestiture proceeds and incremental debt — total debt rose $1,319M in FY2025 alone, almost exactly matching the $1,321M repurchased. This retired 24% of the share count and mechanically produced the “10%+ EPS growth” that re-rated the equity from ~$45 to ~$61.

Two judgments follow. First, on price discipline: the bulk of the repurchase was executed in 2024–25 at prices averaging well above the $41–50 range where the stock traded in 2023 — Ball bought least when its stock was cheapest (share count fell only ~1.2–1.6%/year through 2023, when it traded at $41–50) and most when it was dearer. That is the standard procyclical error, and it is expensive at a company whose ROIC is ~9%: repurchasing stock is only value-creative below intrinsic value, and Ball’s own returns say intrinsic value is not generous. Second, on sustainability: the funding source was non-repeatable. There is no second aerospace to sell.

The engine has now stopped, and this is the most important forward-looking fact in the report. The 2026 plan cuts buyback to ~$600M — less than half the 2025 pace — as leverage takes priority (targeting ~2.7x net debt/EBITDA). More tellingly, shares outstanding rose from 265,374,186 at YE2025 to 266,246,511 at 2026-05-01, as option-exercise issuance outran repurchase in Q1’26. The share-count tailwind that supplied essentially all of FY25’s EPS growth has gone to zero and briefly reversed. Ball still guides to “10%+ comparable EPS growth” for 2026 — but with the buyback contributing only ~3.8%, that guide now requires ~6.2% from operations, against Q1-2026 volume of +1%. That is a large, unexplained gap and it is the load-bearing assumption of the bull case.

Dividends. ~$220M/year, ~$0.80/share, a ~1.3% yield at a ~24% payout. Conservative, well-covered, unremarkable. Not the reason to own this.

M&A discipline post-Rexam: good. Recent deals are small and sensible — Benepack (Jan 2026, $218M/€184M for 80% of ORG’s European can business, only $95M in cash), Florida Can (Q1 2025), Alucan Entec (2024, PHC). At ~$159M/year of acquisition cash against a $16B market cap, this is footprint maintenance. Management has not repeated the Rexam-scale swing, which is to its credit.

Capex discipline: improving, and a genuine positive. Ball spent $474M in FY25 against $622M of D&A, and guides 2026 capex to ~D&A ($600–650M). For a business at ~1% volume growth, in an industry whose defining error is capitalizing every boom, spending at maintenance levels is exactly right — and it is meaningful evidence that management learned from 2020–23. This is the strongest single item in the capital-allocation column: after building ~30% of new NA capacity into a fad and then closing plants, Ball is now refusing to chase. Whether this discipline survives the next boom is the open question, but the current behavior is correct.

Incentives: EVA-based, with a reset worth scrutinizing. Compensation is EVA-based, which is philosophically the right metric for a ROIC≈WACC business — it charges management for capital, and Ball is a company that has historically destroyed value by deploying capital carelessly. Credit where due. But two details temper it:

• In 2025 Ball redefined EVA’s invested capital from the “Net Assets Method” to a “Financing Approach (Net Debt + Equity)” post-aerospace. Any redefinition of an incentive metric’s denominator during a period of large buybacks deserves scrutiny: shrinking equity via repurchase mechanically reduces the “Financing Approach” capital base, which raises EVA without any operating improvement. A buyback-funded EVA improvement is not value creation, but under this definition it can pay out like one. I flag this as a genuine misalignment risk rather than a proven abuse — I have no evidence it was designed for that purpose.
• The 2025–2027 PSU target is $131M of “New EVA” dollars for 100% payout (+$22M/+20% over the reset 2024 baseline), with stretch at +$74M/+68%. On a $22.1B enterprise, a $131M EVA target is thin — it is consistent with, and effectively institutionalizes, a business expected to earn roughly its cost of capital.

Management and insider behavior. CEO transition: Daniel Fisher stepped down 2025-11-10; Ronald J. Lewis (59), previously Chief Supply Chain & Operations Officer, became CEO the same day; CFO is Daniel J. Rabbitt (57). Lewis is an operations insider, not an outside change agent — a signal of continuity, which for a business needing a return-profile change is neutral-to-negative. Compensation optics are poor at the margin: Lewis’s FY25 total comp was $4,218,225 (earned mostly as COO, and below his 2024 pay as an SVP), while departing CEO Fisher took $13.49M — above his 2024 $12.33M in the year he stepped down.

The insider read across the full Form 4 corpus (182 filings, Jan 2024–Jul 2026) is neutral, mildly constructive: codes break down A=148, M=155, F=39, S=5 sales, P=3 purchases, with zero transactions flagged 10b5-1. Of the three P-codes, two are not conviction buys (Fisher’s 3,869 @ $64.56 was an automatic 401(k) DRIP; Glew’s 563 @ $68.15 was an EMEA share-plan purchase). The only genuine discretionary open-market purchase in two and a half years is Villatoro (SVP South America), 1,551 shares @ $64.51 on 2026-03-04 — roughly $100K, at the divisional level. No open-market buying by CEO Lewis, CFO Rabbitt, or any director. Offsetting that, selling is negligible — ~$2.3M across 2.5 years, mostly exercise-and-sell, and the sitting CEO and CFO have sold nothing. Ownership is modest: Lewis 74,201 shares, Rabbitt 27,709. Read fairly: insiders are not signaling distress, and they are not signaling conviction either. Nobody with a view is buying this stock at $61–68.

Verdict: MIXED, trending toward disciplined. The honest scorecard: Rexam was strategically successful and economically expensive, and its cost is still on the balance sheet. The aerospace sale was excellent. The buyback was the right instrument — for a no-growth, ROIC≈WACC business, returning capital is genuinely better than reinvesting it — but it was funded by a one-time windfall and debt rather than by FCF, executed procyclically, and marketed as “10%+ EPS growth” when the underlying earnings were flat at +0.8%. Current capex discipline and post-Rexam M&A restraint are real positives that deserve credit. The verdict is not that management has misallocated capital — on balance it has allocated reasonably given a weak hand — but that management has allowed a mechanical, non-repeatable, debt-and-divestiture-funded share-count reduction to be presented as an operating growth algorithm. With that funding exhausted and the share count now rising, the presentation and the reality are about to converge.

8. Changes and Headwinds — Last Two Years

Portfolio transformation (2024–2026). The defining change is the completed pivot to a pure-play. Aerospace closed to BAE on 2024-02-16 ($5.6B; ~$4.5B after-tax cash; $3,584M booked as a discontinued-operations gain, distorting FY24 GAAP EPS to $13.00 against $1.37 of continuing earnings). Alongside it: the cups business was deconsolidated 2025-03-21; the Saudi operation was deconsolidated 2025-08-27 (selling 41% of a 51% stake, retaining 10%); and Benepack was acquired in January 2026 (80% of ORG’s European can business; Belgium and Hungary; $218M/€184M, only $95M cash). Net: a simpler, more focused, more geographically European business. Thesis effect: strengthening. Focus is good, the aerospace price was full, and the Saudi/cups exits removed distractions.

Leadership change (Nov 2025). Fisher out, Lewis (ex-COO) in, effective 2025-11-10. Thesis effect: neutral. Continuity, not change; an operator promoted into a business whose problem is not operations.

Capital-structure shift. Total debt rose from $5,693M (2024) to $7,012M (2025) to $7,860M (Q1’26), against $3.0B of 2024–25 buyback. Ball refinanced actively in 2025 — issuing €850M 4.25% 2032 and $750M 5.5% 2033; redeeming 6.875% 2028 ($750M) and 4.875% 2026 ($256M); repaying 5.25% notes ($189M), the $600M revolver, and Term A — extending maturities at a ~4.5% blended coupon with revolvers undrawn to June 2030. Competent liability management. But the direction is unmistakable: Ball levered up to buy stock, and it did so as a BB+/Ba1 credit. Thesis effect: weakening. The 2026 pivot to deleveraging (~2.7x target, buyback cut to ~$600M) is the correct response and confirms the constraint is now binding.

Tariffs and input costs (2025–26). Section 232 aluminum tariffs doubled 25% → 50% effective 2025-06-04; the Midwest Premium exceeded $1/lb for the first time in late January 2026. Ball’s CFO expects “some direct tariff cost in 2026,” with metal cost largely passed through. Thesis effect: modestly weakening. The pass-through holds the margin, but two second-order effects bite: a recovery lag on cost timing, and — strategically — a higher delivered can price versus PET and glass, eroding precisely the substrate-shift advantage the bull case leans on. The tariff protects the converter and taxes the product.

The capacity cycle turned favorable (2025–26). After the 2022–23 bust — plant closures at St. Paul, Kent WA, and Phoenix, and $420M of FY24 business-consolidation charges — NA operating rates recovered to the mid-90s%, Europe outperformed in 2025, and supply discipline returned across the industry, with Ardagh Metal Packaging financially constrained at ~7.1x net debt/EBITDA. Thesis effect: strengthening — the most genuinely positive development in the file. This is the favorable half of the Marathon cycle, and it drove ROIC’s recovery from 7.4% (2024) to 9.1% (2025). The caveat is what it reveals: mid-90s% utilization with disciplined competitors still only yields 9.1%.

Business-consolidation accounting swing. The line ran -$133M (2023) → -$420M (2024) → +$41M credit (2025) — a $461M year-over-year swing into a credit in a year of strong reported results. Thesis effect: neutral, but worth flagging. Not aggressive, but it means “comparable” has been excluding large charges for years and then booked a reversal when the cycle turned.

Headwinds into 2026, ranked by materiality.

1. The buyback cliff. ~$600M planned versus $1,321M (2025) and $1,712M (2024) — with shares outstanding already rising in Q1’26. The EPS algorithm’s engine is off.
2. Volume at +1% (Q1’26 global), at or below the bottom of Ball’s own +1–3%/+3–5%/+4–6% algorithm, while the company guides 10%+ EPS growth.
3. Sub-IG rating (BB+/Ba1) with seasonal net debt at ~$7.1B (~3.3x) — leverage constrains the return of capital exactly when the multiple depends on it.
4. Tariff pass-through timing and the competitive-substrate effect of a >$1/lb Midwest Premium.
5. Customer renewal risk on the 2027+ book, unquantifiable from disclosure.

Verdict: on balance NEUTRAL, with the composition deteriorating. The portfolio is cleaner and the industry backdrop is the best in five years — both genuine positives that strengthen the near-term earnings picture. But the two changes that actually drove the stock’s re-rating (the aerospace windfall and the debt-funded buyback) are complete and non-repeatable, the balance sheet is more levered and rated below investment grade, and the growth algorithm now has to be delivered by an operating business growing volume at 1%. The last two years improved the business; they did not improve the returns, and they consumed the balance-sheet capacity that made the last two years’ EPS growth possible.

9. Risk Analysis

Reading the matrix. The risks that matter are #1, #2, and #4 — and they are correlated, which is the point. Ball’s equity is priced on a 10%+ EPS algorithm; that algorithm was ~90% share-count shrink; the shrink is over; and if the operating business cannot fill the gap at +1% volume, the market re-rates a 9.1%-ROIC converter toward the ~9.1x multiple its higher-returning competitor already trades at. Risk #4 is the tail that makes the cycle rhyme: the very tightness that supports today’s earnings is the signal that historically triggers the industry’s next overbuild.

What is genuinely low-risk. Ball will not be disrupted, displaced, or destroyed. The moat is real; the product is essential and consumable; the customers are permanent; the balance sheet, while sub-IG, is well-termed with no near-term wall and undrawn revolvers to 2030. Risk of permanent capital impairment is low; risk of permanent capital stagnation — a decade of ~9% ROIC and no multiple support — is the real exposure. The five-year record (-4.4% annualized, -55% max drawdown) is what that stagnation looks like when purchased at the wrong price.

10. Valuation Discussion

No price target and no recommendation appears in this section. It analyzes embedded expectations and scenarios only.

Striking the numbers correctly — a necessary correction first. Widely-circulated figures for Ball’s EV (~$20.3B, ~10x EBITDA) are stale: they were struck at the $52.97 year-end-2025 close, not at the current $61.30. Restruck properly:

Against FY25 EBITDA of $2,011M, that is ~11.0x; against comparable EBITDA of ~$2.15B, ~10.3x. On comparable diluted EPS of $3.57, the P/E is ~17.2x. These are the correct multiples, and they are ~8–10% higher than the figures the stale EV implies.

The own-history percentile is a trap — this is the report’s key valuation gotcha. AZI’s valuation index places Ball at the 20.9th percentile of its own ten-year range (composite; P/E 13.7th, P/B 20.3rd, P/S 28.7th). Read naively, Ball is in the cheapest quintile of its decade. Three reasons that reading is wrong:

1. The distribution’s upper half is a bubble. The 2020–21 seltzer peak saw Ball at 52x/36x P/E and 19.2x/18.3x EV/EBITDA — multiples the 2022–23 bust proved were built on a fad. Measuring today against a discredited peak manufactures cheapness. A percentile is only meaningful if the reference distribution is legitimate.
2. The P/B component is meaningless. Tangible common equity is ~$60M (~$0.23/share) and was negative in 2022–23 (see Financial Quality). A P/B percentile computed on ~99%-intangible equity carries no information, and it contaminates the composite.
3. The pre-bubble comparison is damning. Ball’s pre-Rexam 2014 multiple was ~10.5x EV/EBITDA — essentially where it trades today. Strip out the bubble years and Ball is not at the cheap end of its history; it is at its normal level, for a business earning less than it did then.

Embedded expectations — what the market is underwriting at $61.30. This is the core analysis. Normalized EBIT is $1,555M (reportable comparable operating earnings less “Other”; cross-checks to GAAP operating income of $1,390M after Rexam amortization), giving NOPAT of ~$1,182M at a 24% tax rate. Reverse-solving EV of $22.1B on the Greenwald growth form yields three mutually exclusive readings:

That third line is the finding. At a house-convention 8% WACC and a realistic 2.5% terminal growth rate, $61.30 is arithmetically unreachable at any level of profitability whatsoever. For comparison, earnings power value — the zero-growth capitalization of normalized NOPAT — is $33.81/share at 8% WACC, meaning the market pays an 81% premium to EPV; even at a generous 6.5% WACC, EPV is $46.66 and the premium is 31%. Ball is nowhere near EPV support.

The bull case is a discount-rate assertion, not an operating one — and Ball’s own credit refutes it. The honest steel-manning: a bottom-up CAPM at Ball’s 0.68 beta computes to ~6.4% — almost exactly the market-implied 6.38%. At a 6.4% WACC, ROIC of 9.1% clears the hurdle by ~270bp, growth does create value, and the price is defensible. So the entire valuation debate reduces to a single question: does Ball deserve a ~6.4% cost of capital?

It does not, and the evidence is Ball’s own rating. Ball is BB+ (S&P) / Ba1 (Moody’s) — sub-investment-grade. A high-yield credit does not earn a 6.4% weighted cost of capital. Ball’s low equity beta reflects the demand defensiveness of selling cans for beer and Coca-Cola — people drink through recessions — and not the risk of the balance sheet financing those cans. Conflating demand stability with capital-structure safety is the specific error underpinning the bull case. A sub-IG issuer with ~3.3x seasonal leverage, ~99%-intangible equity, and 40% customer concentration is not a 6.4%-WACC asset. This article adopts ~7.5% as the base WACC, with 6.5% and 8.0% as explicit bookends — and notes plainly that the cheap-versus-expensive verdict flips on this choice, which is why it is stated rather than buried.

Comparables — the cheapness story does not survive contact with Crown.

The Crown comparison is the single most damaging fact in the valuation. Same product, same customers, same industry structure. Crown earns 350bp more ROIC and 230bp higher operating margin than Ball, and trades at a ~1.9-turn discount. Ball is the lower-quality asset at the higher price. There is no scale argument that rescues this — Ball is the scale leader, and it is the one under-earning. Amcor — a cost-plus converter earning ROIC ≈ WACC, and Ball’s closest economic twin — clears at ~9.5x. Ball trades ~1.5 turns above its twin and ~1.9 above its better.

Scenario analysis (3-year to FY2028E, PV discounted at 8%).

Bear assumes NA volumes flat-to-negative, tariff-lag margin compression, buyback constrained by deleveraging, and a de-rate to Amcor/Crown-like converter multiples. Base assumes Ball delivers its volume algorithm’s midpoint with flat margins, a ~$650M buyback, and a modest de-rate toward — but still above — Crown. Bull assumes the 2027+ renewals reprice favorably, margins expand 100bp, the tight capacity backdrop persists, and the market sustains a premium multiple.

Two things dominate. First, the exit multiple: one turn of EV/EBITDA is worth ~$8/share of PV, more than three years of operating improvement in the base case. This is a multiple story, not an earnings story — which is uncomfortable for a business whose multiple is already above a better competitor’s. Second, the skew is negative: base is -5%, bear is -37%, bull is +25%. Ball needs to hold ≥10x — a sustained premium to a higher-ROIC Crown — merely to return 8%. The skew inverts in the mid-$40s to low-$50s, where EPV at a 6.5% WACC ($46.66) and the converter multiple band (~8.5–9.5x) converge — the zone where an investor is paid for the risk rather than paying for the narrative.

On sell-side targets. RBC (Outperform, $77), Citi (Buy, $72) — both 2026-07-09 — and BofA (Buy, raised to $73, 2026-07-14) all sit well above the current price. These are recorded as market context only. They are recorded as context only: they are not this author’s view, they are not a target, and they are not evidence.

Verdict: NOT cheap — and the appearance of cheapness is an artifact of a contaminated reference distribution. Ball trades at ~11.0x EBITDA and ~17.2x comparable EPS, at an 81% premium to its 8%-WACC earnings power value, at a ~1.9-turn premium to a direct competitor earning 350bp more ROIC, and at roughly its own pre-bubble 2014 multiple while earning less than it did then. A reverse-DCF at a defensible cost of capital cannot reach the price at any level of profitability. What the market is underwriting correctly: the industry’s favorable capital-cycle position, restored supply discipline, real volume stability, and a genuine (if narrow) moat. What the market is underwriting incorrectly: that a 10%+ EPS algorithm — ~90% of which was a debt-and-divestiture-funded share-count reduction that has now stopped — is an operating property of the business; and that a sub-investment-grade, 40%-customer-concentrated, 9.1%-ROIC converter deserves a 6.4% cost of capital.

11. Variant Perception

The consensus belief. Sell-side is uniformly constructive (RBC $77, Citi $72, BofA $73 — all Buy/Outperform, all July 2026, versus a $61.30 price). The consensus narrative runs: aluminum cans are structurally winning share from glass and PET; the industry has consolidated to a rational oligopoly and rebuilt supply discipline after the 2022–23 washout; utilization is back to the mid-90s%; Ball is the global #1 with contracted volume “sold out” through 2026 and 90%+ through 2027; management guides 10%+ EPS growth with >$900M of FCF; and the stock trades at the 20.9th percentile of its own ten-year multiple range — a quality franchise at a cyclically depressed price.

Where consensus is right. Genuinely: the moat is real, the capital cycle is favorable right now, volumes are stable, the balance sheet has no near-term wall, capex discipline has returned, and the aerospace sale was well-executed. This is not a broken company and it is not a short. Ball will still be the world’s #1 can maker in a decade.

The strongest bull case, stated fairly. Ball is the scale leader in a consolidated, rational, capacity-constrained oligopoly at the favorable point of the capital cycle, selling a product with a durable secular tailwind. Its 0.68 beta and defensive demand profile genuinely justify a lower cost of capital than an 8% house convention — at ~6.4%, ROIC of 9.1% clears the hurdle, growth creates value, and the stock is fairly-to-attractively priced. The 2027+ contract book is being renewed into the tightest capacity backdrop in five years, which should reprice conversion spreads upward and finally break ROIC above its historical ceiling. Buyback plus 2–4% volume delivers 10%+ EPS growth, and the multiple holds because supply discipline is now structural rather than cyclical.

The strongest bear case, stated fairly. Ball is a commodity converter earning ~9% ROIC — its cost of capital — that has never cleared 10.4% even at the greatest demand peak in its history, selling a specification product to three customers who are 40% of sales and who can compute its cost base from the pass-through contract. Its reported growth is a fiction of arithmetic: comparable net earnings grew 0.8% in FY25 while comparable EPS grew 12.6%, the difference being an 11% share-count reduction funded by a one-time $4.5B divestiture and $1.3B of incremental debt — an engine that is now off, with shares rising in Q1’26. It trades at 11.0x versus Crown at 9.14x despite earning 350bp less ROIC, at an 81% premium to EPV, at a price a reverse-DCF cannot reach at an 8% WACC even with infinite profitability — and it is a BB+/Ba1 credit being valued as if it deserved a 6.4% cost of capital. The “cheap” percentile is measured against a seltzer bubble the company itself proved was fake.

The 3–5 assumptions that actually matter.

1. The cost of capital. ~6.4% (market-implied) versus ~8% (house). Everything turns on this. My read: the sub-IG rating settles it against the bull. Demand defensiveness ≠ balance-sheet safety.
2. Whether the 2027+ book reprices. The only mechanism that could break ROIC above ~10.4%. Unverifiable from disclosure — Ball publishes no backlog, contracted volume, utilization, or pricing. This is the bulls’ best argument and it is unfalsifiable, which is itself informative.
3. Whether ~6.2% of operating EPS growth can replace the buyback. The 2026 guide requires it against +1% Q1 volume. My read: the gap is large and unexplained.
4. Whether supply discipline is structural or cyclical. The industry converted the last boom into ~30% of new NA capacity. My read: Marathon’s prior applies — discipline at the bottom of the cycle is not evidence of discipline at the top.
5. Whether the multiple holds. One turn ≈ $8/share of PV. Ball must sustain a premium to a better-returning Crown just to return 8%.

The factor and positioning read — what the tape says about consensus. (Source: FactorsToday, pull 2026-07-13/14.) Ball is empirically a low-volatility, value-tilted defensive: market beta 0.68, with strongest loadings to Value +0.41, Sector:Materials +0.40, SmallSize +0.17, DividendYield +0.095, and — tellingly — Quality only +0.050; negative to Momentum -0.22, BetaFactor -0.22, and Growth -0.12 (r² 0.446, idiosyncratic vol 20.9%). The risk-adjusted record confirms a lost half-decade: y5 return -4.4% annualized, Sharpe -0.21, max drawdown -55%; y3 +3.6%; y1 +6.7%; lifetime +11.2%. The stock is -33.2% from its multi-year high but above both its 50-day ($59.12) and 200-day ($57.26) EMAs, with rs_6m +10.95 and rs_12m +6.66.

Read as evidence about consensus positioning, this cuts both ways, and honestly:

• It supports the bull’s framing: Ball is not a crowded momentum trade (Momentum loading is negative) and not a falling knife (6m and 12m relative strength positive, price above both EMAs, uptrend structure). A contrarian buyer here is early, not catching a knife, and is not fighting a mob. The tape is neutral-to-supportive.
• But it corroborates the bear’s economics: the market assigns Ball a near-zero Quality loading (+0.050). The factor model — which has no opinion about cans — independently classifies Ball as a value stock, not a quality stock. Investors are not paying a quality premium for this asset, which is precisely what a 9.1%-ROIC converter should command. The variant perception is therefore not “the market wrongly thinks this is low-quality.” The market has the quality right. The mispricing is that a correctly-identified low-quality asset is trading at a premium multiple to its higher-quality direct competitor.

The variant perception, stated plainly. Consensus and I agree on the business: a real moat, a favorable cycle, stable volumes, competent management. We disagree on what the reported growth was. Consensus treats “10%+ comparable EPS growth” as an operating characteristic of a franchise and capitalizes it accordingly. The primary-source evidence says it was ~90% share-count arithmetic, funded by a non-repeatable asset sale and incremental debt on a sub-investment-grade balance sheet, executed procyclically — and it has now stopped, with the share count actually rising in Q1’26. The stock re-rated from ~$45 to ~$61 on an algorithm whose fuel supply is exhausted. The bull case does not rest on an operating claim; it rests on the assertion that a BB+/Ba1 credit deserves a 6.4% cost of capital, and on a 2027+ repricing that the company declines to quantify. Consensus is not wrong about the can. It is wrong about the arithmetic.

What would falsify each side. Bear falsified: ROIC prints durably above ~12% on the 2027+ book — proving the repricing thesis and that pricing power has genuinely shifted to the converter. Bull falsified: FY2026 comparable EPS growth lands materially below the 10% guide on a ~$600M buyback, exposing the operating base as flat. Both are testable within four quarters, as set out in What Must Be True.

12. Fact vs. Interpretation

13. Open Questions

1. How is the 2027+ contract book being priced? The single most important unanswerable question in this file. Ball says it is >90% contracted for 2027 and >50% through the decade — but “contracted” is not “priced well,” and Ball discloses no backlog, no contracted volume, no utilization, and no pricing. If renewals into a mid-90s%-utilization market are lifting conversion spreads, ROIC breaks above its historical ceiling and the bull case is right. If they merely extend current spreads with customers who can compute Ball’s cost base, nothing changes. No outside party can distinguish these from disclosure. That a company would emphasize being “sold out” while publishing nothing that would let anyone verify or price it is itself a data point.
2. What explains the ~6.2% operating EPS growth implied by the 2026 guide against +1% Q1 volume and flat comparable net earnings? Is it price/mix, tariff recovery, Benepack accretion, cost-out, or optimism? Management has not bridged it.
3. Why does Ball, the scale leader, earn 350bp less ROIC and 230bp lower operating margin than Crown? This article’s answer is that scale in cans stops paying above the co-location threshold — but a mix, footprint, or contract-structure explanation could exist. Resolving this would materially sharpen the moat verdict either way.
4. Was the 2025 EVA redefinition to the “Financing Approach (Net Debt + Equity)” analytically motivated (post-aerospace simplification) or incentive-motivated (a buyback-shrunk denominator flattering payout)? Flagged as a risk; not proven.
5. What is the true normalized capex for a business at ~1% volume growth? FY25’s $474M was ~$150M below D&A and flattered FCF; the 2026 guide of $600–650M ≈ D&A. Which is the run-rate — and is the 2020–23 overbuild still absorbing what would otherwise be growth capex?
6. How much of FY25 OCF is a sustainable payables position versus a stretch that reverses? The +$716M AP build against a -$453M inventory build is a large swing in a year Ball marketed “record” FCF.
7. Does supply discipline survive the next upswing? Marathon’s prior says no; the 2020–21 evidence says no; management’s current capex behavior says maybe. Unresolvable ex ante, but it is the difference between a 9x and an 11x multiple.
8. What is the tariff’s net competitive effect on substrate share? The pass-through protects Ball’s margin, but does a >$1/lb Midwest Premium push marginal volume back to PET? No disclosure quantifies can-vs-PET price elasticity at the filler.
9. Is the FY25 comparable-EPS growth rate cleanly comparable? FY24’s $3.17 base includes aerospace through 2024-02-16. The +0.8% net-earnings/+12.6% EPS divergence is directionally unambiguous, but the precise rate is not.
10. Why has no director, the CEO, or the CFO bought a single share in the open market in two and a half years, at prices from $41 to $68 — while the company was repurchasing $3.0B of stock on their authorization? The company’s conviction and the individuals’ conviction are not aligned.

14. What Must Be True

The Bull Case — What Must Be True

1. Ball deserves a cost of capital near ~6.4%. This is the load-bearing assumption; at 8% the price is unreachable at any ROIC. It requires that Ball’s 0.68 beta reflects genuine low risk rather than demand defensiveness sitting atop a sub-IG balance sheet.
2. The 2027+ book reprices upward, breaking ROIC durably above its ~10.4% historical ceiling for the first time in the company’s modern history.
3. The operating business delivers ~6%+ EPS growth without the buyback — roughly 6x the +1% volume currently being delivered.
4. Supply discipline is structural, not cyclical — the industry does not convert the current tight market into capacity, as it did in 2020–21.
5. The multiple holds at ≥10x, sustaining a premium to a higher-returning Crown.

Falsification test (bull): FY2026 comparable diluted EPS growth comes in materially below the 10% guide — call it under ~7% — on a ~$600M buyback, with ROIC failing to exceed ~10%. That would demonstrate the operating base is flat and the algorithm was arithmetic. Testable within four quarters, at the FY26 print (early February 2027), with a first read at Q2’26 (late July 2026) and a clear one by Q3’26. Corollary tests: share count fails to fall meaningfully from 266.2M; or net debt/EBITDA fails to reach the guided ~2.7x, proving the deleveraging is crowding out the return of capital.

The Bear Case — What Must Be True

1. ROIC stays at ~9–10% — the 2025 recovery is cycle, not structure, and the ceiling holds as it has for eight years.
2. The share-count tailwind stays off. The ~$600M pace persists or shrinks under the sub-IG deleveraging priority, and shares do not resume a meaningful decline.
3. Customers retain the bargaining power — top-3 at ~40% of sales continue to capture the surplus at each renewal, as the pass-through structure and the ROIC record both imply.
4. The multiple converges toward converter norms (~9–9.5x) — Crown’s and Amcor’s levels — as the EPS algorithm’s arithmetic becomes visible.
5. Volume remains GDP-linked at ~1–2%, with the tariff-inflated can price blunting substrate-shift gains.

Falsification test (bear): ROIC prints durably above ~12% — Crown’s level — for four consecutive quarters, accompanied by comparable operating-earnings growth above ~6% on flat-to-modest volume. That combination could only come from genuine conversion-spread expansion on renewals, which would prove pricing power has shifted to the converter and that the moat has begun producing rent rather than merely stability. Testable across FY2026–27 as the 2027 book is priced; an early tell would be segment comparable operating margin breaking above ~13.5% (versus 12.1% blended today) without a commensurate volume surge.

The pivot both cases share. Bull and bear reduce to the same question asked two ways: is Ball’s ~9% ROIC a property of the business or a property of the cycle? Eight years of data — including a peak of 10.4% at the top of the greatest boom in the industry’s history, when it claimed to be sold out for three years — say it is a property of the business. If that is right, the current price requires a discount rate the company’s own credit rating contradicts, and the growth that justified the re-rating already happened, in the form of a share count that has stopped falling.

15. Source Appendix

The complete source appendix — every source with title, URL, date accessed, and the claims it supports, grouped by SEC primary filings, company materials, industry/trade sources, market data, and news — is reproduced as Appendix B below.

Primary sources relied on most heavily:

• Ball Corporation FY2025 Form 10-K (filed 2026-02-19) — segments, customers, market shares, balance sheet, cash flow, risk factors, Note 4 (acquisitions/dispositions)
• Ball Corporation Q1-2026 Form 10-Q (filed 2026-05-05) — share count (266,246,511 at 2026-05-01), total debt $7.86B, Q1 segment results
• Ball Corporation Q4-2025 earnings release, Ex-99.1 to Form 8-K (2026-02-03) — comparable diluted EPS $3.57, comparable net earnings $985M, 2026 guidance
• Form 8-K Item 2.01 (2024-02-16) — completion of the BAE aerospace disposition
• DEF 14A proxy (filed 2026-03-17) — EVA compensation design, invested-capital redefinition, Summary Compensation Table, CEO transition
• Form 4 corpus (182 filings, January 2024 – July 2026) — insider transaction analysis
• ROIC.ai — profitability ratios, enterprise value, valuation multiples, Crown Holdings comparables (third-party aggregated; reconciled to filings; ROE and book-value-per-share confirmed erroneous and discarded)
• Public market data — 5-year price/OHLCV history, valuation-multiple percentiles, news
• FactorsToday — factor loadings, leaderboard, stock info, related stocks

All analysis in this article derives from public primary sources.

Sections 1–15 contain no recommendation and no price target; the Claude's Take block above is the author’s own labeled opinion, offered as general information and not as investment advice.

APPENDIX A — Standard Diligence Questionnaire

Ball Corporation (NYSE: BALL) — Standard Diligence Questionnaire

Report date: 2026-07-14 · Price as-of: $61.30 (2026-07-13 close) · Supplemental to the main article.

Answers are grounded in the primary filings. Fact / Interpretation / Assumption labels are applied where the distinction matters. No price target or recommendation appears in this appendix.

General

What thoughtful questions have other investors asked about this company?

The sell-side conversation (RBC Outperform $77, Citi Buy $72, BofA Buy $73 — all July 2026) is overwhelmingly focused on three things: the durability of the substrate-shift story, the pace of volume recovery post-2023, and whether supply discipline holds. Those are the right topics and, in my judgment, the wrong emphasis.

The genuinely thoughtful questions — the ones that actually determine the outcome — are asked less often:

1. “Why does the scale leader earn less than the #2?” Ball is #1 in all three regions and earns 9.1% ROIC and a 10.6% operating margin against Crown’s 12.6% and 12.9%. Almost nobody asks this out loud, and it is the single most diagnostic fact about the business (Interpretation: scale in cans stops paying above the plant-co-location threshold, because the competitive unit is the plant, not the company).
2. “What was the EPS growth actually made of?” FY25 comparable net earnings grew 0.8% while comparable EPS grew 12.6% (Fact). The gap is share count. Very little published work makes this distinction, and the entire re-rating from ~$45 to ~$61 rests on the conflation.
3. “How is the 2027+ book priced, not just contracted?” Management emphasizes being “sold out”; investors rarely press on spread. Ball discloses no backlog, volume, utilization, or pricing (Fact), so the question cannot be answered from outside — which is itself informative.
4. “Why has no insider bought a share?” Across 182 Form 4s and two and a half years spanning prices from $41 to $68, the only genuine discretionary open-market purchase is ~$100K by a divisional SVP (Fact).

The bear questions that deserve more airtime: whether tariffs blunt the substrate advantage they are meant to protect, and whether the 2025 EVA redefinition rewards buyback-driven capital shrink.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

Neither extreme — but closer to a cyclical high than the “depressed” framing suggests (Interpretation). ROIC of 9.1% (FY25) sits against a 2018–25 range of 7.2–10.4%: above the 2022–24 trough (7.2–7.4%), below the 2021 peak (10.4%). North American operating rates are back to the mid-90s% and supply discipline has returned — this is the favorable half of the capital cycle, not the unfavorable one (Fact/Interpretation).

The crucial framing: this is roughly as good as it gets for Ball. The company’s peak ROIC in the greatest demand boom in its industry’s modern history — when it claimed to be sold out for 24–36 months — was 10.4%. Earnings are cyclically decent and structurally capped.

Driven by the external environment or internal actions?

Predominantly external (Interpretation). The FY25 improvement traces to industry capacity tightening, the post-destock volume rebound, and aluminum pass-through — not to margin expansion Ball engineered. Gross margin has been flat at ~19–21% for eight years through an $8.4B transformational acquisition and $5B of revenue growth (Fact). The one genuinely internal driver of per-share results was the buyback — an allocation decision, not an operating one, and funded externally by divestiture proceeds.

The internal action that is real and creditable: capex discipline, now at ~D&A ($600–650M guided vs. $622M D&A) after the 2020–23 overbuild (Fact).

How stable are revenues?

Volume is very stable; reported revenue is not, and revenue is a low-information number here (Interpretation). Net sales ran $12,062M (2023) → $11,795M (2024) → $13,161M (2025). Those swings are destock/restock, aluminum pass-through, and FX — not demand. Roughly half of FY25’s segment growth was price/mix, substantially pass-through, i.e. revenue arriving with its own cost attached (Fact, from MD&A decomposition).

Underlying demand is genuinely stable — cans for beer and Coca-Cola are consumed through recessions, which is why the equity beta is 0.68. But stable ≠ growing: Q1’26 global volume was +1% (Fact), at or below the bottom of Ball’s own +1–3% / +3–5% / +4–6% regional algorithm.

Outlook for products/services?

Structurally sound, modestly growing. Cans benefit from recyclability (aluminum’s global recycling rate ~75%; Ball’s cans average 74% recycled content), the billboard effect, and energy-drink growth, which is can-native (Fact). Against that: 50% Section 232 tariffs and a >$1/lb Midwest Premium raise the delivered can price versus PET and glass, blunting the substrate advantage (Interpretation). Net: a modest tailwind worth perhaps a point of volume.

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

The global aluminum can market is sized at ~$42B (2025) growing ~4.7% CAGR to ~$57B by 2034 (Coherent Market Insights) — but this is vendor sizing with no published methodology and should be treated as an Assumption, not evidence. The reliable anchor is the 10-K’s unit data (Fact): NA ~139B units, EMEA ~97B, SA ~43B.

International, and increasingly so. US sales are $6,163M of $13,161M (46.8%); Brazil $1,494M; other markets $5,504M (Fact). EMEA is 30.3% of sales and South America 16.4% — and SA carries the highest margin (15.1%) and fastest algorithm (+4–6%). Growing, but at GDP-plus-a-bit.

Business Quality & Competitive Moat

Is the industry getting more or less competitive?

Less competitive right now; structurally prone to becoming more so (Interpretation). Near-term, discipline has improved: Ardagh Metal Packaging is financially constrained at ~7.1x net debt/EBITDA and cannot add capacity, utilization is in the mid-90s%, and the industry is not building (Fact).

But the 2020–23 episode is the base rate. The industry declared itself “sold out 24–36 months” (Credit Suisse, Oct 2020), then added ~30% to North American capacity, then collapsed — Ball closing St. Paul, Kent WA, and Phoenix and taking $420M of FY24 consolidation charges (Fact). The 10-K’s own risk factor concedes overcapacity “may develop… even if demand continues to grow.” Marathon’s prior applies: discipline at the bottom of the cycle is not evidence of discipline at the top.

How profitable is the business (ROIC, ROE)?

ROIC: 9.1% (FY25); 7.2–10.4% across 2018–25; never above ~10.4% (Fact) — against a WACC of ~7.5–8%. This is a business earning approximately its cost of capital.

ROE requires a warning. ROIC.ai reports 6.87% ROE and $49.70 book value per share — both garbled; discard them (Fact of the error). Real equity is $5,421M / 265.4M shares = $20.43/share (AZI independently shows $20.94), and ROE on $912M of continuing net income is 16.8%.

But 16.8% is not a quality signal — it is a leverage artifact (Interpretation). It is high only because the denominator has been repurchased away: tangible common equity is ~$60M. A high ROE produced by shrinking equity with borrowed money measures financial engineering, not profitability. ROIC (9.1%) is the honest measure.

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

Competitors: five firms make substantially all NA volume; four in EMEA; four in SA. Globally: Ball, Crown, Ardagh Metal Packaging, CANPACK (Fact).

Barriers to entry are real and are the freight economics. An empty can is mostly air — cheap per unit, bulky per unit — so shipping it any distance destroys the economics. Plants are therefore built adjacent to customer filling lines, and the relevant market is “cans within economic truck distance of this filler.” Within that radius the co-located incumbent has a cost advantage no distant rival can match (Interpretation, strongly supported). The proof is share stability: Ball has been #1 in all three regions for decades with less than two points of drift (Fact) — close to dispositive evidence of a barrier.

And yet industry profitability is mediocre. Crown at 12.6% ROIC is the best of the group; Ball 9.1%; Amcor (an economic twin in cost-plus converting) ~8%; Ardagh distressed. A tight oligopoly behind real barriers earning ~8–12.6% ROIC tells you the barrier protects incumbents from each other but not from customers.

Can the business be easily understood?

Yes — one of its genuine merits. Ball buys aluminum coil, converts it to cans, and sells them under multi-year contracts to beverage fillers, passing metal cost through. The model fits in a sentence. There is no accounting exotica, no financialization, no opaque segment. The complications are three normalizations (the FY24 aerospace gain, Rexam amortization, and the FCF add-back), all disclosed and all straightforward once identified (Fact).

The irony worth noting: the same transparency that makes Ball easy for us to understand makes it easy for AB InBev to understand. A pass-through contract publishes your cost base to your counterparty.

Can it be undermined by foreign low-cost labor?

No — and this is the moat’s most durable feature (Interpretation). Freight economics make imports uneconomic: you cannot ship empty cans from a low-wage country and compete with a plant attached to the filling line. Labor is also a modest share of conversion cost in a highly automated process. Section 232 tariffs (50% since 2025-06-04) reinforce the barrier further.

This is genuine — Ball’s moat is safe from the offshoring risk that has hollowed out much of industrial manufacturing. It is simply not worth much, because the threat Ball faces is not a foreign entrant; it is the customer across the table.

Do brands matter?

No. Ball’s brand is irrelevant to the end consumer, who sees Coca-Cola’s or Red Bull’s branding on the can. To the customer — a professional procurement organization — Ball’s name carries reputational weight for reliability and quality but zero pricing power (Interpretation). A 12-ounce can is a specification. If the brand mattered, Ball’s margins would exceed Crown’s; they are 230bp lower (Fact).

What is the nature of competition?

Bilateral bargaining over conversion spread, not price competition for share (Interpretation). Share barely moves, so competition is not fought by underpricing to win accounts — it is fought at contract renewal, where a concentrated seller meets a more concentrated buyer. Ball’s top three are ~40% of sales: AB InBev 15%, Coca-Cola Bottlers’ Sales & Services 14%, Red Bull 11% (Fact, verified verbatim).

The asymmetry is structural: Ball’s plant is sunk and immobile and worthless without that customer’s volume, while the customer can credibly shift incremental volume to Crown or CANPACK at renewal. The co-location that creates Ball’s advantage over rivals simultaneously creates the customer’s advantage over Ball. Differentiation exists at the margin — printing quality, specialty formats, delivery reliability — and is worth basis points, not multiples.

Customers’ switching costs?

Real but modest, and frequently overstated (Interpretation). Qualifying a new supplier means new logistics, specifications, and supply risk on a line that cannot stop — genuine friction, and it underpins the 10-K’s “high customer retention.”

But retention is volume recurrence, not price lock-in, and the distinction is everything. The customer must come back; it comes back with leverage. And for a customer buying enough volume to fill a plant, the switching cost is small relative to the spread at stake — and it can always move incremental volume without moving the base. The ROIC record is the proof: if switching costs conferred pricing power, eight years would have shown it above 10.4%.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet?

Few of consequence (Interpretation). The plant network’s co-location value — the actual moat — is an economic asset carried at depreciated cost, arguably understated. Long-standing customer relationships are not capitalized (except as acquired Rexam intangibles).

Against that, the balance sheet arguably carries assets it should not: $4,379M of goodwill and $982M of other intangibles from the 2016 Rexam deal — ~99% of the $5,421M of equity (Fact). The honest read is not hidden value but the opposite: tangible common equity is ~$60M (~$0.23/share), and was negative in 2022 (-$1.9B) and 2023 (-$1.6B). P/B is meaningless for Ball and should not be used — which also means AZI’s P/B percentile (20.3rd) is uninformative and contaminates its composite.

Off-balance-sheet liabilities?

Nothing alarming surfaced in the corpus (Fact, as far as reviewed). Pension obligations are modest and shrinking: $43M contributed in 2025, ~$29M guided for 2026. Operating leases are ordinary-course. Aluminum derivative positions are disclosed and hedge a passed-through exposure.

The genuine off-balance-sheet item is contractual: Ball’s multi-year customer contracts commit it to volumes at spreads it does not disclose. Whether the 2027+ book is an asset or a liability is unknowable from disclosure (Open Question).

How conservative is the accounting?

Fair — not aggressive, but the “comparable” framing does real work (Interpretation). Three observations:

1. The business-consolidation line swung $461M — from -$420M (FY24) to a +$41M credit (FY25), in a year of strong reported results (Fact). Not aggressive, but it means “comparable” excluded large charges for years and then booked a reversal when the cycle turned.
2. “Record $956M adjusted FCF” = $788M actual + a $168M aerospace cash-tax add-back — a real cash outflow (Fact). Disclosed and arguably non-recurring, but “record” is doing work the underlying $788M does not support.
3. FY25 OCF of $1,262M includes a +$716M accounts-payable build against a -$453M inventory build (Fact). A meaningful slice of OCF is a payables stretch, which reverses.

Nothing here is deceptive. All of it is disclosed. But three separate normalizations are required to see the business, and all three cut the same way.

How CapEx-hungry is the business?

Moderately, and honestly so (Fact/Interpretation). Capex was $474M in FY25 against D&A of $622M — Ball under-invested relative to depreciation, flattering FY25 FCF. The 2026 guide takes capex to $600–650M, i.e., ~D&A — a ~$150M/year FCF headwind versus the 2025 run-rate.

Capex/sales of ~3.6–4.9% is not asset-light, but it is not brutal either. The right way to read it: for a business at ~1% volume growth, spending at maintenance is exactly correct, and it is the strongest evidence management learned from the 2020–23 overbuild. The risk is not that Ball spends too much today; it is that discipline may not survive the next boom.

Capital Allocation & Management

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

FCF is volatile and weaker than headlines suggest (Fact):

FY23 was inflated by the destock releasing working capital; FY24 was crushed by $766M of aerospace cash taxes into negative FCF; FY25’s $788M included the payables stretch.

The philosophy is explicit and, for this business, correct in principle: return capital rather than reinvest, because at ~9% ROIC reinvestment creates little value. Ball’s stated goal is >10%/yr comparable EPS growth, growing EVA dollars, and returning capital.

The execution is where it breaks. In FY24 Ball returned $1,932M against negative FCF; in FY25, $1,541M against $788M FCF while total debt rose $1,319M — almost exactly the $1,321M repurchased (Fact). Ball’s own FY24 10-K concedes the buyback used “cash on hand, OCF, proceeds from the sale of businesses and available borrowings.” The capital returned did not come from the business; it came from selling aerospace and from borrowing (Interpretation, well-supported).

Significant acquisitions recently?

Post-Rexam discipline has been good (Interpretation). Recent deals are small and sensible: Benepack (Jan 2026 — 80% of ORG’s European can business, Belgium and Hungary, $218M/€184M total but only $95M cash, the rest assumed debt); Florida Can (Q1 2025); Alucan Entec (2024, PHC). Cash for acquisitions was $159M in FY25 and $75M in Q1’26 — footprint maintenance against a $16.3B market cap.

The one that matters is a decade old. Rexam (2016, $8.4B) made Ball the global #1 and is the reason the balance sheet is ~99% intangible. Strategically it worked; economically it is hard to defend — ten years on, ROIC has never exceeded 10.4%. Ball bought the industry’s leading position and never earned a franchise return from it (Interpretation). The price capitalized the benefits to the seller.

Also relevant: disposals, where management did well. Aerospace to BAE (announced Aug 2023, closed 2024-02-16, $5.6B, ~$4.5B after-tax) was a full price for an unrelated business — the best capital decision in the file (Fact/Interpretation). Cups deconsolidated 2025-03-21; Saudi deconsolidated 2025-08-27.

Buying back shares?

Yes — massively, and that is the story. $1,712M (FY24) and $1,321M (FY25) — ~$3.0B across two years against combined FCF of ~$558M. A $4.0B authorization was approved 2025-01-29 with $2.93B remaining at YE2025 (Fact). Share count: 349M (2016) → 315.7M (2023) → 289.4M (2024) → 265.4M (2025), with the shrink heavily back-loaded at -8.3% in each of 2024 and 2025 versus ~-1.2–1.6%/yr before.

Two criticisms (Interpretation):

• Procyclical execution. Ball repurchased least when the stock was cheapest ($41–50 in 2023) and most in 2024–25 at materially higher prices. That is expensive at a company whose own ROIC says intrinsic value is not generous.
• Non-repeatable funding. There is no second aerospace to sell.

The engine has now stopped — the decisive forward fact. The 2026 plan cuts buyback to ~$600M, and shares outstanding rose to 266,246,511 by 2026-05-01 (from 265,374,186 at YE2025) as option issuance outran repurchase (Fact). Ball still guides 10%+ EPS growth, which now requires ~6.2% from operations against +1% volume.

Issuing large amounts of new shares to insiders?

No — dilution is modest and ordinary (Fact). The Form 4 corpus (182 filings, Jan-24–Jul-26) breaks down A=148 (grants), M=155 (option exercises), F=39 (tax withholding), S=5 (sales), P=3 (purchases). Grant activity is routine equity comp, not egregious issuance.

Worth noting: option-exercise issuance was large enough to outrun a ~$150M/quarter buyback in Q1’26 and push the share count up. At the current reduced repurchase pace, ordinary comp dilution now offsets a meaningful fraction of the program (Interpretation).

Compensation policy of directors/management?

EVA-based — philosophically right for this business, with a caveat worth watching (Fact/Interpretation). Charging management for capital is exactly correct at a company that has historically destroyed value deploying it. Credit where due.

Two tempering details:

1. In 2025 Ball redefined EVA’s invested capital from the “Net Assets Method” to a “Financing Approach (Net Debt + Equity)” post-aerospace. Any redefinition of an incentive denominator during a period of large buybacks warrants scrutiny: repurchasing equity mechanically shrinks the “Financing Approach” capital base, raising EVA with no operating improvement. A buyback-funded EVA gain is not value creation but can pay out like one. Flagged as a misalignment risk, not a proven abuse — I have no evidence of intent.
2. The 2025–27 PSU target is $131M of “New EVA” for 100% payout (+$22M/+20% over the reset 2024 baseline; stretch +$74M/+68%). On a $22.1B enterprise, that is a thin target — consistent with, and effectively institutionalizing, a business expected to earn roughly its cost of capital.

Optics are poor at the margin: Lewis’s FY25 comp was $4,218,225 — earned mostly as COO, and below his 2024 pay as an SVP — while departing CEO Fisher took $13.49M, above his 2024 $12.33M, in the year he stepped down (Fact).

Motivations of management?

Continuity, competence, and modest personal stakes — no evident empire-building, and no evident conviction (Interpretation).

Daniel Fisher stepped down 2025-11-10; Ronald J. Lewis (59), previously Chief Supply Chain & Operations Officer, became CEO the same day; CFO is Daniel J. Rabbitt (57) (Fact). Lewis is an operations insider, not an outside change agent — a continuity signal, which for a business whose problem is its return profile rather than its operations is neutral-to-negative.

The insider read is neutral, mildly constructive: across 182 Form 4s, zero transactions were 10b5-1, and of three P-codes two are not conviction buys (Fisher’s 3,869 @ $64.56 was an automatic 401(k) DRIP; Glew’s 563 @ $68.15 an EMEA share-plan purchase). The only genuine discretionary open-market buy in 2.5 years is Villatoro (SVP South America), 1,551 shares @ $64.51 on 2026-03-04 — ~$100K, divisional level (Fact). No open-market buying by CEO Lewis, CFO Rabbitt, or any director. Offsetting this, selling is negligible (~$2.3M over 2.5 years, mostly exercise-and-sell) and the sitting CEO and CFO have sold nothing. Ownership is modest: Lewis 74,201 shares, Rabbitt 27,709.

Read fairly: nobody is signaling distress, and nobody with a view is buying at $41–68 either — while the company spent $3.0B of shareholders’ money doing exactly that on their authorization (Interpretation).

Valuation & Market Data

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

No. Ball Corporation is a US-domiciled (Indiana-incorporated, Westminster CO-headquartered) C-corporation listed on the NYSE under BALL (formerly BLL). Ordinary common stock; standard Form 1099 dividend treatment; no K-1, no ADR mechanics, no partnership structure (Fact).

Dividend policy?

Conservative and unremarkable: ~$220M/year, ~$0.80/share, a ~1.3% yield at ~24% payout (Fact). Well-covered even on the honest $788M FCF. This is not an income story and it is not the reason to own the shares. Capital return runs through buyback, not dividend — roughly 6:1 in FY25.

How profitable is the business?

Covered above; the essentials: ROIC 9.1% (FY25), range 7.2–10.4% (2018–25), never above ~10.4% — against a ~7.5–8% WACC (Fact). Blended comparable segment operating margin ~12.1% (NCA 12.3%, EMEA 12.4%, SA 15.1%); GAAP operating margin 10.6%; gross margin ~19–21% and flat for eight years through an $8.4B acquisition and $5B of revenue growth.

The comparison that matters: Crown earns 12.6% ROIC and a 12.9% operating margin — 350bp and 230bp better than the larger, #1-everywhere Ball (Fact).

Is net income diverging from cash from operations?

Yes, materially and in both directions — and this is a key QoE issue (Fact):

FY2024 is the extreme case and it is entirely explicable: net income of $4,008M contained a $3,584M non-cash aerospace disposal gain, while OCF was simultaneously crushed by $766M of cash taxes on that same disposal. The gain was non-cash income; the tax was real cash out. Headline FY24 diluted EPS was $13.00 against $1.37 of continuing-operations EPS — a ~10x distortion that will corrupt any screen keying on GAAP EPS.

FY25’s OCF > NI is benign but partly a +$716M payables build that reverses. The divergences are explained, disclosed, and non-sinister — but they mean GAAP net income is unusable for Ball without normalization, and comparable EPS ($3.57 for FY25) plus actual FCF ($788M) are the right lenses.

Risks & Downside

What factors would cause the stock to decline?

In rough order of probability × impact (Interpretation, from the risk matrix above):

1. The EPS algorithm fails as the buyback steps down. The 10%+ guide needs ~6.2% from operations against +1% volume, with the share count already rising. The most likely disappointment and the one the market is least positioned for.
2. Multiple convergence toward converter norms. Ball at 11.0x versus Crown at 9.14x and Amcor at ~9.5x, while earning 350bp less ROIC than Crown. One turn ≈ $8/share of PV; a de-rate to peer levels is most of the bear case without any operating deterioration.
3. Customer renewal repricing. Top-3 at ~40% of sales, all able to compute Ball’s cost base from the pass-through contract.
4. Capital-cycle repeat — the industry converts tight utilization into capacity, as it did in 2020–21 (~30% of new NA capacity).
5. Volume disappointment if substrate shift proves GDP-linked rather than secular, aggravated by tariff-inflated can prices versus PET.
6. Leverage constraint — BB+/Ba1 with ~$7.1B seasonal net debt (~3.3x) crowding out capital return.

Risk of a catastrophic loss?

Low (Interpretation). No plausible mechanism: hard assets across three diversified regions; an essential, consumable, contracted product; ~4.5% blended coupon with maturities laddered to 2033 and revolvers undrawn to June 2030; no near-term wall; ~$1.2B of cash. A sub-IG rating (BB+/Ba1, both stable) raises refinancing cost at the margin but is nowhere near distress — Ardagh at ~7.1x is what distress looks like in this industry; Ball at ~2.7–3.3x is not.

The realistic downside is the bear scenario at ~$38/share (-37%) — a de-rate to 8.0x on flat volumes and compressed margins. That is a valuation drawdown, not an impairment of the business.

Chance of a total loss?

Effectively nil (Interpretation). Ball is a 146-year-old company holding the #1 position in all three of its regions behind a genuine freight-economics barrier, selling a product people consume through recessions, to customers who cannot stop buying it. There is no technological obsolescence path, no offshoring path, and no credible entrant. It will still be the world’s largest can maker in a decade.

The real exposure is not loss — it is stagnation. A decade of ~9% ROIC with no multiple support looks exactly like the last five years: -4.4% annualized, -55% max drawdown, Sharpe -0.21. That is what buying a good-not-wonderful business at a franchise price produces, and it is the risk actually being underwritten here.

Recent News & Events

Has the business environment changed recently?

Yes — favorably at the industry level, and it is the strongest genuine positive in the file (Fact/Interpretation). After the 2022–23 bust, North American operating rates recovered to the mid-90s%, Europe outperformed in 2025, supply discipline returned, and Ardagh is financially incapable of adding capacity. This drove Ball’s ROIC from 7.4% (FY24) to 9.1% (FY25). This is the favorable half of the capital cycle.

Unfavorably: Section 232 aluminum tariffs doubled 25% → 50% effective 2025-06-04, and the Midwest Premium exceeded $1/lb for the first time in late January 2026. The CFO expects “some direct tariff cost in 2026.” Metal cost is largely passed through, so the margin is protected — but the tariff raises the delivered can price versus PET and glass, blunting the substrate-shift advantage it ostensibly protects (Interpretation).

The caveat that reframes the good news: mid-90s% utilization with disciplined competitors still only produces a 9.1% ROIC. The best industry backdrop in five years is not enough to make this a franchise.

Significant acquisitions?

Yes, but small: Benepack (Jan 2026 — 80% of ORG’s European can business, $218M/€184M, only $95M cash, Belgium and Hungary); Florida Can (Q1 2025); Alucan Entec (2024). Cash for acquisitions $159M (FY25), $75M (Q1’26). Disposals were the larger story: aerospace to BAE closed 2024-02-16 ($5.6B); cups deconsolidated 2025-03-21; Saudi deconsolidated 2025-08-27 (Fact).

Change in accounting policies?

No accounting-policy change identified (Fact, as far as the corpus was reviewed). Two adjacent items are worth flagging:

1. The EVA incentive metric’s invested-capital definition changed in 2025 from “Net Assets Method” to “Financing Approach (Net Debt + Equity).” This is a compensation definition, not a GAAP policy — but it changes what management is paid to optimize (see above).
2. The business-consolidation line swung from -$420M (FY24) to a +$41M credit (FY25). Not a policy change — a reversal as the cycle turned — but it flatters FY25’s reported bridge.

Recent changes — new markets, facilities, management?

• Management (the big one): Fisher stepped down as Chairman & CEO 2025-11-10; Ronald J. Lewis (ex-Chief Supply Chain & Operations Officer) became CEO the same day; CFO Daniel J. Rabbitt. Continuity, not change (Fact).
• Facilities: the arc runs from building (Glendale AZ, Pittston PA in 2020–21) to closing (St. Paul, Kent WA, Phoenix in 2022–23) to maintaining (2026 capex guided to ~D&A). Benepack adds Belgium and Hungary.
• Markets: portfolio simplification to a pure-play packaging converter — aerospace, cups, and Saudi all exited within ~18 months, with European exposure added via Benepack.
• Capital structure: total debt $5,693M (FY24) → $7,012M (FY25) → $7,860M (Q1’26), with active 2025 refinancing (issued €850M 4.25% 2032 and $750M 5.5% 2033; redeemed 6.875% 2028 and 4.875% 2026) extending maturities at a ~4.5% blended coupon. Competent liability management — in service of levering up to buy stock (Fact/Interpretation).

Supplemental appendix to the Ball Corporation research article (report date 2026-07-14). Contains no recommendation and no price target.

APPENDIX B — Source Appendix

SOURCE APPENDIX — Ball Corporation (NYSE: BALL)

Report date: 2026-07-14 · Verification pass: 2026-07-17 Price basis: $61.30, close of 2026-07-13 (AZI CSV). Later closes: $60.46 (2026-07-14), $63.15 (2026-07-16). CIK: 0000009389

1. SEC Primary Filings

EDGAR base: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000009389

2. Company Materials (transcripts, IR, releases)

3. Industry / Trade Sources

Methodology caveat. Items 20–21 are vendor market-sizing studies published as marketing abstracts with no disclosed methodology. They are labeled ASSUMPTION / low-confidence and are not load-bearing. The memo’s market-size claims should rest on the 10-K’s own primary unit figures (139B / 97B / 43B containers), which are company estimates from “publicly available information and company estimates” and are disclosed as such.

4. Market Data

5. News

6. Verification status of load-bearing claims

Verified clean against the FY2025 10-K (exact reconciliation): net sales $13,161M and the $6,286 / $3,983 / $2,162 / $730 segment split (reportable $12,431); comparable segment operating earnings $772 / $495 / $327 → reportable $1,594M; the full segment→EBT bridge (Other −$39M, business consolidation +$41M, acquired-intangible amortization −$135M, interest −$314M, debt refi −$19M → EBT $1,128M); net income $912M; FY2024 discontinued-ops gain $3,584M (pretax disposal gain $4,634M); total equity $5,421M; goodwill $4,379M; other intangibles $982M → tangible common equity $60M; shares 265,374,186 (685,107,438 issued − 419,733,252 treasury); total debt $7,012M; cash $1,212M → net debt $5.80B; interest expense $314M; OCF $1,262M; capex $474M; buyback $1,321M; dividends $220M; $4.00B authorization (2025-01-29, through end-2027) with $2.93B remaining at 2025-12-31; customer concentration AB InBev 15% / Coca-Cola Bottlers’ Sales & Services 14% / Red Bull 11%; unit shares NA ~36% of 139B (~50B shipped), EMEA ~39% of 97B (38B), SA ~46% of 43B (~20B); PHC 1.5B units = 21% of 6.8B; cups deconsolidation 2025-03-21; CEO transition dates.

Corrected (did not reconcile) — see the 2026-07-17 research-log entry for full detail:

1. Aerospace close = 2024-02-16, NOT “Aug 2024.” Announced Aug 2023 (SPA filed as Ex-2.1 to the Q3’23 10-Q); closed 2024-02-16 per the 8-K Item 2.01, the 10-K MD&A, and the Q4’25 release.
2. FY2025 continuing-ops diluted EPS = $3.30, not $3.32 (basic continuing $3.33; discontinued = nil, so continuing = total).
3. FY2024 continuing-ops diluted EPS = $1.37, not $1.40 (basic continuing $1.39). QoE wedge is $13.00 vs $1.37.
4. FY2023 continuing diluted = $1.53 ($1.54 is the basic figure).
5. Credit rating is BB+ (S&P) / Ba1 (Moody’s), both stable — NOT “BBB.” One notch below investment grade. Not disclosed anywhere in the 10-K.
6. D&A basis mismatch (not an error): continuing-ops $622 / $611 / $605 (income statement + segment note) vs cash-flow-statement $622 / $620 / $686 including discontinued ops (disc-ops D&A $0 / $9 / $81 reconciles exactly). Label the basis.
7. EV $20.33B / ~10x is stale — that is the 2025-12-31 EV struck on the $52.97 year-end price. At the memo’s $61.30 the EV is ~$23.1B (ROIC Q1’26 TTM EV $23.05B; EV/EBITDA 11.2x).
8. Net debt $5.8B / ~2.9x is a 2025-12-31 point-in-time; at 2026-03-31 debt $7,807M and cash $730M → net debt ~$7.08B (~3.3x) on seasonal working-capital build plus Benepack assumed debt.

Late-disclosed items: Benepack (Jan 2026, 80% stake, $218M/€184M) and the Saudi/UAE deconsolidation (2025-08-27: sold 41% of a 51% stake in Ball United Arab Can Manufacturing Company, retained 10%) are disclosed in FY25 10-K Note 4 and are easily missed.

Open questions resolved: FY2025 comparable diluted EPS $3.57 (log derived ~$3.6–3.7); credit rating BB+/Ba1; CEO Lewis FY2025 total compensation $4,218,225.

Flagged as management framing, not evidence: the Q1-26 “sold out 2026, >90% 2027, >50% through decade” claim is unverifiable from any filing — the string “sold out” appears zero times in the FY2025 10-K, and Ball discloses no backlog, no contracted volume, and no utilization rate. The 10-K’s own risk factor cuts the other way: “overcapacity, which often leads to lower prices, may develop over time in certain regions in which we operate even if demand continues to grow.” Must be labeled INTERPRETATION / management hypothesis, never a moat or durability datapoint.

All figures reconcile to the cited primary filing unless expressly flagged above. Where a third-party aggregator and a filing disagree, the filing governs.