Coca-Cola: A Wonderful Company at an Unwonderful Price
The Short Version
Coca-Cola is the textbook case of a great business. It is also, at current prices, a textbook case of a great business that the market has already paid for. Three of our four gates pass with room to spare; the fourth — price — fails clearly, and fails on arithmetic rather than judgment.
Gate 1: Circle of Competence
Coca-Cola is not a beverage manufacturer. It is the owner of a brand and a concentrate. The company sells concentrate and syrup to independent bottlers in some 200 countries, licenses its trademarks, and lets the bottlers carry the capital-intensive work of bottling and logistics. The result is an asset-light earnings structure: a 61.6% gross margin, a 27.3% net margin, and capex under 5% of revenue ($2.1 billion against $47.9 billion in sales for 2025).
The portfolio extends beyond sparkling colas — Dasani and BODYARMOR in water and sports drinks, Costa in coffee, fairlife in dairy, a stake in Monster in energy — but the core is the same licensing machine.
Is this knowable? Two layers here. Information access: yes, trivially — SEC filings, earnings releases, call transcripts, even the public record of the IRS litigation. Predictability: also yes. The claim that people will still drink Coca-Cola-branded beverages in ten years, and that the company will keep the concentrate margin, is analysis, not a probability bet. A 137-year-old brand, a global distribution network, and a bottler system are not structures that rearrange themselves on a decade timescale. Whether growth comes in at 2% or 4% is uncertain — GLP-1 drugs and sugar taxes may shave volumes — but that is a question about the pace of growth, not survival.
This sits at the center of our circle: consumer monopoly, brand-based, sticky demand.
Gate 2: Moat
The moat has two sources. First, the intangible: the brand, the original specimen of what Buffett called share of consumer mind. Second, a blend of cost advantage and efficient scale — the largest beverage distribution system on earth, spanning 200 countries, is infrastructure that cannot be replicated.
The pricing power evidence is unambiguous. Q3 2025 delivered price/mix of +6% (roughly 4 points price, 2 points mix); Q1 2026 showed organic revenue growth of +10% with volume up +3% — the company raised prices and sold more. Through the inflationary stretch of 2021–2025, price increases stuck and the 61.6% gross margin held. Can it raise prices without holding a prayer meeting? Evidently.
Direction: holding, perhaps modestly expanding. Sparkling volume rose 3.5% in Q4 2025 (led by zero-sugar and flavored variants); the still category grew 8.7%. Zero-sugar grew 13% in Q1 2026, absorbing health-driven defectors inside its own portfolio — evidence the brand moat works even across category shifts. The caveat: developed-market sparkling volumes are stagnant in aggregate. The moat is not widening so much as redeploying territory within the castle.
On AI: neutral to slightly positive. This moat rests on physical distribution, brands, and consumer habit — none of the surfaces AI erodes. AI is a modest margin tool in demand forecasting and bottler logistics, available to competitors too. The real threat is GLP-1, not the machines.
Gate 3: Management
The capital allocation record is disciplined. Dividends have been raised for 63 consecutive years, with $8.8 billion paid in 2025. Buybacks since 1984 total 3.6 billion shares at an average price of $18.43 — emphatically cheap money — though recent net repurchases are small ($0.4 billion in 2025), which reads as sensible restraint at current prices rather than timidity. M&A has been measured: Costa ($4.9 billion, 2019) was mediocre but not destructive; fairlife (2020) was a success — the $6.1 billion contingent payment made in 2025 is, paradoxically, proof of how well that deal outperformed. No meaningful acquisitions in 2025, no signs of empire building.
The retained-earnings test passes. Most of the roughly $13 billion in net income goes out as dividends; what stays funds high-return categories like a $650 million fairlife plant. High payout, limited reinvestment needs — the correct posture for a mature business.
On candor: the Quincey team has acknowledged weak volume quarters (first-half 2025 in North America) without evasion and guided conservatively. The IRS dispute and its potential liability — up to $14 billion in additional payments — is disclosed explicitly in the 10-K. One deduction: the handling of the fairlife supply-chain (animal welfare) litigation was inadequate from a brand-stewardship standpoint.
Gate 4: Price
First, a data note: a reported debt-to-equity of 0.0 is an XBRL tagging omission. Actual total debt as of December 31, 2025 was roughly $45.5 billion against equity of $32.2 billion — a real D/E of about 1.4, well above our quantitative threshold of 0.5. Mitigants: double-digit interest coverage, investment-grade credit, net debt/EBITDA around 2x. The reported 41.4% five-year ROE is leverage-assisted and should be discounted accordingly. Not a disqualifier, but worth the record.
Owner earnings come to roughly $12.0 billion — cross-checked against 2025 free cash flow of $5.3 billion reported, or $11.4 billion excluding the one-time $6.1 billion fairlife payment, and the company’s 2026 guidance of ~$12.2 billion ($14.4 billion operating cash flow less $2.2 billion capex). Accounting profits and cash agree; the asset-light structure is confirmed.
Valuation: at a conservative 12x, $144 billion. At 15x for a certain-growth franchise — which this is — $180 billion. The market capitalizes it at $354.7 billion, a P/E of 26.1 and an owner-earnings yield of 3.4%. Against a ~4.3% ten-year Treasury, that yield is below the risk-free rate. The current price stands at roughly a 97% premium to the base-case intrinsic value. A 30%-discount buy price would be a $126 billion market cap — about $29 per share against a current ~$82. Even relaxing the framework to require merely an owner-earnings yield equal to the ten-year yield gets you only to $280 billion, about $65 per share.
Gate 4 fails. This is a question of price, not quality.
A historical footnote is instructive. Buffett’s original 1988 thesis — global brand moat, pricing power, asset-light bottler model — remains structurally valid in 2026. But he paid roughly 15 times earnings, and never bought another share after 1994. The price at which one holds is not the price at which one buys; 26 times is a holding price. Berkshire’s continued ownership is not an argument for purchasing here.
The Verdict
Watch. Gates 1 through 3 pass — a textbook circle-of-competence business, a brand-plus-distribution moat with pricing power demonstrated annually, and disciplined capital allocation. Gate 4 fails: a 3.4% owner-earnings yield below the risk-free rate, and a market price roughly double our base-case intrinsic value. Target buy zones: about $29 per share on strict criteria, up to about $65 on the relaxed yield test.
What would change it: sustained evidence that GLP-1 adoption is a structural rather than noise-level demand shock — two consecutive years of negative global case volume growth, currently +2–3% — or an IRS appellate loss in the 11th Circuit carrying up to $14 billion in additional payments, would force a rethink. So, in the other direction, would evidence that 15x understates certainty of this quality; we note the tension and keep our multiple.
This analysis is AI-generated, educational, and not investment advice. Figures may contain errors or be delayed. Disclaimer