Coca-Cola

IThe business

Coca-Cola sells a flavour, a name and a promise that the name tastes the same in Lagos and in Ohio. It makes the concentrate, the dark syrup that becomes a cola once someone adds water, sweetener and gas. It sells that concentrate to bottlers, who turn it into drinks and put them on the shelf. Then it spends about $5.4B a year on advertising to make people ask for the finished product.² By the company’s own count, drinks carrying its brands make up 2.2 billion of the roughly 65 billion servings of every kind that people drink in a day.²

The unit the company uses is the unit case, 24 servings of eight ounces. The system sold 33.8 billion of them in 2025.² Sparkling drinks were 69% of that, and Trademark Coca-Cola, meaning Coke, Diet Coke, Coke Zero Sugar and their variants, was 47% of the total.² Revenue of $47.9B over 33.8B cases is about $1.42 a case, or six cents a serving C (calculated).² That is a rough figure, since part of the revenue is finished drinks from bottlers the company owns. It still shows how little of the shelf price the company keeps, and why it matters that it keeps nearly all of it as profit.

The franchise, in plain words

Think of a restaurant franchise, with one difference. A franchisor sells the right to use a name and takes a fee. Coca-Cola sells something every month: the concentrate. A bottler signs a bottler’s agreement that gives it a territory, the right to make and distribute the company’s drinks in cans and bottles there, and the duty to buy all of its concentrate from the company.² In return the company agrees to stay out of that territory for those packages. Those who sign are independent businesses. The 10-K says plainly that a bottler is neither a partner nor an agent.²

The price of the concentrate is not set by a formula that ignores the shelf. In most markets the company uses what it calls incidence-based pricing. The price it charges depends on the bottler’s own prices, on the channel the drink goes through and on the pack size.² If the bottler raises its price or sells more single-serve cans, the company earns more per case. If it sells more big bottles in a discount store, the company earns less. So the two sides share the same signal: how much a drink fetches in the shop. In the United States the agreements run ten years and the bottler can renew them indefinitely; the company keeps the right to terminate for non-performance.² The company makes fountain syrup for restaurants itself and sells it through wholesalers.²

The second table is the one that explains the share price. Look at the segment margins for the first nine months of 2026. Latin America made 61.5% on its sales and EMEA 40.0%. North America made 32.3%. Bottling Investments, the part where the company owns the trucks and the plants, made 10.4%.⁴ A dollar of sales from a plant the company owns brings in about 10 cents of profit. A dollar across the whole company brings in 34 cents. That is why management has spent a decade selling its own bottlers.

Selling the trucks

The company calls this refranchising. It buys or runs a bottler when a market needs fixing, then sells it to a bigger local operator and keeps a stake. The pattern shows up in the ten-year record: revenue fell from $41.9B in 2016 to $34.3B in 2018 as North America and China bottling went, and the operating margin climbed from 20.7% to 28.7%.¹⁰¹¹ Employees went from 86,200 at the end of 2019 to 65,900 at the end of 2025.⁹²

It continues. The company refranchised parts of India in May 2025, and sold its finished-product operations in Nigeria in the fourth quarter of 2025.⁴ It has agreed a deal with Coca-Cola HBC for much of its Africa bottling business. The press reports a price of $2.6B for 75% of the business, with $1.3B in cash going to Coca-Cola for the 41% it holds and an option over the last 25% within six years.¹² The sale is expected to close by the end of 2026 and the company has classified the assets as held for sale. Singapore, Malaysia and Myanmar are held for sale too, with closings expected by mid-2027.⁴ For 2027, the company told investors to expect an 8% drag on revenue and about 2% on earnings per share from these sales.⁵ I read that as the shape of the deal: a lot of revenue leaves and very little profit does.

What it keeps instead: stakes

When the company sells a bottler, it often keeps a minority stake and books its share of the profit as equity income. It owns 18% of Coca-Cola Europacific Partners, 28% of Coca-Cola FEMSA, 22% of Coca-Cola HBC, 24% of Coca-Cola Bottlers Japan and 20% of AC Bebidas, and 21% of Monster Beverage.² Equity income was $2.0B in 2025, and the dividends received from these companies were $993M.² The rest, about $1.0B, was earnings the bottlers kept inside. It adds to net income. The company cannot spend it.

The Monster line is the one I would not have guessed. Coca-Cola owns about 21% of a company that is worth more than $95B at its own price, and the stake sits in the accounts at $5.6B.²¹³ Monster pays no dividend, so none of that value shows up in the cash flow. I count it separately in the valuation, after a tax charge, and the Monster page on this site covers the company itself: see Monster Beverage.

Where the volume comes from

Outside the United States is 84% of unit cases, and four countries, Mexico, China, Brazil and India, are a third of the total.² The United States is 16% of volume, but $19.1B of $47.9B of revenue in 2025 and the segment that earns the most, North America, at $5.1B of operating income in nine months.²⁴ So the profit pool is American and the volume growth is not.

The third quarter of 2026 showed the split clearly. Global unit case volume rose 5%. Asia Pacific rose 6%, Latin America 6% and EMEA 5%. North America was flat.⁵ India, China, Mexico and Brazil led the world total.⁵ In Latin America, price/mix added 7 points and currency 9, so revenue there rose 25%, of which 16 points was organic.⁵ In Asia Pacific, price/mix fell 2 points, which means growth came from selling more at lower prices per case.⁵ On the call the CEO said India delivered strong volume growth, that Brazil kept strong momentum and that the environment in Mexico remains challenging.¹⁴

The table shows what the company has been living on. In 2023 and 2024, price/mix did 10 and 11 points of the work while unit cases grew 2% and 1%. That was a period of inflation when the whole drinks industry raised prices. In 2025 it slowed to 4 points and volume was flat. In the first nine months of 2026 volume is up 4% and price/mix only 1%.⁸⁶²⁴ The mix has flipped. For a company that charges for syrup by the case, volume is the healthier source of growth, because it does not need to raise the price every year. But 1% price/mix is thin. Analysts on the call asked about affordability, and the CFO said consumer behaviour was expected to stay largely unchanged.¹⁴

Currency

Currency matters for a company that earns 60% of its revenue outside the US.² In 2025 a strong dollar took 5 points off comparable EPS growth. In 2026 a weaker dollar adds about 3 points, in the guidance, and 2 points in the third quarter.³⁵ For 2027 the company expects a slight headwind.⁵ The company hedges: its foreign currency cash flow hedges had a notional value of $11.0B at the end of the third quarter.⁴ Hedging delays currency effects; it does not remove them. So the 11% EPS growth guided for 2026 contains about 3 points that the next year may not repeat.

Brands: Coke, flavours, water, dairy, coffee

Sparkling soft drinks grew 5% in the third quarter, with Trademark Coca-Cola up 5% and Coca-Cola Zero Sugar up 16%. Sparkling flavours, such as Fanta and Sprite, grew 3%. Water, sports, coffee and tea grew 5%, led by water at 8%, with sports drinks down 3%. Juice, value-added dairy and plant-based beverages were flat, because the Nigerian sale removed volume that fairlife, Maaza and Minute Maid added.⁵ Zero Sugar growing three times as fast as the core brand matters for a reason I come back to in the risks: the company is moving its own customers out of full-sugar drinks, at the pace it chooses.

fairlife

fairlife makes ultra-filtered milk with more protein and less sugar. Coca-Cola bought it in 2020 with a promise to pay more if it hit financial targets. It did. The last payment, $6.2B, went out in the first quarter of 2025.² Before that, the accounting for the promise had cost the income statement $1.7B in 2023 and $3.1B in 2024, as the estimate of what the company would owe rose.²⁸ Those charges are in reported operating income and they are not cash. The $6.1B of the payment that counted as operating cash flow explains why reported 2025 cash flow was $7.4B and not $13.5B.³

I read this as the price of a success. The company paid a lot because the business grew faster than anyone assumed. On the call the CEO said fairlife continues to do really well, and that supply was still behind demand while the Webster plant ramps up.¹⁴ The company does not report fairlife’s sales or profit. I can tell you it is a leading source of growth in North America and that I cannot size it. That is a gap in the evidence.

Costa

Costa is a UK coffee chain that Coca-Cola bought in 2019 as part of $5.5B of acquisitions that year.⁹ It now sits inside EMEA. The pay-ratio table in the proxy gives a small clue to its scale: the median employee of the whole company is a part-time hourly barista at Costa in the UK, paid $17,947 in 2025.⁷ The Financial Times reported that the company ended a first sale process in December 2025 after private equity bids fell short of about £2B, and Semafor reported in October 2026 that it was looking again. The company did not comment.¹⁵¹⁶ Press reports put the 2018 price at £3.9B, about $5.1B at the time.¹⁷ I treat Costa as an asset likely to fetch less than was paid, and I put no number on it.

The third quarter of 2026

The most recent quarter, reported on 27 October, was strong. Net revenues rose 7% to $13.4B; organic revenue also rose 7%, with concentrate sales running two points ahead of unit cases for shipment timing. Operating margin was 33.4% against 32.0% a year earlier. Comparable EPS rose 11% to $0.91, with a 2-point currency tailwind. The company gained value share in non-alcoholic ready-to-drink drinks.⁵ For the first nine months, cash from operations was $12.5B and free cash flow $11.4B.⁵

The company raised its 2026 guidance: organic revenue growth of at least 5%, comparable EPS growth of at least 11%, and free cash flow of at least $13.0B.⁵ The shares closed at $95.80, a rise of 37% since the last day of 2025.¹

IIThe moat

An advantage that lasts is a cost or a habit that a rival with money and talent can’t copy in a reasonable time. Coca-Cola has several. The question is how much each one is worth today and whether it is growing or shrinking.

The loop is simple. People ask for the drink, so shops stock it. Shops stocking it makes bottlers want the territory. Bottlers who invest in trucks and coolers move more product. Coca-Cola then has more to spend on marketing and more reason to keep the bottlers healthy. Competitors face the reverse: without volume, a bottler can’t justify the fleet.

Evidence it is holding

• The company gained value share in non-alcoholic ready-to-drink drinks in 2025 and in the third quarter of 2026.³⁵
• Unit cases rose 5% in the quarter, with Zero Sugar up 16%, Trademark Coca-Cola up 5% and every geographic segment except North America up at least 5%.⁵
• Comparable operating margin for the first nine months of 2026 was 34.3%, against 33.5% a year earlier.⁵
• Dividends have risen for 64 straight years, to $2.12 a share.²

Evidence it is thinner than it looks

• Volume was flat in 2025, +1% in 2024 and +2% in 2023. All of the growth in revenue in those years was price.⁸⁶²
• North America, the profit centre, was flat in the third quarter. The CEO blamed the timing of the July 4 holiday and said performance got better each month. The CFO said the restaurant industry continues to be under pressure.¹⁴
• The head of North America left in July 2026 and the CFO took the job on an interim basis.¹⁸ That is a short-term gap in the part of the business that matters most.
• BODYARMOR, bought in 2021 for about $5.6B, has lost $1.7B of its value in two write-downs, in 2024 and 2025, because the category slowed and competition grew.⁸² A brand name does not carry into every drink.

On balance I think the advantage is wide and not widening. The company can raise price, add water and dairy, and grow in places where it has a third of its volume. It cannot change how many cola drinkers there are in a rich country. Most of what this advantage protects is a steady stream of cash that grows slowly.

IIIThe competitive advantage

The cleanest way to see what kind of business this is, is to put it next to the others on this site. Two pages here sit closest. Monster Beverage is a leading energy-drink brand, a stake Coca-Cola holds 21% of, and a product some Coca-Cola bottlers distribute. Costco is the other consumer name here, valued on the same terms. The nearest large company in the way it earns, with habitual products, global brands, low capital needs and steady payouts, is Procter & Gamble, which has its own page.

Two things in the table matter. First, Coca-Cola’s economics are nearer to Monster’s than to Costco’s: a 29% margin against Costco’s 4%, and a business that needs little capital for each dollar of sales. Second, among the three, Coca-Cola’s price is the least stretched against my base value, and it is still above. The three pages say much the same thing about the market in late 2026. It is paying up for businesses that are good and steady, and none of them returns my required rate at the price.

Monster shows what a young brand with a similar structure can do. On its page its sales grew 11.8% a year over the last nine years. Coca-Cola grew revenue 1.5% a year over the same span, with its own bottler sales removed along the way C (calculated).¹¹ Costco is the reverse case. It earns pennies on each dollar sold and makes up for it with turnover. Coca-Cola earns about thirty cents on each dollar and turns over slowly. Both ideas work. They are different ideas, and the market now charges a premium for both.

Where Coca-Cola wins

• Everywhere with a bottler of scale and a hot climate: India, Mexico, Brazil, Indonesia, parts of Africa, where volume is growing 5% to 8% in a quarter.⁵
• Zero sugar and water, where the company can move customers away from sugar without losing them. Zero Sugar grew 16% in the third quarter and water 8%.⁵
• Fountain and restaurants, where it makes the syrup itself and has contracts with chains.² That channel is also the one the CFO said was under pressure.¹⁴

Where it loses, or is tested

• Rich countries, where cola volume is flat or falling and growth depends on price.
• Categories without a Coke connection: energy drinks, where the company owns a minority stake in the leader, and sports drinks, where Powerade was down 3% and BODYARMOR was written down.⁵
• Dairy and coffee, where the brand is new to the customer and the economics are different. fairlife is working. Costa is for sale.¹⁴

IVManagement

Coca-Cola changed chief executive this year, for the first time since 2017. James Quincey ran the company from May 2017 and stepped aside on 31 March 2026 to become executive chairman. Henrique Braun, who had been chief operating officer since January 2025, took over. The board also split the chairman and CEO jobs, which Quincey had held together.⁷ The new CEO is 57 and has spent his career in the company’s international business: Greater China and Korea from 2013, Brazil from 2016, Latin America from 2020, International Development in 2023 and 2024.⁷

Two things in that table. The CEO is an insider with long tenure, which suggests continuity more than a new plan. And the CFO is running the largest operating unit alongside his own job for now. That works for a few quarters and I would not want it to last.

Pay, and what it pays them to do

Quincey’s reported pay for 2025 was $31.2M. The median employee, a part-time barista at Costa in the UK, earned $17,947, a ratio of 1,739 to 1.⁷ The pay has two engines. The annual bonus is weighted half on organic revenue growth and half on operating income growth, with targets set at the midpoint of the company’s stated long-term plan: 4% to 6% for organic revenue and 6% to 8% for comparable currency-neutral operating income. For 2025 the payout factor was 150% of target.⁷ The long-term award is performance shares for three years, weighted equally on revenue growth, EPS growth and free cash flow, plus stock options. A relative total return modifier lifts the payout if the shares beat peers. The 2023 to 2025 performance shares paid out at 229%, because every financial measure beat the maximum and the stock finished above the 75th percentile of its peers.⁷

So follow the money. The structure pays executives for growing revenue, profit, earnings per share and cash, and for the share price compared with peers. That is sensible, and it has paid out well for results that were good. But two things are missing. Nothing measures the return on the capital employed. A company can raise EPS by buying other companies at a price that never earns its cost, and the pay would not notice. And revenue growth is organic, which leaves out the effect of selling bottlers. That is reasonable. It also means a sale of a bottler that shrinks revenue by 8% in 2027 will not cost anyone in the pay plan anything.

The record shows why the question matters. BODYARMOR cost about $5.6B in 2021 and has lost $1.7B of book value since then.⁸² The fairlife purchase worked, but the earn-out ended up costing far more than the original estimate. Costa cost about $5.1B in 2018 by press reports, and is for sale.¹⁷¹⁵ Each is a normal cost of buying brands. Each is also part of the record the pay plan does not look at.

Their own money

Quincey is listed with 4,461,496 shares, of which 3,879,031 are options he can exercise. That leaves about 582,465 shares and share units, worth about $56M at $95.80 C (calculated).⁷¹ Braun is listed with 643,876, of which 529,668 are options, so about 114,208 shares and units, worth about $11M C (calculated).⁷¹ Options are worth something only if the price rises, which fits the pay design, and they do not count toward the share ownership guidelines.⁷ Quincey’s stake is nearly two years of his pay and about 0.01% of the company. Braun’s is smaller. Both are real money. Neither is a stake big enough to steer a $413B company.

The company’s largest holder is better known. Berkshire Hathaway reported 400,000,000 shares at 30 June 2026, about 9.3% of the shares outstanding, with a value of $32.5B at that quarter’s end.¹⁹ The count was the same at 31 March 2026.¹⁹

What they did with the money, over a full cycle

I take 2016 to 2025 as the cycle. It covers a Tax Act, a pandemic, a burst of inflation, a Tax Court loss and a series of refranchising deals. Over those ten years the company produced $93.2B of operating cash as reported, or $105.3B once the $6.0B IRS deposit and the $6.1B fairlife payment are added back. It spent $17.7B on capital projects, paid $72.9B in dividends and bought back $16.9B of stock C (calculated).¹¹³

Three things stand out. First, the dividend: it took $73B, about 69% of adjusted operating cash, and rises every year. Second, buybacks swung from $3.7B a year in 2016 and 2017 to $0.1B in 2020 and 2021, and to $0.7B in 2025. A company that buys when it has cash and not when the stock is cheap shows no clear sense of value. In 2026 it repurchased 10.0M shares in nine months at an average of $78.77.⁴ That average is 18% below today’s price, and the buyback is small enough that it does little either way. The 10-Q says the 2026 buybacks are meant to offset the shares paid to staff. Third, the acquisitions have been mixed. fairlife worked. BODYARMOR did not. Costa did not.

The return on the capital employed is respectable and not extraordinary. Operating income after tax, over equity and debt less cash and securities, was about 18% in 2025 C (calculated).² That includes $20.2B of bottler stakes, whose $2.0B of earnings sits below operating income. Without them it is about 26% C (calculated). The business itself, the syrup, needs almost no capital. The returns are held down by what the company has chosen to own around it.

VFinancials

Coca-Cola’s accounts are large and tangled by choice. The company consolidates some bottlers and not others, reports its stakes in many more through equity income, and has booked big non-cash items in recent years for fairlife, BODYARMOR and the Africa sale. The simplest honest way to read them is to find the cash.

Margins

Operating margin as reported rose from 20.7% in 2016 to 28.7% in 2025 C (calculated).¹¹ Part of that is the 2025 comparison with 2024, when fairlife’s earn-out cost $3.1B. The company’s own comparable measure, which strips out impairments and the like, was 31.2% in 2025 against 30.0% in 2024, and 34.3% for the first nine months of 2026 against 33.5% a year earlier.³⁵ In the third quarter the comparable margin was 32.9%, up 106 basis points, of which 60 were currency.⁵ Underlying, currency neutral and without deals, the margin rose 21 basis points.⁵ So margin growth is real and slow, and about a fifth of the reported gain of 140 basis points.

The reason it can rise is the mix. Each bottler sold takes a low-margin business out of the numbers. The reason it will stop is also the mix: once the last of the plants is gone, there is nothing left to dispose of and the margin has to grow from price and volume alone. Marketing investment also rose in the third quarter and operating expenses were higher.⁵

Earnings quality

Net income was $13.1B in 2025 and EPS $3.04, up 23%. Comparable EPS was $3.00, up 4%.³ The gap between 23% and 4% is the one thing to understand about the numbers. Reported EPS rose 23% mostly because 2024 carried $3.1B of fairlife remeasurement charges that 2025 did not. Reported 2025 also had a $331M gain on selling part of the CCEP stake, and an 8-point currency headwind that comparable EPS shows as 5.³⁵ I use the comparable figure for growth and the cash flow for value.

The tax line is its own story. The effective rate fell to 16% in the third quarter of 2026 because of $172M of discrete benefits, including interest the IRS owes on the deposit.⁴ The company guides to an underlying rate of 19.7% for the year, and that figure leaves out the effect of losing the IRS litigation.⁵ If the Tax Court method were applied going forward, the company estimates the rate would rise by about 3.8 points (the 10-K said 3.5, before the 2026 tax law changes).⁴ I deal with this in the valuation.

Cash flow

Reported operating cash flow was $6.8B in 2024 and $7.4B in 2025. Those look weak for a company of this size. Each year held one large payment: $6.0B to the IRS in September 2024, which stays an asset until the Eleventh Circuit rules, and $6.1B of the final fairlife payment in 2025.² Adding both back gives $12.8B and $13.5B. For the first nine months of 2026, operating cash flow was $12.5B, against $3.7B a year earlier, which had the fairlife payment in it.⁴ The 10-Q puts part of the rise down to the factoring of receivables, where the company sells customer invoices to banks at a discount, $10.6B of them in the nine months, and to the timing of working capital.⁴ Some of that can reverse. The company guides to at least $15.2B for 2026 and $2.2B of capital spending.⁵ I treat the add-backs as legitimate for understanding the business and not for estimating what an owner takes home, which is a different number and lower.

Dividends alone took 77% of adjusted free cash flow in 2025. That is steady and leaves room. There is no sign that the dividend is stretched, and no sign that it is a constraint.

Balance sheet

At 2 October 2026 the company had $41.6B of debt and $16.6B of cash, short-term investments and marketable securities, so net debt was $25.0B C (calculated).⁴ Current maturities were $4.9B. Interest expense was $1.1B for nine months and interest income $0.6B.⁴ Net debt is about 1.5 times the last twelve months of operating income plus depreciation, $16.4B E (estimate). That is conservative for a company that makes this much cash. The balance sheet also carries $20.7B of equity-method investments.⁴ The company’s suppliers have 120-day payment terms, and $1.6B of the payables is in a supplier finance programme.²⁴ The supplier arrangement is large enough to be worth watching. Its financial effect is that part of the working capital is funded by suppliers’ banks.

The ten years in one table

VIValuation

Start with what would go wrong

Before any numbers, I ask what would most likely make this a bad investment from here, and work back from that. The likeliest problem is the price. People have wanted the product for 140 years. Here are the three ways it goes badly, in the order I think most likely.

1. The business does fine and the multiple falls. At $95.80 the market pays 34 times my estimate of owner cash. In ten years my base has owner cash per share at $4.72, against $2.82 now. If the market then pays 22 times owner cash, a fair price for a steady business growing 3%, the return on $95.80 is about 4.6% a year, with all cash returned, which is less than a ten-year Treasury E (estimate). Nothing has to break for this to happen. The shares closed at $41.69 in November 2016 and rose 2.30 times in a decade; free cash flow per share after stock pay is guided to rise about 2.06 times from 2016 to 2026 C (calculated).¹¹¹
2. The IRS wins. A federal appeals court heard the case on 25 June 2026 and has not ruled.⁴ If the Tax Court method is upheld, the company says it could owe about $14B in extra tax and interest for 2010 to 2025 as of the end of 2025, and its estimate rises by $1.4B in the nine months to October.⁴ It also says its tax rate would be about 3.8 points higher from 2026. And the $6.0B it deposited in 2024, plus $589M of interest, would not come back.⁴ I can’t judge this case. The company concludes, as its tax accounting requires it to, that it is more likely than not to prevail. And in October 2025 the Eighth Circuit ruled for 3M on the blocked-income regulation, which bears on the part of Coca-Cola’s dispute about its Brazilian licensee, in a case the Tax Court had treated as controlling in Coca-Cola’s.⁴² Neither fact is a ruling in this case.
3. Growth stalls where the profit is. North America earns the most. Volume there was flat in the third quarter, with price/mix at 1%.⁵ The 10-K lists weight-loss drugs, sugar taxes and limits on what food assistance can buy as risks to demand.² One local report counts 18 states with limits on sweetened drinks in the programme.²⁰ I don’t have a study I trust on how much these hurt, and I won’t invent one.

I can show what the first two do, and I do below. The third I size in Section VIII from the company’s own segment figures, since I have no study to lean on.

What the price implies

The Treasury pays 4.90% and the price gives an earnings yield of 3.5% and an owner cash yield of 2.9%. Shares of a steady company can earn more than a Treasury because their cash grows. Here the question is whether it grows enough. The shares were $69.91 at the end of 2025 and rose 37% in ten months. The company guided comparable earnings to grow at least 11% in the same year.⁵ So about two thirds of the share price rise is the market agreeing to pay more per dollar of earnings C (calculated).

Owner cash: what you can take out

Owner cash is what is left after the business has paid for the capital it needs and for the stock it hands to staff. I start from comparable net income for 2026, which is the guided at-least-11% growth on 2025’s $3.00 a share, on 4,313M shares: $14.4B E (estimate).⁵³ I take out the earnings of the bottlers that the company does not receive in cash. Equity income runs about $2.1B a year, and dividends from those companies were $1.0B in 2025.²⁴ That removes about $1.1B. I add back depreciation of $1.05B and deduct all of the $2.2B of capital spending the company guides to, since I can’t tell how much of it is growth, such as the fairlife plant.⁵⁴ Stock pay is already inside net income, so I do not add it back.

So owner cash is about $12.1B against comparable net income of $14.4B, 85% of it C (calculated). The gap is the part of the bottlers’ earnings that stays with the bottlers and capital spending that runs about twice depreciation. The company’s free cash flow guide of at least $13.0B looks higher than mine because it adds back stock pay, the timing of tax and working capital, and a lift from factoring receivables. I use the lower number because it asks the stricter question.

Reverse DCF: what the price needs

Held at my other inputs, the price needs owner cash to grow about 8.8% a year for ten years C (calculated). My base has 5.3%. Over 2016 to 2025 adjusted free cash flow grew 6.4% a year, and that was with a revenue base that fell during the refranchising.¹¹³ The company’s own long-term plan is 4% to 6% organic revenue growth and 6% to 8% comparable currency-neutral operating income growth.⁷ At the top of that range, with margins steady, the price is nearly reached and not exceeded.

There are two other ways to read the same price. One is the rate that equates my base cash flows to the price: 6.8% a year, which is 1.9 points above the Treasury C (calculated). The other is what a buyer pays in 2036. If my base cash flows come through and the market then pays 34 times owner cash, as it does now, the return from $95.80 is about 8.4% a year E (estimate). That is the real bet in this price: a fine business, plus a multiple that never comes down.

The discount rate

I discount at 8%. The capital asset pricing model with the 4.90% ten-year Treasury and Damodaran’s 4.20% implied equity risk premium gives 6.9% on Coca-Cola’s beta against the market over 119 months, 6.2% over 60 and 5.4% over 36 C (calculated).²¹²²²³ Those are low because the betas are low: 0.48, 0.32 and 0.11. On my Costco page I took the highest of the three. Here that would give a required return barely above a Treasury, which I will not use for a share. A beta measured against a market led by a few giant growth stocks says less about a steady company than it looks. I take 8%, a little over a point above the highest CAPM figure and a rate that sits between it and the 9.1% the same inputs give for a beta of one E (estimate). The page shows what other rates do.

At 7.0%, close to the CAPM rate on the longest beta, the base value is $92 E (estimate). Only one corner of the table, 7.0% with 3.5% growth for ever, gets above the price, at $101.

Three cases

My base case is Coca-Cola doing what its plan says, a little slower. Owner cash grows 3.5% in 2027, because the Africa sale takes about 2% of earnings with it, then 7% for two years, easing to 4% by 2035.⁵ That is volume of 2% and price/mix of 3%, which gives revenue growth near 5%, plus 1.5 to 2 points a year from margin and cash conversion, the lower half of the company’s 6% to 8% operating income plan.⁷ Buybacks are not in the totals: owner cash is treated as returned and the share count stays flat. It gives owner cash of $20.3B in 2036 against $12.1B now E (estimate). The bull case has the company hit the top of its plan for most of a decade: volume of 3%, price/mix of 3% and 2.5 to 3 points a year from margin, with a win in the IRS case. The bear case has volume at 1%, price/mix at 2% and no margin gain, and the IRS wins the appeal E (estimate). Even the bear is a business that pays a growing dividend. It is only a poor purchase at $95.80.

The IRS case, on its own

My base leaves out both outcomes. That is not the same as saying they balance. The win adds $1.41 a share and the loss takes away $7.59, so the case is lopsided: the downside is five times the upside. I would need odds of about 84% that the company wins to call it even. The accounting test is only “more likely than not”, which is above 50% and not close to 84%. I can’t read an appeals court. So I show the range and leave my base where it is.

The Monster stake

Coca-Cola owns about 21% of Monster Beverage.² With roughly 1,959M Monster shares and a close of $48.75, that stake is worth about $20B E (estimate).¹³ I take off tax on the gain over the $5.6B carrying value, at 21%, and add the rest, $17.0B, or $3.94 a share, to every case E (estimate). Monster pays no dividend, so none of this is in owner cash. If it were sold, the money would have to come out of an investment the company has held for a decade. I treat it as a store of value, no more.

What I can’t judge

Some inputs can’t be known, and I would rather say so. I don’t know how the Eleventh Circuit will rule. I don’t know how much of the $2.2B of capital spending is upkeep; I show a range of $11.1B to $13.3B for owner cash. I don’t know what weight-loss drugs do to soda volume over ten years, because I could not find a study I trust and the company does not publish it. I don’t know fairlife’s profit; the company does not report it. And I can’t know what multiple the market will pay in 2036. That last input matters more than the rest together, which is why I show the returns at several multiples and not one.

Why the market could be right

Most of the gap between $74 and $95.80 is one choice, the discount rate. My base cash flows are worth $94.50 at 6.9%, the capital asset pricing rate on the longest beta, and the price is what they return at 6.8% E (estimate). A buyer who accepts about 2 points over a Treasury for a steady business is not being foolish. I ask for 8% because betas measured against this market understate the risk of a share, and because the IRS case is open. Other things lean the market’s way. I leave out the $1.1B a year of bottler earnings that stay inside the bottlers; counted at half value, that adds about $3 a share E (estimate). My base leaves out the IRS case, and a win adds $1.41. Put 6.9% and the half-value bottler earnings together and the base is about $99, above the price E (estimate). I don’t use that, because it takes the lowest discount rate and the kindest reading of everything else at once. But it is the market’s case, and it is not a foolish one: the business is no better than I think, and safe cash simply costs less than I assume.

Multiply it out

My base needs several things at once. Volume growth of 2% a year or better to 2031: I would give that 60%. Price/mix of 2% or better: 65%. Operating margin holding after the bottler sales: 80%. The IRS case no worse than my base: 60%. No lasting dent from weight-loss drugs and sugar rules: 75%. Multiplied, that is about 14% E (estimate). They overlap, since good volume helps price/mix and margin, so the true figure is higher, perhaps a quarter. But the price needs more than my base. It needs my base and a multiple near today’s in 2036. If I give that a one-in-five chance, the joint odds are about 3% before overlap E (estimate). The argument for this price is not simple. It needs many things to go right at once.

Quality and price are two questions

How good is the business? Among the best I have studied. It earns about 29 cents of operating profit on each dollar of sales, it needs little capital to do so, its customers form the habit young, and it has paid a rising dividend for 64 years.² How good is the price? Poor, on my numbers. At $95.80 the price is above my bull case. A great business can be a poor purchase at the wrong price, and on this evidence that is what this looks like.

It helps to compare it with the next best use of the same dollar. A ten-year Treasury pays 4.90% with no risk to the dollars promised.²¹ Coca-Cola returns about 6.8% a year on my base, 1.9 points more, with all the risk of a share and a pending IRS ruling. On this site, Monster Beverage returns 7.0% on its page’s base case and Costco 6.2%. Coca-Cola sits between them. Procter & Gamble has a base value of $125 against a price of $167.32 on its page. Of the businesses I cover, Coca-Cola is one of the steadiest and, at this price, returns less on my numbers than Monster or Costco do.

The margin of safety

My margin-of-safety price is $52, 30% below my base value. At $52 my bear case, which includes losing the IRS case, returns about 7.5% a year, close to the 8% I require, and the base returns about 10.2% C (calculated). A 30% cushion is what I use for a business this predictable. The price is 84% above it. The stock last closed below $52 on 24 March 2021, and last closed below my $74 base value, at $73.43, on 29 January 2026 C (calculated).¹

VIIThe case for the price

What most people miss

First, the company owns more than its balance sheet shows. The seven listed stakes were worth $34.3B at the 2025 year-end market prices against a carrying value of $14.0B, a hidden $20.3B C (calculated).² Second, the cash flow is not what it looks like. Reported operating cash flow was $6.8B in 2024 and $7.4B in 2025, but each year held one big one-time payment. Adjusted, the figures were $12.8B and $13.5B C (calculated).²³ Third, only 3.6% of operating income comes from the plants the company owns, which are 12% of revenue C (calculated).⁴ That is why selling them costs so little profit.

The bull case

In plain words: Coca-Cola is a toll on the world’s thirst. People who drink a Coke today will drink one next year, local owners put up the capital, the company keeps the high-margin syrup, and the dividend has risen 64 years in a row.²

Management’s own version has numbers. Its long-term plan is 4% to 6% organic revenue growth and 6% to 8% comparable currency-neutral operating income growth, and the 2025 bonus targets were set at the midpoint of both.⁷ In the third quarter the company raised guidance, and the CFO told investors that developing markets will continue to outgrow.⁵¹⁴ The CEO said fairlife continues to do really well, with supply still lagging demand.¹⁴ Among funds I checked, the largest holder with a published position is Berkshire Hathaway, with 400,000,000 shares at 30 June 2026, unchanged since the end of 2025.¹⁹ That shows what it holds. I did not search its letters for a stated case, so I put none in its mouth.

Against reality

Much of the bull case is true today: volume is up 4% in nine months and the company raised its guidance in October.⁵ My bull case gives it more: volume of 3% and price/mix of 3% for most of a decade, and a win in the IRS case. That is worth $93 E (estimate). So if Coca-Cola does all that and hits its targets, it is a better business in 2036 than today. It still doesn’t reach $95.80 at 8%. The price needs owner cash to grow about 8.8% a year at my other inputs, against 5.3% in my base and 6.4% in the last decade, or a buyer in 2036 who pays 34 times owner cash again C (calculated).

So at this price the bull case is about the multiple more than the business. The shares rose 37% this year while comparable earnings are guided up at least 11%.¹⁵

What a 10x would take. Ten times $95.80 is $958 a share, about $4.1T of market value on today’s share count C (calculated). At 30 times earnings that needs about $138B of net income, 9.6 times the $14.4B I assume for 2026. At the same 28% net margin it needs about $490B of revenue, against about $51B this year E (estimate). In ten years that is revenue growth near 25% a year with a multiple close to today’s. The company’s plan is 4% to 6%. From $413B, a tenfold rise is close to impossible, and I would not build a case on it. The shares rose 2.3 times in the last ten years, from about $182B C (calculated).¹

Where I come down

The adjusted cash flow and the small profit share of owned bottlers moved me: the business is steadier than its reported numbers suggest, and my base gives it credit for that. The bull case didn’t move me on value, because even my bull falls short of the price. What would raise my values most is a clean win in the IRS case together with price/mix back near 3% while volume holds at 4%, which would show growth coming from both the number of drinks and what each one earns.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Weight-loss drugs. The 10-K lists them among the factors that may cut demand for sweetened drinks.² I could not find a study I trust that sizes the effect on soda, and the company does not publish one. Zero Sugar and water are growing faster, which helps the company more than a bottler of regular cola. It is still a risk I can’t measure.
• Sugar taxes and food-aid limits. The 10-K names taxes on sweetened drinks, packaging taxes and limits on what food assistance can buy.² One local report counts 18 states with limits on sweetened drinks in the programme.²⁰ Mexico is a top-four volume market. The Rio Times reports that its excise on sugary drinks rose about 87% on 1 January 2026, to 3.08 pesos a litre, and now also covers drinks with non-caloric sweeteners; I did not check that against the official gazette.²⁴ The CEO called the Mexican environment challenging, and yet Latin America volume rose 3% in nine months and 6% in the third quarter.¹⁴⁵ Congress has considered a federal excise tax on sweetened drinks in the past.²
• Bottlers. The company depends on partners it does not control. Five of them carry 44% of volume.² If one runs into trouble with debt or a currency, volume and the concentrate price suffer.
• Currency. The dollar’s direction moved 2026 comparable EPS by about 3 points up and 2025 by 5 points down.⁵³ A rise in the dollar would take back the gain.
• Costs. The CFO said cost pressures are elevated in aggregate, while the CEO said commodities have been manageable.¹⁴ The company’s market-risk note names sweeteners, metals, juices, PET plastic and fuels.⁸
• Management bench. A new CEO, a departed head of North America and a CFO covering both jobs on an interim basis.⁷¹⁸ None of that is unusual. The three things together in one year is.
• Costa, BODYARMOR and fairlife. Costa may fetch less than was paid, BODYARMOR may need another write-down, and fairlife is a bright spot whose profit the company does not disclose.²¹⁶

How much the US matters

I can’t size the weight-loss and sugar-rule risk with a study, so I size it from the company’s own numbers. The United States was 16% of unit case volume in 2025, but $19.1B, or 40%, of revenue.² Sparkling drinks were 61% of US volume.² North America made $5.1B of the $14.4B segment operating income in nine months, 35% of it C (calculated).⁴ So a US problem is a problem for a third of profit and a sixth of volume.

Even a 10% fall in North America revenue costs about 3.8% of operating income if the margin holds. That is a large drop in sales and a limited hit to value, because about two thirds of the profit is earned outside North America. The weight of the risk is elsewhere. The United States is where price/mix has to come from, and a category under pressure tends to lose pricing power before it loses volume.

What would raise my base value: the 2027 guidance in February showing comparable EPS growth of about 6% even with the Africa drag, volume growth above 3% with price/mix near 3%, and a ruling for the company. What would lower it: volume slipping below 2%, price/mix going negative, or a ruling for the IRS.⁵⁴