The Coca-Cola KO

> ARCHIVE NOTE — WHAT THE COMPLETE PULL ADDED, AND IT IS NOT SMALL. The earlier dive stated in its own provenance section that all of its 8-Ks contained "a list of listed debt securities and nothing else." The complete archive carries three 8-Ks with full Item text, and two of them disclose events absent from every field of the vendor payload:

>

> 1. fairlife ransomware, disclosed 2026-07-16 — nineteen days before this dive. "fairlife, LLC ('fairlife'), a dairy company owned by the Company, identified unauthorized access by a third party to a portion of its systems, including its production-related systems, in connection with a ransomware event... as a result of the incident, production operations at fairlife in the United States are temporarily suspended... The full scope, nature and impacts of the incident are not yet known. Accordingly, the Company has not yet determined whether the incident is reasonably likely to materially affect the Company." fairlife is the business for which Coca-Cola paid a $6,173 million final contingent-consideration milestone in March 2025.

>

> 2. Departure of the head of North America, disclosed 2026-06-25. "Jennifer Mann, Executive Vice President and President, North America Operating Unit, will be departing the Company. Effective July 31, 2026, Ms. Mann will step down... Effective August 1, 2026, John Murphy, President and Chief Financial Officer, will assume responsibility for the North America Operating Unit on an interim basis." North America was 40.4% of second-quarter revenue. The Chief Financial Officer took on that second job four days before this dive.

>

> What did NOT change: the 2026-07-28 earnings 8-K still carries only the Item 2.02 pointer to Exhibit 99.1, and the exhibit is still absent. That does not matter here, because the 10-Q filed 2026-07-29 carries the complete June-quarter financial statements, and every quarterly figure in this dive comes from it. The verdict did not change either. The base fair value moves from $89 to $88, and the reason is the unquantified fairlife shutdown.

Reference table

1. The business and the five segments — from the filing, because the payload's version is broken

Coca-Cola manufactures and sells beverage concentrates and syrups, and finished beverages, worldwide. Chief executive Henrique Braun (who signs the certification on the 10-Q filed 2026-07-29); President and Chief Financial Officer John Murphy; approximately 65,900 employees; incorporated in Delaware in September 1919, succeeding a Georgia corporation organised in 1892; common stock, $0.25 par value, NYSE-listed, 11,200 million shares authorised, 7,040 million issued, 2,737 million in treasury.

Five operating segments plus Corporate, per the 10-Q: "Europe, Middle East and Africa ('EMEA'), Latin America, North America, Asia Pacific, and Bottling Investments."

Second-quarter 2026 third-party revenue by segment, filing-verified ($M):

Read the margin column and the strategy explains itself. Latin America earns 59.1%, EMEA 39.7%, Asia Pacific 38.3% and North America 25.9% — while Bottling Investments, the segment where Coca-Cola still owns the trucks, earns 7.4%. Bottling Investments has fallen to 11.4% of revenue from 13.2% a year earlier. Every further disposal is arithmetically margin-accretive, and the July balance sheet carries $5,438M of assets and $2,442M of liabilities held for sale, so the process is not finished. This is the mechanism behind a knowledge-base claim from 2024 that is graded in Section 7.

Volume, which is the leg of the algorithm price cannot fake. Second-quarter worldwide unit case volume +5% and concentrate sales +4%; six months +4% and +6%. By segment for the quarter: EMEA +4%, Latin America +3%, North America +3%, Asia Pacific +8%, Bottling Investments +5%. That is a genuine acceleration: fiscal 2025 worldwide unit case volume grew 1%.

Revenue growth decomposed by the company, second quarter 2026 against 2025: volume +4, price/mix +2, foreign currency +2, acquisitions and divestitures −1, total +7. For the six months: volume +6, price/mix +2, currency +2, acquisitions and divestitures −1, total +9. Currency is a tailwind for the first time in several years, and the company states that after hedging, exchange rates increased operating income by 5% in the quarter and 4% over six months.

2. The two events the vendor payload has never heard of

fairlife ransomware — disclosed 2026-07-16, unresolved

The 8-K, in full on the substance:

> "On July 16, 2026, The Coca-Cola Company (the 'Company') announced that fairlife, LLC ('fairlife'), a dairy company owned by the Company, identified unauthorized access by a third party to a portion of its systems, including its production-related systems, in connection with a ransomware event. After detecting the issue, the Company promptly activated its incident response and business continuity protocols... Product quality and safety have not been impacted. However, as a result of the incident, production operations at fairlife in the United States are temporarily suspended. fairlife's Canada production operations are not currently impacted. The Company is working diligently to complete the investigation and restore the systems and impacted operations. The full scope, nature and impacts of the incident are not yet known. Accordingly, the Company has not yet determined whether the incident is reasonably likely to materially affect the Company."

Why this matters more than a generic cyber headline. fairlife is not a small line item. Coca-Cola acquired it in 2020 with a contingent-consideration structure whose final milestone payment was $6,173 million, paid in March 2025 — a payment so large it drove fiscal 2025 operating cash flow down to $7,408M from $11,599M in fiscal 2023, and drove the first half of fiscal 2025 to a negative $1,391M of operating cash flow. A company does not pay $6.2 billion for a business that cannot move the needle.

Coca-Cola does not disclose fairlife revenue separately. It sits inside the "juice, value-added dairy and plant-based beverages" category and the North America segment. We therefore cannot size the impact and we will not estimate one. What is stated is what is knowable: US production suspended as of the disclosure, Canada unaffected, quality unaffected, impact undetermined, nineteen days elapsed as of this dive, and the first financial disclosure that could quantify it is the 2026-10-20 third-quarter report.

One observation that is arithmetic rather than speculation. Consensus for fiscal 2026 revenue is $49,728M. First-half actual revenue was $25,852M. That implies a second half of $23,876M against $24,277M in the second half of 2025 — a 1.7% decline, immediately after a first half that grew 9.2%. The vendor's quarterly estimate for the September quarter is $12,895M (+3.5%), which implies a December quarter of $10,981M against $11,822M — down 7.1%. We do not know what is in that number. It could be the fairlife shutdown, it could be the pending disposal represented by the $5,438M of assets held for sale, or it could be an estimate file that has not been refreshed consistently. It is flagged as an unexplained shape in the consensus and is one reason the base case is not rounded up.

The President of North America departs — disclosed 2026-06-25

> "On June 25, 2026, The Coca-Cola Company... announced that Jennifer Mann, Executive Vice President and President, North America Operating Unit, will be departing the Company. Effective July 31, 2026, Ms. Mann will step down from her current positions. Beginning on August 1, 2026, she will continue with the Company as a senior advisor until April 30, 2027. Effective August 1, 2026, John Murphy, President and Chief Financial Officer, will assume responsibility for the North America Operating Unit on an interim basis."

North America is 40.4% of second-quarter revenue and the lowest-margin of the four branded segments at 25.9%. It is also the segment that absorbed the $960 million BodyArmor trademark impairment in fiscal 2025. From 2026-08-01 — four days before this dive — it is run on an interim basis by the Chief Financial Officer, who retains the finance role. The separation agreement provides severance under the standard plan, an eligible 2026 annual incentive if she remains employed through 2026-12-31, no 2027 incentive, and no additional equity grants.

We do not read a departure as a signal on its own. What we do read is capacity: a Chief Financial Officer running the largest operating unit is a temporary arrangement by construction, and it arrives in the same six weeks as an unquantified production shutdown inside that same segment. Both facts are in the filings and neither is in the payload.

And a third thing the filings settle: who runs the company

The vendor profile names Henrique Braun as chief executive. The insider file lists James Quincey as "director, officer: Chairman." The 10-Q settles it: the Rule 13a-14(a) certification for the June 2026 quarter is "executed by Henrique Braun, Chief Executive Officer of The Coca-Cola Company," with the second certification "executed by John Murphy, President and Chief Financial Officer." Braun is chief executive; Quincey is Chairman. The archive contains no 8-K describing the handover, so the transition itself is outside this window. The filing wins and the payload's profile is right.

3. The numbers — three years of annual, five quarters of recent

Annual, from inc_a, in $M:

The fiscal 2025 earnings jump is mostly a base effect and the dive says so. Fiscal 2024 carried $4,163M of other operating charges — of which $3,277M sat in Corporate, principally the fairlife contingent-consideration remeasurement — and fiscal 2025 carried $1,261M, of which $960M was the impairment of the BodyArmor trademark in North America, $97M the productivity and reinvestment programme, $47M a final fairlife remeasurement and $44M a Latin America trademark impairment. Strip both and the operating-margin path is far flatter than 21.2% → 28.7% suggests.

Quarterly, from inc_q and cross-checked against the 10-Q, in $M:

The December 2025 quarter is the outlier and the BodyArmor impairment is why — a 15.6% operating margin in a business that runs at 34.9%. Every trailing multiple on Coca-Cola is distorted by it, and this dive uses the FY2027 forward estimate as the valuation anchor for exactly that reason.

The 10-Q's own June-quarter statement, which the vendor reproduces correctly: cost of goods sold $4,965M, gross profit $8,415M, selling/general/administrative $3,720M, other operating charges $23M, operating income $4,672M, interest income $198M, interest expense $369M, equity income $604M, other income $370M, income before taxes $5,475M, income taxes $1,037M, consolidated net income $4,438M less $13M attributable to noncontrolling interests, net income attributable to shareowners $4,425M, diluted EPS $1.03 on 4,313M diluted shares. Every figure matches inc_q to the dollar.

One reporting nuance worth naming: earn_cal reports June-quarter revenue of $13,373M; the 10-Q reports $13,380M. A $7M, 0.05% difference. The filing wins and $13,380M is used throughout.

Equity income of $604M in one quarter is 12.9% of operating income and is easy to miss. Coca-Cola holds equity-method stakes in its largest bottlers, whose aggregate revenue the 10-K puts at $102,800M in fiscal 2025 with $13,426M of operating income. The economic system Coca-Cola sits at the centre of is roughly three times the size of the accounts you can see.

4. Cash generation — recomputed, with a fully verified capex line

The capital-expenditure check on Coca-Cola is clean in both directions and it is reported as a finding.

The diagnostic is also arithmetically sound, which is the second half of the test. capexToDepreciationTTM reads 1.978, and the direct computation gives trailing capex of $2,045M against trailing depreciation and amortisation of $1,034M = 1.978 — identical. A ratio near 2x is high, but it is not the impossible sub-zero reading or the implausible sub-0.15 reading that flags corruption; it reflects a company whose net property, plant and equipment ($9,636M) is small relative to revenue and which is investing ahead of it. Both the datum and the diagnostic pass. This is a verified-clean capex check.

Free cash flow, recomputed from filed numbers, and the trap in the annual figure:

This is the single most important correction on the page and it runs in the company's favour, which is unusual. A reader taking the vendor's fiscal 2025 free cash flow of $5,296M at face value would compute a 1.42% yield and conclude that the $8,779M of dividends paid that year were funded 1.66x by borrowing. The 10-K says otherwise in terms: "the activity in 2025 included $6.1 billion of the $6.2 billion final milestone payment for fairlife. The activity in 2024 included the $6.0 billion IRS Tax Litigation Deposit." Two consecutive years each absorbed roughly $6 billion of non-recurring operating outflow. Normalised, fiscal 2025 free cash flow was approximately $11,365M, and the trailing figure — which now excludes both — is $14,297M, a 3.84% yield that covers the $2.12 forward dividend 1.57x.

We use the trailing figure and label the annual one. freeCashFlowYieldTTM of 0.038389 reconciles exactly to our recomputation, so on this name the vendor's trailing cash-flow block is correct; it is the annual rows that mislead if read without the filing.

Capital return. Dividends of $8,779M in fiscal 2025 and $4,562M in the six months to 2026-07-03 — the 10-Q explains the year-on-year jump from $2,283M as timing: "the Company paid all of the 2026 second quarterly dividend in the second quarter and paid substantially all of the 2025 second quarterly dividend in the third quarter." The July 2026 board meeting approved $0.53 per share, payable 2026-10-01 to holders of record 2026-09-15 — $2.12 annualised, a 2.45% forward yield, against dividendPerShareTTM of $2.08 which is trailing. Buybacks are immaterial: $746M in fiscal 2025 against $1,795M in fiscal 2024 and $2,289M in fiscal 2023 — a declining programme, and at 0.2% of market capitalisation it is not a component of return. In the six months to July, $663M of treasury purchases against $251M of stock issuances is a net $412M. Total shareholder yield is approximately 2.45% dividend plus roughly 0.2% net buyback ≈ 2.65%.

5. Balance sheet

From the 10-Q at 2026-07-03 ($M), against 2025-12-31:

Leverage is moderate and improving. Net debt of $30,014M (total debt less cash and short-term investments) is 1.51x trailing EBITDA of roughly $19,831M; including the $2,842M of marketable securities it is 1.36x. Total debt fell $1,949M in six months. Interest expense is falling: $369M in the June quarter against $445M a year earlier, and $744M over six months against $832M — a 10.6% reduction while interest income rose to $420M from $368M. Net interest cost is now $324M for the half-year, roughly 3.6% of operating income.

Two items deserve naming. $5,438M of assets held for sale against $2,442M of liabilities held for sale — a net $2,996M disposal in progress, consistent with the continuing refranchising of Bottling Investments, which the 10-Q elsewhere connects to bottling operations in India. The payload contains no description of it. And $20,782M of equity method investments, the carrying value of the bottler stakes that generate $604M of quarterly equity income; the 10-K notes the aggregate carrying value exceeds Coca-Cola's proportionate share of investee net assets by $8,744M, which "is not amortized."

6. Valuation — priced in or room?

At $86.56 (market cap $372,428.6M, ~4,302.5M shares):

Estimate coverage is good on the near years and collapses after. FY2026 rests on 13 analysts and FY2027 on 14 — the anchor. FY2028 rests on 7. FY2029 rests on 2 and is excluded from every conclusion; FY2030 rests on 5. The FY2029 row is the thinnest in this batch and the fact that FY2030 has more coverage than FY2029 is itself a reason to distrust the far end of the array.

est.ebitdaAvg and est.ebitAvg are rejected on BOTH available grounds, which is unusual. First, the arithmetically impossible variant: ebitAvg exceeds ebitdaAvg in every single forward year — FY2026 $19,044M against $17,794M, FY2027 $19,109M against $17,855M, FY2028 $20,027M against $18,713M, FY2029 $20,649M against $19,294M, FY2030 $21,807M against $20,375M. Earnings before interest and taxes cannot exceed earnings before interest, taxes, depreciation and amortisation. Second, the fixed-ratio fabrication signature: ebitdaAvg is exactly 35.78% of revenueAvg and ebitAvg exactly 38.30% in every one of those years. All forward valuation uses epsAvg, which verifies against actuals — the FY2025 estimate of $2.99006 against a delivered $3.04 (1.6% low).

Enterprise value — checked, and the noncontrolling-interest omission is present but small. enterpriseValueTTM of $403,064.6M is market capitalisation plus $30,636M, which is total debt less cash only, omitting the $622M of short-term investments and the $2,165M of noncontrolling interests. Corrected enterprise value is $404,608M (market cap + net debt $30,014M + NCI $2,165M) — 0.4% higher. The NCI omission this dive was directed to look for is present and is immaterial at Coca-Cola's scale. netDebtToEBITDATTM of 1.545x becomes 1.51x on corrected net debt.

6a. What today's price assumes (the inversion)

At $86.56 — 26.0x trailing, 24.5x FY2027 consensus, a 3.84% free cash flow yield — the price embeds:

• EPS reaches $3.532 in 2027 and $3.802 in 2028. (Consensus; 14 and 7 analysts.) Six-month 2026 EPS was $1.94, so FY2026's $3.305 needs $1.365 in the second half against $1.39 delivered in the second half of 2025 — a 1.8% decline after a first half that grew 17.6%. Whatever the sell-side has marked down, it is in the second half of this year, and the filings do not explain it.
• Revenue growth of 0.3% in 2027. (Consensus; 14 analysts.) Against a June quarter that grew 6.7% with unit case volume up 5%. This is the most fragile assumption in the price — not because it is too high, but because it is so low that it makes the 24.5x multiple hard to defend if it is right, and makes the consensus internally inconsistent if it is wrong.
• The 34.9% operating margin holds. (Our derivation from the 10-Q.) It is 6.2 points above the fiscal 2025 reported figure and 13.7 above fiscal 2024's, and both of those years were depressed by charges. The clean comparison is quarter-on-quarter, where 34.9% has now printed twice in a row.
• The market keeps paying 22–26x. (Our number.) At 20x FY2027E the stock is $71; at 28x it is $99. The band matters more here than anywhere else in this batch, because with 7% earnings growth and a 2.45% yield the multiple is not a rounding item — it is roughly half the ten-year outcome.
• fairlife costs nothing material. (Our number, and it is the only genuinely unknown input.) The company has expressly declined to say. The price contains no visible discount for it.

6b. The return bridge (why the multiple moves)

Expected return over the next twenty-four months decomposes as: EPS growth (+15.0% cumulative, from FY2026E $3.305 to FY2028E $3.802) + multiple drift (HELD, roughly 24x on the then-forward year throughout) + shareholder yield (+2.65% per year) ≈ +9% to +10% a year.

Our base of $88 applies 24.9x FY2027E, which is a hold rather than a compression or an expansion, and that is a deliberate choice we should defend. There is an argument for compression: 24.5x for 7% growth is a price-to-earnings-to-growth ratio above three, and the knowledge base's own historical observation is that this stock "stayed expensive for 25 years and underperformed since" after reaching 60x. There is an argument for expansion: a 0.349-beta, 43%-return-on-equity, sixty-year dividend grower is exactly what buyers pay up for when they are worried, and the stock has matched the index over twelve months on a third of its volatility.

We assume neither, and the consequence is that essentially the whole expected return is earnings growth plus the dividend. That is a robust decomposition and a boring one. It is also, precisely, why the verdict is Hold: a return that requires nothing from the crowd is a good thing to own and a poor thing to buy at 1.22:1.

6c. Variant perception (where we differ, what would surprise)

• We have essentially NO variant perception on the level and say so plainly. Our base of $88 is 8.2% below the street's $95.82, and that gap is the multiple, not the earnings — our EPS numbers are the consensus. We are not seeing anything the market is missing about the business. What we are doing is declining to pay 24.5x for 7% growth, which is a preference, not an insight.
• We do differ on the second half of 2026, and the difference is in the market's favour, not ours. Consensus implies second-half revenue 1.7% below last year's after a first half 9.2% above, and second-half EPS 1.8% below after a first half 17.6% above. Either the sell-side has priced the fairlife shutdown and a disposal that the company has not sized, or the estimate array is stale. Watchable number: the 2026-10-20 revenue print against $12,895M. A number materially above it says the second-half estimates are simply wrong and the stock is cheaper than 26.2x FY2026 suggests.
• We think the fairlife event is under-discussed relative to its cost basis. Coca-Cola paid a $6,173 million contingent milestone for this asset in March 2025 — an outlay large enough to turn a half-year's operating cash flow negative — and its US production has been stopped since 2026-07-16 with the impact expressly undetermined. The stock is 2.8% off its 52-week high.
• Positive surprise that would force a re-rate: a September quarter with revenue above $13.2B and unit case volume above 4%, which would falsify the second-half consensus outright; or completion of the held-for-sale disposal at a price that lifts group margin visibly, since Bottling Investments earns 7.4% against a 34.9% group figure.
• Negative surprise that would break the thesis: a quantified fairlife charge or an extended shutdown into the fourth quarter; a second impairment following BodyArmor's $960M; unit case volume returning to fiscal 2025's +1%; or a permanent North America appointment from outside, which would signal the interim arrangement is longer than intended.

Synthos fair values

All three anchors are multiples of the FY2027 consensus EPS distribution (mean $3.53205, low $3.44174, high $3.61232, 14 analysts), cross-checked against trailing free cash flow of $14,297M.

• Bear ~$69 — 20.0x the FY2027 consensus LOW of $3.442. Cross-check: 5.1% above the 52-week low of $65.67; a 4.8% free cash flow yield; a 3.07% dividend yield on $2.12. The scenario: the second-half consensus proves right, volume reverts to +1%, fairlife costs real money, and a 7%-growth staple re-rates to 20x. −20.3%.
• Base ~$88 — 24.9x the FY2027 consensus MEAN of $3.532. Cross-check: 26.6x FY2026E; 23.1x FY2028E; a 3.78% free cash flow yield; 8.2% below the street's $95.82. The scenario: growth settles at the consensus 7% a year, the margin holds near 34.9%, fairlife is a nuisance rather than a charge, and the multiple does not move. +1.7%.
• Bull ~$108 — 29.9x the FY2027 consensus HIGH of $3.612. Cross-check: 21.2% above the 52-week high; a 3.08% free cash flow yield; 12.7% above the street's high target of $104. The scenario: the volume acceleration of the June quarter proves durable, the remaining bottling disposals complete and lift the group margin toward 38%, and a defensive compounder re-rates in a nervous market. +24.8%.

Base is 1.7% above spot; asymmetry roughly 1.22:1 (20.3% down, 24.8% up), against a shareholder yield of 2.65%. A base case under 2% and a payoff ratio of 1.22:1 is not a Buy and is not close to one. It is a Hold — a security worth owning if owned, at a price that offers no reason to start.

7. Knowledge base — a deep bullish moat lane, a recent bearish price lane, and six claims we can now grade

Raw hits: 42. After a case-sensitive entity re-run: 39. Text-only, discarded: 3. Name-level claims about Coca-Cola: approximately 24. Channels: 13.

The homograph test was run first, and it passed. The sweep ran entity tokens KO, Coca-Cola, Coca Cola, The Coca-Cola Company, Coke, fairlife, BodyArmor and Costa Coffee across all 52,021 distilled claims. The two-letter token "KO" produced ZERO collisions — every claim carrying "KO" as an entity resolves to Coca-Cola on content inspection, and there were zero case-insensitive-only matches, which is the signature that would indicate a casing-variant collision. Given that this programme has confirmed a three-letter ticker matching a Taiwanese computer maker at conviction 85 and another matching a foundation rather than a company, a clean two-letter sweep is worth recording rather than assuming. Separately: no claim in this lane originates from the eli_lilly_ceo slug, so the confirmed cross-company contamination on that slug — which has carried Pfizer, Uber, Starbucks and other companies' claims — does not touch this name. Both checks were run because they were directed, and both came back clean.

The three discards are text-only and correctly discarded: a Josh Brown claim about insurance and Berkshire that mentions Coca-Cola as one of Berkshire's holdings (2026-07-28); a we_study_billionaires claim about Coloplast that carries "Coca-Cola FEMSA" as an entity (2023-10-12); and a we_study_billionaires claim about American Express that mentions Coca-Cola in passing (2022-10-17). None is a view on this company.

The concentration problem, named

Thirteen of the thirty-nine surviving claims — 33% — come from a single business_breakdowns episode dated 2024-02-18. That is the highest single-source concentration in this batch and it must be stated before the lane is used. Remove that one episode and the lane is 26 claims from 12 channels, and its centre of gravity shifts markedly toward neutral and negative. The remaining bullish claims are mostly about capital structure and brand principle (lyn_alden, we_study_billionaires) rather than about the stock at today's price.

The moat lane — deep, bullish, and two and a half years old

> 2024-02-18 · bullish · conviction 85 · thesis · business_breakdowns

> "Coke's real moat is its unrivaled bottler distribution network plus dominant marketing scale, creating a durable network effect competitors can't match."

> 2024-02-18 · bullish · conviction 80 · principle · business_breakdowns

> "Coke proved (New Coke) that the brand, not the product, is the asset — marketing and mindshare, not taste, sustain the moat."

> 2024-02-18 · bullish · conviction 76 · thesis · business_breakdowns

> "On beverages Coke is twice Pepsi's operating income and 4x its nearest pure-beverage peer, out-spending rivals 3-5x on marketing and taking value share."

These are descriptions of a competitive position, they are almost certainly right, and they are entirely consistent with a 61.9% gross margin and a 34.9% operating margin. They say nothing about whether $86.56 is a good price.

Six claims from that episode are now testable. We grade them.

> "Steady ~3% volume growth plus ~2-3% inflationary pricing yields 5-6% system revenue growth, well above the consumer-staples peer group." (conviction 78)

>

> VERDICT: FALSIFIED on the delivered years, MET in the current one. Fiscal 2024 revenue grew 2.9%, fiscal 2025 1.9% with unit case volume +1%. But the June 2026 quarter grew 6.7% with volume +5%, and the six months grew 9.2% with volume +4% — comfortably inside the claimed range. Consensus then says FY2027 revenue grows 0.3%. The claim was wrong for two years and is right today; that is a genuine split verdict and it is reported as one.

> "Post-refranchising, EBIT margins rise from ~30% toward 35-40% as capital-heavy bottling exits the balance sheet, lifting cash conversion." (conviction 70)

>

> VERDICT: MET at the low end. Second-quarter and first-quarter 2026 operating margins were both 34.9%. Bottling Investments has fallen to 11.4% of revenue and earns 7.4% against a group 34.9%, and $5,438M of assets held for sale says the exit continues. This claim has worked.

> "Growing FCF ($10B toward $12B), an unleveraged balance sheet and 60+ year dividend support a stable ~75% payout, more buybacks and further acquisitions." (conviction 72)

>

> VERDICT: BEATEN on cash flow, FALSIFIED on buybacks and on "unleveraged." Trailing free cash flow is $14,297M, above the $12B target. But repurchases fell from $2,289M (FY2023) to $1,795M (FY2024) to $746M (FY2025) — the opposite of "more buybacks" — and net debt is $30,014M at 1.51x EBITDA, which is not an unleveraged balance sheet. The payout is $8,779M against normalised free cash flow of roughly $11,365M, about 77% — the claim's one accurate leg.

> "Franchising bottling/distribution lets Coke earn ~30% ROIC as a capital-light concentrate franchisor while bottlers handle capital-heavy local execution." (conviction 80)

>

> VERDICT: NOT SUPPORTED on the reported figures. The payload's returnOnInvestedCapitalTTM is 13.9%, returnOnCapitalEmployedTTM 18.0% and returnOnEquityTTM 43.0%. The gap is goodwill and intangibles — $15,456M of goodwill and $12,500M of indefinite-lived trademarks against $107,922M of assets, 25.9% of the balance sheet — which a 30%-ROIC framing implicitly excludes. The 30% figure is defensible on a tangible-capital basis and is not what the reported ratios say. We record the discrepancy rather than picking whichever number flatters.

> "Costa can grow via international vending expansion, but the big prize — a successful ready-to-drink coffee — remains unproven; not a near-term needle-mover." (neutral, conviction 50)

>

> VERDICT: STILL TRUE. Costa is not separately disclosed in any filing in this archive. Two and a half years on, the claim's caution is unrefuted.

> "The 'Total Beverage Company' strategy — 26 billion-dollar brands plus acquisitions in coffee, water, sports and alcohol — extends growth beyond mature Cola." (conviction 62)

>

> VERDICT: MIXED, tilting negative on the acquisition leg. Fiscal 2025 impaired the BodyArmor trademark by $960 million and a Latin America trademark by $44 million. fairlife, the successful one, has had US production suspended since 2026-07-16. The organic portfolio is doing the work; the acquired portfolio is not yet.

The price lane — recent, sparse, and negative

> 2025-10-03 · BEARISH · conviction 65 · thesis · compound_and_friends

> "Coca-Cola is a spectacular company but, based on current trajectory, the stock is priced for its next decade of growth and may be flat in 10 years."

> 2025-10-03 · BEARISH · conviction 60 · thesis · compound_and_friends

> "Coca-Cola is a great company but the stock likely sits at the same price in 10 years — already priced for a decade of growth."

> 2025-04-29 · neutral · conviction 50 · thesis · invest_like_the_best

> "Coca-Cola is obviously a great brand-loyal business, but that quality is widely understood and priced in — little edge remains."

> 2025-10-30 · neutral · conviction 65 · fact · we_study_billionaires

> "Buffett bought Coke at 12-14x earnings for international expansion; it rose ~14x in a decade to 60x earnings, then stayed expensive for 25 years and underperformed since."

These four are the most recent, most specific and most directly relevant claims in the lane, and every one of them is a warning about price rather than about business quality. The two compound_and_friends claims are the same view stated twice on the same day, so they count as one independent source, not two — and this dive counts it that way. The arithmetic in Section 6 arrives at the same place independently: 24.5x for 7% growth, with the whole return being earnings plus a 2.45% dividend.

The remaining lane, briefly

Bullish on capital structure: lyn_alden 2026-03-31 (conviction 75, skill 1.1) — "The best product Coca-Cola sold was its bonds, not Coke — profitable firms issue cheap long-dated debt to short the currency and buy scarcer assets" — and a 2023-12-02 restatement of the same principle. Coca-Cola does carry $43,543M of debt at falling interest cost, so the mechanism described is real.

Bullish on disruption-resistance: compound_and_friends 2026-02-13 (conviction 65) — "physical, non-replicable businesses like Coke and Pepsi are safe from chatbots" — and 2026-02-03 (conviction 60) — "asset-heavy, low-margin staples are ideal AI beneficiaries; an AI layer streamlines operations and drives margin improvement even without volume growth." The second is testable and the six-month margin of 34.9% is consistent with it, though the charge-driven base makes attribution impossible.

Bullish with weakened attribution: ray_dalio 2026-01-28 (conviction 70, skill 0.7, speaker: uncertain, speaker_role: independent) — "Allocate 30-35% to businesses with pricing power that raise prices faster than inflation — dividend aristocrats in staples (PG, KO, CL)... payout below 60%." A detail worth flagging: Coca-Cola's payout is above that threshold — $8,779M of dividends against $13,107M of fiscal 2025 net income is 67%, and against normalised free cash flow roughly 77%. On the claim's own criterion, KO does not qualify. A second ray_dalio claim from 2023-07-15 (conviction 45, skill 0.7) recommends KO alongside Cisco and Starbucks.

Excluded on content: nassim_taleb 2023-04-07 (conviction 60, speaker: uncertain) carries KO as its only entity but the claim is "There's no platonic 'true' worth of a share; price is just where buyers and sellers met and is always wrong." That is a statement about price discovery in general, using KO as a placeholder. It is excluded from the name-level count and reported here so the exclusion is visible.

Attribution across the lane. Thirty-six of thirty-nine claims carry speaker: null and are channel-attributed only — the weakest sourcing the four-lane policy admits. Three name a speaker: two ray_dalio claims marked uncertain and one Josh Brown claim among the discards. speaker_role reads independent on the three that carry it and is not relied on. NO claim in this lane is management voice, so the half-weight rule does not engage anywhere.

Conclusion. Breadth 39, name-level claim count ~24, net conviction mixed — positive on the moat, negative on the price. The knowledge base's view of Coca-Cola as a business is emphatic and favourable and two and a half years old. Its view of Coca-Cola as a stock is recent, sparse and unfavourable. The verdict follows the second, because the second is the question.

8. Data integrity — what we rejected and why

Eight findings: two clean checks and six defects. The two segment blocks are the worst of them, and they fail in a way this programme has not previously recorded.

CLEAN 1 — capitalExpenditure verified EXACTLY against three filed years. FY2023 $1,852M, FY2024 $2,064M, FY2025 $2,112M — the vendor and the 10-K agree to the dollar in all three, and the 10-K states them twice, in the cash flow statement and in the liquidity discussion. The diagnostic passes too: capexToDepreciationTTM of 1.978 reconciles exactly to trailing capex of $2,045M over trailing depreciation of $1,034M. Both the datum and the diagnostic are clean, and the resulting free cash flow of $14,297M is used throughout.

CLEAN 2 — the June-quarter income statement, balance sheet and share count all tie to the 10-Q. Revenue $13,380M, gross profit $8,415M, operating income $4,672M, net income attributable $4,425M, diluted EPS $1.03, diluted shares 4,313M — every one matches inc_q. Market capitalisation divided by price implies 4,302,548,010 shares against the 10-Q's 4,303M basic average — a 0.01% match. km_ttm is present and complete; the truncation defect found on other names does not appear here.

DEFECT 1 — seg_prod reports 77.8% of revenue under two labels, one of which is a mislabelled sum of three geographic segments. The FY2025 block contains exactly two lines: "Pacific" $31,558M and "Bottling investments" $5,735M, summing to $37,293M against revenue of $47,941M. Coca-Cola has five operating segments. The "Pacific" figure is not Asia Pacific — Asia Pacific was 11.1% of fiscal 2025 revenue, roughly $5,321M. $31,558M is exactly the sum of the three lines in the seg_geo block (North America $19,586M + Latin America $6,334M + Pacific $5,638M). So seg_prod's largest line is a geographic aggregate wearing a segment label, and EMEA — 22.6% of revenue, roughly $10,835M — is absent from the block entirely, as are Corporate and eliminations. The FY2024 block is structurally different again (five lines summing to exactly $47,061M, with "Pacific" at $38,776M and a "Global Ventures" line that no longer exists in the current segment structure). The block is rejected in full. All segment figures in this dive come from the 10-K and 10-Q.

DEFECT 2 — seg_geo omits EMEA, the second-largest segment, and covers only 65.8% of revenue. The FY2025 block is North America $19,586M, Latin America $6,334M, Pacific $5,638M — $31,558M against $47,941M. The FY2024 block contains Europe at $8,122M; the FY2025 block does not. This is the geography-block-missing-a-market defect, and although the omitted market is not the home market — North America is present and correct — the omitted segment is 22.6% of revenue. The block is rejected. Note the interlock with DEFECT 1: the same $31,558M appears as a single line in seg_prod, so the two blocks are contaminated by one another and neither can be repaired from the other.

DEFECT 3 — est.ebitAvg EXCEEDS est.ebitdaAvg in every forward year, and both carry the fixed-ratio signature. FY2026 through FY2030: $19,044M vs $17,794M, $19,109M vs $17,855M, $20,027M vs $18,713M, $20,649M vs $19,294M, $21,807M vs $20,375M. This is the arithmetically impossible variant. Separately, ebitdaAvg is exactly 35.78% of revenueAvg and ebitAvg exactly 38.30% in all five years. Both rows rejected; all forward valuation uses epsAvg.

DEFECT 4 — the annual free cash flow rows are real cash but structurally misleading without the filing, and the error runs against the company. freeCashFlow of $5,296M for fiscal 2025 and $4,741M for fiscal 2024 both absorb roughly $6 billion of non-recurring operating outflow — the $6,069M fairlife final milestone in 2025 and the $6.0 billion IRS Tax Litigation Deposit in 2024, both identified by name in the 10-K. A reader taking $5,296M at face value computes a 1.42% free cash flow yield and a dividend covered 0.60x. The trailing figure, which excludes both, is $14,297M — a 3.84% yield covering the forward dividend 1.57x. This is not a vendor error; the annual numbers are correct. It is a trap for anyone who does not read the filing, which is exactly what the standing rule exists to catch.

DEFECT 5 — capitalLeaseObligations reads zero in fiscal 2025 after $923M in fiscal 2024 and $1,001M in fiscal 2023. Lease obligations do not vanish. totalDebt is therefore not comparable across years: FY2024's $45,735M is short-term debt $2,437M plus long-term $42,375M plus leases $923M, while FY2025's $45,492M is $3,373M plus $42,119M alone. This is the lease-inclusion inconsistency rather than the double-counting defect, and it is the same signature found on another name in this batch. We use the 10-Q's filed components — $48M of loans and notes payable, $6,494M of current maturities and $37,001M of long-term debt, total $43,543M at 2026-07-03.

DEFECT 6 — netDebt and enterpriseValueTTM omit short-term investments AND noncontrolling interests. netDebt of $35,222M for fiscal 2025 is total debt less cash only, ignoring $3,602M of short-term investments. The trailing enterprise value of $403,064.6M implies net debt of $30,636M — total debt of $43,543M less cash only of $12,907M, omitting $622M of short-term investments — and adds nothing for the $2,165M of noncontrolling interests. Corrected enterprise value is $404,608M, 0.4% higher. The noncontrolling-interest omission this batch was directed to look for is confirmed present and is immaterial here at 0.5% of enterprise value, in contrast to the $50,729M (35.6%) omission confirmed on an alternative-asset manager.

Also checked and noted: earn_cal reports June-quarter revenue of $13,373M against the 10-Q's $13,380M — a 0.05% difference; the filing wins. dividendPerShareTTM of $2.08 is trailing; the July 2026 board declaration of $0.53 per quarter gives a forward $2.12, and we use the filing. researchAndDevelopmentExpenses reads 0 in every year — Coca-Cola does not report a separate research line, so this is an absence rather than an error. quote.yearHigh/yearLow ($90.92/$65.35) against tech.hi52/lo52 ($89.08/$65.67), a 2.0% and 0.5% discrepancy; tech is used throughout.

Vendor composite rating — noted. B / 3 overall, with 5 out of 5 on both return on equity and return on assets, 1 out of 5 on debt-to-equity and price-to-book, and 2 out of 5 on price-to-earnings. The block's own arithmetic reaches this dive's conclusion: the returns are excellent and the price is not. The discounted-cash-flow sub-score of 3 rests on a free cash flow figure this dive has recomputed and carries no weight.

Non-equity tripwire — checked and passed. KO is common stock, $0.25 par value, NYSE-listed, per the 10-Q balance sheet (11,200M authorised, 7,040M issued, 2,737M in treasury). Price of $86.56 is not par-like; beta is 0.349; the dividend is variable and was raised at the July 2026 board meeting; volume was 15.1M shares (roughly $1.31B of turnover); the 52-week band of $65.67 to $89.08 is a 36% range. This is common equity.

9. Technicals

• Price $86.56, down 0.35% on the day on 15,119,412 shares. Opened $85.79, ranged $85.16 to $86.68.
• −2.8% from the 52-week high of $89.08; +31.8% above the 52-week low of $65.67. Position within the annual range: 89th percentile — the highest in this batch.
• Above both moving averages, with the shorter above the longer: +5.3% above a 50-day average of $82.23 and +13.5% above a 200-day average of $76.28. This is a clean, intact uptrend.
• RSI 61.3 — strong, not overbought. MACD +1.50, positive.
• Maximum drawdown from peak over the trailing year: −2.8% — by a wide margin the shallowest in this batch, which is what a 0.349 beta looks like in practice.
• Relative performance: 3-month +10.7% against SPY +7.6% and QQQ +7.7%; 6-month +14.9% against SPY +11.1%; 12-month +25.7% against SPY +24.3% and QQQ +30.8%. A stock with a third of the market's volatility has beaten the market over twelve months. That is an unusual and under-remarked fact, and it is also the reason the multiple is 24.5x rather than 20x.
• Sentiment: 29 buy, 16 hold, 3 sell across 48 analysts; consensus target $95.82 (+10.7%), median $95, high $104 (+20.1%), low $86 — 0.6% below spot, i.e. essentially at the price. The lowest analyst target on the street is the current price, which is a fair description of a fully-valued defensive.

Today's move and what it does to the entry

KO closed 2026-08-04 at $86.56, down $0.30 from $86.86, on 15.1 million shares — a 0.35% move in a stock whose annual range is 36% wide. There is no company-specific filing dated 2026-08-04 in this archive. The most recent filings are the 10-Q of 2026-07-29 and the earnings 8-K of 2026-07-28.

The honest read: this is a fully-priced defensive sitting 2.8% below its high, and there is no entry discount of any kind. The base case is +1.7%. A position taken here is a decision about what to own, not a decision about what to buy, and the dive is graded on that distinction.

10. Insiders — two officers, both exercise-and-sell, both immediately after the print

Total open-market sales: 679,570 shares for approximately $61.0 million, by two officers, across four days.

The reading, stated carefully because the transaction structure matters. Every sale is paired with an M-Exempt option exercise at a strike of $44.475 or $45.435 — roughly half the sale price. These are exercise-and-sell transactions, the mechanically cheapest way to monetise expiring options, and they are the least informative category of insider activity after tax withholding. They are not the same thing as an executive selling shares he already owned.

Three observations that survive that qualification.

First, the timing. Quincey's two sales were executed on 2026-07-28 and 2026-07-29 — the day of the second-quarter earnings release and the day of the 10-Q filing — at $90.04 and $90.09, 4.0% above today's close. Murphy's was 2026-07-31 at $87.31.

Second, Quincey sold his entire exercised position both times and ended at 122,833 shares held. From 503,973 to 122,833 across two days is a 75.6% reduction, and the second exercise-and-sell on 2026-07-29 brought him back to the identical 122,833. The Chairman's directly-held position is now roughly $10.6 million, against $61 million of stock monetised in four days.

Third, Murphy's M-Exempt disposition line ends at zero securities owned in that reporting chain, and he retains 279,917 shares — approximately $24.2 million. He took on the North America operating unit two days after this transaction.

What the file does not contain: a single purchase, by anyone, and no transaction at all by Henrique Braun, the chief executive. On a name where the chief executive is new and the payload and the insider file disagree about who holds which title, the absence of any transaction by the chief executive is uninformative rather than reassuring.

11. Verdict, kill-criteria and flip conditions

Hold.

The business is not in question and this dive spends no words pretending otherwise. A 61.9% gross margin, a 34.9% operating margin in each of the last two quarters, a 43.0% return on equity, $14,297M of trailing free cash flow on a verified capital-expenditure line, 1.51x net leverage with interest expense falling 10.6% year on year, a 0.349 beta, and a dividend raised for more than sixty consecutive years and most recently declared at $0.53 per quarter. Second-quarter revenue +6.7% with unit case volume +5% and all five operating segments growing. The knowledge base's assessment of the moat — an unrivalled bottler network, marketing spend three to five times its rivals', twice PepsiCo's beverage operating income — is almost certainly correct.

The price is where the argument ends. At $86.56 that is 26.0x trailing earnings and 24.5x the FY2027 consensus of $3.532, for 7% annual earnings growth, a 2.45% forward dividend and a buyback of $746M — 0.2% of the market capitalisation — which is not a capital-return programme. Our base of $88 is 1.7% above spot with 1.22:1 asymmetry. The expected return is the earnings growth plus the dividend, roughly 9-10% a year, entirely contingent on the market continuing to pay 24x.

Three specific reservations beyond the multiple.

First, fairlife. US production has been suspended since 2026-07-16 by a ransomware event whose impact the company expressly says is "not yet known." This is an asset for which Coca-Cola paid a $6,173 million final milestone in March 2025. The stock sits 2.8% below its 52-week high with no visible discount for it, and the first quantifying disclosure is 77 days away.

Second, the second-half consensus is inexplicable from the filings. Consensus implies second-half 2026 revenue 1.7% BELOW last year after a first half 9.2% above, and second-half EPS 1.8% below after a first half 17.6% above. Either something is priced that the company has not disclosed, or the estimate array is stale. We do not know which, and we say so.

Third, the leadership of the largest segment. North America — 40.4% of revenue, the lowest-margin branded segment at 25.9%, and the one that absorbed a $960M impairment last year — has been run on an interim basis by the Chief Financial Officer since 2026-08-01, four days ago.

And one that is not a reservation but a discipline: the knowledge base agrees with the arithmetic. The three most recent, most specific claims about this stock — from two independent channels — say the quality is understood, priced in, and that the shares "may be flat in 10 years." We reached the same conclusion from the estimate file without reference to them, which is the strongest form of corroboration available here.

Pre-registered KILL criteria — what would take this to Avoid:

• A quantified fairlife charge above roughly $500 million, or an extended US production suspension running into the December quarter.
• A second trademark impairment following the $960M BodyArmor and $44M Latin America write-downs of fiscal 2025.
• Unit case volume returning to fiscal 2025's +1% for two consecutive quarters, which would falsify the current acceleration and leave 24.5x indefensible.
• The operating margin falling below 32% on a clean, charge-free quarter.
• Net leverage above 2.0x EBITDA while the dividend continues to grow.

Pre-registered UPGRADE conditions — what would take this to Buy — Tactical:

• A price below roughly $74, at which the base case is +19% and the payoff ratio moves through 2:1 with no change in our numbers. This is the cleanest path and it requires nothing from the company.
• A September-quarter revenue print materially above the $12,895M consensus with volume above 4%, which would falsify the second-half estimates outright and mean the stock is cheaper on FY2026 than 26.2x implies.
• Completion of the held-for-sale disposal at a price that lifts the group operating margin toward 38%, since the exiting business earns 7.4% against a group 34.9%.
• A material acceleration in the buyback from $746M — the one lever that would convert a 2.65% shareholder yield into something that changes the return arithmetic.

Where KO sits in the Synthos Framework Portfolio. No new position. Where held, hold: a 0.349-beta compounder producing 9-10% expected annual returns with the shallowest drawdown in this batch is a legitimate ballast holding, and nothing in this analysis argues for selling it. On the watch list at $74 for a first tranche of 1.5% in the consumer-defensive sleeve. Logged as a tracked Synthos call (Hold) as of 2026-08-04 at $86.56, with the fair-value anchors, kill criteria and upgrade conditions all gradeable. The distinction being graded here is between a good business and a good purchase, and Coca-Cola at $86.56 is unambiguously the first and not the second.

Single biggest risk: the multiple, not the business. At 24.5x the FY2027 consensus on 7% earnings growth and a 2.45% dividend, the price-to-earnings-to-growth ratio is above three and the entire ten-year outcome turns on whether buyers still pay 24x in 2036. The knowledge base holds the historical precedent and it is not encouraging: this stock reached 60x earnings, "then stayed expensive for 25 years and underperformed since." A de-rating to 20x FY2027 consensus is $71 — a 18.4% decline with no deterioration in the business whatsoever. That is the shape of the downside here: not a broken company, a re-priced one.

Provenance & disclosures

• Traceability: approximately 24 name-level knowledge-base claims name Coca-Cola out of 52,021 distilled claims. Raw hits 42; surviving a case-sensitive entity re-run 39; discarded as text-only 3; breadth 39 across 13 channels; net conviction mixed — positive on the moat, negative on the price. The homograph test on the two-letter token "KO" was run FIRST and PASSED with zero collisions and zero case-insensitive-only matches, and no claim in this lane originates from the eli_lilly_ceo slug, so the confirmed cross-company contamination on that slug does not touch this name; both checks were directed and both came back clean. Thirteen of the 39 claims — 33% — come from a single business_breakdowns episode dated 2024-02-18, a concentration named in Section 7 and tested by removing it. Six claims from that episode are now testable against filings and are graded individually: one FALSIFIED then MET, one MET at the low end, one BEATEN on cash flow but FALSIFIED on buybacks and leverage, one NOT SUPPORTED on the reported return ratios, one STILL TRUE, one MIXED. The most recent and most specific claims are negative: compound_and_friends 2025-10-03 (BEARISH, conviction 65 and 60 — the same view stated twice on one day, counted as ONE independent source), invest_like_the_best 2025-04-29 (neutral, 50) and we_study_billionaires 2025-10-30 (neutral, 65). Thirty-six of 39 carry speaker: null; two ray_dalio claims carry speaker: uncertain with skill 0.7 and are weighted down; one nassim_taleb claim carrying KO as its only entity is EXCLUDED on content because it is a statement about price discovery in general. No claim in this lane is management voice. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals — income statement, balance sheet, cash flows and segment results through 2026-07-03, all filing-verified from the 10-Q filed 2026-07-29; annual figures for FY2025, FY2024 and FY2023 from the 10-K filed 2026-02-20 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873602 = 2026-08-04T20:00:02Z ($86.56, −0.35%; 50-DMA $82.23; 200-DMA $76.28; RSI 61.3; MACD +1.50) · knowledge-base claims 2026-08-04. Coca-Cola's fiscal year ends 31 December; its quarters end on a 52/53-week convention, so the June 2026 quarter ended 2026-07-03 and the June 2025 quarter ended 2025-06-27. All figures come from the Synthos vendor data file for KO or from the SEC filings in the KO archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents: 10-K filed 2026-02-20 (fiscal 2025 — the source of the three-year capital-expenditure series, the other-operating-charges detail including the $960M BodyArmor impairment, the segment revenue and margin percentages, and the identification of the $6.0 billion IRS deposit and $6.1 billion fairlife milestone); 10-Q filed 2026-04-30 (March 2026 quarter); 10-Q filed 2026-07-29 (June 2026 quarter — the primary source for this dive: income statement, balance sheet, cash flows, segment revenue, volume decomposition, dividend declaration and the officer certifications establishing that Henrique Braun is chief executive); 8-K filed 2026-06-25, Item 5.02 — the departure of Jennifer Mann as President of the North America operating unit and John Murphy's interim assumption of the role from 2026-08-01; 8-K filed 2026-07-16, Item 8.01 — the fairlife ransomware event and the suspension of US production; 8-K filed 2026-07-28, Item 2.02 — second-quarter results, cover page and exhibit pointer only; Exhibit 99.1 is not in the extract. All carry preserved [TABLE] statement data.
• What the complete archive changed versus the earlier truncated one. The earlier 2026-08-04 dive recorded in its own provenance section that its 8-Ks were "present as XBRL cover-page tags only — they contain a list of listed debt securities and nothing else," and that its pull included a stale 8-K from 2021-12-20. The complete archive replaces those with three current 8-Ks carrying full Item text, and two of them disclose material events that the earlier dive could not have seen and that appear nowhere in the vendor payload: the fairlife ransomware event of 2026-07-16, which has suspended US production of a business acquired with a $6,173 million final milestone payment, and the 2026-06-25 departure of the President of North America — 40.4% of revenue — with the Chief Financial Officer assuming the unit on an interim basis from 2026-08-01. What did NOT change: the 2026-07-28 earnings 8-K still carries only the Item 2.02 pointer and no Exhibit 99.1, which is immaterial because the 10-Q filed the following day carries the complete statements. The verdict did not change — Hold, then and now. The base fair value moves from $89 to $88, and the reason is the fairlife shutdown, which is real, dated, disclosed and expressly unquantified by the company. The earlier dive's knowledge-base work and its claim-grading were struck against the same complete claim store and are carried forward and extended here rather than replaced.
• Where the filings contradicted or corrected the vendor (detailed in Section 8): seg_prod, which for fiscal 2025 reports only two lines totalling $37,293M against revenue of $47,941M (77.8%), of which the larger — labelled "Pacific" at $31,558M — is arithmetically the sum of the three lines in seg_geo and is therefore a geographic aggregate wearing a segment label, with EMEA (22.6% of revenue) absent entirely; seg_geo, which covers only 65.8% of revenue and omits EMEA, present in the FY2024 block at $8,122M and gone from FY2025; the annual free cash flow rows, which are correct as cash but misleading without the 10-K's identification of $6,069M of fairlife milestone in fiscal 2025 and the $6.0 billion IRS Tax Litigation Deposit in fiscal 2024 — the trailing figure of $14,297M is the run rate, not the annual $5,296M; capitalLeaseObligations, reading zero in FY2025 after $923M in FY2024 and $1,001M in FY2023, making totalDebt non-comparable across years, against the 10-Q's filed components totalling $43,543M; netDebt and enterpriseValueTTM, which omit short-term investments and the $2,165M of noncontrolling interests, giving a corrected enterprise value of $404,608M against the vendor's $403,064.6M; est.ebitAvg, which exceeds est.ebitdaAvg in every forward year — the arithmetically impossible variant — while both carry a fixed-ratio signature of exactly 35.78% and 38.30% of revenue; and earn_cal, reporting June-quarter revenue of $13,373M against the 10-Q's $13,380M. Where vendor and filing AGREED — recorded, because these were the checks this batch was directed to run: the capex check PASSED on all three filed years EXACTLY ($1,852M / $2,064M / $2,112M) and on the capexToDepreciationTTM diagnostic (1.978, reconciling to $2,045M/$1,034M); the June-quarter income statement ties to the 10-Q on every line; the implied share count of 4,302,548,010 matches the 10-Q's 4,303M basic average to 0.01%; km_ttm is complete, not truncated; and the profile chief-executive field naming Henrique Braun is confirmed by the 10-Q certification, against an insider file that shows only James Quincey as Chairman.
• Basis note: analyst EPS estimates for Coca-Cola track a comparable-operations basis that runs close to GAAP diluted EPS — the FY2025 estimate of $2.99006 against a delivered $3.04 — so all forward multiples are struck on epsAvg and the trailing multiple of 26.0x uses netIncomePerShareTTM of $3.328, which reconciles to the sum of the last four quarterly diluted EPS figures ($0.86 + $0.53 + $0.91 + $1.03 = $3.33). Note that the trailing figure is depressed by the $960M BodyArmor impairment in the December 2025 quarter, which is why the FY2027 forward estimate is the valuation anchor.
• Estimate coverage: 13 analysts on FY2026 EPS and 14 on FY2027, the anchor for all three fair values, with an FY2027 range of $3.442 to $3.612 (4.9% — the tightest in this batch); 7 on FY2028; 2 on FY2029, excluded from every conclusion; 5 on FY2030. That FY2030 carries more coverage than FY2029 is itself a reason to distrust the far end of the array.
• Peer note: the vendor peer set — Anheuser-Busch InBev, Coca-Cola Europacific Partners, Coca-Cola Consolidated, Keurig Dr Pepper, Coca-Cola FEMSA, Monster Beverage, PepsiCo, Procter & Gamble, Philip Morris International and Unilever — is unusually well constructed for this programme and contains three of Coca-Cola's own bottlers. But no earnings or estimate data for any peer is present in this file, only price and market capitalisation, so no peer-multiple comparison is drawn anywhere in this dive.
• Fair-value caveat: the $69 / $88 / $108 anchors are multiples of the FY2027 consensus EPS distribution — 20.0x the low of $3.44174, 24.9x the mean of $3.53205, and 29.9x the high of $3.61232 — cross-checked against trailing free cash flow of $14,297M and against the $2.12 forward dividend. Stated arithmetic, not a discounted cash flow. The base assumes the multiple HOLDS: no expansion and no compression, with the entire expected return coming from earnings growth and the dividend.
• Timing: second-quarter 2026 results were released 2026-07-28, seven days before this dive, and beat the consensus EPS estimate by 5.4% ($0.97 against $0.92) and revenue by 1.5%. The 10-Q followed on 2026-07-29. The next print is 2026-10-20, 77 days away, against consensus of $0.87 and $12,895,220K — the first quarter that will carry the fairlife shutdown. The most recent insider filings are dated 2026-07-28, 2026-07-29 and 2026-07-31, and all three sale transactions are exercise-and-sell pairs. 2026-08-04 carried no company-specific SEC filing in this archive.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full-rewrite.