CVS Health CVS

⚠ EARNINGS IN ONE DAY — 2026-08-05

CVS Health reports second-quarter 2026 results on 2026-08-05, the morning after this dive is struck. Vendor consensus is adjusted EPS $1.86 and revenue $100.028B. The company has beaten the consensus adjusted EPS estimate in each of the last five reported quarters — by 32.4%, 24.0%, 16.8%, 9.0% and 17.9%, averaging approximately +20%. Management raised full-year guidance on 6 May to adjusted EPS of $7.30-7.50 and operating cash flow of at least $9.5 billion. The number that matters is not the headline but the medical benefit ratio, which fell from 87.3% to 84.6% in the March quarter and is the sole reason the guidance moved. This verdict is set with tomorrow's print as an explicit input.

Reference table

1. What the business is, and the segment data the payload does not carry

Chairman and chief executive J. David Joyner; approximately 300,000 employees. Four reportable segments, per the 10-Q: "Health Care Benefits, Health Services, Pharmacy & Consumer Wellness and Corporate/Other."

The March 2026 quarter by segment, from the 10-Q ($M):

Read the adjusted-operating-income row. Health Care Benefits — Aetna — produced 59.0% of segment operating income on 31% of gross segment revenue. Health Services, which contains both Caremark and the impaired care-delivery assets, produced $1,489 million on $48,237 million of gross revenue — a 3.1% margin. Pharmacy & Consumer Wellness produced $1,197 million on $31,989 million, 3.7%. Corporate/Other cost $577 million.

The reconciliation from adjusted operating income to GAAP operating income is small and clean, which is worth recording: $5,150M adjusted less $442M of intangible amortisation, $16M of net realised capital losses and $12M of acquisition-related integration costs gives $4,680M of GAAP operating income. A 9.1% adjustment is modest by the standards of this batch — Danaher's is 26%, Bristol Myers Squibb's is 26%.

The vendor's seg_prod block contains none of this. Its FY2025 entry is a revenue-type disaggregation:

These are useful figures — premiums grew 35.8% in two years while front-store revenue fell 4.4% — but they are not segments, and the FY2025 total of $389,161M is $12.9 billion short of reported revenue of $402,067M. Note also the "Product and service, other" line falling from $15,966M to $3,926M in one year, a 75% drop with no corresponding disclosure, which suggests reclassification rather than economics.

The vendor's seg_geo block is EMPTY. There are no entries at all. CVS is a domestic business, so the omission costs little analytically, but it is a null field rather than a disclosure.

2. The wreckage in the trailing numbers, and why the 45x P/E is not information

The annual series shows a company whose revenue compounded and whose earnings collapsed:

Revenue rose 49.6% across five years; GAAP earnings per share fell 74% from the FY2023 peak. The 10-K explains FY2025 directly:

> "Operating income decreased $3.9 billion, or 45.3%, in 2025 compared to 2024. The decrease in operating income was primarily due to a $5.7 billion goodwill impairment charge related to the Health Care Delivery reporting unit and the $1.2 billion of legacy litigation charges described above, both recorded during the year ended December 31, 2025. These decreases were partially offset by improved operating performance in the Health Care Benefits segment and the absence of approximately $1.2 billion of restructuring charges recorded in the prior year."

And the impairment's cause, which is worth reading carefully:

> "The results of the impairment test showed that the fair value of the Health Care Delivery reporting unit was lower than its carrying value, resulting in a $5.7 billion goodwill impairment charge... The fair value of the Health Care Delivery reporting unit was determined using a combination of a discounted cash flow method and a market multiple method. In addition to the lower financial projections, lower market multiples of the peer group companies contributed to the amount of the goodwill impairment charge."

Two drivers, and only one is within management's control. Lower financial projections is an operating failure; lower peer-group market multiples is not. $85,478 million of goodwill remains, 32.7% of total assets, and the same mechanism can produce another charge without anything changing inside the business.

The quarterly series, which shows exactly where the damage sits:

Three observations.

First, the trailing twelve months in the vendor payload run from the June 2025 quarter to the March 2026 quarter and therefore CONTAIN the $5.7 billion impairment. Trailing GAAP net income is $2,932M and trailing GAAP EPS is $2.30, giving a headline price-to-earnings ratio of 45.2x. That number is arithmetic, not information, and it is not used for any conclusion in this dive. Trailing adjusted EPS across the same four quarters is $7.07, giving 14.8x.

Second, the December 2025 quarter carried a tax BENEFIT of $1,569 million against pre-tax income of $1,354 million — producing net income of $2,943 million from a pre-tax figure less than half that. The 10-K explains: "The Company's effective income tax rate decreased to 19.1% in 2025... due to a worthless stock deduction associated with a subsidiary that filed for bankruptcy in 2025." That is a genuine, non-recurring, non-operating benefit and it flatters the FY2025 GAAP figure that the impairment had already crushed.

Third, and most importantly: adjusted EPS beat consensus in all five quarters shown, by 9.0% to 32.4%, averaging approximately 20%. Estimates have been persistently too low throughout the recovery, which is the single most useful pattern in the file going into tomorrow's print — and equally, a pattern the market has now had five quarters to learn.

3. The medical benefit ratio, and the guidance raise

From the 8-K filed 2026-05-06, which carries the full first-quarter release:

> "First quarter total revenues increased to $100.4 billion, up 6.2% year-over-year · First quarter GAAP diluted EPS of $2.30 and Adjusted EPS of $2.57 · Raising full-year 2026 guidance: GAAP diluted EPS guidance range to $6.24 to $6.44 from $5.94 to $6.14; Adjusted EPS guidance range to $7.30 to $7.50 from $7.00 to $7.20; Cash flow from operations guidance to at least $9.5 billion from at least $9.0 billion."

And the reason, in management's own framing — half-weighted as management voice:

> "The Company is increasing its full-year 2026 GAAP diluted EPS, Adjusted EPS and cash flow from operations guidance to reflect increases in the Health Care Benefits and Pharmacy & Consumer Wellness segments, while maintaining a cautious view for the remainder of the year in light of continued elevated cost trends and the potential for macro headwinds."

The mechanism is one ratio:

The 10-Q defines it precisely: "Medical benefit ratio is calculated by dividing the Health Care Benefits segment's health care costs by premium revenues and represents the percentage of premium revenues spent on medical benefits for the segment's insured members."

The arithmetic is direct and it is the whole investment case. Health Care Benefits health-care costs were $28,579 million in the March quarter. At an 84.6% ratio, premium revenue was approximately $33.8 billion for the quarter, roughly $135 billion annualised. Each point of medical benefit ratio is therefore approximately $1.35 billion of annual pre-tax income — and the 2.7-point improvement is worth roughly $3.6 billion pre-tax, or approximately $2.20 per share after tax.

That single number explains the guidance raise, the 67.2% twelve-month share-price return, and the shape of the consensus estimate path. It also explains the risk: a 2.7-point move in the wrong direction erases the entire recovery, and the medical benefit ratio is driven by utilisation patterns among Medicare Advantage and commercial members that management observes with a lag.

Management's own guidance table, from the same release:

Note that consensus FY2026 adjusted EPS of $7.451 sits at the top of management's raised range — the estimate block has fully absorbed the raise, and there is no stale-estimate cushion here of the kind found on Bristol Myers Squibb elsewhere in this batch.

4. The balance sheet, which is the reason this is a 7 on risk

Four facts, in order of importance.

First, net debt rose $10.7 billion in FY2025 — from $74.3 billion to $85.1 billion — in a year the company earned $1.8 billion. Long-term debt rose $12.4 billion. That is the shape of a business funding its dividend and its operations with borrowing.

Second, interestCoverageRatioTTM is 1.92x. On FY2025's figures, operating income of $10,385 million against interest expense of $3,119 million is 3.3x, but the trailing figure including the impairment quarter is 1.92x. Either way, this is the thinnest interest coverage in this batch by a wide margin — Chubb's is 14.1x, Lockheed's 8.2x, Bristol Myers Squibb's 8.0x.

Third, tangible book value is NEGATIVE $25.85 per share, against book value of $60.99, because $110,986 million of goodwill and intangibles sits against $75,214 million of equity. The vendor reports the two as properly distinct, which is correct. intangiblesToTotalAssetsTTM is 43.7%. Nearly half the balance sheet is the accounting residue of the Aetna, Caremark, Oak Street and Signify acquisitions, and $5.7 billion of it was written off nine months ago.

Fourth, the dividend was not covered by GAAP earnings. dividendPayoutRatioTTM is 1.161 — the $2.66 dividend costing $3,397 million against trailing GAAP net income of $2,932 million. On adjusted earnings the payout is 37.6% and comfortable; on GAAP it was 116%. Both are true, and which one matters depends entirely on whether the impairment recurs.

Cash flow and capital allocation:

Operating cash flow fell from $16.2 billion in FY2022 to $9.1 billion in FY2024 and has recovered to $10.6 billion. Management guides at least $9.5 billion for FY2026 — a 7.1% yield on the market capitalisation before capital expenditure.

The buyback was ZERO in FY2025, against $3,023 million in FY2024 and $3,500 million in FY2022. That is the clearest signal in the file about how management viewed its own balance sheet: it stopped buying stock and kept paying the dividend while net debt rose $10.7 billion. The $16,612 million of FY2023 acquisitions — Oak Street Health and Signify Health — is the spend that produced the reporting unit written down in 2025. Total shareholder yield is therefore 2.55%, entirely dividend, and the share count has drifted up from 1,262M to 1,279M.

5. Valuation — priced in or room?

At $104.42 (market cap $133.23B, approximately 1,275.9M shares):

Estimate coverage is good on the near years — 17-18 analysts on FY2026 and FY2027 EPS — and thins to 11 on FY2028 and 4 on FY2029. The FY2027 range is remarkably tight: $8.245 to $8.798, a 6.7% spread across 18 analysts, which says the sell-side regards the recovery path as close to arithmetic. That tightness means, as with Danaher elsewhere in this batch, that the entire outcome distribution sits in the multiple rather than the estimate.

est.ebitdaAvg and est.ebitAvg are rejected. In every forward year ebitdaAvg is exactly 4.208% of revenueAvg and ebitAvg is exactly 2.919% — the fixed-ratio fabrication signature. Additionally, netIncomeAvg of $15,502M exceeds ebitAvg of $14,555M in FY2030, which is impossible. All forward valuation uses epsAvg and revenueAvg.

Peer context. The vendor set — Bristol Myers Squibb, Cigna, Cencora, Elevance, GSK, HCA, Humana, McKesson, Medtronic, Vertex — contains four genuine managed-care comparables (Cigna, Elevance, Humana, and by extension UnitedHealth which is absent) alongside manufacturers and distributors. BMY is also in this batch and sits on the opposite side of the same industry. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn.

5a. What today's price assumes (the inversion)

At $104.42 — 14.0x FY2026 consensus and 12.4x FY2027 — the price embeds:

• Adjusted EPS of $7.45 in 2026, at the top of management's raised $7.30-7.50 range. (Consensus; 17 analysts. Guidance, half-weighted.) The March quarter delivered $2.57, so the balance of the year needs $4.88 across three quarters — and tomorrow's consensus of $1.86 implies a steep sequential decline that is the normal seasonal shape of the Health Care Benefits segment.
• The medical benefit ratio holds at or below roughly 85%. (Our derivation from the 10-Q's disclosure.) This is the most fragile assumption in the price. At approximately $1.35 billion of annual pre-tax income per point on a $135 billion premium base, a return to 87.3% would remove roughly $3.6 billion pre-tax — about $2.20 per share after tax, or 29% of the FY2026 consensus.
• Adjusted EPS reaches $8.44 in 2027 and $9.69 in 2028 — 13.2% then 14.8% growth on 4-5% revenue growth. (Consensus.) All of that is margin, and margin in this business is the medical benefit ratio plus pharmacy-benefit-manager economics.
• No further goodwill impairment against the remaining $85,478 million. (Our number.) The 10-K attributes part of the 2025 write-down to "lower market multiples of the peer group companies" — a driver that has since reversed, which helps, but which nobody controls.
• The market keeps paying roughly 12-14x forward adjusted earnings. (Our number.) At 10x FY2027E the stock is $84; at 16x it is $135. CVS has traded across that whole band within twelve months — the 52-week range of $61.45 to $110.60 is 80% wide.

5b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: adjusted EPS growth (+13.2%, from FY2026E $7.451 to FY2027E $8.436) + multiple drift (mild COMPRESSION, from 14.0x on the forward year to roughly 13.5x on the then-forward year, about −4%) + shareholder yield (+2.55%) ≈ +9% to +14%.

Our base assumes mild compression, for a reason specific to recovery stories. A company earning its way back from an operating loss re-rates fastest early, when the market is repricing from distress toward normal. CVS has already done that: +67.2% in twelve months and +23.1% above its own 200-day average. The easy part of the re-rating is behind, and from 14x forward the remaining return has to come from earnings.

That is not a bad place to be. Roughly 90% of the expected return is earnings growth plus dividend, and only the residual depends on the crowd. If the multiple simply HELD at 14.0x on FY2027E the price would be $118 (+13.2%); at 16x, $135 (+29.3%); at 10x, $84 (−19.6%).

5c. Variant perception (where we differ, what would surprise)

• We differ from the street by 4.7%, upward — $114 against $108.86 — and we regard that as a small edge, not a thesis. With 35 of 41 analysts at buy or better and NO sell ratings, this is the most positively-covered name in the batch, and the consensus target of $108.86 implies only 4.3% upside. When 85% of the coverage is at buy and the average target is 4% away, the crowd has already made the call and priced it.
• We differ from the vendor data on what the trailing multiple means. The payload's priceToEarningsRatioTTM of 45.2x is arithmetically correct and completely misleading, because the trailing window contains the September 2025 quarter's $5.7 billion impairment and $3,207 million operating loss. Any screen ranking CVS on trailing earnings is ranking a write-down. The comparable figure is 14.8x on trailing adjusted earnings. That is a real, if unglamorous, informational edge.
• We think the pharmacy-benefit-manager reform risk is the under-priced item, and the knowledge base is unusually consistent about it. Eleven independent claims across four channels argue the model faces transparency legislation, rebate pass-through requirements or manufacturer disintermediation: "PBMs are at the end of their S-curve"; "separating formularies from PBMs would collapse their rent-extraction model"; "PBM reform gaining momentum (transparency, 100% rebate pass-through)"; and direct-to-consumer manufacturer channels. None names CVS — which is why none enters the conviction pool — but Caremark sits inside a Health Services segment producing $48.2 billion of quarterly gross revenue. Watchable event: any federal pharmacy-benefit-manager legislation reaching a vote.
• We think the goodwill balance is a live risk that the recovery narrative has crowded out. $85,478 million remains, 32.7% of assets, after a write-down whose stated cause was partly "lower market multiples of the peer group companies." Peer multiples have since risen, which reduces the near-term risk — but a company that impaired $5.7 billion nine months ago and still carries eleven times that amount is one bad projection away from repeating it.
• Positive surprise that would force a re-rate: a June-quarter medical benefit ratio below 84%, with the FY2026 adjusted EPS guidance raised a second time; or resumption of the buyback, absent for a full year, which would signal management considers the balance sheet repaired; or Health Services adjusted operating income growing materially, which would say Caremark is holding its economics.
• Negative surprise that would break the thesis: a medical benefit ratio back above 86%; a guidance cut; a further goodwill impairment; an adverse development in the securities class actions over Health Care Benefits profitability; or federal pharmacy-benefit-manager legislation with teeth. Any of these hits a stock that has risen 67.2% in twelve months.

Synthos fair values

All three anchors are multiples of the FY2027 consensus adjusted EPS distribution (mean $8.436, low $8.245, high $8.798, 18 analysts) — the last year with strong coverage — cross-checked against management's FY2026 guidance range of $7.30-7.50.

• Bear ~$87 — 10.5x the FY2027 consensus LOW of $8.245. Cross-check: 11.7x the FY2026 guidance midpoint of $7.40; 41.6% above the 52-week low of $61.45; a 3.06% dividend yield. The scenario: the medical benefit ratio backs up toward 86%, guidance is cut, a further goodwill charge lands, and the multiple returns to where distressed managed-care trades. −16.7%.
• Base ~$114 — 13.5x the FY2027 consensus MEAN of $8.436. Cross-check: 15.4x the FY2026 guidance midpoint; 11.8x FY2028E; 3.1% above the 52-week high of $110.60. Sensitivity, stated openly: 12.4x — today's multiple, held — gives $105, essentially spot; 15x gives $127. The scenario: the medical benefit ratio holds near 84-85%, consensus lands, cash flow reduces net debt, and the multiple drifts down modestly as the recovery matures. +9.2%.
• Bull ~$145 — 16.5x the FY2027 consensus HIGH of $8.798. Cross-check: 15.0x FY2028E; 31.1% above the 52-week high. The scenario: the medical benefit ratio keeps falling toward 83%, the buyback resumes at scale, leverage falls, and the market pays a full managed-care multiple for a business it had written off. +38.9%.

Base is 9.2% above spot; asymmetry roughly 2.33:1 to the upside (16.7% down, 38.9% up), plus a 2.55% dividend. That ratio is genuinely good, and it is the strongest argument here for a higher tier. What holds it at Watch is that the print is tomorrow, the coverage is 85% buy with a target 4% away, and the conviction lane is one low-conviction claim that contains its own valuation warning.

6. Knowledge base — twenty-five hits, three claims, and nine of them are about light therapy

Raw hits: 25. Entity matches on a case-sensitive re-run: 3. Discarded: 22. Used as name-level conviction: 1 (the other two are a competitor comparison and a half-weighted counterparty-management observation).

The sweep ran entity terms CVS, CVS Health, Aetna, Caremark, Oak Street Health and Signify Health, plus free text on CVS, Aetna, Caremark, pharmacy benefit and PBM, across all 51,928 distilled claims. It returned 25 hits, and 22 are noise — the richest homograph field in this batch, across three distinct collision classes.

Discard class 1 — "PBM" means PHOTOBIOMODULATION (9 claims)

Nine claims from the michael_levin channel, all dated 2021-11-19, use PBM to mean photobiomodulation — red and near-infrared light therapy. They concern:

> "PBM mobilizes hypoxic-niche stem cells and switches macrophages/microglia from M1 to M2 anti-inflammatory phenotype" (conviction 75)

> "PBM drives neurogenesis and BDNF in hippocampus at 7 days, then synaptogenesis in cortex at 28 days, improving learning and memory in TBI mice" (conviction 70)

> "Transcranial/intranasal PBM improved dementia cognition ~7x more than Aricept's trial effect" (conviction 68)

> "Transcranial PBM raises brain oxygenation and produces significant improvements in treatment-resistant depression and anxiety" (conviction 60)

> "PBM is being approved as one of the few treatments for dry age-related macular degeneration" (conviction 58)

> "Cheap harmless LEDs replaced lasers; whole-body beds, wearable pads and brain helmets are proliferating — every household will soon own a PBM device" (conviction 65)

Plus three more on dose-response curves, mitochondrial metabolites and cell-free mitochondria. Not one of them has anything to do with CVS Health, pharmacy benefit management, or finance. This is the same defect class as "Chevron deference" and "Amex Exploration" recorded elsewhere in this programme, and it is the largest single instance of it we have found.

Discard class 2 — "CVs" means curriculum vitae (1 claim)

> 2025-11-04 · invest_like_the_best · entity: Bending Spoons — "Heavy AI investment predicting future performance from CVs, cover letters and tests pays off only at high hiring volume, giving large recruiters a selection edge."

A hiring-algorithm claim about an Italian software company. The ticker CVS matched the plural of curriculum vitae.

Discard class 3 — "ULCVs" means ultra-large container vessels (1 claim)

> 2022-03-23 · odd_lots · entity: Evergreen — "Container ships keep getting bigger because insurers insure, ports accept, operators order them; economies of scale dominate (Evergreen ordered three 24,000-box ULCVs mid-grounding)."

A shipping claim. The ticker CVS matched the last three letters of "ULCVs".

Discard class 4 — genuine PBM-industry claims that do not name CVS (11 claims)

These are thematically relevant and are reported as sector colour, but they fail the entity test and enter no score. Across eli_lilly_ceo (5), all_in (3) and biotech_hangout (3), the argument is consistent and hostile:

> 2025-09-29 · eli_lilly_ceo · bearish 65 — "PBMs are at the end of their S-curve — original IT/bulk-negotiation rationale is obsolete and service has 'enshittified' toward their own benefit; transparent PBMs rising."

> 2025-09-09 · all_in · bearish 75 — "Vertically-integrated insurers/PBMs game an opaque system and face antitrust risk; separating formularies from PBMs would collapse their rent-extraction model."

> 2026-01-23 · biotech_hangout · bullish 50, horizon FACT — "PBM reform gaining momentum (transparency, 100% rebate pass-through) could benefit biotech innovation, though not yet locked in."

> 2025-08-26 · biotech_hangout · bullish 55 — "Increasing likelihood of real PBM/middlemen reform this year; strong White House sense that middlemen harm patients and the system."

> 2025-11-11 · eli_lilly_ceo · bullish 65 — "LillyDirect sells directly to consumers online, upending the traditional middleman/PBM distribution model."

Four independent channels, spanning 2019 to 2026, arguing that the layer CVS's Health Services segment occupies is under structural attack. None enters the conviction rating. All of it informs the exponential score of 3.

The three entity-matched claims

> 2025-07-12 · bullish 80 · thesis · entities: UnitedHealth Group, CVS Health, Cigna, Elevance, Humana · channel: business_breakdowns · no named speaker

> "UNH out-earns peers on integration depth; CVS is retail-heavy, Elevance late, Humana over-concentrated in Medicare Advantage — none match Optum's synergy."

> This is a bullish claim about UnitedHealth in which CVS appears as an also-ran. Its stance field reads bullish; its content, with respect to CVS, is mildly negative. We record it as a competitor comparison, not as conviction on CVS.

> 2025-07-17 · neutral 65 · horizon: FACT · entities: Eli Lilly, Novo Nordisk, CVS, Zepbound, Wegovy · channel: eli_lilly_ceo · no named speaker

> "CVS formulary loss to Wegovy is on low-opt-in templated small-employer book; Lilly rejects one-of-one access-restricting PBM deals, focuses on innovation."

> QUOTED AT HALF WEIGHT AND EXCLUDED FROM THE CONVICTION POOL. The eli_lilly_ceo channel slug is named for a pharmaceutical chief executive, and the attribution policy specifically flags this slug as carrying other companies' management voice. This claim reads as a manufacturer's chief executive characterising a formulary decision by a pharmacy benefit manager he is negotiating against — counterparty management voice on both sides. The substance is useful (a Wegovy formulary win at CVS affects a small book) and it is not conviction.

> 2026-03-01 · bullish 48 · thesis · entities: MCD, CVS · channel: jordi_visser · role: independent · skill 2.0 — the most recent and highest-skill claim in the lane

> "Rotating into three-year survivors like McDonald's and CVS that he knows will still exist, though the market is now overpaying multiples for that certainty."

> Conviction 48 is the lowest band we record, and the claim carries its own qualification. It is a survivorship rotation, not a valuation thesis — and the second clause is explicitly a warning that the multiple has run ahead. On a stock that has since risen further, that qualification has aged into the more useful half of the claim.

Conclusion. Breadth 3, claim count 3, net conviction negative-low. No concentration sensitivity test is meaningful across three claims from three channels. The Synthos knowledge base's only usable view on CVS Health is that it is a survivor whose survival the market may be overpaying for, from a speaker at the lowest conviction band we record. The far larger contribution is negative and sector-level: eleven independent claims arguing the pharmacy-benefit-manager model is structurally threatened. Both feed the Watch verdict.

7. Data integrity — what we rejected and why

Seven findings. CVS's file is not badly broken so much as badly framed: the headline multiple is a write-down, the segment block is a revenue-type table, and the geography field is empty.

1. priceToEarningsRatioTTM of 45.2x is arithmetically correct and analytically void — REJECTED. The trailing twelve-month window (June 2025 quarter through March 2026 quarter) contains the September 2025 quarter, which had an operating LOSS of $3,207M and a net loss of $3,975M on a $5.7 billion goodwill impairment. Trailing GAAP EPS is $2.30; trailing adjusted EPS across the same quarters is $7.07, giving 14.8x. Every valuation figure in this dive uses adjusted earnings and says so. Note also that dividendPayoutRatioTTM of 1.161 and netProfitMarginTTM of 0.72% inherit the same distortion.

2. seg_prod is a revenue-type disaggregation, not a segment table, and it does not foot. The four lines — Pharmacy Revenue $229,025M, Premiums $134,751M, Front Store Revenue $21,459M, Product and Service Other $3,926M — sum to $389,161M against reported FY2025 revenue of $402,067M, a $12.9 billion shortfall. CVS's four actual segments (Health Care Benefits, Health Services, Pharmacy & Consumer Wellness, Corporate/Other) appear nowhere in the payload. Note also the "Product and service, other" line falling from $15,966M to $3,926M in one year, a 75% drop with no corresponding event — a reclassification. All segment analysis comes from the 10-Q.

3. seg_geo is EMPTY. There are no entries at all — not a stale year, not a partial list, an empty array. This is the same defect class as UNH's empty geographic block. CVS is a domestic business so the analytical cost is small, but the field is null.

4. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature — REJECTED. In every forward year ebitdaAvg is exactly 4.208% of revenueAvg and ebitAvg is exactly 2.919%. Corroborating defect: netIncomeAvg of $15,502M exceeds ebitAvg of $14,555M in FY2030, which is impossible. All forward valuation uses epsAvg and revenueAvg.

5. The FY2029 and FY2030 estimate rows rest on 4 analysts each — EXCLUDED. Against 17-18 on FY2026 and FY2027 and 11 on FY2028. Neither supports a conclusion here.

6. returnOnEquityTTM is NULL and epsdiluted is null in every income-statement row — worked around. Return on equity is computed here from the filings; all per-share figures are computed and labelled, and cross-checked against the 8-K's stated GAAP diluted EPS of $2.30 and adjusted EPS of $2.57 for the March quarter.

7. The June 2026 quarter is absent from inc_q — correctly, and it is worth saying so. The most recent quarterly row is 2026-03-31. This is NOT the missing-quarter defect found on ANET and MRK: CVS has not reported the June quarter, and does so tomorrow morning. The consequence is that every trailing figure in the payload — and therefore in this dive — ends at 31 March 2026 and includes the impairment quarter. Stated so no reader mistakes the trailing window for a current one.

Correct and worth recording: bookValuePerShareTTM of $60.987 is correctly distinct from tangibleBookValuePerShareTTM of −$25.853, a genuinely negative tangible book properly reported, with shareholdersEquityPerShareTTM of $60.845 correctly excluding the $168M of noncontrolling interests; intangiblesToTotalAssetsTTM of 43.7%; incomeQualityTTM of 2.507, correctly reflecting $10,639M of operating cash flow against $1,768M of net income in an impairment year; the dividend of $2.66; and quote.yearHigh/yearLow ($110.68/$61.35) within 0.2% of tech.hi52/lo52 ($110.60/$61.45), with tech used throughout.

Non-equity tripwire — checked and passed. CVS is common stock, NYSE-listed. Price of $104.42 is not par-like; beta is 0.596; the dividend is variable; volume was 5.63M shares (~$588M of turnover); the 52-week band of $61.45 to $110.60 is an 80% range — the widest in this batch, and the opposite of a fixed-income-like instrument. This is common equity.

8. Technicals

• Price $104.42. −5.6% from the 52-week high of $110.60; +69.9% above the 52-week low of $61.45. Position within the annual range: 87th percentile.
• Above both moving averages, and the 200-day is a long way below: +2.8% above a 50-day average of $101.57; +23.1% above a 200-day average of $84.82. The gap between the two averages is 19.7% — the widest in this batch, and the signature of a stock that has re-rated violently over roughly six months.
• RSI 46.1 — neutral, and notably NOT overbought despite the 67% run, because the last three months have been a consolidation. MACD +1.04.
• Maximum drawdown from peak over the trailing year: −5.8%.
• Relative performance, and it is the strongest in this batch: 3-month +27.3% against SPY +7.6%; 6-month +38.5% against SPY +11.1%; 12-month +67.2% against SPY +24.3% and QQQ +30.8%. A 42.9-point twelve-month outperformance from a 0.596-beta healthcare company.
• Sentiment is the most positive in this batch: 2 strong buy, 33 buy, 6 hold, zero sell across 41 analysts — 85% at buy or better; consensus target $108.86 (+4.3%), median $110.50, high $123 (+17.8%), low $92 (−11.9%). Near-unanimous coverage with a target only 4.3% away is a crowded position, and we treat it as a caution rather than a comfort.

Today's move and what it does to the entry

CVS closed 2026-08-04 at $104.42, down 0.90% or $0.95 from $105.37. It opened at $104.87, traded $103.275 to $105.00, and closed near the low on 5.63M shares. On a day most of this batch rose, CVS fell — consistent with positioning ahead of tomorrow morning's report.

The honest read: this is a consolidation, not an entry. The stock has done nothing for three months in price terms while the 200-day average catches up 23% below. What matters is the print.

What would change this:

• A price near $92-95 — roughly 11x FY2027E, at the street's low target — which would put our base $114 at a 20-24% premium and lift the dividend yield toward 2.9%. That is 9-12% below today's close, and a disappointing print would deliver it in a session.
• Or the 2026-08-05 print showing a medical benefit ratio below 84% with FY2026 adjusted EPS guidance raised a second time. At that point paying 14x is defensible, because the one variable the whole thesis rests on would have been confirmed twice.

9. Insiders — a director sold $261 million

The reading. Three of the eight entries are F-InKind — shares withheld by the issuer for tax on vesting, mechanical and signal-free, including both the chief financial officer's and the chief medical officer's. Two are D-Return, a disposition to the issuer, likewise not an open-market sale.

The three entries that matter are all Larry Robbins, and they total 2,781,166 shares sold at approximately $94 — roughly $261 million. His holding fell from approximately 8,196,799 to 5,415,633 shares, a 33.9% reduction, executed across two days in May.

Two qualifications and one observation. Robbins is a director whose position is an investment-fund holding rather than compensation-derived equity, so a partial monetisation after a large run is ordinary portfolio management rather than an executive cashing out. The residual holding of 5,415,633 shares is worth approximately $565 million at today's close — still one of the largest individual stakes on any board in this batch. And the sales printed at approximately $94, which is 10% BELOW today's close of $104.42 — so whatever the motivation, it was not well-timed against the subsequent move.

What the file contains is therefore: zero purchases by anyone, three mechanical tax withholdings, and one director reducing a very large position by a third. We report that as a mild negative, weighted down by the fund-holding context. No transaction by chief executive J. David Joyner appears anywhere in the file.

10. Verdict, kill-criteria and flip conditions

Watch.

What is genuinely good: a medical benefit ratio down 2.7 points to 84.6%, worth roughly $3.6 billion of annual pre-tax income; full-year guidance raised on 6 May across all three headline metrics — adjusted EPS to $7.30-7.50, GAAP EPS to $6.24-6.44, operating cash flow to at least $9.5 billion; five consecutive adjusted-EPS beats averaging approximately 20%; revenue of $100.4 billion in the March quarter, +6.2%; Health Care Benefits adjusted operating income of $3,041 million in a single quarter; a modest 9.1% GAAP-to-adjusted reconciliation, better than most of this batch; consensus adjusted EPS growing 13.2% then 14.8% with an FY2027 range spanning only 6.7%; and $114 of base fair value against $104.42, with 2.33:1 asymmetry.

What holds it at Watch: earnings land tomorrow morning; the stock has already returned +67.2% in twelve months and sits 23.1% above its 200-day average; 85% of the coverage is at buy with a consensus target only 4.3% away; net debt of $85.1 billion with interest covered 1.92 times, the thinnest in this batch; tangible book value of negative $25.85 per share and $85,478 million of goodwill remaining — 32.7% of assets — after a $5.7 billion write-down nine months ago; securities class actions alleging misrepresentation about the profitability of the exact segment that is now recovering; no buyback at all in FY2025; a knowledge-base lane of three claims, the most conviction-weighted of which warns the market is "overpaying multiples for that certainty"; and eleven independent claims arguing the pharmacy-benefit-manager layer is structurally threatened.

The distinction that matters. The recovery in this business is real, measurable and driven by a number we can watch. What we are declining is the timing: a 2.33:1 payoff struck the evening before a print, on a name where 85% of analysts are already positive and the average target is 4% away, is a payoff that will look completely different in twenty-four hours. In the Synthos frame, that is a Watch with two named triggers — a price near $92-95, or a June-quarter medical benefit ratio below 84% with guidance raised again. Either would convert this to a position.

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

• The medical benefit ratio above 86% in any quarter. That is the recovery reversing, and it is worth roughly $2 billion of pre-tax income against the March-quarter level.
• A cut to the FY2026 adjusted EPS guidance of $7.30-7.50, raised only three months ago.
• A further goodwill impairment against the remaining $85,478 million, after the $5.7 billion Health Care Delivery charge.
• An adverse development in the securities class actions over Health Care Benefits profitability, or a material legacy-litigation charge beyond the $1.2 billion already taken.
• Federal pharmacy-benefit-manager legislation with binding rebate pass-through or formulary separation, which four independent knowledge-base channels regard as increasingly likely.
• Interest coverage falling below 1.5x, or net debt rising above roughly $90 billion without a corresponding earnings recovery.

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

• A price in the $92-95 band — roughly 11x FY2027E, at the street's low target — with the guidance intact. This is the single most likely path to a Buy and a disappointing print tomorrow would deliver it in one session.
• The 2026-08-05 print showing a medical benefit ratio below 84% with FY2026 guidance raised a second time, confirming the one variable the whole thesis rests on.
• Resumption of the buyback at scale, absent for a full year, which would signal management considers the balance sheet repaired and would add materially to shareholder yield.
• Net debt below roughly $78 billion with interest coverage above 3x, which would remove the principal balance-sheet objection.
• Genuine name-level knowledge-base conviction appearing. At present the lane is three claims, one of which is a competitor comparison and one of which is half-weighted management voice.

Where CVS fits in the Synthos Framework Portfolio. The healthcare-services sleeve, at 0% today with a 1.5-2.5% target on a fill in the $92-95 band or a confirming June-quarter medical benefit ratio. On batch overlap: BMY is also in this batch and is in CVS's own vendor peer set — but they sit on opposite sides of the same industry, with Bristol Myers Squibb selling the drugs whose prices Caremark negotiates down. Holding both is a partial hedge rather than a concentration; holding CVS alongside any other pharmacy-benefit-manager exposure is not. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $104.42.

Single biggest risk: the medical benefit ratio stops improving. It fell from 87.3% to 84.6% in the March quarter, and that 2.7-point move — worth approximately $3.6 billion of annual pre-tax income and roughly $2.20 per share after tax — is the entire reason guidance was raised, earnings are recovering and the stock has risen 67.2%. It is one ratio, in one segment, driven by the medical utilisation of Medicare Advantage and commercial members, which management observes with a reporting lag and does not control. It blew out once already, in 2024 and 2025, and the consequence was an operating margin of 2.28%, a $5.7 billion impairment, a securities class action and a share price of $61.45. At 12.4x forward adjusted earnings with 1.9x interest coverage and negative tangible equity, there is no balance-sheet cushion if it happens again. That is the $87 bear case, and the next observation of it is tomorrow morning.

Provenance & disclosures

• Traceability: 3 tagged knowledge-base claims name CVS Health out of 51,928 distilled claims (raw hits 25, entity matches on a case-sensitive re-run 3, discarded 22, used as name-level conviction 1; breadth 3, net conviction negative-low). The 22 discards fall into four classes and the first is the most instructive homograph catch in this programme: NINE claims from the michael_levin channel, all dated 2021-11-19, use "PBM" to mean PHOTOBIOMODULATION — red and near-infrared light therapy — and discuss neurogenesis in traumatic-brain-injury mice, dementia cognition versus Aricept, treatment-resistant depression, dry age-related macular degeneration and household LED devices. ONE claim (invest_like_the_best, 2025-11-04, on Bending Spoons) uses "CVs" to mean curriculum vitae; ONE (odd_lots, 2022-03-23, on Evergreen) matches "ULCVs", ultra-large container vessels. Eleven further discards are genuine pharmacy-benefit-manager industry claims across eli_lilly_ceo (5), all_in (3) and biotech_hangout (3) that never name CVS — consistently hostile to the model ("PBMs are at the end of their S-curve", "separating formularies from PBMs would collapse their rent-extraction model", "PBM reform gaining momentum... 100% rebate pass-through") — reported as sector colour, entering no score, and informing the exponential rating of 3. The three entity matches: business_breakdowns 2025-07-12 (conviction 80) is a claim about UnitedHealth in which CVS appears as the also-ran ("CVS is retail-heavy... none match Optum's synergy"), recorded as a competitor comparison; eli_lilly_ceo 2025-07-17 (conviction 65, horizon FACT, on the Wegovy formulary decision) sits on the channel slug the attribution policy specifically flags as carrying other companies' management voice and is quoted at half weight and EXCLUDED from the conviction pool; and jordi_visser 2026-03-01 (conviction 48, the lowest band we record, skill 2.0) is the only usable claim — "Rotating into three-year survivors like McDonald's and CVS that he knows will still exist, though the market is now overpaying multiples for that certainty" — nominally bullish and carrying its own valuation warning. All 25 hits carry speaker: null and are channel-attributed. No concentration test is meaningful across three claims from three channels. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals — segment results, the medical benefit ratio, balance sheet and litigation disclosure through 2026-03-31, filing-verified from the 10-Q filed 2026-05-06; first-quarter results and the raised FY2026 guidance from the 8-K filed 2026-05-06, which carries the full earnings release; the $5.7 billion Health Care Delivery impairment, the $1.2 billion legacy litigation charges, the worthless-stock tax deduction and the store-closure programme from the 10-K filed 2026-02-10 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873599 = 2026-08-04T19:59:59Z ($104.42, −0.90%; 50-DMA $101.57; 200-DMA $84.82; RSI 46.1; MACD +1.04) · knowledge-base claims 2026-08-04. CVS's fiscal year is the calendar year. The June 2026 quarter has not been reported and is correctly absent from the payload; every trailing figure here ends at 2026-03-31 and includes the September 2025 impairment quarter. All figures come from the Synthos vendor data file for CVS or from the SEC filings in the CVS archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents: 10-K filed 2026-02-10 (fiscal 2025 — the source of the impairment disclosure and its stated causes); 10-Q filed 2025-10-29 (September 2025 quarter); 10-Q filed 2026-05-06 (March 2026 quarter — the source of the segment table, the four-segment structure and the securities-litigation disclosure); 8-K filed 2026-05-06, which carries the complete first-quarter earnings release including the medical benefit ratio and the raised guidance table; 8-K filed 2026-03-19 (officer transition); 8-K filed 2026-05-18 (annual meeting results and the 2026 incentive compensation plan). All carry preserved [TABLE] statement data.
• Where the filings supplied what the vendor could not, and where they contradicted it (detailed in Section 7): priceToEarningsRatioTTM of 45.2x, rejected because the trailing window contains a $5.7 billion goodwill impairment and a $3,207M operating loss — the comparable figure is 14.8x on trailing adjusted earnings, and dividendPayoutRatioTTM of 1.161 and netProfitMarginTTM of 0.72% inherit the same distortion; seg_prod, a revenue-type disaggregation summing to $389,161M against reported revenue of $402,067M — a $12.9 billion shortfall — with CVS's four actual segments absent entirely and the "Product and service, other" line falling 75% year on year on reclassification; seg_geo, which is EMPTY; est.ebitdaAvg/ebitAvg, rejected for a fixed-ratio signature (exactly 4.208% and 2.919% of revenue in every forward year) with netIncomeAvg exceeding ebitAvg in FY2030; returnOnEquityTTM, which is NULL; and epsdiluted, null in every income-statement row. The medical benefit ratio, the four-segment adjusted operating income table, the guidance raise, the impairment's stated causes and the securities litigation appear in NO vendor field and come entirely from the filings. Where vendor and filing AGREED: bookValuePerShareTTM of $60.987 correctly distinct from tangibleBookValuePerShareTTM of −$25.853, with shareholdersEquityPerShareTTM of $60.845 correctly excluding noncontrolling interests; the Q1 2026 revenue of $100,426M and GAAP EPS of $2.30 against the 8-K; incomeQualityTTM of 2.507; and quote versus tech 52-week ranges within 0.2%.
• Basis note — this determines the entire valuation and is disclosed for that reason. CVS reports GAAP and adjusted earnings per share, with the adjustment excluding amortisation of intangible assets, net realised capital gains and losses, and acquisition-related integration costs. The March-quarter reconciliation is unusually clean: $5,150M of segment adjusted operating income less $442M of intangible amortisation, $16M of net realised capital losses and $12M of integration costs gives $4,680M of GAAP operating income — a 9.1% adjustment, modest by the standards of this batch. But the ANNUAL gap is enormous because of impairments and litigation: FY2025 GAAP EPS was $1.40 against roughly $6.60 adjusted. Analyst estimates and management guidance are on the adjusted basis, and every forward multiple in this dive uses it; the trailing multiple is shown on both (45.2x GAAP, 14.8x adjusted) and the GAAP figure is explicitly rejected as a lens.
• Estimate coverage: 17-18 analysts on FY2026 and FY2027 adjusted EPS — FY2027 is the anchor for all three fair values, with a range of $8.245 to $8.798, a 6.7% spread that is among the tightest in this batch; 11 on FY2028; 4 on FY2029 and 4 on FY2030, both excluded from every conclusion.
• Peer note: the vendor peer set — Bristol Myers Squibb, Cigna, Cencora, Elevance, GSK, HCA, Humana, McKesson, Medtronic, Vertex — contains four genuine managed-care comparables (Cigna, Elevance, Humana, and by extension the absent UnitedHealth) among manufacturers and distributors. BMY is also in this batch and sits on the opposite side of the same industry. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn.
• Fair-value caveat: the $87 / $114 / $145 anchors are multiples of the FY2027 consensus adjusted EPS distribution — 10.5x the low of $8.245, 13.5x the mean of $8.436, and 16.5x the high of $8.798 — cross-checked against management's FY2026 guidance range of $7.30-7.50. Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 12.4x — today's multiple, held — gives $105, essentially spot; 15x gives $127.
• Management-voice note: the raised FY2026 guidance (revenue at least $405.0B, GAAP EPS $6.24-6.44, adjusted EPS $7.30-7.50, operating cash flow at least $9.5B), the attribution of the raise to the Health Care Benefits and Pharmacy & Consumer Wellness segments, and the chief executive's characterisation of "nearly 185 million people" served are all management's own claims and are half-weighted throughout, and labelled wherever quoted.
• Timing: second-quarter 2026 results are released 2026-08-05, ONE DAY after this dive, against consensus adjusted EPS of $1.86 and revenue of $100.028B. This is the dominant near-term fact and the verdict is set with it as an explicit input. The last report was 2026-05-06 (adjusted EPS $2.57 against $2.18 estimated), the fifth consecutive beat in the file, all five by 9.0% to 32.4%. 2026-08-04 carried no company-specific filing, and CVS fell 0.90% on a day most of this batch rose.
• 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.