Bristol-Myers Squibb BMY

Reference table

1. What the business is — and the two-portfolio disclosure that is the whole thesis

Board chair and chief executive Christopher S. Boerner; approximately 32,500 employees. Bristol Myers Squibb reports as a single segment and discloses revenue by product, which the vendor's seg_prod block reproduces in full and — unusually — sums to $48,195 million for FY2025 against reported revenue of $48,195 million, an exact match.

The 10-Q's own two-portfolio split for the June 2026 quarter ($M):

Read the Legacy column and the pattern is complete. Every drug that has lost exclusivity is collapsing at 25-71% a year. Every drug. Revlimid is down 49%, Pomalyst 71%, Abraxane 48%, Sprycel 27%. And Eliquis — the one Legacy drug that has not lost exclusivity — grew 22%. That contrast is not a coincidence; it is the mechanism, and it tells you exactly what happens to $4,481 million a quarter when 2028 arrives.

The FY2025 product table, top ten, from seg_prod (which reconciles exactly to revenue):

Eliquis and Opdivo together are 50.9% of the company. Opdivo Qvantig — the subcutaneous formulation, $238M in its first full year — is the company's own answer to Opdivo's eventual biologic exclusivity, and it is worth noting the company created a separate line for it.

Geography. The 10-Q gives the June quarter as United States $8,991M (69.3%), International $3,664M (28.2%), Other $318M (2.5%). The vendor's seg_geo block for FY2025 reports only two lines — United States $33,279M and "Other Region" $1,087M, summing to $34,366M against reported revenue of $48,195M. The "Rest of World" line of approximately $13.8 billion, present in the FY2024 and FY2023 entries, has been dropped. 28.7% of revenue is missing from the vendor's geographic block and no geographic figure in this dive uses it.

2. The cliff, with dates

From the 10-K's exclusivity table, the single most important row in the filing:

> Eliquis (apixaban) — United States: 2028

Supporting disclosures from the same 10-K, all verbatim:

> "for Eliquis, generics have challenged the composition of matter patents and related SPCs in various jurisdictions, and trials have taken place, or are scheduled to take place, in certain European countries. While these legal proceedings are pending, generic manufacturers have begun marketing generic versions of Eliquis in certain EU countries and may seek to market generic versions of Eliquis in other EU countries prior to the expiration date of applicable patents."

> "following certain adverse judicial decisions in the UK and Finland, generic manufacturers have begun marketing generic versions of Eliquis in these countries."

> "In November 2025, BMS received a Notice Letter from Azurity Pharmaceuticals, Inc. notifying BMS that Azurity had filed a 505(b)(2) application containing a paragraph IV certification seeking approval to market apixaban products in the U.S. and challenging a formulation patent listed in the Orange Book for Eliquis but not the composition of matter patent."

> "As is common in the pharmaceutical industry, BMS expects that sales of its branded products like Orencia, Eliquis and Opdivo will decline after the loss of market exclusivity."

And the government-pricing layer on top, which is already in force:

> "In August 2024, as part of the first round of government price setting pursuant to the IRA, the HHS announced the 'maximum fair price' for a 30-day equivalent supply of Eliquis, which applies to the U.S. Medicare channel effective January 1, 2026. In November 2025, the HHS announced the 'maximum fair price' for a 30-day supply of Pomalyst, which applies to the U.S. Medicare channel effective January 1, 2027. In January 2026, the HHS selected Orencia as a medicine subject to 'negotiation' for government-set prices beginning in 2028."

Three things follow, and the third is the one most likely to be missed.

First, the near-term evidence is better than the framework predicts. Eliquis grew 22% in the June 2026 quarter — its first full quarter with the Inflation Reduction Act maximum fair price in effect for US Medicare. Volume growth is more than offsetting the mandated price cut. That is a genuine positive and it is a large part of why the guidance was raised.

Second, the consensus path takes the cliff seriously and dates it correctly. Revenue falls $6.0 billion in FY2028 alone ($47,684M to $41,685M), which is the year the 10-K names. The sell-side is not ignoring this; it is modelling it.

Third — and this is the part worth pausing on — three of Bristol Myers Squibb's largest products face government price-setting on a schedule that is independent of patent expiry. Eliquis (30.0% of revenue) from January 2026, Pomalyst (5.7%) from January 2027, Orencia (7.7%) from 2028. That is 43.4% of FY2025 revenue subject to administratively-set prices within three years, and the 10-K warns that further selections "could, among other things, accelerate revenue erosion prior to expiry of intellectual property protections." The patent cliff and the pricing regime arrive together.

3. The near-term print, and the guidance raise

The June 2026 quarter, from the 8-K earnings release filed 2026-07-30:

Chief executive Christopher Boerner, quoted in the release — management voice, half-weighted: "The Growth Portfolio continues to deliver, achieving 15% growth in the quarter, and represents an expanding share of our overall business... we are raising our 2026 full-year outlook."

The guidance change, verbatim and material:

A roughly $3 billion revenue raise — 6.3% at the midpoint — in a single quarter, attributed to "broad-based and continuing momentum across the portfolio." That is a large move for a company of this size and it is the proximate cause of the 49% twelve-month return.

Note the estimate-block lag. The vendor's FY2026 consensus non-GAAP EPS is $6.744, which sits below the floor of management's own $6.75-7.00 range. The estimate block has not fully absorbed the raise. At the guidance midpoint of $6.875 the forward multiple is 9.6x rather than 9.8x — a small difference, but it means every forward multiple in this dive is, if anything, conservative.

The quarterly series, GAAP, showing the volatility that non-GAAP smooths:

Gross margin stepped up from roughly 56% to roughly 71% between 2024 and 2025 as amortisation on the Celgene-era intangibles ran off — depreciation and amortisation fell from $9,600M (FY2024) to $4,011M (FY2025), a $5.6 billion reduction that flows almost entirely to reported profit. That is the largest single driver of the reported earnings recovery and it is mechanical rather than operational. FY2024's net loss of $8,948M and FY2020's of $8,995M were both driven by acquired in-process research and development charges and amortisation, not by trading.

4. Balance sheet, cash flow and the capital-allocation record

Equity of $18,473M against $40,857M of goodwill and intangibles gives TANGIBLE BOOK VALUE OF NEGATIVE $8.23 PER SHARE, against book value of $10.93 — the vendor reports the two as properly distinct, which is worth recording given how often that test has failed elsewhere in this programme. intangiblesToTotalAssetsTTM is 44.6%.

Net debt of $36,930M against roughly $16.7 billion of trailing EBITDA is about 2.2x, down from $40,854M a year earlier. Interest coverage is 8.0x. priceToBookRatioTTM of 6.03x and returnOnEquityTTM of 46.8% are both distorted by the small equity base and are rejected as lenses; returnOnInvestedCapitalTTM of 15.5% is the honest return measure.

The capital-allocation record is the clearest thing in the file. Bristol Myers Squibb bought back $8.0 billion of stock in FY2022 and $5.2 billion in FY2023, then spent $20,722 million on acquisitions in FY2024 — producing a net loss of $8,948 million that year — and has bought back nothing since. Free cash flow is being used to pay the $5 billion dividend and to reduce debt from $51.2 billion to $47.1 billion.

Trailing free cash flow is $11,443 million, an 8.5% yield at the current price, against a dividend costing $5,045 million — 44% of free cash flow, comfortably covered. The dividend is $2.51 per share, a 3.81% yield, at a 54.9% payout of trailing GAAP earnings.

Total shareholder yield is 3.81% and it is entirely dividend. With no buyback, the share count has drifted up slightly — 2,027M diluted (FY2024) to 2,041M (FY2025) to 2,048M in the June quarter.

5. Valuation — priced in or room?

At $65.895 (market cap $134.60B, enterprise value approximately $169.0B):

This is the most unusual estimate block in the batch and it needs to be read carefully. The forward multiple RISES every year — 9.8x, 10.2x, 11.9x, 13.3x, 13.6x — because earnings fall faster than the calendar advances. That is the opposite of the normal rolldown, and it is the single most informative fact in the valuation section. A stock that gets more expensive the further out you look is a stock where the market is paying for the near years and discounting the far ones, and the question is whether it is discounting them enough.

Coverage collapses beyond FY2027: 8 analysts on FY2028 EPS, 6 on FY2029 and 6 on FY2030. Those rows are shown because the shape they describe is the thesis, but no fair value is anchored on them.

est.ebitdaAvg and est.ebitAvg are rejected. In every forward year ebitdaAvg is exactly 33.24% of revenueAvg and ebitAvg is exactly 14.20% — the fixed-ratio fabrication signature. Corroborating defect: netIncomeAvg exceeds ebitAvg in every single forward year — $14,013M against $6,959M in FY2026, $13,213M against $6,773M in FY2027 — which is arithmetically impossible. All forward valuation uses epsAvg and revenueAvg.

Peer context. The vendor set — Cigna, CVS, Elevance, GSK, HCA, McKesson, Pfizer, Sanofi, Vertex, Zoetis — mixes payers and distributors with pharmaceutical manufacturers. The genuine comparables are GSK, Pfizer, Sanofi and Vertex; CVS is also in this batch. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn.

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

At $65.895 — 9.8x FY2026 consensus, 10.2x FY2027, 13.6x FY2030 — the price embeds:

• Revenue of approximately $49.5 billion in 2026, at the midpoint of management's raised guidance. (Company guidance, half-weighted.) The first half delivered $24,462M, so the second half needs approximately $25.0B — a 2.2% sequential step. This is the least fragile assumption in the price.
• Non-GAAP EPS of $6.75-7.00 in 2026, then a decline to roughly $6.44 in 2027 and $5.56 in 2028. (Guidance, then consensus.) The market is explicitly paying for a declining earnings stream, which is what a 9.8x multiple on a company with a 70% gross margin and 3.8% dividend yield means.
• Eliquis loses US exclusivity in 2028 and erodes on the pattern Revlimid, Pomalyst, Sprycel and Abraxane have already demonstrated — 25% to 71% in a single year. (10-K date; our extrapolation from the company's own Legacy Portfolio experience.) This is the most fragile assumption in the price, and it is fragile in both directions: an earlier loss (through the Azurity 505(b)(2) challenge or European spillover) is materially worse, and a later or slower loss is materially better.
• The Growth Portfolio grows fast enough to leave FY2030 revenue at roughly $37.3 billion. (Consensus, 11 analysts.) Here is the arithmetic the consensus implies and it is worth stating explicitly. The Growth Portfolio is running at roughly $30 billion annualised and compounding 14.6%. Four years of that alone reaches roughly $52 billion. Consensus FY2030 total revenue of $37.3 billion therefore requires the Growth Portfolio's compound rate to collapse to low single digits — because Legacy will be nearly gone by then either way. That is a strong assumption and it is the variant perception available in this name.
• The dividend is maintained through the cliff. (Our number.) At $5.0 billion against $11.4 billion of free cash flow it is well covered today; on FY2030 consensus earnings of $4.86 per share it would be a 52% payout, still covered. The dividend is not obviously at risk, and at 3.81% it is a real part of the return.

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

Expected return over the next twelve months decomposes as: EPS growth (NEGATIVE 4.5%, from FY2026E $6.744 to FY2027E $6.442) + multiple drift (EXPANSION required, from 9.8x to roughly 10.5x on the respective forward year, +7%) + dividend yield (+3.81%) ≈ +3% to +7%.

Note what that decomposition says and how unusual it is: the earnings contribution is negative, and the base case therefore REQUIRES a modest multiple expansion to produce any return at all. That is the fragile leg of this thesis and we name it as such. The justification for expansion is that the market is currently paying 9.8x for a company whose near-year guidance was just raised 6.3% and whose largest product is growing 22% — a multiple more typically applied to a business already in decline. If Eliquis keeps growing through 2027 the market may reasonably decide 9.8x was too pessimistic. If it does not, there is nothing here.

Sensitivity: if the multiple simply HELD at 9.8x on FY2027E the price would be $63.13 (−4.2%). At 12x it is $77.30 (+17.3%). At 8x it is $51.54 (−21.8%).

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

• We differ from the street by only 4.0% ($68 against $65.38) but the sign matters: the street's consensus target is BELOW the current price, with 21 of 42 analysts at hold and a consensus rating of Hold — the only Hold consensus in this batch. When the crowd's target is below the price and our number is barely above it, that is a name that has run past its own coverage, and the correct response is to wait rather than to argue.
• We differ from consensus on the shape of the outer years, and this is the genuine variant perception. Consensus FY2030 revenue of $37.26 billion implies the Growth Portfolio — currently $30 billion annualised and compounding 14.6% — essentially stops growing. We do not think that is the base case; we think it is the conservative case that outer-year pharmaceutical models default to. But we are not paying for it either, because the coverage on those rows is 6 to 11 analysts, the Cobenfy setback in the knowledge base cuts directly against it, and four-year compound assumptions in oncology and neuroscience are not a basis for a position taken at a 52-week high.
• We think the Eliquis growth rate is being under-appreciated in the near term and correctly discounted in the long term. Watchable number: Eliquis revenue in the 2026-10-29 print. It grew 22% in the June quarter against a year-ago base that already included pre-Inflation-Reduction-Act pricing. A second consecutive 20%-plus quarter would say volume is overwhelming the mandated price cut, which is a meaningfully better setup into 2027 than consensus assumes. A deceleration below 10% would say the price cut is starting to bite and would pull the whole path forward.
• We think the knowledge base's Cobenfy claims deserve more weight than their sourcing warrants, because there is nothing else. The lane records the adjunctive schizophrenia trial as having failed with only a modest signal, and peak sales cut to $2.8 billion from $3.4 billion, with Alzheimer's psychosis the route back to roughly $5 billion. Cobenfy is the single most important asset in the post-2028 story — it is the reason the Karuna acquisition cost $14 billion and produced a company-level net loss in 2024 — and our only outside read on it is a downgrade. We flag the sourcing weakness in Section 6 and weight it accordingly, but we do not ignore it.
• Positive surprise that would force a re-rate: Eliquis sustaining 20%-plus growth into 2027; a positive Alzheimer's-psychosis readout for Cobenfy; a favourable outcome in the European apixaban patent litigation that slows the erosion; or resumption of the buyback, which has been zero for two years and would signal management believes the balance sheet is repaired.
• Negative surprise that would break the thesis: an adverse ruling accelerating US Eliquis genericisation before 2028 — the Azurity 505(b)(2) filing is the live vector; a further Inflation Reduction Act selection hitting Opdivo or Reblozyl; a Growth Portfolio quarter below 10% growth; or another acquired-IPRD charge at Karuna scale. Any of these hits a stock at its 52-week high with an RSI of 87.

Synthos fair values

All three anchors are multiples of the FY2027 consensus non-GAAP EPS distribution (mean $6.442, low $5.724, high $7.129, 16 analysts) — the last year with adequate coverage — cross-checked against management's FY2026 guidance midpoint of $6.875.

• Bear ~$46 — 8.0x the FY2027 consensus LOW of $5.724. Cross-check: 6.7x the FY2026 guidance midpoint; 8.0% above the 52-week low of $42.60; a 5.5% dividend yield at that price. The scenario: Eliquis genericises earlier than 2028 or erodes faster, the Growth Portfolio decelerates, further Inflation Reduction Act selections land, and the market applies a terminal-decline multiple. −30.2%.
• Base ~$68 — 10.6x the FY2027 consensus MEAN of $6.442, and 9.9x management's FY2026 guidance midpoint of $6.875. Cross-check: 12.2x FY2028E; a 3.69% dividend yield at that price. Sensitivity, stated openly: 9.8x — today's multiple, held — gives $63.13, which is BELOW spot; 12x gives $77.30. The scenario: the guided year lands, Eliquis holds up through 2027, the Growth Portfolio compounds, and the market pays a modestly higher multiple for a company whose near-year earnings are visibly beating. +3.2%.
• Bull ~$90 — 12.6x the FY2027 consensus HIGH of $7.129. Cross-check: 13.1x the FY2026 guidance midpoint; 16.2x FY2028E; a 2.79% dividend yield. The scenario: Cobenfy reads out positively in Alzheimer's psychosis, the Growth Portfolio's compounding is recognised as capable of replacing Eliquis, and the market re-rates the whole outer-year path. +36.6%.

Base is 3.2% above spot; asymmetry roughly 1.21:1 to the upside (30.2% down, 36.6% up), plus a 3.81% dividend. That is a thin base case and an unremarkable payoff ratio, offered at the exact 52-week high with the most overbought technical reading in the batch and a street target below the price. It is a Watch, and the price we want is nearer $57.

6. Knowledge base — eleven claims, one channel, and a degenerate concentration test

Raw hits: 11. Entity matches: 10. Discarded: 1. Used as name-level conviction: 0 — every claim is asset-level or basket-level. Channel concentration: 100%.

The sweep ran entity terms BMY, Bristol Myers Squibb, Bristol-Myers Squibb, Bristol Myers, Eliquis, Opdivo, Cobenfy and Karuna, plus free text on the same, across all 51,928 distilled claims. It returned 11 hits, and all eleven come from a single channel: biotech_hangout.

The concentration sensitivity test the house rules require is therefore not merely triggered — it is degenerate. Stripping the dominant channel leaves ZERO claims. There is no independent corroboration of anything in this lane from any other source in the knowledge base, and no way to test whether the channel's view is idiosyncratic. We state that as the headline finding. All eleven also carry speaker: null and are attributed only to the channel — the weakest sourcing the 4-lane policy admits.

One discard: a 2025-11-25 claim entity-tagged to Seaport Therapeutics and Karuna Therapeutics about starting with molecules that have already shown human efficacy. It is about Seaport, and Karuna appears only as a reference. Not a BMS claim.

The remaining ten, grouped by what they actually say.

On Cobenfy — the two most decision-relevant claims in the lane, and both are cautionary:

> 2025-08-21 · bullish 62 · thesis · entities: Cobenfy, Bristol Myers Squibb, Karuna

> "Cobenfy adjunctive schizophrenia trial failed with only modest signal, but drug will still be widely used as combination therapy; bull case intact."

> 2025-08-21 · neutral 55 · FACT · entities: Cobenfy, Bristol Myers Squibb

> "Peak sales cut to $2.8B from $3.4B after removing adjunctive schizophrenia; Alzheimer's psychosis could restore upside to ~$5B."

Cobenfy is the asset the entire post-2028 story rests on — it is why Bristol Myers Squibb paid roughly $14 billion for Karuna, a transaction that contributed to a net loss of $8,948 million in FY2024 — and the knowledge base's only word on it is a failed trial and an 18% peak-sales downgrade. The filings corroborate the small current scale: $10 million of FY2024 revenue, $155 million of FY2025 revenue, and $63 million in the June 2026 quarter (+80% year on year). That is a real launch trajectory and a very long way from $2.8 billion.

On business development — three claims, all bullish, all specific:

> 2026-01-23 · bullish 65 · FACT · entities: Bristol Myers Squibb, Janux Therapeutics

> "BMS acquisition of Janux is a vote of confidence in conditionally-active/masked T-cell engager antibodies for oncology."

> 2025-08-21 · bullish 65 · thesis · entities: Bristol Myers Squibb, BioNTech

> "BMS partnering BioNTech's PD-L1xVEGF (~$3.5B guaranteed to 2028) de-risks PDL1-vs-PD1 uncertainty and enables potential PD1 combos; good for BioNTech."

> Corroborated by the 10-Q, which describes the June 2025 global strategic collaboration on pumitamig (BNT327/BMS986545) with equal profit-and-loss sharing.

> 2025-08-21 · bullish 65 · thesis · entities: Bristol Myers Squibb, Amgen, Tango Therapeutics

> "PRMT5 is an emerging pancreatic-cancer target; BMS ASCO data ~25% response rate, exciting potential combined with KRAS inhibitors."

On setbacks — one claim:

> 2025-08-27 · neutral 50 · FACT · entities: Moderna, Bristol Myers Squibb

> "Bristol's LAG-3/PD-1 adjuvant melanoma failure lowers the competitive bar for Moderna's key readout but flags adjuvant melanoma as hard to hit."

On the sector — three claims, one of which is the only thing resembling a view on the stock:

> 2026-01-16 · bullish 68 · thesis · entities: Eli Lilly, AbbVie, Merck, Bristol Myers Squibb, Pfizer, Gilead, Amgen

> "Generalists are rotating into pharma/biotech because it looks cheap versus an expensive AI-driven S&P; interest is broadening beyond mega-caps to Merck, Bristol, Pfizer, Gilead, Amgen."

> A basket rotation call, not name-level conviction — and, dated January 2026, it has been right: the stock is up 49% over twelve months.

> 2025-08-21 · neutral 50 · FACT · entities: Bristol Myers Squibb, Gilead, Teva

> "Tariff pressure drove US investment pledges (BMS $40B, Gilead +$11B/3000 jobs, Teva $30B), though much may be repackaged spend rather than truly additive."

> 2025-08-27 · bullish 65 · thesis · entities: Johnson & Johnson, Intra-Cellular Therapies, Karuna, Bristol Myers Squibb

> "CNS is a hot M&A theme; J&J's Intra-Cellular buy is largest since Karuna/BMS, the fourth multi-billion CNS acquisition in a year."

Conclusion, stated as a finding. Breadth 1. The Synthos knowledge base has a well-informed scientific view of Bristol Myers Squibb's pipeline and no view whatsoever on Bristol Myers Squibb the security — there is not one claim about valuation, the Eliquis cliff, the dividend, or the share price. And because every claim comes from one channel with no named speaker, there is no way to test the view against an independent source. The lane is useful as colour on Cobenfy, Janux, BioNTech and PRMT5, and it is not conviction. Net conviction is recorded as positive-low-unverifiable, and the Cobenfy downgrade is a direct input to the exponential score of 4.

7. Data integrity — what we rejected and why

Seven findings. BMY's product-level revenue disclosure is the cleanest in this batch; its geographic block is missing 28.7% of revenue; and the forward EBIT/EBITDA rows are impossible.

1. seg_geo for FY2025 omits the "Rest of World" line — 28.7% of revenue missing. The vendor reports United States $33,279M and "Other Region" $1,087M, summing to $34,366M against reported revenue of $48,195M. The FY2024 and FY2023 entries include a "Rest Of World" line ($13,199M and $12,752M); the FY2025 entry does not. The gap is approximately $13,829M. All geographic figures in this dive come from the 10-Q's own regional table (Q2 2026: United States $8,991M, International $3,664M, Other $318M).

2. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature AND are arithmetically impossible — REJECTED. In every forward year ebitdaAvg is exactly 33.24% of revenueAvg and ebitAvg is exactly 14.20%. Additionally, netIncomeAvg exceeds ebitAvg in EVERY forward year — $14,013M against $6,959M (FY2026), $13,213M against $6,773M (FY2027), $11,261M against $5,921M (FY2028) — net income roughly double operating income, which cannot happen. All forward valuation uses epsAvg and revenueAvg.

3. The FY2026 consensus EPS of $6.744 sits BELOW management's guidance floor of $6.75 — the estimate block is stale relative to the 2026-07-30 raise. Guidance was lifted from an implied lower range to $6.75-7.00 and revenue guidance from ~$46.0-47.5B to ~$49.0-50.0B. est FY2026 revenueAvg of $48,993M sits at the bottom of the new revenue range. Every forward multiple in this dive is therefore, if anything, slightly conservative; at the guidance midpoint of $6.875 the FY2026 multiple is 9.6x rather than 9.8x.

4. The FY2028, FY2029 and FY2030 estimate rows rest on 8, 6 and 6 analysts — used for shape, never for a fair value. They are shown because the shape they describe — revenue down 24% and EPS down 28% from the 2026 base — is the central fact about this company. No anchor is computed on them.

5. epsdiluted is null in every row of inc_a and inc_q — worked around. All per-share figures here are computed and labelled, and cross-checked against the 8-K's stated GAAP EPS of $1.62 and non-GAAP EPS of $2.04 for the June quarter.

6. The FY2024 income statement is dominated by non-operating items and is not comparable — flagged. FY2024 shows a net loss of $8,948M on revenue of $48,300M, driven by acquired in-process research and development charges from $20,722M of acquisition spending, plus $9,600M of depreciation and amortisation against $4,011M in FY2025. The $5.6 billion year-on-year fall in amortisation is the largest single driver of the FY2025 earnings recovery and is mechanical, not operational. No growth rate in this dive is computed across the FY2024/FY2025 boundary on a GAAP earnings basis.

7. quote.yearHigh/yearLow and tech.hi52/lo52 agree to within 0.2% — a clean pass, and it matters here. The quote block reports $66.31 and $42.52; the technical block reports $65.895 and $42.60. The technical block's 52-week high equals the closing price exactly, giving pct_from_hi of 0.0% — the stock closed at its high. The quote block's $66.31 is the intraday high of the session. We use tech throughout and state that the closing price is the 52-week closing high.

Correct and worth recording — this is the cleanest product disclosure in the batch: seg_prod for FY2025 lists twenty individual products summing to exactly $48,195M against reported revenue of $48,195M, with full four-year history and no double-counting — better than any other segment table in this batch; bookValuePerShareTTM of $10.930 is correctly distinct from tangibleBookValuePerShareTTM of −$8.231, a genuinely negative tangible book properly reported; dividendPerShareTTM of $2.51 and dividendPayoutRatioTTM of 54.9%; interestCoverageRatioTTM of 8.04x; and incomeQualityTTM of 1.378, correctly reflecting $14,156M of operating cash flow against $7,054M of net income.

Non-equity tripwire — checked and passed. BMY is common stock, NYSE-listed. Price of $65.895 is not par-like; beta is 0.235, low but explained by the defensive sector rather than by a fixed-income structure; the dividend is variable and has been raised over time; volume was 17.29M shares (~$1.1B of turnover); the 52-week band of $42.60 to $65.895 is a 55% range. This is common equity.

8. Technicals

• Price $65.895 — the 52-week closing high. tech.pct_from_hi reads exactly 0.0%. +54.7% above the 52-week low of $42.60. Position within the annual range: 100th percentile.
• Above both moving averages by wide margins: +12.9% above a 50-day average of $58.39; +18.1% above a 200-day average of $55.79. Both rising.
• RSI 87.1 — the most overbought reading in this batch by a wide margin, and 17 points above the conventional 70 threshold. MACD +2.10.
• Maximum drawdown from peak over the trailing year: −18.8%, against a current drawdown of zero.
• Relative performance: 3-month +14.8% against SPY +7.6%; 6-month +17.9% against SPY +11.1%; 12-month +49.0% against SPY +24.3% and QQQ +30.8%. A 24.7-point twelve-month outperformance from a stock with a beta of 0.235 — which means essentially the entire return is idiosyncratic.
• Sentiment is the most cautious in this batch: 19 buy, 21 hold, 2 sell across 42 analysts, with a consensus rating of Hold — the only Hold consensus in the batch; consensus target $65.38, 0.8% BELOW spot; median $67.50; high $75 (+13.8%); low $40 (−39.3%). The $35 spread between the high and low targets, on a $66 stock, is a 53% range — the widest dispersion here, and it is the patent cliff being modelled differently by different analysts.

Today's move and what it does to the entry

BMY closed 2026-08-04 at $65.895, up 0.65% or $0.425 from $65.47. It opened at $64.89, traded $63.87 to $66.31, and closed near the high on 17.29M shares. No company-specific filing is dated 2026-08-04; the last event was the 2026-07-30 earnings release and guidance raise, five days earlier.

The honest read: this is not a gap to avoid, it is a level to avoid. A 0.65% move is noise. What matters is that the close IS the 52-week high, RSI is 87.1, the stock is up 49% in twelve months, and the average analyst target is below the price.

What would change this:

• A price near $57 — roughly 8.8x FY2027E, close to the 50-day average of $58.39 — which would put our base $68 at a 19% premium and lift the dividend yield to 4.4%. That is 13% below today's close and well inside a normal consolidation after a 49% run.
• Or Eliquis growing 20%-plus again in the 2026-10-29 print, which would say volume is beating the Inflation Reduction Act price cut decisively and would push the whole revenue path a year to the right. At that point paying up is defensible.

9. Insiders — eight transactions, all mechanical

Zero open-market purchases. Zero open-market sales. Five routine director restricted-stock-unit grants dated 2026-06-30; one executive unit conversion; and two entries for director Julia Haller coded M-Exempt at $53.76 — an option or unit conversion at the stated exercise reference, recorded on two ownership lines, not a sale.

This is a neutral file and it is reported as neutral. What is notable by absence: with the stock at a 52-week high after a 49% run, not a single insider sold. That is a mild positive — mildly, because the absence of selling at a high is more often a reflection of trading-window policy than of conviction. And no transaction by the chief executive or chief financial officer appears anywhere in the file.

10. Verdict, kill-criteria and flip conditions

Watch.

What is genuinely strong: a Growth Portfolio at $7,560 million in the June quarter, +14.6%, now 58.3% of revenue, with Breyanzi +41%, Camzyos +60% and Cobenfy +80%; Eliquis growing 22% in its first full quarter under Inflation Reduction Act pricing; guidance raised on 30 July by roughly $3 billion of revenue to ~$49.0-50.0B and to $6.75-7.00 of non-GAAP EPS; a 70.2% gross margin and $11,443 million of trailing free cash flow at an 8.5% yield; a 3.81% dividend costing 44% of that free cash flow; net debt down from $40.9B to $36.9B; and a product-revenue disclosure that reconciles exactly.

What we are declining to pay for: a close at the exact 52-week high with an RSI of 87.1 after +49.0% in twelve months; a street consensus target of $65.38 that is BELOW the closing price, with a Hold consensus rating and 21 of 42 analysts at hold; a consensus earnings path that declines in every forward year — $6.74, $6.44, $5.56, $4.94, $4.86 — on revenue falling 24% from 2026 to 2030; a forward multiple that RISES with time (9.8x, 10.2x, 11.9x, 13.3x, 13.6x); Eliquis at 30.0% of revenue with US exclusivity recorded at 2028 and generics already selling in the UK and Finland; 43.4% of FY2025 revenue subject to government price-setting within three years; tangible book value of negative $8.23 per share; a knowledge-base lane that is 100% one channel and whose only word on the key asset is a downgrade; and a base case that requires multiple expansion to produce a positive return.

The distinction that matters. This is not a bad company and the recent execution has been genuinely good — the guidance raise was large and the Growth Portfolio is doing what it needs to do. It is a company whose near-year earnings are visibly beating and whose far-year earnings are visibly falling, bought at the exact price at which the market has most enthusiastically celebrated the first fact. In the Synthos frame, a name where the base case needs a re-rating to work, the crowd's target is below the price, and the conviction lane cannot corroborate itself, is a Watch with a number attached — and the number is roughly $57.

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

• Any development accelerating US Eliquis genericisation before 2028 — the Azurity 505(b)(2) formulation-patent challenge filed November 2025 is the live vector, and European spillover is the second.
• Eliquis growth below 10% in a quarter, which would say the Inflation Reduction Act price cut is overwhelming volume and would pull the whole revenue path forward.
• Growth Portfolio growth below 10%, which would remove the only offset to the cliff.
• A negative Alzheimer's-psychosis readout for Cobenfy, which the knowledge base identifies as the route from a $2.8 billion peak back toward $5 billion.
• A further Inflation Reduction Act selection hitting Opdivo ($10,049M) or Reblozyl ($2,327M).
• Another acquired-IPRD charge at Karuna scale without a corresponding asset, or any dividend action.

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

• A price near $57 — roughly 8.8x FY2027E with a 4.4% dividend yield — with the guidance intact. This is the single most likely path to a Buy and it is 13% below today's close.
• A second consecutive 20%-plus Eliquis quarter in the 2026-10-29 print, which would materially change the shape of the 2027 and 2028 estimates.
• Cobenfy exceeding roughly $100 million in a quarter with a positive Alzheimer's-psychosis signal, which converts the post-2028 story from an assertion into a number.
• Resumption of the buyback, absent for two years, which would signal management considers the balance sheet repaired and the stock undervalued.
• Independent knowledge-base corroboration from a second channel — at present the lane is 100% one source and cannot be stress-tested, and that alone is worth roughly a tier.

Where BMY fits in the Synthos Framework Portfolio. The large-cap pharmaceutical sleeve, at 0% today with a 1.5-2.5% target on a fill near $57 or a confirming October print. On batch overlap: CVS is in this batch and in BMY's own vendor peer set, but they are opposite sides of the same industry — BMY sells the drugs, CVS distributes and pays for them — and holding both is a hedge rather than a concentration. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $65.895.

Single biggest risk: Eliquis. $14,443 million of FY2025 revenue — 30.0% of the company — with United States exclusivity recorded in the 10-K's own patent table as 2028. Generic apixaban is already marketed in the United Kingdom and Finland following adverse judicial decisions; a 505(b)(2) challenge on a US formulation patent was filed in November 2025; and the Inflation Reduction Act maximum fair price has applied to US Medicare since 1 January 2026. The company's own Legacy Portfolio shows exactly what follows: Revlimid −49%, Pomalyst −71%, Abraxane −48%, Sprycel −27%, all in a single year, all after exclusivity ended. Consensus models a $6.0 billion revenue decline in FY2028 alone. At a 9.8x multiple on the guided year the market appears to have priced it; at 13.6x the 2030 number it appears not to have. Both statements are true, and the difference between them is the whole investment question.

Provenance & disclosures

• Traceability: 10 tagged knowledge-base claims name Bristol Myers Squibb out of 51,928 distilled claims (raw hits 11, entity matches 10, discarded 1, used as name-level conviction 0; breadth 1, net conviction positive-low-unverifiable). ALL ELEVEN HITS COME FROM A SINGLE CHANNEL, biotech_hangout — 100% concentration. The mandatory concentration sensitivity test is therefore degenerate: stripping the dominant channel leaves ZERO claims, and no independent corroboration of any assertion in this lane exists anywhere in the knowledge base. All eleven carry speaker: null and are attributed only to the channel — the weakest sourcing the 4-lane policy admits; no claim carries a management speaker_role. The two most decision-relevant claims are both cautionary and both concern Cobenfy, the asset the post-2028 story depends on: 2025-08-21 (conviction 62) recording that the adjunctive schizophrenia trial "failed with only modest signal", and 2025-08-21 (conviction 55, horizon FACT) recording peak sales "cut to $2.8B from $3.4B... Alzheimer's psychosis could restore upside to ~$5B." Three business-development claims are bullish and specific: Janux (2026-01-23, conviction 65), BioNTech PD-L1xVEGF at ~$3.5B guaranteed to 2028 (2025-08-21, conviction 65 — corroborated by the 10-Q's pumitamig collaboration note), and PRMT5 in pancreatic cancer at a ~25% response rate (2025-08-21, conviction 65). One is a setback (LAG-3/PD-1 adjuvant melanoma failure, 2025-08-27, conviction 50). Three are sector- or basket-level, including a rotation call naming Bristol alongside Merck, Pfizer, Gilead and Amgen (2026-01-16, conviction 68) which has been correct. One discard: a 2025-11-25 claim entity-tagged to Seaport Therapeutics and Karuna Therapeutics which is about Seaport, not BMS. No homograph collisions occurred. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals — product-level revenue, portfolio split and geographic detail through 2026-06-30, filing-verified from the 10-Q filed 2026-07-30; second-quarter results and the raised FY2026 guidance from the 8-K filed 2026-07-30, which carries the full earnings release; exclusivity dates, Inflation Reduction Act detail and litigation from the 10-K filed 2026-02-11 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873599 = 2026-08-04T19:59:59Z ($65.895, +0.65%; 50-DMA $58.39; 200-DMA $55.79; RSI 87.1; MACD +2.10) · knowledge-base claims 2026-08-04. Bristol Myers Squibb's fiscal year is the calendar year. All figures come from the Synthos vendor data file for BMY or from the SEC filings in the BMY archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents: 10-K filed 2026-02-11 (fiscal 2025 — the source of the Eliquis 2028 exclusivity row, the European generic-launch disclosure, the Azurity 505(b)(2) notice and the Inflation Reduction Act pricing schedule); 10-Q filed 2026-04-30 (March 2026 quarter); 10-Q filed 2026-07-30 (June 2026 quarter — the source of the Growth/Legacy portfolio split, product-level revenue and the BioNTech and Hengrui collaboration notes); 8-K filed 2026-07-30, which carries the complete second-quarter earnings release including the raised guidance table; 8-K filed 2026-04-30 (first-quarter release with reaffirmed guidance); 8-K filed 2026-05-08 (annual meeting voting results). All carry preserved [TABLE] statement data.
• Where the filings supplied what the vendor could not, and where they contradicted it (detailed in Section 7): seg_geo for FY2025 omits the "Rest of World" line of approximately $13,829M — 28.7% of revenue — leaving two lines summing to $34,366M against reported revenue of $48,195M, so all geographic figures come from the 10-Q's regional table; est.ebitdaAvg/ebitAvg rejected for a fixed-ratio signature (exactly 33.24% and 14.20% of revenue in every forward year) and for netIncomeAvg exceeding ebitAvg in every forward year by roughly 2x, which is arithmetically impossible; the FY2026 consensus EPS of $6.744 sits below management's guidance floor of $6.75, so the estimate block is stale relative to the 2026-07-30 raise and every forward multiple here is conservative; the Eliquis 2028 US exclusivity date, the UK and Finland generic launches, the Azurity challenge and the Inflation Reduction Act price-setting schedule appear in no vendor field and come entirely from the 10-K; and the FY2024 income statement, dominated by acquired-IPRD charges and $9,600M of amortisation, is flagged as not comparable with FY2025's $4,011M. Where vendor and filing AGREED — recorded because these defect classes usually bite: seg_prod for FY2025 lists twenty products summing to EXACTLY $48,195M against reported revenue of $48,195M — the cleanest product-revenue disclosure in this batch; bookValuePerShareTTM of $10.930 correctly distinct from tangibleBookValuePerShareTTM of −$8.231; the Q2 2026 revenue of $12,973M and GAAP EPS of $1.62 against the 8-K; and quote.yearHigh/yearLow within 0.2% of tech.hi52/lo52.
• Basis note: analyst estimates and company guidance are on non-GAAP EPS, which excludes amortisation of acquired intangibles, acquired in-process research and development charges, integration and restructuring costs, impairments, and contingent-value-right fair-value adjustments. The gap is large and lumpy: June-quarter GAAP EPS was $1.62 against non-GAAP $2.04, a 26% difference; FY2024 produced a GAAP net LOSS of $8,948M against a positive non-GAAP year. Forward multiples in this dive use consensus non-GAAP EPS and management's non-GAAP guidance; the trailing multiple of 14.5x is on GAAP and is labelled as such. Acquired IPRD charges recur every year and are excluded from every non-GAAP figure — $1.5 billion for the Hengrui transaction in the June 2026 quarter alone — which is a genuine, repeating economic cost that the headline number does not carry.
• Estimate coverage: 15-18 analysts on FY2026 and FY2027 — FY2027 is the anchor for all three fair values, with a range of $5.724 to $7.129 (24%); 8 on FY2028 EPS, 6 on FY2029 and 6 on FY2030. The outer rows are shown because the shape they describe (revenue −24%, EPS −28% from the 2026 base) is the central fact about this company, but no fair value is anchored on them.
• Peer note: the vendor peer set — Cigna, CVS, Elevance, GSK, HCA, McKesson, Pfizer, Sanofi, Vertex, Zoetis — mixes payers and distributors with manufacturers; the genuine comparables are GSK, Pfizer, Sanofi and Vertex. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn. CVS is also in this batch and sits on the opposite side of the same industry.
• Fair-value caveat: the $46 / $68 / $90 anchors are multiples of the FY2027 consensus non-GAAP EPS distribution — 8.0x the low of $5.724, 10.6x the mean of $6.442, and 12.6x the high of $7.129 — cross-checked against management's FY2026 guidance midpoint of $6.875. Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 9.8x — today's multiple, held — gives $63.13, which is BELOW spot; 12x gives $77.30. The base case explicitly REQUIRES modest multiple expansion, because consensus EPS growth into FY2027 is negative; that is the fragile leg and it is named as such in the return bridge.
• Management-voice note: the raised FY2026 guidance (revenue ~$49.0-50.0B, non-GAAP EPS $6.75-7.00, operating expenses ~$16.5B, gross margin ~69-70%, tax rate ~18%) and the chief executive's characterisation of Growth Portfolio momentum are management's own claims and are half-weighted throughout, and labelled wherever quoted.
• Redaction disclosure: the 2026-07-30 earnings release describes a May 2026 enterprise agreement with a large-language-model provider covering research, clinical development, manufacturing, commercial and corporate functions. The provider's and the product's names are redacted here under the standing house rule against naming AI models or laboratories. The substance — an enterprise-wide deployment as a shared intelligence platform — is reported in full. The separately disclosed NVIDIA infrastructure collaboration is named because NVIDIA is a semiconductor manufacturer, not an AI laboratory.
• Timing: second-quarter 2026 results and the guidance raise were released 2026-07-30, five days before this dive. The next print is 2026-10-29, 86 days away. 2026-08-04 carried no company-specific filing, and the closing price of $65.895 IS the 52-week closing high.
• 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.