Growth Portfolio Clears Half of Revenue While EPS Falls 12%: Initiating Bristol Myers Squibb at Hold Ahead of a Binary Second Half

Results vs. Consensus

Q1 2026 Scorecard

Consensus figures span a provider range of $10.69B to $10.96B on revenue; $10.92B is the most widely cited and is what the beat magnitude above is calculated against. The $1.42 EPS figure is unanimous across every provider we checked and is a non-GAAP number, so it compares to the $1.58 non-GAAP result rather than to GAAP $1.31.

Year-Over-Year Comparison

The Change column is computed from the two dollar columns beside it. The Ex-FX column is the company's own constant-currency figure and is not a recomputation: currency contributed 2 points of the 3 points of reported total revenue growth, and 7 of the 11 points of international growth.

Quarter-Over-Quarter Comparison

Sequential comparisons for a US-weighted pharmaceutical company are noisy by construction. The first quarter carries benefit-design resets, deductible resets, gross-to-net true-ups and wholesaler destocking after the fourth quarter's stocking, so an 8% sequential revenue step-down is a seasonal pattern rather than evidence of deterioration. The sequential figures worth reading are the ones that break the seasonal pattern: Eliquis up 19.8% into a seasonally weak quarter (a price-cut-driven inventory build, which management pre-announced will reverse), and Opdivo IV down 20.3% (a wholesaler drawdown plus subcutaneous conversion). The gross-margin improvement is not a mix story; the fourth quarter carried a $564M intangible impairment inside cost of products sold, and Q1 did not.

Quality of Beat

Assessment on revenue: The top-line beat was the right kind in composition and the wrong kind in durability. Twelve of the thirteen Growth Portfolio lines grew, seven of them at 15% or better ex-currency, and the portfolio's share of company revenue held above half at 54.2%. That is the transition working. But the two largest positive contributors in dollar terms, Eliquis at $4,137M and currency at roughly 2 points, are both explicitly temporary: management told us the Eliquis build reverses next quarter, and a currency tailwind that flatters a reported number does nothing for the constant-currency guidance range.

Assessment on margins: We are more relaxed about gross margin than the 280bps headline suggests, and less relaxed about the operating expense line than the 0.9% growth suggests. Gross margin at 70.3% came in at the top of the 69% to 70% full-year guide, which means the mix headwind management has been signaling has not yet arrived at the pace they modeled. Operating expenses, on the other hand, are running at 23.6% of the $16.3B full-year guide, which means the remaining three quarters must average $4,149M, or 7.7% above the Q1 run rate. That step-up is either a spending ramp behind the pumitamig program and the second-half launches, or it is conservatism in the guide. On the evidence of one quarter we cannot tell which, and it is the difference between the low end and the high end of the EPS range.

Assessment on EPS: This is the quarter where the below-the-line cushion that had been flattering BMY's non-GAAP earnings came off. The diabetes royalty was a genuine $1.1B annual income stream in 2025 that is now zero, and the tax rate normalized toward the guided 18%. Both were disclosed and neither is a surprise to anyone who read the FY25 filings. What matters is what it implies about the shape of the recovery: from here, EPS growth has to come from operating leverage on a shrinking revenue base, which is a much harder ask than it was when royalty income and a 15% tax rate were doing part of the work.

Segment Performance

BMY does not report operating segments in the conventional sense. It reports one commercial business split into a Growth Portfolio of assets early in their life cycles and a Legacy Portfolio of brands facing or already past loss of exclusivity. That split is the entire investment debate, so it is where we start.

Portfolio Mix

Growth Portfolio by Product (worldwide)

Legacy Portfolio by Product (worldwide)

Component lines are as filed and are computed on unrounded amounts, so the Legacy Portfolio components sum to $5,278M against the $5,277M total shown. The total row is the filed figure, not the sum.

Growth Portfolio: the number that matters crossed 50%

At $6,227M the Growth Portfolio is 54.2% of company revenue, against 49.7% a year ago. For a company whose entire equity story is whether a new portfolio can be built faster than an old one erodes, this is the scoreboard, and it moved 4.5 points in a year. The composition is what makes it credible: growth is not concentrated in one asset. Camzyos added $155M, Breyanzi $148M, Opdivo Qvantig $154M, Reblozyl $77M, Other Growth Products $74M, Opdualag $43M and Orencia $48M, and every one of those is a different therapeutic area with a different competitive setup.

"Our growth portfolio continued to perform well with global revenue increasing 9% to $6.2 billion. As Chris mentioned, several products that are still early in their life cycles are driving growth as we intentionally expand our business across a wider range of key assets."
— David Elkins, Chief Financial Officer

Assessment: The breadth is genuine and it is the strongest fact in the print. The caution is arithmetic rather than qualitative: the Growth Portfolio added $664M of revenue year over year while the Legacy Portfolio lost $361M, a net $303M. That 1.8-to-1 coverage ratio is comfortable now, but Revlimid still has $349M of quarterly revenue left to lose and Pomalyst $513M, and the guide implies the coverage ratio goes below 1.0 for the balance of the year.

Eliquis: the strongest line in the quarter, and the one that reverses next quarter

Eliquis grew 16% reported and 13% ex-currency to $4,137M, comfortably above the 10% to 15% full-year growth guide. Management was explicit that part of it is not demand. A US list-price reduction took effect on January 1 and wholesalers built inventory ahead of and around it.

"We continue to see strong demand. And given our U.S. price reduction that took effect at the beginning of the year, we also saw some wholesale inventory build in the first quarter. We anticipate this build to reverse in the second quarter."
— David Elkins, Chief Financial Officer

The same price action has a cash consequence that is larger than the revenue consequence. Operating cash flow of $1,104M was down 43.5% year over year, and the CFO put roughly $1.2B of that on lower net cash collections tied to the Eliquis list-price reduction, with the offset coming later in the year through lower rebate payments.

Assessment: The 16% print should be read as a mid-single-digit-to-low-teens underlying rate with a one-off pull-forward on top. That is still a good outcome for a product this mature, and it is consistent with the full-year guide. The item to watch is not Q2 revenue, which management has already flagged, but whether the promised rebate-payment offset shows up in Q2 and Q3 cash flow. If it does, the Q1 cash shortfall is a timing artifact. If it does not, the dividend coverage question below becomes a real one.

Opdivo and Qvantig: the franchise held while the IV brand fell

Opdivo IV fell 5.3% reported and 8% ex-currency to $2,146M, with most of the decline in the US. Management attributed it to a wholesaler inventory drawdown rather than to demand, and to conversion into the subcutaneous formulation.

"This is primarily driven by an Opdivo inventory drawdown at the wholesaler level, where inventories are at the low end of the typical range. We continue to monitor whether these levels will normalize over the balance of the year."
— David Elkins, Chief Financial Officer

Taken together with Qvantig at $163M, up from $9M in the year-ago quarter, the combined franchise was $2,309M against $2,274M, or 1.5% growth. Management characterized US conversion as having passed 10% of IV volume in just over a year on the market, and reiterated an expectation that physicians convert roughly 30% to 40% of the IV business over the next two years.

Assessment: This is the most reassuring reading available on the quarter's largest single drag, and it is also the least verifiable. "Inventory drawdown, not demand" is what every management says about a soft quarter in a mature oncology brand, and the company itself hedged with "we continue to monitor whether these levels will normalize." What is verifiable is the conversion arithmetic, and it is going well: a 10% conversion rate at roughly a year post-launch, on the way to a stated 30% to 40%, is a genuine defensive moat against the eventual nivolumab biosimilar. We treat the franchise total, not the IV line, as the number to track from here.

Camzyos, Breyanzi and Cobenfy: the three launches carrying the arc

Camzyos nearly doubled to $314M, with management pointing to roughly 25,000 US patients now prescribed and thousands more internationally. Breyanzi grew 56% to $411M on demand across its approved indications in both the US and international markets. Cobenfy grew 107% to $56M, which is real but is still among the smallest lines in the Growth Portfolio and is growing off a base that makes percentages flattering.

Assessment: Camzyos is the asset most exposed to a near-term change in the competitive setup, with a competing cardiac myosin inhibitor now approved and launching. Management's argument is that the four-year-old Camzyos REMS infrastructure, the single-step titration from 5mg to 10mg covering roughly 90% of patients, and the speed of symptomatic improvement combine into a switching barrier. That is a reasonable case, not a proven one, and the 11% sequential decline (partly seasonal) is not yet evidence either way. Cobenfy at $56M is not a commercial story at all yet; it is a call option on the Alzheimer's psychosis readouts, which is a different security.

Legacy Portfolio: the erosion is on schedule, which is the problem

Revlimid fell 63% to $349M, Sprycel 58%, Abraxane 52% and Pomalyst 22%. Nothing here is a surprise, and management did not treat any of it as one. The Legacy Portfolio's 6.4% decline was smaller than it would otherwise have been only because Eliquis, which sits inside it, grew 16%.

Assessment: Ex-Eliquis, the Legacy Portfolio fell from $2,073M to $1,140M, a 45% decline. That is the real shape of the cliff, and it is being masked in the reported total by a single mature product whose settled US generic entry date is 2028. The Growth Portfolio has to be big enough to absorb both, and on the current trajectory the answer is "eventually, but not in 2026 or 2027."

Key Topics & Management Commentary

Overall Management Tone: Confident, procedural and noticeably rehearsed, with the confidence concentrated in the pipeline rather than in the current P&L. Management repeated the same three-priority framework twice on the call and used the phrase "say-to-do ratio" unprompted, which is the posture of a team that believes its credibility problem is behind it. The two places the tone thinned were both financial: the cash-flow explanation was delivered quickly and not returned to, and the below-the-line collapse in other income was never characterized at all.

1. The Growth Portfolio crossing half of revenue

Management led the call with the portfolio transition rather than with the beat, which is the correct emphasis and a change in framing from a company that spent much of 2024 and 2025 defending the Celgene-era balance sheet. The Growth Portfolio at 54.2% of revenue keeps the new business larger than the old one, as it has been since Q2 2025.

"We delivered a solid Q1 and continued to improve our say-to-do ratio, with disciplined execution across the business as we continue to best position the company for long-term sustainable growth."
— Christopher Boerner, Board Chair and CEO

Assessment: The claim is supported by the numbers, and the numbers are the ones investors should be underwriting. What the framing elides is that crossing 50% is a mix milestone, not an earnings milestone: the Growth Portfolio carries a lower gross margin than the Legacy Portfolio it is replacing, which is precisely why gross margin fell 280bps in the same quarter the milestone was reached.

2. Where the guidance language actually changed

Every line of the February guide was reaffirmed at unchanged values. The only new information in the outlook is a sentence, repeated by both the CEO and the CFO, about where within the ranges the company expects to land.

"In summary, based on our performance, we see the business currently tracking towards the upper end of our guidance ranges."
— Christopher Boerner, Board Chair and CEO

There is a currency subtlety buried in the release's footnotes that materially changes how generous this is. The February guide was struck at mid-January exchange rates; the April reaffirmation was struck at mid-April rates. Between those dates the dollar weakened enough to contribute 2 points to Q1 reported revenue growth and 7 points to international growth. Holding the dollar range unchanged across that move means the constant-currency outlook was effectively trimmed, or at minimum that the currency tailwind was absorbed rather than passed through.

Assessment: "Upper end" is doing a lot of work in the market's reading of this print, and it deserves less weight than it received. The company did not raise a single number, absorbed a currency tailwind without passing it through, and issued the qualifier one quarter into a four-quarter year. It is a real signal, but it is the smallest possible version of one.

3. Gross margin, and why the compression is structural

Non-GAAP gross margin fell 280bps to 70.3% on product mix. The CFO framed it as expected rather than as a miss.

"As expected, gross margin declined 280 basis points in the first quarter to 70.3%, which was primarily driven by product mix."
— David Elkins, Chief Financial Officer

The mechanism is arithmetic. Revlimid, Sprycel and Abraxane carried very high gross margins and are collapsing. Cell therapy (Breyanzi), biologics with complex manufacturing, and newly launched small molecules carry lower ones. Each point of Growth Portfolio mix shift is dilutive by construction, and the full-year guide of 69% to 70% embeds further compression from Q1's 70.3%.

Assessment: This is the clearest example on the call of a number that looks like a miss and is actually a plan working. It is also the reason why the "Growth Portfolio crosses 50%" milestone should not be read as a straightforward positive for earnings. The company is trading high-margin declining revenue for lower-margin growing revenue, which is the right trade for durability and a bad trade for near-term margins.

4. The below-the-line step-down nobody discussed

Non-GAAP other income and expense swung from $150M of income in Q1 2025 to $173M of expense in Q1 2026, a $323M adverse move. The GAAP schedule shows the mechanism: "Royalty income - divestitures," which contributed $272M in the year-ago quarter, is zero, because the diabetes-product royalty stream expired at the end of 2025. Royalty and licensing income fell from $259M to $195M on top of that. Interest expense fell from $494M to $411M, partially offsetting.

The written release names the cause in a single clause. On the call, across prepared remarks and eleven analyst questions, it was not raised once.

Assessment: This is the most consequential number in the quarter that received no airtime. Roughly $1.1B of annual pre-tax income disappeared from the P&L on January 1, and it will recur as a headwind in each of the next three quarters. The full-year guide accounts for it (other income and expense is guided to roughly negative $700M for 2026 on a non-GAAP basis), so this is not a guidance risk. It is an earnings-quality fact that the "upper end" framing sits on top of.

5. Cash conversion and the dividend

Operating cash flow was $1,104M against $1,954M a year ago. Capital expenditure was $347M, so free cash flow was $757M. Dividends paid were $1,283M. The dividend was not covered by free cash flow in the quarter, and was not covered by operating cash flow either.

"This quarter's cash flow reflects roughly $1.2 billion and lower net cash collections due to Eliquis list price reductions. We expect this to be more than offset later in the year through lower rebate payments."
— David Elkins, Chief Financial Officer

The explanation is specific, mechanically plausible, and consistent with how a list-price reduction moves gross-to-net through the working capital cycle. It also has to be right, because the balance sheet has $33,607M of net debt and the company repurchased no shares in 2025.

Assessment: We take management's timing explanation at face value for one quarter and no longer. The dividend costs roughly $5.1B a year against FY25 operating cash flow of $14,156M, so annual coverage is not in question. Quarterly coverage going negative is nonetheless the kind of fact that a 4.2% yield holder base notices, and the promised offset is now a testable commitment with a Q2 deadline.

6. The $410M IPRD impairment

GAAP research and development rose 17% to $2,649M while non-GAAP R&D was flat at $2,238M. The gap is a $410M in-process R&D impairment recorded in the quarter, disclosed only in the specified-items schedule. There was no accompanying disclosure of which programme was written off.

Assessment: A $410M write-off of acquired IPRD is a pipeline asset the company has concluded will not deliver. In a quarter whose entire narrative is pipeline breadth and R&D productivity, the absence of any identification of the impaired asset is a gap. It does not change the investment case on its own, but it is the sort of item that a company confident in its disclosure would name.

7. Milvexian: the year's largest binary

The Factor XIa inhibitor milvexian has two Phase 3 readouts guided to the end of 2026, in atrial fibrillation and in secondary stroke prevention. The AFib trial recruited 20,500 patients and is event-driven, testing non-inferiority against apixaban on efficacy with a subsequent superiority test on bleeding.

"In AFib, the study will test noninferiority versus apixaban, and then we will have a superiority testing for bleedings."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Management framed the strategic prize in terms of category leadership rather than in terms of a specific revenue number.

"So BMS will be the only Factor XI company with Factor XI inhibitor in AFIB, of course, with the presence also in SSP, this can allow us to continue to lead in the thrombotic space."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: The design is coherent and the disclosed non-inferiority margin (0.8 to 1.3) is conventional. The strategic logic is also sound: BMY owns the incumbent it is trying to displace, so a milvexian win defends the anticoagulation franchise through the Eliquis loss of exclusivity rather than merely adding to it. But this is a genuine coin-flip on a 20,500-patient event-driven trial in a class where a competitor programme has already failed, and it reads out after the point at which we would need to underwrite it. That asymmetry is the core of the Hold.

8. Cobenfy: four shots on goal in Alzheimer's psychosis

Management described a four-trial programme in Alzheimer's disease psychosis (ADEPT-1, ADEPT-2, ADEPT-4 and a newly disclosed ADEPT-5), of which two are guided to read out by the end of 2026. ADEPT-4 enrolls a biomarker-confirmed population while ADEPT-2, inherited with the Karuna acquisition, does not.

"As you know, ADEPT-2 was a study that was already ongoing in the moment we acquired Cobenfy. And we wanted to have a more predefined patient population to be recruited into an Alzheimer disease psychosis setting. This is why we took this approach of biomarker positivity."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

The commercial framing was sized on the call at roughly 7 million diagnosed Alzheimer's patients in the US, with 30% to 50% experiencing psychosis and no approved treatment for the indication.

Assessment: Running four trials on one hypothesis is expensive insurance, and management said openly that the best case requires at least two to hit. That is a candid statement of how uncertain this is. The commercial prize is real and the safety argument against existing antipsychotics in elderly dementia patients is strong. The catch is that a $56M quarterly product cannot support the valuation of an Alzheimer's franchise until the data land, and the same second-half window that carries milvexian carries this.

9. Admilparant: the readout with the most attractive setup

The LPA1 inhibitor admilparant has Phase 3 readouts in idiopathic pulmonary fibrosis guided to the end of 2026 and in progressive pulmonary fibrosis a few months later. Management's confidence rests on a Phase 2 dataset with a clear dose response.

"We have more than 60% improvement versus placebo in the lung function decline with 60-milligram BID."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

The commercial argument is a tolerability argument rather than an efficacy one, and it was quantified precisely.

"GI tolerability remains a significant barrier where approximately 50% to 60% of patients on treatment today are discontinuing therapy by 12 months with current standard of care. And even what we're seeing with the newest approved products, the diarrhea rate is roughly 40%."
— Adam Lenkowsky, Chief Commercialization Officer

Assessment: Of the four second-half readouts this is the one with the best risk-adjusted setup, because the Phase 2 effect size was large, the dose response was monotonic across two doses carried into Phase 3, and the Phase 2 and Phase 3 populations were deliberately matched. A market where half of patients abandon standard of care within a year is a market where a tolerable drug wins on tolerability alone. It is also the smallest of the four opportunities.

10. R&D productivity and the $2B cost programme

The CEO devoted a full section of prepared remarks to development-cycle productivity, with two specific numerical targets attached.

"We have set a target to reach lead molecule identification approximately 50% faster while applying greater rigor so that only the most differentiated molecules advance."
— Christopher Boerner, Board Chair and CEO

"Over time, we expect these efforts to deliver a 30% reduction in cycle times versus just a few years ago."
— Christopher Boerner, Board Chair and CEO

Separately, the $2B strategic productivity initiative remains on track for completion by the end of 2027. The visible evidence in this quarter is that non-GAAP R&D was flat at $2,238M while the company funded incremental investment behind pumitamig, which the CFO said was largely offset by productivity savings.

Assessment: Neither the 50% nor the 30% figure has a baseline, a definition or a date attached, which makes them unfalsifiable as stated. The verifiable version is the expense line, and there the story holds: flat R&D while absorbing a major new programme is a real result. We will grade the productivity claim on the OpEx line, not on the targets.

11. Business development posture

Asked whether the year-end readouts change the appetite for deals, the CEO gave the most direct answer of the call.

"BD continues to be a top capital allocation priority for us. It's not impacted by the end-of-year readouts. We have a very strong late-stage pipeline. We certainly don't need to chase deals."
— Christopher Boerner, Board Chair and CEO

"We're size agnostic as we've been, and we certainly have the financial horsepower to go after multiple sized deals."
— Christopher Boerner, Board Chair and CEO

He also signalled where the emphasis moves next: as the post-loss-of-exclusivity growth profile becomes clearer, greater weight goes to expanding the early and mid-stage portfolio to support growth into the 2030s.

Assessment: "Size agnostic" and "financial horsepower" are unusual phrases for a company carrying $33.6B of net debt with a quarter in which the dividend exceeded free cash flow. We read this as positioning rather than as a signal of an imminent large transaction, and we would treat a large deal announced before the second-half readouts as a negative signal about management's own confidence in those readouts rather than as a positive about the deal.

12. Non-obstructive HCM: a new trial disclosed on the call

Unprompted by the question that preceded it, the Chief Medical Officer announced a new Camzyos programme.

"So I want to announce that we are planning now to run a new more focused study in non-obstructive HCM."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: This follows a prior non-obstructive HCM programme that did not deliver, and management was candid that the expected benefit in non-obstructive disease is smaller than in obstructive because the patient population is more heterogeneous. A more focused re-attempt is a sensible use of a marketed asset's life-cycle budget. It is a 2029-and-beyond item and does not affect the current investment case.

Guidance & Outlook

The February figures were struck at mid-January exchange rates and the April figures at mid-April rates. Holding an unchanged dollar range across a period in which currency turned into a 2-point tailwind on reported revenue is the quiet part of this reaffirmation.

Implied quarter-over-quarter ramp: Q1 delivered $11,489M. FY2025 total revenue was $48,194M and Q1 2025 was $11,201M, so Q2 through Q4 2025 was $36,993M. Hitting the $46.0B low end requires $34,511M over the remaining three quarters, a 6.7% decline against that base; the $46.75B midpoint requires $35,261M, a 4.7% decline; and the $47.5B top end requires $36,011M, a 2.7% decline. Q1 grew 2.6%. The reaffirmed guide therefore implies a swing of at least 5 points into negative territory for the balance of the year, and that is at the optimistic end.

Implied EPS phasing: more forgiving than the revenue phasing. FY2025 non-GAAP EPS was $6.15 with $1.80 in Q1 and $1.26 in Q4. Hitting the $6.20 midpoint from Q1's $1.58 requires $4.62 across the remaining three quarters, or $1.54 per quarter, essentially flat against Q1. Against a business whose revenue is guided to fall, holding EPS flat sequentially requires the operating expense guide to be conservative or the gross margin to hold above the guided range.

Street at: consensus sat at roughly $47.12B on revenue and $6.26 on EPS at the time of the print, so both reaffirmed midpoints ($46.75B and $6.20) are below the Street. The "upper end" qualifier is what reconciles a below-consensus midpoint with a positive share-price reaction; taken literally it puts the company at $47.5B and $6.35, above consensus on both.

Guidance style: BMY has guided conservatively through the loss-of-exclusivity transition and has beaten the resulting bar consistently, which is what a $1.42 consensus against a $1.58 result reflects. This reaffirmation is consistent with that pattern: no numbers moved, a directional qualifier was added, and the company retained the option to raise in July with more of the Eliquis inventory reversal and the operating expense ramp visible. It is a well-managed guide. It is not a confident one.

Analyst Q&A Highlights

Eleven analysts asked questions across roughly forty minutes. The distribution is itself the story: nearly all of the questions were about clinical programmes and trial design, and not one was about revenue, margins, cash flow, the expired royalty stream or the dividend. The Chief Medical Officer answered more of the call than the CEO and CFO combined.

How the year-end readout calendar shapes the appetite for deals

The opening question paired a request to handicap the second-half readouts with a question about whether their timing constrains business development. The answer to the second half was unambiguous and separated the two decisions entirely: the deal pipeline is not being held hostage to the data calendar, the late-stage pipeline is strong enough that nothing has to be bought, and both early-stage and near-term-growth transactions remain live simultaneously.

Q: "And then related for Chris, how do the timing of these readouts impact the company's BD strategy as you think about the different outcomes that could unfold with the results of each of these readouts. And where do you see opportunity as you scan the landscape ahead of these readouts?"
— Asad Haider, Goldman Sachs

A: "BD continues to be a top capital allocation priority for us. It's not impacted by the end-of-year readouts. We have a very strong late-stage pipeline. We certainly don't need to chase deals."
— Christopher Boerner, Board Chair and CEO

Assessment: A clean answer that is also a slightly uncomfortable one. Decoupling deal appetite from a data calendar this consequential is defensible in principle, but a company that would transact identically whether or not its largest Phase 3 succeeds is either genuinely indifferent to the outcome or is not underwriting deals against it. Given $33.6B of net debt and a quarter of negative dividend coverage, we would prefer to hear that capacity is being preserved through the readouts.

Whether a partial win on the AFib trial still supports a commercial franchise

The most valuable question of the call asked what happens if the primary endpoint is not met cleanly, and whether a subgroup with a more favorable risk calculus could still carry a commercial opportunity. The response did not engage the partial-win scenario at all. It restated the pre-specified design, confirmed both trials remain blinded and on track, and asserted confidence in hitting both non-inferiority on efficacy and superiority on bleeding.

Q: "Can you help us think through how you'd approach a subgroup analysis, particularly for patients where the risk for or calculate might be more favorable for Factor XI. And to the extent that the top line doesn't meet that primary endpoint cleanly. Is there a path where the this patient population still supports a meaningful commercial opportunity?"
— Alexandria Hammond, Wolfe Research

A: "In AFib, the study will test noninferiority versus apixaban, and then we will have a superiority testing for bleedings."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: This is a non-answer, and it is the single most informative exchange on the call. There is no articulated fallback. The readout is being run as a binary, and investors should model it that way. The commercial framing offered in follow-up, that roughly 40% of patients who should be anticoagulated are untreated, underdosed or have discontinued because of bleeding concerns, describes the prize on a clean win and says nothing about the value of a miss.

What the disclosed trial parameters do and do not tell you about the AFib design

A follow-up pressed on the specifics of the design and dose selection, since the programme was inherited rather than designed by the current development leadership. The answer was the most substantive technical disclosure of the call: the 100mg twice-daily dose was selected from a multi-dose Phase 2 in total knee replacement, the trial recruited 20,500 patients, and the disclosed non-inferiority margin runs to 1.3.

Q: "So you really inherited design in kind of the Milvexian trials. Can you walk me through the aspects of the trial design, patient selection that increased your confidence in a potentially successful readout later this year?"
— Evan Seigerman, BMO Capital Markets

A: "Going specifically to your question, the design of the study is to be able to show a noninferiority apixaban in net to add comparison in a very well sized study, we recruited 20,500 patients."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: A 20,500-patient trial powered against a non-inferiority margin extending to 1.3 is conventionally designed and adequately sized. What the exchange revealed is that the confidence rests on a venous-thromboembolism surrogate population rather than on any atrial fibrillation data, which is exactly the extrapolation that has tripped up this mechanism before. The design is sound; the inference from Phase 2 to Phase 3 is the risk.

Whether falling background stroke rates could push the readout into 2027

An event-driven trial reads out when a pre-specified number of events accrues, and modern background stroke rates are materially lower than in the era when the incumbent was studied. The question asked directly whether that could slip the readout into 2027, and whether a slip would be good news. The answer confirmed the trial is event-driven, declined to disclose baseline characteristics or accrual rates, and reaffirmed the year-end timing with an explicit caveat.

Q: "If you take a look at the recent AFib results, you can see that the stroke rates are quite historically a lot lower than they were back in the original days of Eliquis. So just thinking about whether you've taken that into consideration and to what extent you think that impacts the study design and whether you think that there's any chance that the study results will ultimately end up in 2027, which could end up being more positive for you than in 2026."
— Michael Yee, UBS

A: "We are recruiting the events as expected. And the events predefined number is for efficacy and, of course, for safety. So we are on track by year-end."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: The caveat that followed ("we will see later in the year if this event rate will change or not") is the honest part, and it is a soft hedge on the single most-anchored date in the company's 2026 narrative. A slip is not a negative in itself, but a company whose second-half equity story rests on four readouts cannot afford one of them to move without the market re-rating the whole set.

What magnitude of bleeding reduction payers will actually reimburse for

The commercially sharpest question asked what bleeding delta the trial is powered for on superiority, or failing that, what delta would be clinically relevant enough to drive coverage. Management declined the first and answered the second with genuine specificity, which is the more useful of the two.

Q: "I appreciate the details on the Milvexian AFIB trial in terms of noninferiority on the efficacy endpoint, but just was wondering if you could elaborate in terms of what differential it's powered for on bleeds for superiority or if you don't want to answer that question, what you think is a clinically relevant delta versus Eliquis that would drive reimbursement coverage?"
— Terence Flynn, Morgan Stanley

A: "What they're looking for is clinically meaningful and statistically credible reduction in major bleeds that translates into fewer hospitalizations and fewer events that are clinically and economically important."
— Adam Lenkowsky, Chief Commercialization Officer

Assessment: The most useful commercial disclosure of the call. Payers are not anchored to a numeric threshold, which means a statistically significant reduction in major bleeding of almost any credible magnitude opens the formulary conversation. That materially widens the definition of a commercial win relative to the binary framing of the endpoint, and it is the strongest argument available for the bull case on this asset.

Why ADEPT-4 added a biomarker requirement that ADEPT-2 does not carry

A pointed question asked whether the biomarker-confirmation criterion added to the newer Alzheimer's psychosis trial was a response to problems observed in the inherited trial's conduct. The answer separated the two cleanly: the older trial was already enrolling when the asset was acquired, and the biomarker requirement is about population homogeneity rather than about remediation, at the cost of a higher screen-failure rate.

Q: "Can you talk about why you added that criteria for the study? And was it based on any issues you saw with the ADEPT II trial conduct?"
— Akash Tewari, Jefferies

A: "As you know, ADEPT-2 was a study that was already ongoing in the moment we acquired Cobenfy. And we wanted to have a more predefined patient population to be recruited into an Alzheimer disease psychosis setting. This is why we took this approach of biomarker positivity."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: A credible answer that also concedes the underlying concern. Running the inherited trial without biomarker confirmation while designing its successor with it is an implicit statement about which of the two is more likely to produce a clean result, and management said openly that approval likely requires two positive studies. The programme's four trials are insurance against exactly this heterogeneity problem, which is a reasonable design and an expensive admission.

Whether PFS gains from VEGF bispecifics translate to survival

The closing question raised the debate now dominating immuno-oncology positioning: whether the progression-free survival advantage seen with VEGF and PD-1 combinations degrades when carried to overall survival, particularly against an active immuno-oncology comparator rather than chemotherapy. The answer defended the bispecific format on mechanistic grounds and set a deliberately modest bar for what counts as success.

Q: "So your competitor has signaled that VEGF PD-1 compared to an IO or PD-1, may be able to show a minimal regression in hazard ratio when you go from PFS to OS."
— Mohit Bansal, Wells Fargo

A: "I truly believe bispecifics are a better way to deliver 2 different mechanisms than using 2 different antibodies because you are much more on target, you are much more selective in delivering them and potentially, you can decrease also the off-target issues like adverse events."
— Cristian Massacesi, Chief Medical Officer and Head of Global Drug Development

Assessment: The base case articulated in the fuller answer, that a good progression-free survival gain with a statistically significant survival gain and no dramatic safety cost is "good enough" to establish a backbone, is a lower bar than the market has been pricing for this class. Setting expectations down ahead of a data presentation is usually the right read on how management sees its own dataset. The strategic logic, that a combinable backbone is worth more than a single winning regimen, is sound and is the same logic that made the incumbent PD-1 valuable.

What They're NOT Saying

1. The identity of the $410M IPRD impairment: the write-off appears only in the specified-items schedule with no programme named, in a quarter whose prepared remarks were substantially about pipeline breadth and R&D productivity. A company confident in the disposal would name it.
2. Any characterization of the $323M below-the-line swing: the expired diabetes royalty stream and lower licensing income drove the entire non-GAAP EPS decline. It is named in one clause of the release and was not mentioned once across prepared remarks or eleven analyst questions.
3. The size of the Eliquis inventory build: management confirmed a build exists and confirmed it reverses in Q2, but quantified neither. Given Eliquis grew 16% against a 10% to 15% full-year guide, the unquantified portion is the difference between a beat and an in-line quarter.
4. Whether the Opdivo wholesaler drawdown will normalize: the language used was "we continue to monitor whether these levels will normalize over the balance of the year," which is a materially weaker formulation than the Eliquis reversal was given. The company is not committing to a recovery in its second-largest product.
5. The bleeding-superiority powering assumption for milvexian: asked directly what magnitude of bleeding reduction the trial is powered to detect, management redirected to the payer conversation. The efficacy non-inferiority margin was disclosed; the superiority delta was not.
6. Anything about the dividend against the quarter's cash flow: free cash flow of $757M against $1,283M of dividends paid went unremarked, and no analyst raised it. The rebate-timing explanation was offered for operating cash flow generally, not for coverage specifically.
7. The operating expense ramp implied by the guide: Q1 non-GAAP operating expenses of $3,854M against a ~$16.3B full-year guide require a 7.7% step-up in the average of the remaining quarters. Management characterized Q1 spending as "slightly above" last year and did not address the implied acceleration.
8. Eliquis loss of exclusivity: the single largest product in the company, at 36% of quarterly revenue, faces US generic entry in 2028 under previously executed patent settlements. It was discussed only as a growth driver and as the milvexian comparator, never as an expiry.

Market Reaction

• Pre-print setup: BMY closed at $57.59 on Wednesday April 29, up 6.8% year to date against the S&P 500's 4.2%, up 14.7% over the trailing twelve months, but down 5.0% over the trailing thirty days. The 52-week closing range entering the print was $42.60 to $62.37, so the stock was roughly 8% below its 52-week closing high and had given back a meaningful part of a strong twelve months in the month before the print.
• Reaction session: the print was released before the open on Thursday April 30 with the call at 8:00 a.m. ET, so the print day is the reaction session. The stock opened at $57.69, a 0.2% gap, traded a range of $57.39 to $60.78, and closed at $60.59, up 5.2% or $3.00.
• Volume: 24.9M shares against an 11.7M thirty-day average, or 2.1 times normal.
• Index context: the S&P 500 rose 1.0% on the same session, leaving roughly 4.2 points of the move idiosyncratic.

The shape of the move matters more than its size. The stock gapped up only 0.2% on the press release and traded to a session low of $57.39, below the prior close, before building through the day to close near the high of the range. That is not a reaction to the headline numbers, which were available before the open. It is a reaction that accumulated across and after the 8:00 a.m. call, which points to the "trending toward the upper end" language and the pipeline discussion rather than to the beat itself as the operative catalyst.

The setup supports that reading. A stock down 5.0% over the preceding thirty days into a print, with a full-year guide the Street had modeled slightly above the company's midpoint, is positioned for disappointment rather than for a beat. The combination of a 5% revenue beat, an $0.16 EPS beat, no cut to any guidance line, and a verbal upgrade to the positioning within the ranges removed the near-term downside case without requiring the market to underwrite anything new. On 2.1 times normal volume, that is a positioning unwind more than a re-rating.

What the move did not price is anything about the second half. The stock closed at 9.8 times the full-year non-GAAP EPS midpoint. A market genuinely re-rating for four second-half Phase 3 readouts, one of which is a category-defining anticoagulant, would not leave the shares at a single-digit multiple. The reaction is best read as relief that the transition is on schedule, not as an advance on the catalysts.

Street Perspective

Debate: Is the Growth Portfolio compounding fast enough to outrun the cliff?

Bull view: The bull case on the Street is that the sustained majority share is the inflection point, and that with the Growth Portfolio adding $664M year over year against $361M of Legacy erosion, the coverage ratio is already comfortably above 1.0 and improving as the newest launches scale. Revlimid is now small enough that its remaining decline is manageable in absolute dollars.

Bear view: The bear camp contends that the coverage ratio is flattered by two things that reverse. Eliquis, which sits in the Legacy Portfolio and grew 16%, is masking a 45% decline in the rest of that book, and Eliquis itself has a settled US generic entry date of 2028. Strip it out and the erosion is far steeper than the reported 6.4%. The company's own guide, which implies Q2 through Q4 revenue falls between 2.7% and 6.7%, concedes the point.

Our take: The bears have the arithmetic and the bulls have the direction. Both are right about different time horizons. Through 2027 the guide's own math says revenue declines, and no amount of Growth Portfolio momentum changes that. Beyond 2028 the question is whether the Growth Portfolio plus whatever survives the second-half readouts is large enough to absorb the Eliquis generic entry as well, and that is genuinely unknowable today. That unknowability is the reason a company with this pipeline trades at 9.8 times.

Debate: How much should milvexian be worth in the stock today?

Bull view: Some on the Street argue the market is assigning close to zero to a Factor XIa inhibitor that would be the only one in atrial fibrillation, in a market where roughly 40% of patients who should be anticoagulated are untreated, underdosed or have discontinued because of bleeding risk. The payer feedback disclosed on the call, that no numeric bleeding threshold is required, widens the definition of a win considerably.

Bear view: The bear framing is that this mechanism has already disappointed at Phase 3 elsewhere, that the dose was selected from a knee-replacement surrogate population rather than from atrial fibrillation data, and that management's refusal to describe any partial-win scenario tells you they see it as binary too. A trial that reads out at the very end of 2026 offers no way to reduce the position ahead of the event.

Our take: The bulls are right that the option is cheap and the bears are right that it is an option. The payer disclosure is the most important new fact from this call on this asset and it does widen the win condition. But an event-driven readout whose timing management would only defend with "we will see later in the year" is not something to underwrite in April, and the stock offers no mechanism to participate in the upside without also owning the downside. We would rather pay up after the data.

Debate: Is a 9.8x multiple with a 4.2% yield cheap, or correctly priced for a shrinking business?

Bull view: The value case circulating is straightforward: single-digit earnings multiple, dividend yield above 4%, a $2B cost programme still running through 2027, and four independent shots at a pipeline re-rating within nine months. On any one of those readouts working, the multiple has substantial room, and the downside is protected by the yield.

Bear view: The skeptics answer that a single-digit multiple on a business the company itself guides to shrink is not cheap, it is arithmetic. The below-the-line support that flattered prior-year earnings is gone, gross margin is structurally compressing with the mix shift, net debt is $33.6B, and the quarter's free cash flow did not cover the dividend that is supposed to be the downside protection.

Our take: The multiple is genuinely low and genuinely deserved, which is an uncomfortable but accurate place to land. Our fair-value framework applies 9 to 11 times to the $6.20 full-year non-GAAP EPS midpoint, giving $55.80 to $68.20 with a $62.00 midpoint, about 2% above the reaction close. Add the 4.2% yield and the expected twelve-month total return is roughly 6.5%, which is a market return, not an excess one. That is the definition of a Hold. The bear argument on cash conversion is the one we would most want to see disproved, and Q2 will do that or not.

Model Framework at Initiation

We are initiating coverage without a prior model, so the table below sets out our working assumptions rather than changes to an existing one.

Valuation: at the $60.59 reaction close, with 2,042.1M shares outstanding, the market capitalization is $123.7B and enterprise value is $157.3B including $33.6B of net debt. That is 9.8 times the full-year non-GAAP EPS midpoint of $6.20, 9.5 times the $6.35 upper end, and 3.4 times enterprise value to the guided revenue midpoint. The annualized dividend of $2.52 yields 4.2%.

Fair value: applying 9 to 11 times to $6.20 gives a range of $55.80 to $68.20 with a $62.00 midpoint. Against the $60.59 close that is roughly 2% of price appreciation plus a 4.2% yield, or about 6.5% of expected twelve-month total return. The multiple range is deliberately wide because the second-half readouts are capable of resolving it in either direction well outside these bounds.

Thesis Scorecard Post-Earnings

This is first coverage, so the pillars below are established here rather than graded against a prior quarter. Each carries the status it holds after this print, and each becomes the standing scorecard against which subsequent quarters are graded.

Overall: The operating thesis and the financial thesis point in opposite directions, which is the whole of the BMY problem. The portfolio transition is demonstrably working and reached a genuine milestone this quarter. The financial statements show a business whose revenue is guided to shrink, whose gross margin is structurally compressing, whose below-the-line earnings support has permanently gone, and whose quarterly free cash flow did not cover its dividend. The pipeline is the reconciling asset and it is entirely unpriced, because it is entirely unresolved.

Action: Hold. The valuation already reflects the loss-of-exclusivity arithmetic and the 4.2% yield pays to wait, but the second-half readouts are unhandicappable from here and the reaction has already taken the shares up 5.2%. We would move to Outperform on any of: milvexian AFib clearing non-inferiority with statistically significant bleeding superiority; two of the four Cobenfy Alzheimer's psychosis studies reading out positive; or Q2 demonstrating that the Eliquis rebate offset restores cash conversion while revenue lands at the upper end. We would move to Underperform if the guide's upper-end language is withdrawn in July, or if the operating expense ramp implied by the full-year guide arrives without the corresponding revenue.