Amgen's Beat-and-Raise Gets Sold: Low-Quality Composition and Two Fresh Overhangs Outweigh a Cheaper Stock

Results vs. Consensus

Year-over-year comparison (Q1'26 vs. Q1'25)

Quality of Beat/Miss

• Revenue: Medium quality. The 6% total-revenue growth flatters the 4% product-sales growth because other revenue jumped 45% to roughly $400M. Product volume growth of 9% is healthy and demonstrates real demand, but 2 points of price erosion and 2 points of inventory drag show the pricing environment and the LOE cliff are live.
• Margins: Stable and reassuring. Non-GAAP operating margin held at 45.3% (down just 0.4pp) even as R&D grew 16%, and management reaffirmed the ~45%–46% full-year margin. After Q4's 3.5pp compression, margin stabilization is a genuine positive and validates the "heaviest R&D step-up is now in the base" read.
• EPS: Lower quality. The non-GAAP tax rate of 13.6% (down a full point YoY) did real work in the $5.15 print, and the 2026 guide lowered the tax-rate assumption further (to 15.0%–16.5% from 16.0%–17.5%). Some of the EPS raise is therefore tax, not operations. GAAP EPS of $3.34 is cleaner and grew a more modest 4%.

Segment Performance

The story is consistent with our initiation: a fast-growing portfolio out-running an accelerating loss-of-exclusivity cliff. The six growth drivers hit ~70% of product sales and grew 24% in aggregate; the drag is concentrated in denosumab.

General Medicine — Repatha extends the inflection, denosumab accelerates down

Repatha delivered $876M (+34%), with US new-to-brand prescriptions up 44% in the quarter as the VESALIUS-CV data pull cardiology depth and primary-care breadth. Evenity grew 27% to $562M and holds 65% US bone-builder share. The offset is stark: Prolia fell 34% to $727M on 17% lower volume, and Prolia and Xgeva combined fell 32% to roughly $1.1B, with management reiterating that erosion accelerates through 2026.

"Repatha is now the only PCSK9 inhibitor with positive outcomes data in both high-risk primary and secondary prevention patients." — Murdo Gordon, EVP, Global Commercial Operations

Assessment: The Repatha/Evenity engine is doing precisely what the thesis requires, and the +44% new-to-brand figure is the clearest evidence yet that VESALIUS-CV converts to volume. But the denosumab cliff is now steep and visible in the numbers; the segment's net growth depends entirely on the two 30%-ish growers staying ahead of a franchise that is shrinking a third year over year.

Rare Disease — Uplizna is the standout, Tavneos turns into a risk

Rare disease grew 25% to $1.2B. Uplizna surged 188% to $262M on the gMG launch (bio-naive and switch), an EU gMG approval, and continued NMOSD leadership. Tepezza rose 29% to $490M and reported positive Phase 3 subcutaneous on-body-injector data that supports a more convenient route. Tavneos grew 32% to $119M but is now an overhang rather than a growth story (below).

"Given this momentum, we see a meaningful opportunity for Uplizna to become the first-line and first-switch choice for appropriate patients living with gMG." — Murdo Gordon, EVP, Global Commercial Operations

Assessment: Uplizna remains the cleanest growth asset in the company and the clearest justification for the Horizon deal, and its multi-indication CD19 runway (autoimmune hepatitis and CIDP pivotals starting by 2H26) extends the story. Tepezza's subcutaneous readout is an underappreciated convenience catalyst. The blemish is Tavneos.

Inflammation — Tezspire steady through an inventory burn

Tezspire grew 20% to $343M despite a channel-inventory headwind, with the new chronic-rhinosinusitis-with-nasal-polyps indication broadening its reach. The legacy inflammation assets (Enbrel, Otezla) continue their managed decline.

Assessment: Underlying Tezspire demand is stronger than the 20% suggests given the inventory burn, and the respiratory label-expansion pipeline (COPD, EoE) keeps it a durable multi-year grower. Inflammation growth is now essentially a Tezspire story with legacy erosion underneath.

Oncology — the BiTE platform keeps compounding

Innovative oncology grew 25% to $1.8B. Imdelltra reached $258M with more than 1,800 US administering sites and encouraging first-line maintenance data (median overall survival of 25.3 months in the Phase 1b DeLLphi-303). Blincyto grew 12% to $415M. The STEAP1 asset xaluritamig is advancing in two Phase 3 prostate studies.

"Imdelltra is emerging as a standard of care in second-line extensive stage small cell lung cancer, delivering an unprecedented survival benefit." — James Bradner, EVP, R&D

Assessment: The bispecific T-cell-engager platform is the most credible home-grown oncology franchise Amgen has, and Imdelltra's move toward first-line plus the FDA real-time-trial pilot are genuine positives. Note the subcutaneous-Blincyto enrollment pause over inflammatory reactions, a reminder that BiTE safety management is ongoing.

Biosimilars — durable, cash-generative, ignored

Biosimilars grew 14% to $835M, with Pavblu at $280M and cumulative biosimilar sales now above $14B since 2018.

Assessment: A quietly compounding, cash-generative business that funds the pipeline and diversifies away from the LOE cliffs. It will not re-rate the stock, but it is a real earnings ballast.

Grading Last Quarter's Commitments

Our Q4 initiation flagged six commitments to grade this quarter. The scorecard:

Net: four of six commitments met or progressing, one confirmed as an expected headwind, one (Tavneos) deteriorated. Management largely did what it said it would do, which is the continuity that matters, but the Tavneos escalation and a brand-new IRS overhang are the wildcards that were not on the Q4 card.

Key Topics & Management Commentary

Overall Management Tone: Management struck the same confident, on-message posture as last quarter, leaning into "springboard year" language and treating the beat-and-raise as validation. The tone was noticeably most energized on MariTide's dosing/tolerability story and Repatha's guideline momentum, and most careful on Tavneos and the newly disclosed IRS matter, where the language was firm but defensive. Compared with Q4's Tavneos "we were surprised," the Q1 framing was more resolved to fight, but the underlying issue had worsened.

1. MariTide — the SWITCH study reframes the commercial story

The headline pipeline news was the expansion of the MariTide Phase 3 program. Amgen initiated MARITIME-SWITCH, a 300-patient study that runs patients in on weekly semaglutide or tirzepatide and then switches them to MariTide on an every-8-week or every-12-week schedule, with a 52-week body-weight primary endpoint. It also added two long-term maintenance extension studies (monthly, Q8W, quarterly) and reaffirmed three Type 2 diabetes Phase 3 starts in 2026.

"We'll evaluate switching from medicines which are injected 52 times a year to one which can be injected as few as 4 or 6 times a year. We expect there'll be a great deal of interest in these data." — Robert Bradway, Chairman & CEO

Assessment: The SWITCH study is a shrewd commercial framing: it explicitly addresses the "I already take a weekly, why change?" objection that a late entrant faces, and it turns dosing frequency from a feature into a switching pitch. But note what it is not: it is not efficacy data. The design (every 8 or 12 weeks, not monthly) reads as a bet on convenience, and the market will not pay for the obesity franchise until pivotal weight-loss numbers arrive.

2. MariTide tolerability — three-step dose escalation

The most substantive de-risking of the quarter was on tolerability. Management reported that three-step dose escalation further reduced nausea and vomiting versus two-step, and argued that MariTide's antibody backbone produces smooth, stable drug exposure that avoids the peak-trough spikes of weekly injectables and oral GLP-1s.

"The level of nausea and vomiting observed with 3-step dose escalation is lower than we've seen before... we await efficacy and tolerability data from the ongoing Phase III studies, but we're quite encouraged by what we've seen." — James Bradner, EVP, R&D

Assessment: This directly addresses the single biggest bear point on AMG 133, its historical GI tolerability at higher doses. If the three-step-titration tolerability carries into Phase 3, it materially improves the risk profile of the whole program. It is the most important incremental positive in the quarter, and it is why the thesis is intact despite the stock reaction.

3. Repatha — VESALIUS-CV keeps compounding the case

A new prespecified VESALIUS-CV subgroup (3,655 high-risk diabetes patients without known significant atherosclerosis) presented at ACC and published in JAMA showed a 31% reduction in three- and four-point MACE, plus nominal reductions in cardiovascular and all-cause death. Updated ACC/AHA guidelines now push earlier risk identification and lower LDL-C targets, and management notes those guidelines do not yet reflect VESALIUS-CV.

"If you look at new-to-brand prescription evolution in the U.S., we were up 44% in the quarter... driven by increased depth of prescribing by cardiologists and increased breadth of prescribing by primary care physicians." — Murdo Gordon, EVP, Global Commercial Operations

Assessment: The primary-prevention expansion is real and accelerating, and there is still guideline upside once VESALIUS-CV is formally incorporated. This is the most reliable growth pillar in the company and it strengthened this quarter.

4. The guidance raise — modest and partly tax-driven

Amgen raised 2026 revenue to $37.1B–$38.5B and non-GAAP EPS to $21.70–$23.10, and lowered its non-GAAP tax-rate assumption to 15.0%–16.5%. The operating-margin and capex guides were unchanged.

"Our strong first quarter performance reinforces that outlook, and we're raising our 2026 guidance ranges for both revenue and non-GAAP earnings per share." — Peter Griffith, EVP & CFO

Assessment: A raise is a raise, and it signals confidence. But the magnitude (~$0.1B revenue, $0.10 EPS) barely exceeds the size of the Q1 beat, and the lower tax-rate assumption means the operating raise is smaller than the headline. Investors who wanted evidence of an accelerating trajectory did not get it, which is part of why the stock fell.

5. Tavneos — the FDA dispute worsened

The Tavneos situation deteriorated from Q4. The FDA has now formally proposed to withdraw the approval, and Amgen submitted a label change (CBE-30) amending the hepatotoxicity warning to add vanishing bile duct syndrome cases, including fatal outcomes, and modifying liver-testing and discontinuation language.

"As publicly disclosed, the FDA proposed to withdraw the approval of Tavneos. We continue to believe that Tavneos is an important medicine... We are confident in the benefit-risk profile of this medicine and expect to engage further with the FDA." — James Bradner, EVP, R&D

Assessment: Tavneos is small (roughly $460M annualized), so the direct revenue at risk is modest. But the escalation to a formal withdrawal proposal plus a hepatotoxicity/fatal-VBDS label change turns this from a procedural surprise into a genuine safety-and-regulatory overhang with real headline risk. This is a bear point that moved the wrong way.

6. IRS draft NOPA — a new, potentially material tax overhang

The CFO disclosed that in April 2026 Amgen received a draft Notice of Proposed Adjustment from the IRS for tax years 2016–2018, primarily on the allocation of profits between the US and Puerto Rico, similar in nature to the pending 2010–2015 Tax Court dispute.

"If sustained in full, the adjustments set forth in the draft NOPA could have a material impact on our financial statements... We firmly believe that the IRS positions are without merit, and we also believe that our tax reserves are appropriate." — Peter Griffith, EVP & CFO

Assessment: This is the most consequential new disclosure of the quarter for the risk profile. The 2010–2015 dispute already carried multi-billion-dollar tail risk; extending the same Puerto Rico theory to 2016–2018 enlarges the potential exposure. Management disputes it and a Tax Court decision on the earlier years is not expected before 2H26, but the overhang is now larger and the resolution is not near.

7. Olpasiran and the Lp(a) opportunity

Olpasiran continues in the ~7,300-patient OCEAN(a) outcomes study (95% Lp(a) reduction on quarterly dosing), and Amgen initiated OCEAN(a)-CCTA to measure noncalcified plaque by CT angiography. Updated ACC/AHA guidelines now recommend broader Lp(a) testing.

Assessment: The guideline push toward Lp(a) testing is a quiet tailwind that seeds the future market, and management sounded more confident on trial design than in Q4 (where accrual pace was the concern). Still an event-driven, later-dated catalyst, but the setup improved at the margin.

8. Oncology real-world trial pilot and BiTE safety

Amgen is collaborating with the FDA on a real-time, real-world clinical-trial pilot for Imdelltra that leverages electronic health records and real-time data capture, potentially compressing development timelines. Separately, it paused enrollment in some subcutaneous-Blincyto studies over inflammatory reactions while in dialogue with the FDA.

Assessment: The real-world-trial pilot is a genuinely novel regulatory experiment that, if it works, is a durable structural advantage for Amgen's oncology development. The Blincyto pause is a routine safety-management item, not a red flag, but worth monitoring.

9. AI/data and the R&D leadership change

Management devoted unusual airtime to AI, citing a 50% acceleration in antibody lead optimization, a proprietary trial-site-selection model improving enrollment up to threefold, and AI automation cutting a manufacturing line-clearance step from ~30 minutes to ~2 minutes. It also announced that David Reese, the first CTO and former R&D head, will retire at the end of Q2, with Jay Bradner adding enterprise AI/data leadership to his R&D role.

Assessment: The AI narrative is early and hard to underwrite in the model, but the specific, quantified productivity examples are more credible than generic "we use AI" claims. The leadership transition concentrates more responsibility on Bradner; execution risk is modest but worth noting.

10. Capital allocation and the balance sheet

Free cash flow rose to $1.5B (from $1.0B), Amgen retired debt via open-market repurchases (a ~$90M gain in OI&E), and cash was $12.0B against $57.3B of debt. The dividend rose 6% to $2.52, and there were no buybacks in the quarter.

Assessment: Deleveraging continues but the debt load ($57.3B) is still the elephant in the room, and the absence of buybacks plus a rising capex bill confirm that capital priorities remain pipeline and balance sheet first. The new IRS exposure makes preserving balance-sheet flexibility more important, not less.

11. BD and M&A appetite

Asked about the active sector M&A backdrop and Amgen's "needs," the CEO pushed back on the framing but confirmed the company remains active in business development in its areas of strategic interest.

"Not sure I'd use the word needs the way you have in your question. But we're very active in business development, as we always have been, looking for innovation that we think we can add value to." — Robert Bradway, Chairman & CEO

Assessment: Guidance explicitly excludes BD, so any deal is incremental to the model, and the CEO's tone suggests bolt-ons rather than a transformational deal. Given the leverage and the fresh tax overhang, a large deal would likely be received cautiously.

Guidance & Outlook

The raise reinforces the "springboard year" narrative, but its composition is the issue: revenue midpoint moved only from $37.7B to $37.8B, the EPS midpoint from $22.30 to $22.40, and the tax-rate cut does part of that lifting. Management held the operating-margin and capex guides, and pointed to Q2 operating margin in line with Q1.

Implied ramp: The revenue midpoint still implies only ~3% full-year growth over 2025, back-half-weighted as denosumab laps and the growth brands compound. The guide's credibility improved (management raised after one quarter), but it did not signal acceleration.

Street at: Consensus was already near the prior midpoint; the small raise nudges numbers marginally higher. The more notable Street move was estimate revisions lower for the upcoming period from some desks, reflecting the low-quality composition.

Guidance style: Classic Amgen: raise modestly and early, keep operating assumptions conservative, exclude BD. The tax-rate reduction is the one place the guide got a non-operational tailwind.

Analyst Q&A Highlights

MariTide SWITCH study design

The opening question sought specifics on the newly announced switch study: the dosing arms, and whether it targets superiority or non-inferiority. Management laid out the design without committing to a statistical framing beyond the endpoint.

Q: "First question on the MariTide SWITCH studies. Are you switching sort of 1:1:1 to monthly, every 2 months and every 3 months? And are you looking at superiority or non-inferiority? And what's the non-inferiority margin?"
— Yaron Werber, TD Cowen

A: "There will be 300 subjects on study with obesity or overweight. There will be a run-in on weekly semaglutide or tirzepatide, and then they'll switch to MariTide... every 8-week or a quarterly basis. And the primary endpoint of this trial will be a change from baseline body weight after 52 weeks of MariTide treatment."
— James Bradner, EVP, R&D

Assessment: The design confirms the commercial intent (prove weekly-to-MariTide switching works at convenient intervals) but management sidestepped the superiority/non-inferiority question, which matters for how strong a switching claim the eventual label supports. A change-from-baseline endpoint at 52 weeks is a reasonable bar; the read-through depends on the comparator context.

Why the switch study skips monthly dosing

A follow-up probed why the switch study evaluates only every-8-week and every-12-week dosing rather than monthly, and how large the maintenance opportunity really is. Management explained the design logic and framed the maintenance/switch commercial rationale.

Q: "Could you comment on the MariTide switching study and why it only evaluates every 2 months and 3 months or not every 1 month? And how significant do you expect the maintenance opportunity to be?"
— Salveen Richter, Goldman Sachs

A: "We have a lot of experience with monthly MariTide in this program... the long-term extensions... give us a chance to explore less frequent dosing after effective dosing. And comparably, in the SWITCH study, we're focusing that trial on the learnings of going from weekly to an every 8-week and every 12-week treatment regimen, which can make MariTide quite attractive to patients if successful."
— James Bradner, EVP, R&D

Assessment: The answer is internally consistent, monthly is already covered in the core pivotal program, so the incremental studies logically test the less-frequent tail. It also reveals the commercial thesis clearly: Amgen expects to enter a market where most patients are already on weeklies and needs data to justify a switch. Sensible, but again convenience, not efficacy.

Three-step titration tolerability specifics

A question pushed for quantitative detail on the vomiting rates and duration seen with three-step titration versus incumbents. Management declined specific numbers but was directionally emphatic and drew a contrast with weekly and oral GLP-1s.

Q: "Are you able to provide any more color on what levels of vomiting and duration of vomiting you're seeing with the 3-step titration... or directionally where that's shaking out versus Wegovy or Zepbound?"
— Christopher Schott, JPMorgan

A: "The level of nausea and vomiting observed with 3-step dose escalation is lower than we've seen before... When we do observe nausea and vomiting, it tends to be quite short in its duration, over the course of one or several days, no different than the weekly GLP-1s. But different than the weekly GLP-1s... this trough to peak spike... can be avoided with a steady, stable, long-acting medicine like MariTide."
— James Bradner, EVP, R&D

Assessment: The refusal to quantify is a modest disappointment, but the mechanistic argument (stable exposure avoids peak-trough-driven GI intolerability) is coherent and, if validated in Phase 3, is the differentiator. This exchange is the crux of the bull case on tolerability.

Olpasiran trial design and background therapy

A question asked whether background GLP-1 or PCSK9 therapy could confound the olpasiran outcomes trial or a competitor's. Management argued the Lp(a)-specific mechanism insulates the design.

Q: "I'm wondering if you think that background therapies such as GLP-1 or PCSK9 either would impact your trial design or your competitor trial design and how you think about that impacting the overall results."
— Michael Yee, UBS

A: "Unfortunately, to your question, you can't take a GLP-1 medicine or a statin or even Repatha and meaningfully reduce levels of Lp(a). This independent risk factor maps to a very atherogenic and inflammatory characteristic... We're very confident in the study as defined, which focuses on a high-risk, high-leverage elevated Lp(a) population."
— James Bradner, EVP, R&D

Assessment: A strong, mechanistically-grounded answer. Because no approved therapy meaningfully lowers Lp(a), background therapy is less of a confounder for olpasiran than for LDL-focused trials. It reinforces that olpasiran is a differentiated shot, though still event-driven and later-dated.

Repatha primary-prevention runway and Lp(a) crossover

A question asked how far Repatha's primary-prevention penetration can run and how the primary-prevention opportunity looks in the Lp(a) segment. Management quantified the new-to-brand momentum and connected it to the guideline and testing tailwinds.

Q: "Repatha has been consistently strong, but I want to get your perspectives on penetration into primary prevention and where it could go. And as you look to the olpasiran data, how do you think primary prevention looks as a key market within the Lp(a) segment?"
— Geoffrey Meacham, Citi

A: "In the quarter, we had strong overall growth in Repatha globally. But if you look at new-to-brand prescription evolution in the U.S., we were up 44%... we still have a huge opportunity ahead of us in primary prevention promotion of Repatha, the only PCSK9 with that data generation now."
— Murdo Gordon, EVP, Global Commercial Operations

Assessment: The 44% new-to-brand figure is the single most encouraging commercial data point in the quarter, and the primary-prevention runway is genuinely large (roughly 100 million patients not at LDL goal). This is the pillar with the clearest visibility.

Imdelltra's frontline expansion and the FDA real-time-trial pilot

A question asked what comes next for Imdelltra as it moves toward earlier lines, and how the FDA real-time clinical-trial pilot would actually work. Management framed both the frontline opportunity and the novel regulatory experiment.

Q: "Can you tell us more about what's next for Imdelltra, not only in terms of the opportunity once you move to the earlier lines, but also now with Imdelltra selected as part of the real-time clinical trial pilot program? Could you help us understand the process?"
— Luca Issi, RBC Capital Markets

A: "Medicines that work in later stages of the disease tend to confer even more clinical benefit when they've moved to earlier lines of therapy... we are advancing Imdelltra quite actively and aggressively into frontline induction as well as frontline induction and maintenance... we have a very fine design, coming together with the FDA, that will give us a chance to characterize Imdelltra in a clinical trial setting but in the real world, leveraging things like electronic health records and real-time data capture."
— James Bradner, EVP, R&D

Assessment: The frontline push (DeLLphi-305 maintenance, DeLLphi-306 limited-stage) is the durable oncology growth story, and the 25.3-month first-line-maintenance overall-survival signal is encouraging. The real-time-trial pilot is a genuinely novel structural bet: if EHR-based real-world trial capture works, it compresses development timelines and is a durable advantage. Both are positives, though neither is a near-term needle-mover at the corporate multiple.

AmgenNow cash-pay uptake

A question asked for quantification of the direct-to-patient AmgenNow channel's uptake for Repatha and whether the model extends internationally. Management sized the program but downplayed its mix.

Q: "Could you provide some quantification on the uptake of Repatha by cash-pay patients, and possibly the current mix of sales between cash pay and covered given the strong ramp of AmgenNow?"
— David Risinger, Leerink Partners

A: "Repatha is offered at a $239 a month price point... We are in the kind of the 8,000 to 9,000 patient range of patients moving through the AmgenNow program... as a percentage of total Repatha, as you'll note, it's relatively small."
— Murdo Gordon, EVP, Global Commercial Operations

Assessment: AmgenNow is a useful access and brand-defense tool but not yet a material revenue channel (8,000–9,000 patients). Its bigger significance is strategic optionality (direct-to-patient, Trump Rx) as the drug-pricing landscape shifts; the near-term P&L impact is small.

What They're NOT Saying

1. Still no MariTide pivotal efficacy: Every MariTide disclosure this quarter was about dosing schedule and tolerability. The weight-loss efficacy that determines the franchise's value remains unread out, and management would not quantify even the tolerability improvement.
2. Superiority vs. non-inferiority for SWITCH: Management declined to state whether the switch study targets superiority or a non-inferiority margin, which shapes how strong a switching label it can ultimately support.
3. The size of the IRS exposure: Management said the 2016–2018 NOPA could be "material" but would not quantify potential exposure or reserve adequacy beyond asserting the reserves are appropriate.
4. Tavneos worst case: Management reaffirmed conviction but did not address what a forced withdrawal would mean operationally or for the ChemoCentryx-related intangibles.
5. Denosumab trajectory specifics: "Accelerated erosion" was reiterated without a quantified 2026 exit rate for Prolia/Xgeva, the largest moving part inside the guide.
6. Merck oral PCSK9 impact: Management defended Repatha's data advantage but did not address the convenience threat from an oral competitor now nearing market.

Market Reaction

• Pre-print setup: AMGN closed at $346.25 on 2026-04-30 entering the print, up ~5.8% YTD, down ~1.6% over the trailing 30 days, and up ~19.0% over the trailing 12 months. The stock sat well within its 52-week closing range of $262.28–$388.16, with positioning full after the run.
• After-hours move: Despite the beat-and-raise, the stock fell roughly 2.4% in after-hours trading on the evening of the print, an early signal that the market read the composition as low-quality.
• Reaction-day session: On 2026-05-01 the stock opened at $332.52 (a 4.0% gap down), traded between $324.63 and $337.00, and closed at $329.82, down 4.7% (-$16.43) on ~3.8M shares versus a 2.6M 30-day average (1.4x).
• Relative: The S&P 500 rose ~0.3% on the reaction day, so AMGN underperformed the tape by roughly 5 percentage points.

This is a textbook "sell the news" on a low-quality beat. The market looked past the headline double-beat-and-raise to the composition (4% product growth, a 13.6% tax rate, a 45% jump in other revenue, a minimal and partly tax-driven guide raise) and priced in the two overhangs that got worse this quarter: the Tavneos withdrawal proposal with a fatal-VBDS label change, and a fresh, potentially material IRS exposure. With the stock entering the print up 19% over twelve months and near the top of its range, there was little cushion for a quarter that reassured on the base but not on the trajectory.

Street Perspective

Debate: Was the sell-off an overreaction or a fair repricing?

Bull view: A 5% drop on a beat-and-raise, into a ~15x forward multiple with a 24%-growing driver portfolio and de-risking MariTide tolerability, is an overreaction to headline noise (IRS/Tavneos) that will fade.

Bear view: The repricing is fair: product growth is only 4%, the beat was tax-flattered, the raise was cosmetic, and two overhangs got materially worse. The multiple should compress until the quality of growth improves.

Our take: Both are partly right, which is why we hold. The valuation is more attractive after the drop, but the reasons for the drop (composition, tax dependence, IRS, Tavneos) are legitimate, not noise. Net neutral.

Debate: Is MariTide meaningfully de-risked, or just repackaged?

Bull view: Three-step titration lowering nausea and vomiting attacks the core AMG 133 bear point, and the SWITCH plus maintenance studies show a company confident enough to build the full commercial evidence base.

Bear view: New trials and tolerability color are not efficacy. The switch study's focus on every-8-week/12-week dosing (not monthly) hints at hedging, and the whole program is still pre-pivotal-readout.

Our take: The tolerability progress is a real, thesis-relevant positive and the most important thing that happened this quarter. But it is not the efficacy readout that would let us pay for the franchise. De-risked at the margin, not resolved.

Debate: How much do the IRS and Tavneos overhangs matter?

Bull view: Tavneos is tiny and the IRS matter is years from resolution and vigorously disputed; both are manageable tail risks that do not change the earnings power of the franchise.

Bear view: The IRS extending the Puerto Rico theory to 2016–2018 enlarges an already multi-billion-dollar exposure on a $57.3B-levered balance sheet, and a fatal-VBDS label plus an FDA withdrawal proposal is a real safety event, not a technicality.

Our take: The direct P&L stakes are modest, but the tail risk is genuine and grew this quarter. On a levered balance sheet, tax tail risk deserves weight. These are the reasons a cheaper stock did not earn an upgrade.

Model Update Needed

Valuation impact: At the $329.82 reaction-day close, AMGN trades at roughly 14.7x the 2026 non-GAAP EPS midpoint of $22.40, a discount to both the market and its own recent range. That is inexpensive for a franchise growing its core drivers 24% with genuine obesity optionality, but the discount is warranted by 4% headline product growth, a tax-dependent beat, and enlarged tax/regulatory tail risk. We frame fair value in the low-to-mid $340s on the base business, implying modest upside from here, with MariTide the source of asymmetric upside. The improved valuation and the fresh overhangs roughly offset, which keeps the risk/reward balanced.

Thesis Scorecard Post-Earnings

Scoring this quarter against the standing thesis established at initiation. Status tags reflect the direction of travel this quarter.

Overall: Thesis unchanged in shape. The growth engine and Repatha pillars strengthened, margin compression eased, MariTide de-risked on tolerability, but denosumab erosion is confirmed and, more importantly, the leverage/policy bear escalated with a new IRS exposure and a worse Tavneos posture. The positives on quality of growth are offset by the negatives on tail risk.

Action: Maintain Hold. The de-rating improves the entry and the base business is executing, but a tax-flattered beat and enlarged tail risk keep the risk/reward balanced. The upgrade trigger is unchanged: pivotal MariTide weight-management efficacy data, or clear resolution of the IRS and Tavneos overhangs at a still-discounted multiple.