Beat, Raise, Sell-Off: Sanofi Cuts Three Programmes and Concedes the 2031 Base Case

Results vs. Consensus

Sanofi reports in euros per ordinary share on a "business" (non-IFRS) basis. One SNY American depositary share represents one-half of one ordinary share, and US consensus is compiled in dollars per ADS. The two bases are kept separate throughout this note and no figure is converted between them. No euro-basis consensus is retrievable for this print, so those rows read n/a and the euro figures are judged against the prior-year base and the company's own guidance.

Two independent US sources put the dollar consensus at $1.09 and $1.10 per ADS and the actual at $1.21 and $1.22. Rather than pick one, we print the range: a beat of roughly eleven percent, roughly three times the size of the three-and-eight-tenths percent surprise delivered in the first quarter.

Year-over-Year: the Business (non-IFRS) P&L

Every line moved the right way at once, which has not been true of a Sanofi quarter in some time. Sales grew 17.8% at constant currency; gross profit grew faster at 23.6%; operating expenses grew slower at 13.1%; and the result was 35.8% operating-income growth and a margin 3.8 points wider. The currency drag that dominated the first quarter narrowed sharply, from 7.4 points on sales to 1.8.

The composition of the beat

Management's prepared remarks did the unbundling themselves, which is to their credit and is also the most important disclosure in the release. Three separate restatements were volunteered:

Revenue. The 17.8% is real but is not the run rate. Strip the two acquisitions and it is about 15%; strip the gross-to-net true-up and it is lower again. The more useful figure is the one embedded in the raised guidance: for the full year to grow around 10% at constant currency against a first half at 15.7%, the second half needs roughly 5%. That is a ten-point deceleration management has now formally guided to, driven by annualising the Dupixent indication launches, lapping the July 2025 Ayvakit consolidation, and a vaccines line expected to fall by low-to-mid teens in the third quarter.

Margins. The 3.8-point operating-margin expansion is the headline number and about half of it is durable. Underlying gross margin improved 1.7 points on genuine mix, as specialty care and rare disease take a larger share, and selling and general expenses fell 1.6 points as a share of sales even while absorbing two acquisitions. That is real operating leverage of a kind the first quarter did not show, and it is the clearest evidence yet against our standing concern that cost discipline above gross profit was being consumed below it. The offset is the alliance line, discussed below, which is consuming more of the operating income than a year ago rather than less.

EPS. Business EPS grew 33.3% at constant currency, ahead of business net income's 31.0% on a share count down 1.7% from the completed buyback, but behind operating income's 35.8%. Working against it were higher net financial expenses, up from €59m to €104m on higher net debt, and a tax rate 2.2 points higher at 21.7% against 19.5%. The full-year tax rate guide moved up to around 21% from around 20% precisely because the impairments taken this quarter are not deductible, which is a small, real, and easily missed cost of the pipeline decisions.

The IFRS-to-Business Bridge

The gap between the two earnings measures is wider this quarter than at any point in recent memory, and it is worth walking because it is where the pipeline decisions actually landed.

The headline decline of 91.3% in IFRS net income is not a like-for-like comparison: the prior-year quarter carried €2,707m of discontinued-operations income from the Opella disposal. The honest measure is continuing operations, where basic EPS fell from €1.02 to €0.29, a decline of 71.6%. That decline is real and it is the pipeline: a €1,031m impairment, of which €952m is the amlitelimab intangible, sitting alongside €552m of amortisation and €444m of restructuring.

Assessment: Investors who mark this business on business EPS saw a 33% growth quarter. Investors who mark it on IFRS saw €0.29. Both are correct, and the reconciliation is the point: €2,158m of the gap is non-cash or non-recurring, but €952m of it is the write-off of an asset the company spent years and a great deal of money developing. That is not a bookkeeping adjustment. It is the cost of the strategy being reset, disclosed in the quarter it was taken, and it will not be the last of its kind if the portfolio review continues as management indicated.

Segment Performance

By geography

The geographic split is the quarter in one table. The United States grew 33.5% and now represents 54.7% of sales, up from 50.3% in the first quarter. Europe grew 1.3%. Rest of World grew 3.7% and China shrank 4.9%. Sanofi is becoming, at an accelerating pace, a US business with a European listing and a euro reporting currency.

Assessment: This concentration cuts two ways and both are material. It amplifies the currency exposure that took 7.4 points off first-quarter growth and only 1.8 off this quarter's, meaning reported results will stay hostage to the dollar in either direction. And it raises the stakes on US drug pricing policy, a topic that has now gone two consecutive earnings calls without being raised by management or by a single analyst, even as the vaccines business openly attributes a cut in its 2030 ambition to changes in US immunisation policy.

By franchise

Dupixent

Sales of €5,154m, up 37.6% at constant currency, the first quarter above five billion euros, and more than 1.5 million patients on therapy. US sales grew 42.8%, Europe 19.4%, Rest of World 26.9%. Nine years after launch this is an extraordinary growth rate for an asset of this size, and it is predominantly volume.

"Dupixent continued its very strong growth at 37% in Q2, and that was driven largely by volume, that means more patients receiving it." — Belén Garijo, Chief Executive Officer

It is not entirely volume, and the difference matters for how the second half is modelled. Management confirmed a favourable gross-to-net adjustment in the US worth a couple of hundred million euros in the quarter, of which a portion is non-recurring and a portion reflects tighter control of 340B discounted channels that should persist. Growth is guided to moderate in the second half against tougher comparatives.

Assessment: The franchise is executing beyond any reasonable expectation, and that is precisely the problem the equity has. At 44.4% of sales, up from 39.7% in the first quarter, every additional point of Dupixent outperformance concentrates the 2031 exposure rather than diluting it. Sanofi is winning the race it does not want to be running.

Pharma launches

The eleven-medicine launch cohort reached €1,305m, up 48.3%, now 11% of group sales. The 13%-of-sales and better-than-60% growth figures quoted on the call belong to the wider launch line that also counts vaccine launches, €1,530m and up 61.1%; the chief financial officer was explicit that the disclosed pharma-launch scope excludes vaccines. Ayvakit, ALTUVIIIO and Sarclisa led the pharma cohort. On the call the chief financial officer put the cohort at €5.2bn to €5.3bn for the full year and framed the path to the €10bn 2030 ambition against the cohort's current like-for-like growth rate. He added that he is "not worried at all" about reaching it, on the grounds that the products are already approved and selling, leaving what he called "essentially a commercial risk."

"To get to the EUR10 billion by 2030, we need to move to grow on average CAGR by 15%, given that we are on the trend of 27% today." — François-Xavier Roger, Chief Financial Officer

Assessment: This is the most credible number in the 2030 framework and the only one that improved on a like-for-like basis without being revised. An ambition requiring 15% growth from a portfolio compounding at 27%, with the products already approved and selling, carries commercial rather than scientific risk. It was also confirmed as already profitable at the operating line. This is the pillar of the growth story that this quarter strengthened.

Rare disease

Sales reached nearly €1.9bn, up 24%, led by Ayvakit at €190m and ALTUVIIIO at €349m. The chief executive singled the franchise out as an area of expansion, citing a top-two or top-three market position, and flagged it as a target for business development. Within the franchise the switch dynamics are working as designed: Nexviazyme grew 15.6% while Myozyme fell 22.9%, and ALTUVIIIO grew 23.7% as patients moved off legacy factor products.

Assessment: Rare disease is the one area where a new chief executive has explicitly promised to spend, and the base is genuinely strong. The caution is that the same call described an intention to expand beyond rare genetic disorders, which is a broadening rather than a deepening, and broadening is how the last decade of Sanofi capital allocation is usually criticised.

Vaccines

Sales fell 4.7% to €1,150m against a difficult influenza comparative, with the polio, pertussis and Hib franchise flat at €700m as newly acquired Heplisav-B in the US offset a 21.2% decline in Rest of World driven by falling Chinese birth rates. Beyfortus grew 54.2% but off a small seasonal base. The 2030 ambition was cut from about €10bn to about €9bn.

"This EUR1 billion difference for 2030 is coming on half of it on a US exchange rate, so US to euro, it's pure financial extent rate. The second half is driven mostly by US VC, so US VCI evolution, which has turned out to be weaker than expected following the new US administration." — Thomas Triomphe, Executive Vice President, Vaccines

Assessment: Half the cut is currency and half is a policy-driven decline in US vaccination coverage. The second half is the one that matters, because it is the first time Sanofi has quantified US immunisation policy as a durable headwind to a long-term target rather than a quarterly nuisance. A business bought into via Dynavax in February is being guided down for 2030 five months later, and the reason given is demand destruction that no amount of commercial execution addresses.

Key Topics & Management Commentary

Overall Management Tone: This was the most candid Sanofi call in years and the least reassuring, which is the same fact stated twice. Where the interim administration in April defended the status quo and declined to date or size anything, the new chief executive opened by conceding that the company has over-promised and under-delivered, attributed the pipeline failures to weak internal decision-making rather than bad luck, and had already cancelled three programmes before being asked about any of them. The confidence was concentrated in the commercial results and the near-term guidance; on strategy, the repeated answer was that a fuller account arrives by year-end.

1. A new chief executive who leads with the diagnosis rather than the plan

Belén Garijo joined in May and used her first quarterly call to describe twelve weeks of listening rather than to unveil a strategy. That is a defensible choice and it was executed with unusual directness. Asked what investors are underappreciating about Sanofi, she did not list assets. She answered that the company had spent years damaging its own credibility:

"we have, for a period of time, over promise and under deliver. That's the bottom line." — Belén Garijo, Chief Executive Officer

She paired that with a second observation, that as a newcomer she finds it "really shocking that the strong performance of this company is really not recognized by investors." The two statements sit slightly awkwardly together: the market's refusal to pay for the performance is largely a function of the credibility deficit she had just described. A full strategic account was promised "in the coming quarters" and "before the year-end".

Assessment: Naming the credibility problem out loud is worth something, and it is more than the prior administration offered. But an investor is being asked to fund a further two quarters of diagnosis on a stock whose entire discount is about what happens in 2031. The year-end strategy day is now the single most important scheduled event in this name.

2. Three programmes cancelled, and what they cost

Amlitelimab will not progress to regulatory submission in atopic dermatitis despite the ESTUARY long-term extension study showing maintained response for up to 72 weeks with no new Kaposi's sarcoma cases. Itepekimab is discontinued across COPD and chronic rhinosinusitis. Balinatunfib is discontinued after interim Phase 2 analyses in Crohn's disease and ulcerative colitis missed internal efficacy expectations. The financial cost was a €1,031m impairment, of which €952m was the amlitelimab intangible, plus more than €200m of wind-down expense inside research and development.

The reasoning offered on amlitelimab is notable for what it concedes. The release records that the ESTUARY extension showed maintained clinical response without relapse for up to 72 weeks and a safety profile building on previous data, and then discontinues the programme anyway, on the stated ground that the asset would not represent a meaningful improvement over the standard of care. The asset worked, and worked for long enough, but not well enough to matter.

"Sanofi has determined that the totality of efficacy and safety evidence generated to date does not support further development of amlitelimab in AD." — Q2 2026 results press release

Our April note flagged itepekimab's absence from the company's own newsflow table across two consecutive releases and recorded a working assumption of termination. That assumption was correct. Amlitelimab we had flagged for a different reason, two cutaneous Kaposi's sarcoma cases across roughly 4,500 patients, and the safety signal turned out not to be what killed it.

Assessment: Killing three programmes in a quarter is good capital allocation and bad news, and the market is entitled to treat it as both. The saved development spend is real and the impairment is sunk. What is not recoverable is the four-year gap: these were the assets meant to be generating revenue as Dupixent's patent runs out, and the replacements have to be bought rather than developed. That is precisely the sequence in which acquirers overpay.

3. The chief executive's own explanation for why the pipeline failed

Pressed repeatedly on whether the R&D problem is structural, the answer was consistent and specific: the failures are decision-making failures, located at the Phase 2 to Phase 3 transition, and therefore fixable faster than a capability deficit would be.

"When you make a decision to go from Phase 2 to Phase 3 and engage a significant and get ready to engage a significant amount of capital, you really need to challenge whether or not your data in Phase 2 are justifying the move to Phase 3, right? And you're going to start by dreaming on a target product profile that is not based on rigorous assumption." — Belén Garijo, Chief Executive Officer

She was explicit that this "is nobody's fault" and that decisions on critical scientific questions had not clearly sat where they belonged, adding that she expects "the scientists to make scientific decisions." A new head of research and development, Paulo Fontoura, arrives in September; a reduced eight-member executive committee takes effect on 1 September with three new members, all internal.

Assessment: This is a coherent and testable diagnosis, and it is more useful than a generic promise of R&D excellence. It also has an uncomfortable implication: if the problem was that Phase 2 data were being over-read, then the remaining late-stage pipeline was selected by the same process. Management effectively conceded this by declining to say the review is finished.

4. The portfolio review is not over

Asked directly whether the cuts announced represent the end of the process, the answer was no, with the caveat that there is no quota.

"So the pipeline review is ongoing. So I'm not excluding that we continue some additional assets. But we don't have a target number to discontinue, right?" — Belén Garijo, Chief Executive Officer

The remaining late-stage catalysts are frexalimab in relapsing multiple sclerosis reading out in 2027 and in secondary progressive disease in 2028, riliprubart in the VITALIZE study in the second half of 2026, brivekimig in hidradenitis suppurativa in 2027, and pneumococcal and yellow fever vaccines. Riliprubart already carries a warning: the MOBILIZE study in refractory chronic inflammatory demyelinating polyneuropathy was stopped early on independent monitoring committee advice, though the same committee recommended VITALIZE continue.

Assessment: An open-ended review with a new research head arriving in September means every remaining asset carries discontinuation risk through at least the next two prints. For a company whose valuation gap is about what replaces Dupixent, an open-ended review is a rational management position and an expensive one for the equity.

5. Management now plans on losing the Dupixent patent in 2031

This is the most consequential disclosure of the quarter and it arrived almost in passing. The general counsel gave the expected defence, extending the stated patent runway by a year against the April account.

"You mentioned patents going to '45. Actually, this quarter, we can say it's patent expiration dates going up to 2046. We have a very strong patent portfolio around the many years of R&D, multiple innovations, indications to date. And of course, we intend to be vigorously defending." — Roy Papatheodorou, Executive Vice President, General Counsel

He added that it remains too early to speculate on biosimilar entry dates, that Sanofi does expect Dupixent to be protected beyond March 2031, and that no specific patents have yet been challenged. Then the chief executive closed the exchange with a sentence that says something quite different about how the company actually plans.

"What I can tell you, Simon, is that we are taking a base case that is associated to the patent to the loss of the product pattern." — Belén Garijo, Chief Executive Officer

Read plainly, and in context of the question it answered, the company's internal base case assumes exclusivity ends when the product patent does, in March 2031. The formulation and method-of-treatment estate running to 2046 is upside to that case, not the case itself.

Assessment: In April we wrote that the bull case on the patent estate rested on an assertion and the bear case on a base rate. Management has now told us which one they use internally, and it is the bear case. This is a genuine improvement in disclosure and a deterioration in the fundamental picture: an investor can no longer argue that the company privately expects to hold the franchise past 2031, because the company has said it does not.

6. Dupixent's 2030 ambition goes up, and barely clears where the Street already was

The 2030 ambition for Dupixent was raised to about €25bn from about €22bn. The launch cohort ambition was held at about €10bn. Vaccines was cut to about €9bn from about €10bn. The chief financial officer described the three together as upgraded "by almost 10% on a like-for-like basis at constant exchange rates", and noted the constant-currency baseline itself was reset to today's less favourable rate from the 2023 rate used when the ambition was first set.

Two observations follow. First, the €3bn Dupixent raise against a €1bn vaccines cut is a net €2bn improvement in which the growth is more concentrated in the asset with the 2031 problem: Dupixent moves from roughly 52% of the three-item ambition to about 57%. Second, sell-side consensus for Dupixent in 2030 already sat at €24.2bn, so the upgraded company ambition cleared standing Street numbers by well under a billion euros.

Assessment: A long-term target raised to approximately where analysts already modelled it does not create value, and one raised by concentrating further into the disputed asset arguably destroys some. The vaccines cut is the more informative half of the announcement.

7. The Regeneron step-down lands on schedule, and 2027 is much better than we modelled

In April management quantified a roughly €400m operating-income hit in 2026 from the end of Regeneron's development-cost reimbursement, with a further roughly €700m in 2027. We carried both into our model. The second quarter completed the reimbursement, as promised, and the 2027 arithmetic was materially revised.

"Q2 marked the end of the Regeneron development balance reimbursement. For the full year 2026, the increase of Amvuttra royalties will fully offset this negative BOI impact. Next year, in '27, we will see a negative EUR200 million BOI gap between these two items. Indeed, we expect a final negative impact from the general reimbursement of about EUR600 million year on year from '26 to '27, while Amvuttra royalties income is expected to increase at the same time by about EUR400 million year on year as well." — François-Xavier Roger, Chief Financial Officer

Amvuttra income was €217m in the quarter against €57m a year ago. The company also introduced a projection of Amvuttra royalties to 2030 based on the company-compiled consensus.

Assessment: This is the single largest positive revision in the quarter and it received almost no attention. A 2027 headwind we carried at roughly €700m now nets to about €200m, a €500m swing worth roughly €0.42 of business EPS before tax, or about €0.33 after tax at the guided rate, on the current 1,196.6m share count. Both halves of the Q1 commitment on this line were met: the reimbursement ended when management said it would, and the offset materialised.

8. The alliance takes a larger share of its own success, again

The Regeneron alliance line cost €1,850m in the quarter against €1,265m, up 46.2%. Within it, profit sharing rose 45.0% in reported euros to €1,972m while Dupixent sales grew 37.6% at constant currency; because reported growth ran below constant-currency growth group-wide, the real gap is wider than the seven points those two published figures imply. The alliance now absorbs 56.2% of business operating income against 51.4% a year ago.

Assessment: This was our second bear pillar in April and it is unambiguously worse, not better. The mechanism is structural: as Dupixent becomes a larger share of a more profitable business, the counterparty's claim on group operating income grows faster than the group does. The reimbursement relief described above is a timing benefit within a relationship whose underlying economics continue to move against Sanofi.

9. An alliance being renegotiated, without a timeline

The chief executive devoted more airtime to Regeneron than to any other single topic and was strikingly frank about the state of the relationship.

"To be honest, my fresh impression is that as for any partnership, trust is absolutely essential and for different reasons that I'm not going to judge, I didn't have the feeling that, that was, at this time, one of the environment in which we have been operating." — Belén Garijo, Chief Executive Officer

She described conversations as productive, covering both the addition of assets to the collaboration and improvements to how it operates, and declined to commit to any timing. Separately, litigation between the two companies over information sharing was characterised from the Sanofi side as narrow, and management confirmed the alliance is not a barrier to unrelated business development.

Assessment: An admission that trust was absent in the partnership governing 44% of revenue is remarkable candour and a material risk disclosure. It is also, for now, unpriceable: there is no timeline, no framework, and no indication of which direction the economics would move. Any renegotiation that improves Sanofi's profit share would be the largest single positive catalyst available in this name outside the patent estate.

10. The gross-to-net tailwind, and how much of it repeats

A recurring line of questioning tried to establish how much of the Dupixent beat was price or channel rather than demand. Management's answer separated the two components without sizing the recurring half.

"When we talk about operational improvements, operational effectiveness, fully agree that it is linked to 340B partially. So as I said, is a portion that is recurring. There's a portion that is nonrecurring. The nonrecurring portion a couple of hundred million. The recurring portion exactly, as you say, is linked to us really looking at how do we ensure patient benefits that reached intended recipients." — Manuela Buxo, Head of Specialty Care

Assessment: Tighter control of 340B discounted channels is a genuine and durable margin lever that several US-exposed manufacturers are pulling, and management confirmed some of it continues into the second half. The unsized recurring portion is the reason the second-half guide can be read as conservative. It is also why the underlying growth rate of this franchise is currently unknowable to an outside modeller within a few points.

11. Capital allocation: buyback finished, net debt higher, and an appetite for larger deals

The €1bn buyback completed in April, so the share-count tailwind fades from here, a point the chief financial officer made explicitly when framing the second half. Net debt rose from €10,988m at 1 January to €15,513m at 30 June, driven by the €1,635m Dynavax acquisition, the €4,923m dividend and the buyback, partly funded by a €2.3bn note issue in April. Free cash flow was strong, at €2,670m in the quarter and €3,724m in the half.

On mergers and acquisitions the posture shifted. Sanofi has historically done bolt-ons; the chief executive opened the door wider.

"Traditionally, Sanofi has been on smaller deals, bolt-ons to eventually increase innovation of buildups. I think our appetite for bigger deals is an option." — Belén Garijo, Chief Executive Officer

Business development now reports to the chief financial officer. Management also indicated a shift toward later-stage and potentially commercial assets rather than the early-stage focus of recent years, and stated there is no urgency to transact by any particular date. The dividend policy was reaffirmed.

Assessment: The strategic logic is sound: a company that has just written off its internal answer to the 2031 problem must buy one. The risk is equally clear. A large-cap pharmaceutical with a publicly acknowledged pipeline hole, a five-year clock, and stated openness to bigger deals is the textbook profile of a buyer that pays too much. The absence of a stated deadline is the single most reassuring thing management said on this subject.

12. Opella surfaces at last, and it loses money

In April management deferred disclosure of the retained Opella stake to the half-year. That commitment was met. Sanofi holds 48.2% of OPAL JV Co, and the summary financials show first-half sales and other revenues of €2,657m and a net loss of €61m, against total assets of €19,598m, total equity of €6,394m and non-current liabilities of €11,575m.

Assessment: The disclosure is welcome and the content is not. A leveraged consumer-health business running a small net loss makes Sanofi's 48.2% interest a drag on IFRS earnings rather than a source of value, and the €11.6bn of non-current liabilities against €6.4bn of equity explains why. Sanofi received €10.4bn of cash on the way out, which was the point of the transaction, but the retained stake should be marked as an equity-method loss line rather than as hidden value.

13. Research and development spending will grow more slowly than sales

Asked whether fixing the pipeline requires spending more, in a year when the research budget already absorbed €200m of wind-down costs, the answer was that it does not, at least not yet.

"You asked a question about the R&D as well. Is it going to grow faster or in line with our sales in the short term anyway, given the challenges that we have in decision that we have made lately, it will grow to a moderate level of R&D, which means at a lower level than sales. But as we grow over time and as we gain confidence in our capabilities as well, it will certainly increase." — Olivier Charmeil, Executive Vice President, General Medicines

Sanofi's research-to-sales ratio was noted on the call as sitting at the bottom end of the sector. Management's position is that allocation matters more than quantum, and that spending is driven by contribution to growth and returns rather than by fixed budgets.

Assessment: Defensible in theory and awkward in practice. A company that has just attributed its pipeline failures to poor internal decision-making is proposing to fix them while letting the research budget grow more slowly than the top line, with the difference to be made up through acquisitions. That is a bet that the market for late-stage assets is more efficiently priced than Sanofi's own laboratories, which may well be true, and is an expensive way to be right.

Guidance & Outlook

The headline is a raise, and it is a genuine one. But the interesting content sits in the second and third columns rather than the first. Guidance for the year went up because the first half beat, not because the second half improved: alongside the upgrade came a first-time vaccines guide of slightly negative for the year, a third quarter guided to a low-to-mid-teens vaccines decline, a higher tax rate caused by the non-deductibility of this quarter's impairments, and an explicit warning that the gross-margin one-offs, the Regeneron reimbursement and the buyback benefit all stop contributing.

Implied second-half ramp: with the first half at 15.7% constant-currency growth and the full year guided to around 10%, and applying the 2025 half-year weightings, the second half needs roughly 5% constant-currency growth. That is a deceleration of more than ten points, formally guided. Management named the causes: annualising Dupixent's newly launched indications, lapping the July 2025 Ayvakit consolidation, and the vaccines seasonality above.

Implied second-half earnings: the first half delivered €3.97 of business EPS against €3.39. On our full-year estimate of about €8.50, the second half carries roughly €4.53 against €4.44 a year ago, or about 2% reported growth, after a first half that grew 17.1%. The guidance is a full-year raise built entirely on a first half that has already happened.

Street position: the raise takes the company to approximately where the sell-side had already migrated on the year, and the 2030 Dupixent ambition of about €25bn compares with consensus that already sat at about €24.2bn. Neither number moved the debate.

Guidance style: this management has one quarter of history, so the pattern is not yet established. What can be observed is that the interim administration guided conservatively in April and the new administration has raised into strength while simultaneously introducing three new negative disclosures. That combination reads as a deliberate reset of the baseline rather than an expression of confidence in the second half.

Analyst Q&A Highlights

What is actually blocking progress with Regeneron

The opening question of the call went to the alliance, and specifically to what an incoming chief executive with no history in the relationship could see that others could not. The answer was more direct than the question anticipated, naming an absence of trust as the obstacle rather than any commercial or legal issue, and it set the tone for the rest of the hour.

Q: "So can you give us an idea of, from your point of view, what are the key factors that may be blocking faster progress here in terms of adding assets into the collaboration? And you say you're in early stages of the discussions. But what should we think in terms of timelines? Could we see an update here in 2026? Or is that too early?"
— James Quigley, Goldman Sachs

A: "My main objective together with Manuela, who is the Head of Specialty Care here with me has been to rebuild fast, to be able to be transparent to one another, to be able to create a path forward for this conversation."
— Belén Garijo, Chief Executive Officer

Assessment: The question asked for a timeline and got a description of relationship repair. That is an honest non-answer, and it means the largest single lever on group profitability remains unquantifiable for at least another quarter. Note what was not said: nothing about the economics of the profit split, which is the term that actually matters.

Whether the pipeline plan differs from the last three

The sharpest exchange of the call put the R&D reset in historical context, observing that successive chief executives have arrived with similar-sounding plans. The response declined the premise that anything is structurally wrong, which is a meaningful position to take given the same call had just cancelled three programmes.

Q: "And then secondly, on R&D, Sanofi's been through many iterations of attempting to improve the pipeline. This is, I think, the third CEO sort of my coverage seen come in with a new plan that sounds suspiciously like the old plan. So perhaps could you just talk us through what you think is systemically wrong in the organization and if and how quickly that can change?"
— Graham Parry, Citi

A: "I don't think there is a systemically wrong issue in the organization. I think you have to pull the levers, right, be consistent and never complacent and managing risk and the risk profile of the pipeline is going to be something that will be very much at the top of our priorities."
— Belén Garijo, Chief Executive Officer

Assessment: This is the exchange to revisit in twelve months. Management's position is that the problem is process rather than structure, which is optimistic but at least falsifiable. The tension with the rest of the call is real: if nothing is systemically wrong, it is not obvious why three programmes had to be cancelled at a cost of more than a billion euros within twelve weeks of a new chief executive arriving.

How long an R&D turnaround takes, and what protects Dupixent past 2031

A recurring line of questioning tried to put a clock on the transformation, with reference to sector precedents that took four to five years. The same question then moved to the simplest available fix for the 2031 problem, which is for exclusivity to run longer than the product patent. The answer to the second half is the most important sentence in the transcript.

Q: "Now we know you have a lot of IP beyond March '31 out to 2045, I think. I'd be interested to get your early perspectives on what you think as a fresh pair of eyes on this is what you think the strength of the IP beyond '31 is for Dupixent. And therefore, the more realistic possibilities on when we will face biosimilar competition for that asset?"
— Simon Baker, Redburn

A: "What I can tell you, Simon, is that we are taking a base case that is associated to the patent to the loss of the product pattern."
— Belén Garijo, Chief Executive Officer

Assessment: The general counsel had just finished describing a patent estate running to 2046 and an intention to defend it vigorously. The chief executive then confirmed the company plans on the product patent date. Both can be true, and together they define the shape of the investment case: everything past March 2031 is an option the company itself does not underwrite. Any investor paying for post-2031 Dupixent revenue is taking a more aggressive view than management.

Where the Dupixent upside actually came from

The dominant technical topic in Q&A was the gap between reported Dupixent growth and what volume alone would support, with a specific hypothesis put to management about tighter administration of the US 340B discount programme. Management confirmed the hypothesis and split the benefit into recurring and non-recurring halves, sizing only the half that stops.

Q: "Can I push you on your operational effectiveness comment? Has the alliance changed its policy on 340B claims for to reduce access to heavily discounted drug. And I ask that because that's the only thing I can think of that's the really sign improvement in GTN over the first half of the year."
— Multiple analysts incl. Matthew Weston, UBS; Peter Verdult, BNP Paribas Exane

A: "The nonrecurring portion a couple of hundred million. The recurring portion exactly, as you say, is linked to us really looking at how do we ensure patient benefits that reached intended recipients. 340B, enhanced 340B control is a part of that. It's not the only part, but it's a part of that — a large part of it."
— Manuela Buxo, Head of Specialty Care

Assessment: A clean answer to a well-constructed question, and the most useful modelling disclosure on the call. A couple of hundred million euros of the quarter's revenue does not repeat, which is roughly two points of group growth. The recurring portion was confirmed but never sized, so the underlying Dupixent growth rate remains a range rather than a number for anyone outside the company.

Whether the portfolio review has further to run

With three cancellations already announced and a new head of research arriving in September, the natural question was whether the remaining late-stage assets are safe. They are not, and management said so without hedging, while declining to work to a quota.

Q: "Belén, you've cut a few pipeline programs. Should we think that the review is now complete? Or should we think about further discontinuations when Paulo joins?"
— Richard Vosser, JPMorgan

A: "So the pipeline review is ongoing. So I'm not excluding that we continue some additional assets. But we don't have a target number to discontinue, right? We are going to continue to be operating on the basis of scientific merits and risk, as I mentioned, and potential and some other elements."
— Belén Garijo, Chief Executive Officer

Assessment: Correct governance and a live overhang. Every remaining late-stage asset now carries a discontinuation probability that did not exist a quarter ago, and the arrival of a new research head in September makes the next two prints the likely venue. An investor cannot currently underwrite any specific pipeline asset in this name.

Why the 2030 vaccines ambition came down

The vaccines reduction drew the most pointed challenge on the long-term framework, precisely because it followed an acquisition in the same business five months earlier. The answer split the cut evenly between currency and a policy-driven decline in US immunisation rates.

Q: "And then my second question is about the long-term guidance for vaccines. You bought Dynavax, but you cut the 2030 vaccine guide by 10% to EUR9 billion. I know that some of that — but is it because you're meaningfully more cautious on Beyfortus? Or is it the flu franchise? Or is it something else?"
— Matthew Weston, UBS

A: "But also, you've seen that CDC and other revenues have shown that there has been a decrease on the vaccination coverage rate of US pediatric vaccines. So if you put both together, that explains the difference. It doesn't change anything on our long-term ambition for vaccines."
— Thomas Triomphe, Executive Vice President, Vaccines

Assessment: The closing sentence does not survive contact with the number. A target reduced by ten percent is a changed ambition by definition. The substance of the answer is nonetheless valuable and unusually candid for this sector: a large European manufacturer has put a euro figure on the effect of US vaccination-policy change on its 2030 revenue, and that figure is roughly half a billion euros.

Whether fixing the pipeline requires spending more on it

A recurring theme was that the research budget sits at the bottom of the peer group as a share of sales, which invites the question of whether the transformation is funded. The answer was that spending grows more slowly than sales in the near term, with the gap to be closed by transactions rather than internal investment.

Q: "So we've heard that you think you might need to do a bit more R&D. But is there any way that you might not need to spend quite so much. So are you looking to reallocate as if you might be able to take some spend, save from SG&A and reallocate it to R&D or not. So is there an opportunity there?"
— James Gordon, Barclays

A: "Our choices in terms of spend on — I would rather talk of investment rather than spend are driven by contribution to growth and contribution to return. So if there is a better return on the commercial side of things, what we will do."
— Olivier Charmeil, Executive Vice President, General Medicines

Assessment: A returns-based framework applied to research budgets is defensible and is what most well-run companies say. It is harder to reconcile with a diagnosis that attributes the pipeline failure to insufficient rigour and insufficient capability, both of which cost money to fix. The practical read is that the 2031 gap will be filled by acquisition, and the market should price the risk of that accordingly.

What They're NOT Saying

1. Any quantified erosion curve for Dupixent after March 2031. Unchanged from April, and now more conspicuous because management has confirmed the product-patent date is their base case. Sanofi will tell you it plans on losing exclusivity in 2031 and will not tell you what it thinks happens to a franchise running at more than €20bn a year when it does.
2. A total 2030 sales ambition. Three components were given, Dupixent, launches and vaccines, summing to about €44bn. The company does not disclose what the remainder of the business, including established medicines and rare disease outside the launch cohort, is expected to contribute. The framework describes the growing part and stays silent on the rest.
3. Any 2030 margin or earnings ambition. The long-term framework is stated entirely in sales. In a business where the alliance partner's claim on operating income is rising faster than revenue, a sales ambition without a margin ambition is a partial disclosure.
4. The size of the recurring gross-to-net benefit. The non-recurring half was sized at a couple of hundred million euros. The recurring half was confirmed to exist, confirmed to be substantially 340B-related, and never quantified, which leaves the underlying growth rate of 44% of revenue indeterminate.
5. Any operating-margin guidance for 2026. Sanofi guides sales and EPS and nothing in between. This was our observation in April and it survives a change of chief executive. Note also that the explicit April commitment to grow operating income in both margin and absolute terms in 2026 and 2027 was not repeated on this call.
6. A split of the impairment and wind-down charges by programme. The €1,031m impairment is disclosed as "mainly" €952m of amlitelimab. The remaining €79m, and the more than €200m of wind-down cost inside research and development, are not allocated between itepekimab and balinatunfib.
7. Any timeline, framework or economic direction on the Regeneron renegotiation. Management confirmed conversations are productive and declined to commit to timing, structure, or whether the profit split itself is in scope.
8. Any measurable target for the R&D transformation. No productivity metric, no phase-transition success-rate target, no date. The transformation is described qualitatively and the only scheduled deliverable is a strategy update by year-end.
9. US drug pricing, tariffs or most-favoured-nation policy. A second consecutive call in which the topic is raised by neither management nor a single analyst, with the United States now at 54.7% of sales. The vaccines answer acknowledged US policy as a durable headwind to a 2030 target, which makes the silence on pharmaceutical pricing more striking rather than less.
10. What the €47bn of business development spend over eight years has produced. Management volunteered that €47bn was invested and €7bn written off, and characterised the remaining €40bn as having created substantial value. No return figure or asset-level accounting was offered for a number that large, and it is now the most relevant track record in the story given the stated pivot toward larger deals.

Market Reaction

• Pre-print setup: The Paris line closed at €79.92 on 29 July, down 3.4% year to date against a CAC 40 up 3.2%, down 6.5% over twelve months, but up 6.4% over the trailing thirty sessions, inside a 52-week closing range of €72.49 to €90.44. The ADS closed at $44.85, down 7.4% year to date against an S&P 500 up 6.9%, down 8.9% over twelve months, and up 5.1% over thirty days, inside a 52-week closing range of $41.33 to $52.34.
• The bad news was not all new. The decision not to submit amlitelimab was announced on 24 July, six days before the print, and the executive committee reorganisation on 21 July. Paris fell 0.43% on the amlitelimab announcement and then rose 2.31%, 2.04% and 0.68% across the following three sessions, gaining 5.1% into the print. The market had absorbed the amlitelimab news and was positioned for a beat.
• Reaction session, 30 July: Paris opened at €78.33, traded down to €72.07, and closed at €72.77, down 8.95%, on 7.0m shares against a 30-day average of 2.5m, or 2.8 times normal. The CAC 40 rose 0.9% on the same session. The ADS opened at $42.04, a 6.3% gap, ranged $41.42 to $42.93, and closed at $42.89, down 4.37% on 10.2m shares against a 3.9m average, or 2.6 times normal, against an S&P 500 up 1.7%.
• The two listings did not close on the same number. Euronext Paris closes at 11:30 New York time and the ADS trades a further four and a half hours. The euro appreciated 0.53% on the day, which accounts for about half a point of the 4.6-point gap between the two closes. The rest is session timing: the ADS traded as low as $41.42, down 7.6%, which is the level corresponding to the Paris close, and then recovered through the New York afternoon. Converted at same-day rates, the ADS closed at the equivalent of €74.41 per ordinary share against Paris at €72.77.

The shape of the move is unusually informative because the sequencing rules out the obvious explanation. This was not a market discovering that amlitelimab had failed; it had known that for six days and had bought the stock anyway. What was new on 30 July was the itepekimab and balinatunfib discontinuations, the reduction of the 2030 vaccines ambition, the formal guide to roughly 5% second-half growth, and management's own bridge showing that about half of a spectacular-looking quarter does not repeat.

Set against that, the positives were a raise to a number the Street had largely reached, a 2030 Dupixent ambition that cleared consensus by well under a billion euros, and a 2027 Regeneron improvement that almost nobody discussed. A beat-and-raise sold off 8.9% because the beat was disclosed as half non-recurring and the raise was already in the price, while the incremental news went entirely to the terminal value.

The second element is the divergence between the two listings, which is the closest thing to a real-time verdict available. The European session, which heard the call live and closed immediately after it, marked the shares down 8.9%. The New York afternoon, trading for four and a half hours after Paris had shut, bought back roughly half of that. Two pools of capital looking at the same disclosures on the same day reached materially different conclusions, and the more considered of the two, in the sense of having had longer to read the release, was the less negative.

Street Perspective

Debate: Do the pipeline cancellations make the equity better or worse?

Bull view: The bull case being made on the Street is that cancelling three programmes that were not going to work is straightforwardly value-accretive, that the impairment is sunk cost recognised rather than value destroyed, and that a chief executive willing to take a billion-euro charge within twelve weeks of arrival is demonstrating exactly the capital discipline the company has lacked. On this reading the money that would have funded three failures is now available for assets that work.

Bear view: The bear camp contends that the assets were the answer to the 2031 question and no longer exist, that replacing them requires acquisitions at a moment when every large-cap pharmaceutical is competing for the same late-stage assets, and that a company advertising a pipeline hole and an appetite for bigger deals has just weakened its own negotiating position. The remaining pipeline was selected by the same process that produced the three failures.

Our take: The bull argument is right about the decision and the bear argument is right about the position. Cancelling was correct; needing to cancel was the problem, and the problem is now four years closer to mattering. The decisive variable is not whether these three were killed but what gets bought, at what price, and how soon, and management explicitly declined to attach urgency to that question.

Debate: How much of the 35.8% operating-income growth is the real business?

Bull view: Some desks argue that even the underlying figure management themselves published, roughly 20% operating-income growth on roughly 15% underlying sales growth, would be an excellent result for a large-cap pharmaceutical, that one-offs recur in some form every year, and that the durable half of the margin story, mix improvement and selling-expense leverage, is exactly the operating leverage bears said would never appear.

Bear view: The bear camp points to a gross margin flattered by a provision reversal, a revenue line flattered by a gross-to-net true-up, an operating line flattered by favourably phased capital gains, and a comparison flattered by two acquisitions, and notes that management guided the second half to roughly 5% growth and roughly 2% earnings growth immediately after publishing these numbers. On this view the quarter tells you very little about the run rate.

Our take: We side with the bulls on the underlying business and with the bears on the print. Roughly 20% operating-income growth on 15% underlying sales growth is genuine operating leverage and it directly answers the concern we raised in April that cost discipline above gross profit was being consumed below it. But the second-half guide is the company's own statement of the run rate, and it is a great deal lower. Both facts belong in the model.

Debate: Does a Dupixent franchise growing 37.6% make the 2031 problem better or worse?

Bull view: A growing consensus view is that every year of compounding before the cliff is cash in hand, that a franchise at €25bn in 2030 generates enormously more cumulative profit than one at €22bn regardless of what happens afterwards, and that biologic erosion is slow enough that a larger base in 2031 is worth more even after biosimilars arrive.

Bear view: The skeptics counter that the arithmetic runs the other way for the equity: Dupixent moved from 39.7% of sales in the first quarter to 44.4% in the second, and from roughly 52% to about 57% of the three-part 2030 ambition. The alliance partner's share of operating income rose from 51.4% to 56.2%. Every point of outperformance makes the terminal hole larger in absolute terms while the assets meant to fill it are being cancelled.

Our take: The bear framing is the better description of what a shareholder actually owns, and the sell-off on a 37.6% Dupixent quarter is the market saying so. The bull argument is correct on cumulative cash and irrelevant to the multiple, because the multiple is not a function of cash received before 2031, it is a function of what the market believes exists after it. Concentration risk that increases while the diversification programme is being cancelled is a deterioration, not an achievement.

Debate: Is 8.6 times earnings with a 5.7% yield now too cheap?

Bull view: The value case is that a business compounding at double digits on a constant-currency basis, generating €3.7bn of free cash flow in a half, paying a dividend with thirty-one consecutive years of increases, and trading at under nine times earnings with the dividend covered nearly twice, is priced for a terminal event that is still five years away and may not fully arrive. The de-rating from about 9.8 times in April to about 8.6 times now happened while estimates went up.

Bear view: The bear camp contends that the multiple is a considered judgment rather than an oversight, that it has compressed in each of the last several quarters for the good reason that the post-2031 business keeps getting harder to see, and that the company's own confirmation of a product-patent base case removes the last argument that the market is being unduly pessimistic. On this view eight to nine times is what a shrinking-after-2031 asset is worth.

Our take: The cheapness is real and the reason for it is also real. What changed this quarter is that both got more extreme at once: the shares de-rated by more than a point of earnings while the near-term outlook improved and the terminal outlook deteriorated. That is a wider distribution rather than a better expected value. We would need either the year-end strategy to contain a credible post-2031 revenue bridge, or a transaction that supplies one, before treating a sub-nine multiple as an opportunity rather than a fair price for an unresolved question.

Model Update & Valuation Framework

Valuation. On our c.€8.50 full-year business EPS estimate, the Paris line at its 30 July close of €72.77 trades on 8.6 times, against 9.8 times at the time of our April note. The 2025 dividend of €4.12 is a 5.7% gross yield on the same close, up from 5.0% in April. Because one ADS represents half an ordinary share the multiples are the same in principle, though the ADS closed the reaction session at the equivalent of €74.41 per ordinary share, about 2.3% above the Paris line, a session-timing gap that normally closes.

Fair value framework. We lower our multiple range to 8.5 to 9.5 times our full-year business EPS estimate, from 10 to 11 times in April, and set a twelve-month fair value range of $42 to $47 per ADS. The reduction is deliberate and reflects three things this quarter established: the assets intended to reduce Dupixent dependence have been cancelled, the 2030 vaccines ambition has been reduced, and management has confirmed its own planning base case assumes the March 2031 product-patent expiry. The midpoint of roughly $44.50 implies about +4% from the 30 July close of $42.89, to which a gross dividend yield of roughly 5.7% can be added. That is a market-like total return, which is the arithmetic behind maintaining a Hold. Note that we are lowering the multiple range and raising the earnings estimate at the same time, which is the honest expression of a quarter that improved the near term and impaired the long term.

What would change the framework. Upward: a year-end strategy containing a quantified post-2031 revenue bridge; a Regeneron renegotiation that improves the profit split or adds assets on disclosed economics; a late-stage acquisition at a defensible price that materially diversifies revenue away from Dupixent; or evidence that the recurring gross-to-net benefit is larger and more durable than disclosed. Downward: further late-stage discontinuations as the portfolio review continues; a failure at riliprubart in VITALIZE or frexalimab in relapsing multiple sclerosis; a large acquisition on stretched terms; a first challenge filed against the Dupixent estate; or a further reduction in the vaccines ambition as US immunisation policy develops.

Thesis Scorecard Post-Earnings

The pillars below are the ones established in our April initiation and carried in our standing thesis. They are graded here against what this quarter's print and call revealed, with the status tag movement noted where it occurred.

Overall: Thesis unchanged in direction, wider in dispersion. Two bull pillars were confirmed and the third improved from at risk to on track, which is the best operating quarter this thesis has recorded. Simultaneously the central bear point got worse in the most consequential way available, by management conceding its own planning assumption, and a fifth bear point has been added because the route to filling the 2031 gap has shifted from internal development to acquisition. The near-term earnings power is better than we modelled in April and the terminal value is less defensible.

Action: Hold. Our April standard for an upgrade was explicit: a tested patent estate, or a franchise-scale pipeline asset. This quarter tested nothing and removed three assets, so the standard is not met even though the shares are 11% cheaper and the estimates are higher. We would become buyers on a year-end strategy containing a quantified post-2031 bridge, on a Regeneron renegotiation with disclosed economics, or on a disciplined late-stage acquisition that diversifies revenue. We would become sellers on further late-stage discontinuations without offsetting business development, on a large acquisition at a stretched price, or on evidence that the recurring gross-to-net benefit was a larger share of this quarter's beat than disclosed.