Summary

Sanofi is buying Blueprint Medicines for $9.1B ($129/share plus $2 and $4 CVRs), the year’s second-largest deal after ITCI, at a modest ~27% premium. Eric Schmidt, who took Blueprint public at $18 a decade ago, calls the near-10-bagger “the way that biotech success stories really should play out,” while Sam pushes back on the lost-future-Amgen lament: “we seem in the biotech world to want our cake and eat it.” Sam’s read is that the deal fits Sanofi’s rare-disease and immunology ambitions; if consensus is right on a ~$2B drug by 2030, $9.1B is “not ridiculously stretched.” Eric says that, given Sanofi’s history of paying heavily for risky assets, this looks like a mature, well-timed strike on depressed valuations.
The hosts think Summit’s post-HARMONi-A sell-off punished the wrong thing. The first dataset with US patients showed a 0.52 PFS hazard ratio, with management “quite adamant” that the China and US subsets were consistent—the biggest bear case was addressed in management’s view—but OS missed at p=0.057 with all alpha spent, so the company may not be able to file in second-line EGFR+ lung. Brad dismisses bears claiming the US cohort was too immature to matter: “people are just looking for excuses,” the same reflex seen in the early days of Carvykti.
Brad’s caution on the PD-1xVEGF class stands: there’s still “no real conclusive hypothesis” why the bispecific beats the Roche combination, and nobody has hit OS yet. He’s “in the camp that this is all going to succeed in the end,” but the risk is legitimate; Eric counters that differentiation is visible—squamous activity, no expected systemic VEGF-inhibition side effects, head-to-head pembro wins—and “over time, data will win.”
Bristol’s partnership takes some risk away from BioNTech’s PD-L1xVEGF asset (~$3.5B “pretty much guaranteed”), while 3SBio’s ASCO poster looked “one of the best out there”—with a safety asterisk. ORR eased from ~70% to 67.5% as n grew to 34, with ~100% disease control and deepening spider-plot responses. The flag was 17% grade-3 hypertension, with no proteinuria, versus ~5% grade-3 hypertension in the much larger Avastin dataset. These molecules are different and are not all doing the same thing.
Vera’s phase 3 atacicept (placebo-adjusted proteinuria cuts in the low 40s, filing by year-end) ran straight into Otsuka’s APRIL-only sibeprenlimab days later—but Eric sees a $15B-plus IgAN market with room for both. He disputes STAT News’ claim that Otsuka’s data were better, arguing baseline confounders and noting that Vera has kidney-function data that Otsuka so far lacks. Sam notes that competitive data in a hot nephrology market make you imagine the asset in the hands of a much bigger pharmaceutical company.
Protein degradation is the “first inning” technology Brad wants investors excited about. Kymera’s STAT6 degrader hit 90% degradation from 1.5 mg upward, with doses reaching 200 mg, clean safety and Th2 biomarker drops (TARC, eotaxin-3)—an oral once-daily Dupixent challenger with atopic-dermatitis patient data due in Q4. Arvinas’ first-ever PROTAC phase 3 was technically successful even if the market viewed it as disappointing. The same day, Nurix struck a STAT6 deal with Sanofi; Nurix will have BTK data at EHA after BeiGene presented BTK data at the last ASH.
ASCO had no mega-headline but “innovation kind of creeps up on you”: AstraZeneca’s seventh straight plenary year (MATTERHORN’s “unprecedented” gastric EFS, first-line Enhertu), Gilead’s best ASCO in years (Trodelvy ASCENT-04 PFS HR 0.65 in first-line PD-L1+ TNBC, Kite at ~$2B revenue), PRMT5 posting ~25% response in pancreatic cancer, and Immatics’ PRAME TCR at 55% ORR in refractory melanoma.
The recurring investor error, per Eric: “we always like to think it’s a winner-take-all battle—that’s rarely how these markets play out.” He applies it to Merus (~$4B+ cap) versus Bicara (trading a little above cash) in frontline head-and-neck, where both post 50–60% ORR on Keytruda, and to Fc-silent TIGIT, where AstraZeneca is running 10 phase 3s across ~8,000 patients. AstraZeneca’s GEMINI biliary-tract data combined rilvegostomig with chemotherapy in HER2-negative disease and showed a 31% ORR. On policy, FDA signals are positive (“talk is cheap,” though), while MFN pricing and the Section 232 tariff probe remain the real overhangs.

Deep dive

1. Sanofi–Blueprint: how a biotech story is supposed to end

Eric’s victory lap, earned—he took Blueprint public ~10 years ago at $18 when “systemic mastocytosis sounded like Greek to me”: the team (Jeff Albers, then Kate Haviland, chair Alexis Borisy) found patients in need, executed “to a T,” and built almost a 10-bagger. “This is kind of the way that biotech success stories really should play out.”
Sam’s answer to Adam Feuerstein’s lament about losing another future Amgen: “we seem in the biotech world to want our cake and eat it”—small deals “don’t count,” big ones kill the next Amgen. For Sanofi it fits rare disease and immunology ambitions; if consensus is right on a ~$2B drug by 2030, $9.1B is “not ridiculously stretched”—though no deal this size is a game-changer for large pharma. Eric adds that the deal looks mature and well timed given Sanofi’s history of paying heavily for risky bets and the depressed valuations of even good companies.
The terms puzzle Sam: ~27% premium, $129/share plus a $2 development CVR and $4 regulatory CVR that he assumes are mostly achievable given their size, so “you almost wonder why they’re even there.”

2. Summit/Akeso: the translation box is checked, the OS miss got the blame

Eric’s framing: ivonescimab is “the granddaddy of them all” in PD-1xVEGF, having beaten pembrolizumab head-to-head twice—and the single biggest, most reasonable bear case was whether China data translate to the US. HARMONi-A, the first trial with a US cohort, delivered a 0.52 PFS hazard ratio, with management “pretty adamant” that the geographic subsets were consistent.
Yet the stock fell: OS came in at p=0.057 with all alpha spent—any future crossing of 0.05 is nominal only—and the company conceded it may not be able to file in second-line EGFR+ lung. Eric’s challenge to Brad: an OS hazard ratio “anywhere in the 0.8 range” is a strong trend behind very strong PFS—“please tell me what I’m missing.”
What’s missing, per Brad: a conclusive hypothesis for why the bispecific beats the two-drug approach Roche never made work, plus the old worry that VEGF may shrink tumors early but “accelerate mets over the long term.” He’s “in the camp that this is all going to succeed in the end,” but with J&J’s regimen already approved post-EGFR—itself without an OS signal—commercial value there was never the point; the cross-read was.
Brad’s swipe at the bears arguing the US cohort was too immature to move the hazard ratio: enrollment closed in August after roughly two years open, so plenty of US patients are likely counted in PFS even if OS events lag. “People are just looking for excuses”—the same reflex as in the early days of Carvykti.

3. Bristol–BioNTech and the 3SBio poster: reading the rest of the field

A co-host’s long-standing worry about BioNTech’s molecule (and Instil Bio’s): poll the conference and PD-1 beats PD-L1 on history, and there is a risk PD-L1xVEGF “ends up going down the road of other PD-L1s.” Bristol’s ~$3.5B “pretty much guaranteed” takes some of that risk away—and even enables a hypothetically logical combo, PD-L1 as the tumor-targeting arm plus a PD-1 doing its job on the T cell.
Eric’s synthesis: there’s a lot we don’t know, but differentiation keeps showing up—no systemic VEGF-type side effects, squamous-tumor activity, head-to-head pembro wins, and activity in second-line EGFR+ where PD-1s would not be expected to work. “Buckle up” for Summit’s volatility, but “over time, data will win.”
A co-host’s read of 3SBio’s ASCO update: ORR eased from ~70% to 67.5% as patients went 24→25→34 (March cutoff, responses may deepen), disease control remained near 100%, and spider plots showed deepening responses—suggesting this may not be just a first-scan effect. The flag was 17% grade-3 hypertension, with no proteinuria, versus ~5% grade-3 hypertension in the much larger Avastin dataset. “These molecules are different, they’re not all doing the same thing.”

4. Vera vs Otsuka in IgAN: a $15B market with room for two

“What a round trip” for Vera, in Eric’s telling: Monday’s phase 3 of atacicept, a BAFF/APRIL inhibitor, showed placebo-adjusted proteinuria reductions in the low 40s, supporting a filing toward year-end—then Otsuka’s APRIL-only sibeprenlimab hit the tape with, if anything, slightly larger absolute proteinuria drops. He rejects STAT News’ “Otsuka is better” call: baseline variables confound it, and what doctors actually care about is slowing kidney-function decline—data Vera has and Otsuka so far does not.
The through-line Eric keeps returning to: “there is no usual winner-take-all phenomenon in biotech”—this is a $15B-plus opportunity and both will be quickly and widely embraced. Sam, while careful about M&A talk, notes that competitive data in nephrology—a hot deal area for larger companies—make you imagine the drug in the hands of a much bigger pharmaceutical company.

5. Protein degradation: the next technology to get excited about, “in the first inning”

Brad’s thesis: biotech always needs a new technology, and degradation is it. Arvinas delivered the first-ever PROTAC phase 3 readout—technically a successful phase 3 trial for 40% of this breast-cancer group, though the market called it disappointing because the ER degrader worked only in a subset. Then Kymera’s STAT6 healthy-volunteer data: 90% degradation from 1.5 mg upward, with doses reaching 200 mg, clean safety, and reductions in Th2 biomarkers (TARC, eotaxin-3)—the pitch being a once-daily oral Dupixent competitor, with atopic-dermatitis patient data due toward the end of Q4 and asthma next.
The tape confirmed the theme: the same day, Nurix struck a STAT6 deal with Sanofi. Nurix will have BTK data at EHA next week, after BeiGene presented BTK data at the last ASH. A co-host’s endorsement was that degradation opens “a whole new set of targets” not amenable to standard enzyme inhibition—“a cornerstone of future drug development.”

6. ASCO’s verdict: no mega-headline, but “innovation kind of creeps up on you”

Eric’s calibration: a down year only in the absence of surprise gamechangers. In pancreatic cancer and melanoma, where 7–12% response rates were once celebrated, “we’re kind of sticking our nose up at” 20–40%. PRMT5, led by Bristol with Amgen and perhaps Tango also emerging, posted ~25% response in pancreatic cancer—he wants to see it combined with KRAS inhibitors—and Immatics’ PRAME-directed TCR therapy hit 55% ORR in refractory melanoma, with follow-up beyond a year and KOLs “raving.”
A co-host called it another AstraZeneca ASCO—seven straight plenary years, “lots of clapping” if no standing ovation: MATTERHORN’s “really unprecedented” event-free survival in gastric cancer and DESTINY-Breast09 opening a significant first-line HER2+ opportunity for Enhertu. Gilead’s Trodelvy in ASCENT-04 (first-line PD-L1+ TNBC with Keytruda, PFS HR 0.65) stacked on ASCENT-03 and suggested a meaningful franchise. The bigger point: “there are now people with myeloma or breast cancer who live long enough to die of something else—if you can’t call that cure, I don’t know what you’re looking for.”
A co-host’s Gilead coda: this was the company’s biggest ASCO in years, Kite is quietly at ~$2B in revenue, and the $20B+ Trodelvy roller coaster—through trial setbacks and program changes—is an argument for scale: “I don’t think a smaller company could have done that.”

7. TIGIT’s second act and Arcus’ quiet pivot to HIF-2α

After Roche’s SKYSCRAPER miss and a wave of divestments (BeiGene recently; iTeos literally closing shop about a week ago), only AstraZeneca and Arcus are full speed ahead, and both programs are Fc-silent. A co-host’s simple logic: “why would I want antibody-directed cytotoxicity against the T cell?”
AstraZeneca is “really putting their money where their mouth is”: 10 phase 3s, ~8,000 patients—five in NSCLC, two biliary tract, plus HCC, endometrial and gastric—nearly all rilvegostomig combinations, many with Dato-DXd. Signals cited: 71% ORR in PD-L1-high lung and 40% in negative patients, surpassing the rilvegostomig monotherapy results, though 53% stomatitis “really upsets people.” GEMINI in biliary tract cancer combined rilvegostomig with chemotherapy in HER2-negative disease and showed a 31% ORR. Readouts are expected in 2026 or 2027.
Brad’s Arcus call: HIF-2α—where data look “slightly better than Merck right now” against a validated, marketed target—has displaced TIGIT as the company’s main program. Like IgAN, the market is big enough that second-and-comparable can still win. Arcus trades a little above its cash balance.

8. Bicara vs Merus in head-and-neck—and an FDA saying the right things

Eric’s coverage view (Bicara only): frontline head-and-neck is a $3–4B market where both molecules post 50–60% ORR on top of Keytruda—roughly 3x pembro alone—and physicians are far more even-handed than the Street, which has handed Merus ($4B+ cap) winner-take-all status while Bicara trades a little above cash. Bicara’s HPV-negative selection, he argues, has real mechanistic rationale via the TGF-β arm: “markets have a way of evening out over time.”
Sam’s pushback: Merus appears to work in both HPV-positive and HPV-negative disease, and the company’s apparent abandonment of HPV+ may say something about the asset’s strength. Separately, credit where due—Bicara CEO Claire Mazumdar faced a video interview amid a falling stock and tough analyst comparisons: “that’s what true leaders do.”
The policy close: Eric, fresh off hosting John Crowley on a Cantor webinar, hears positive signals from the FDA—Makary’s pro-innovation, pro-efficiency comments and the cell-and-gene-therapy meeting—but “talk is cheap,” and staffing concerns remain. A co-host also noted RFK’s comment on the base-editing baby who went home. The real fight is MFN drug pricing, which “has not progressed as favorably,” with the Section 232 tariff investigation potentially hanging over the sector for several months. Sam’s parting frame: a week of two deals, strong data and the XBI meaningfully up—don’t mistake the hosts’ exhaustion for gloom.

Goodfire AI’s Bet: Interpretability as the Next Frontier of Model Design — Myra Deng & Mark Bissell SpaceX Buys xAI: What's Behind Elon’s Mega-Merger?