Episode 199 - October 16, 2026
Summary
- Index strength masks weakness in many underlying holdings and a historic de-grossing episode. XBI remained up 20% YTD but had fallen to roughly 146 from an August high of 169, while XLV accelerated from 7% to 15% YTD. Goldman prime-broker data showed hedge funds unwinding biotech risk in six of eight weeks—the largest cumulative de-grossing in more than five years.
- The panel sees fund flows and psychology, not a sector-wide valuation bubble, as the immediate problem. Crowded longs are being sold while crowded shorts are covered, creating a self-reinforcing cycle in which nobody wants to “catch a falling knife” that could drop another 10%-15%. Skorney’s valuation test: markets are far from the days when investors could “IPO a preclinical idea on a napkin for a billion dollars.”
- M&A remains the clearest potential reset, but rates and fund calendars may need to clear first. Matteis argued that acquisitions could provide the external validation and FOMO needed to restart buying; Werber pointed to many funds’ October 31 annual close, the midterms and another rate increase—possibly two, with at least one by year-end—as remaining overhangs. Meanwhile, strong ETF performance is amplifying the pain for specialist funds whose underlying holdings have lagged.
- The IPO window remains open, but it increasingly rewards clinical catalysts and differentiated technology. TRex Bio raised $117 million, Iambic Therapeutics $150 million and City Therapeutics an upsized $200 million, bringing the unofficial 2026 biotech count to about 28. Becraft’s filter was explicit: investors want an IP and execution moat plus clinical evidence that can de-risk it.
- Bristol’s admilparant Phase III supports LPA1 in IPF, but its dose response left the class only partially de-risked. The low dose reached statistical significance while the high dose did not; the panel still saw a likely approval and potentially large market because better tolerability may permit combinations. Contineum’s stock moved from up 20% premarket to down more than 25% as investors questioned whether greater receptor coverage will create differentiation before its 2028 readout.
- Immunology produced both platform validation and a warning against extrapolating Phase II biology. Mirum’s Livmarli generated consistent pruritus reductions across additional pediatric cholestatic diseases, while three competing celiac programs all showed early histologic and symptomatic activity. By contrast, argenx terminated Vyvgart in Sjögren’s after a Phase III futility analysis, reinforcing Werber’s view that “it’s not one specific factor that drives the biology.”
- CNS delivery and cancer vaccines supplied the week’s highest-upside—but still unresolved—technology readouts. Arrowhead achieved more than 80% CSF tau lowering with a transferrin siRNA drug conjugate, though CSF reductions may not predict tau-PET or clinical benefit. Moderna/Merck’s forthcoming ESMO data have put “mRNA back on the menu,” but Becraft stressed the absence of a nonspecific-mRNA control and the need to prove benefit beyond resected melanoma.
Deep dive
1. Forced de-grossing has turned crowded positioning into a vicious cycle
Suvannavejh set the contradiction: XBI was still up 20% YTD, but at roughly 146 it had surrendered about 700 basis points since the prior episode and sat well below its late-August high of 169. XLV, meanwhile, had accelerated from 7% to 15%, suggesting rotation from biotech into other healthcare subsectors.
Skorney defined de-grossing as a long-short fund shrinking total exposure—for example, taking a $2 billion book to $1.5 billion by selling longs and covering shorts, not necessarily for a fundamental reason. Crowded specialist positioning makes the mechanical effect unusually violent.
When one fund reduces risk, its crowded longs fall and shorts rise, damaging similar portfolios and prompting higher-level risk managers to cut their own gross exposure. Even long-only investors can join the selling, while prospective buyers wait rather than “catch a falling knife” with another 20%-30% downside.
The pressure had felt two weeks from ending since June, Skorney joked, like the construction crew in The Money Pit. A Goldman prime-broker chart sharpened the point: hedge funds had unwound biotech risk in six of eight weeks, the largest cumulative de-grossing in more than five years.
2. Cheapness alone cannot break a market governed by fear and fund flows
Matteis described a behavioral stalemate: investors liked holdings 30% higher and now see attractive catalysts, but fear buying before another flow-driven 10%-15% decline. “People are waiting for others to make the first move” because nobody knows when the carnage stops.
M&A could create that first move by validating that quality companies have become bargains and triggering FOMO. Matteis called it a cliché that acquisitions cure biotech bear markets, but saw no comparably obvious catalyst beyond deals or simply allowing time to pass.
Werber saw possible pressure relief after the October 31 annual close for many funds, followed by the midterms, but emphasized interest rates as a recurring source of sector stress. He said another rate increase was likely, possibly even two, and certainly expected one by year-end. Generalists were also asking whether a Democratic sweep might bring AI regulation and potentially improve biotech sentiment; speculation around a $50 billion Isomorphic Labs valuation intensified that debate.
Skorney judged stocks cheaper, though not 2023-cheap, and “very far from a bubble”: the market can no longer “IPO a preclinical idea on a napkin for a billion dollars.” Yet value without a near-term catalyst attracts little sponsorship, while Werber noted that ETF-driven trading is largely valuation-insensitive and good data can still leave a stock flat or down.
Werber also noted that the XBI’s 25 largest stocks were performing well, partly driven by Twist, an AI-driven diagnostics play, even as many underlying investments were not. That left fund managers facing LP questions about why their performance did not match the strong-looking sector index.
3. A functioning IPO market is separating differentiated companies from tourists
TRex Bio raised $117 million around a Phase I Treg program and partnerships with Lilly and J&J. AI-led oncology company Iambic Therapeutics raised $150 million with a Phase I HER2-mutated cancer program and multiple pharmaceutical partnerships.
City Therapeutics upsized its offering to $200 million, bringing Suvannavejh’s unofficial 2026 biotech IPO count to roughly 28 and putting the market on pace for 30-plus. City has partnered preclinical ophthalmology and neurology candidates with Bausch + Lomb and Biogen, respectively, alongside a proprietary Phase I thromboembolic-disease candidate. A new Frazier Life Sciences SPAC also debuted, even as announced strategic M&A had quieted.
Matteis said July and August private-company outreach had started to feel like 2020: very early businesses were asking “why not us?” and considering reverse mergers. Six weeks later, those calls had slowed as the public-market backdrop changed abruptly.
Werber divided the remaining pipeline into hot companies with new data, companies already too far along to stop, and capital-needy issuers forced to test the window. Becraft’s common denominator was a “flight to both quality and innovation”: technological differentiation, defensible IP, an execution moat and a near-term clinical catalyst—not merely an interesting platform.
4. Bristol validated LPA1 in IPF without settling the dose-response debate
Matteis framed Bristol’s admilparant Phase III as positive for the company but scientifically awkward: the low dose reached statistical significance and the high dose did not. Given IPF’s unmet need, he found approval difficult to rule out, but Bristol released no hard clinical data; full judgment waits for a likely spring medical meeting.
The commercial opportunity could be substantial: IPF drugs could potentially sell billions in annual revenue, yet current treatments leave patients progressing and can be difficult to tolerate. An LPA1 drug that is better tolerated could be combined with other agents, where current options such as pirfenidone and nintedanib are hard to stack.
Contineum initially rose 20% premarket, then finished down more than 25%. Its thesis is that greater receptor coverage may improve efficacy, but Bristol’s apparent “inverse dose response”—possibly just a flat response or noise—made investors question that differentiation while Contineum’s own trial will not read out until 2028. Matteis said its once-daily dosing could at least offer an advantage.
Skorney cautioned that IPF’s pathophysiology remains poorly understood and Phase II-to-III translation has repeatedly failed. His pirfenidone regulatory “battle scars,” plus Werber’s recollection of Galapagos’ autotaxin program and its mortality imbalance in two Phase III studies, argued against calling the class fully de-risked. Unlike cystic fibrosis, IPF lacks a comparably clear genetic and pharmacologic map.
5. Mirum expanded a proven symptom franchise as celiac opened a new battlefield
Mirum’s Livmarli produced consistent placebo-adjusted pruritus reductions in a Phase III basket of pediatric cholestatic diseases, led by biliary atresia. Skorney estimated about 2,000 addressable U.S./EU biliary-atresia patients and potentially roughly another 2,000 across the smaller subgroups.
Skorney said the broader evidence is that an IL-10 inhibitor can reduce cholestatic pruritus. Mirum also has a pivotal PBC readout for adult candidate volixibat in the coming months and plans a PSC NDA in the first half of next year, where he said no therapy is approved.
Werber called celiac disease “the best of biotech”: Teva targets the IL-15 ligand, while argenx—which bought Forte Biosciences—and First Tracks target CD122, part of the IL-2 receptor. Small studies across all three showed reductions in intraepithelial lymphocytes, histologic improvement and symptom benefit, setting up a “coming-out party” at November’s International Celiac Disease Symposium in Melbourne. Teva’s drug is administered subcutaneously every three months.
The biology may be cleaner than the market. No approval pathway exists, trial designs use different gluten challenges, and developers must resolve whether histology or symptoms become the endpoint, how long to treat and what payers will spend versus a gluten-free diet. Bulls see a huge market; bears think “no one’s gonna pay for it.” Early vitiligo activity, forthcoming argenx alopecia data and First Tracks’ EoE study broaden the opportunity.
6. Argenx’s Sjögren’s failure exposed the limits of single-pathway reasoning
Werber had just finished a bullish physician panel when argenx terminated Vyvgart in Sjögren’s after a futility analysis—an analysis the panelists had not known about. The surprise followed encouraging Phase II data and sent the stock lower, even though the group still viewed Sjögren’s as a potentially large market.
His biological lesson was that inflammation often depends on B- and T-cell interaction rather than one isolated factor. Vyvgart’s autoantibody reduction failed in Phase III, while Amgen’s CD40 ligand approach, which regulates that interaction, had moved from strong Phase II evidence to positive Phase III results. The panel was also positive on the BAFF/APRIL approach, with Novartis’ drug viewed as weaker than Vera’s.
Matteis noted that argenx was roughly flat year over year while Alnylam and Insmed had fallen sharply. Werber attributed the resilience to strong execution, expected Q3/Q4 performance and pervasive generalist ownership—“literally everybody owns it”—with upcoming MMN data for its C2 antagonist helping determine whether the opportunity is second-line through non-inferiority to IVIG or front-line through superiority.
7. Transferrin shuttles are becoming a platform, but tau remains the real test
Arrowhead’s transferrin siRNA drug conjugate achieved more than 80% tau lowering in CSF, making it the first to show clinical data for an siRNA drug conjugate using transferrin delivery. Matteis found the execution notable because Arrowhead had less transferrin experience than Roche or Denali.
The central caveat is compartmental: CSF tau may not represent pathology in the relevant brain regions, so tau-PET reduction still matters. Biogen previously showed much larger CSF than PET effects; its intrathecal Phase IIb produced clinical benefit at the lowest dose, alongside an inverse dose response and toxicity at the higher dose.
Matteis suspected Biogen’s toxicity was idiosyncratic to the ASO because tau lowering was similar across doses, but conceded that only next-generation programs from Arrowhead, Denali or others can answer the question. Transferrin is beginning to feel like GalNAc in the liver—“I don’t want to say it’s commoditized”—with multiple groups solving the science simultaneously.
Suvannavejh added Alector’s ABC (Alector Brain Carrier) platform: its lead beta-amyloid-targeting antibody is AL-137, and its tau-siRNA program has shown roughly 75%-80% preclinical mRNA knockdown but will not enter the clinic until next year or possibly 2028. The company raised $75 million after its R&D day, following a pivot after late-stage clinical assets failed.
8. Oncology enthusiasm returned, but controls still matter
Suvannavejh highlighted “very, very excellent” early Phase I and II data for Verastem’s mutant-selective KRAS G12D inhibitor across cancers, particularly pancreatic cancer. The readout joined sustained interest in Revolution Medicines’ pan-RAS inhibitor, which had just received a positive CHMP opinion.
Becraft described Moderna/Merck as the first potential major cancer-vaccine breakthrough, with apparent progression-free-survival benefit in melanoma. But the initial setting involves fully resectable tumors, and BioNTech’s cancellation of monotherapy trials in colorectal and head-and-neck cancers cautioned against extrapolating across tumors.
His core pushback was trial design: the study was controlled against placebo, not a nonspecific mRNA vaccine, despite growing evidence that mRNA—including COVID vaccines—may stimulate T cells independently of selected cancer antigens. “I would have loved to see” that control before crediting an AI-derived antigen algorithm.
The ESMO late-breaker would provide the full picture. Becraft, disclosing that he runs an mRNA therapeutics company rather than a vaccine company, said “mRNA is back on the menu again,” while stressing that the larger ambition—moving from vaccines and therapeutic vaccines into successful mRNA therapeutics—remained unfulfilled after roughly a decade of promise.