Episode 193 - August 14, 2026

Summary

• Tess Cameron’s read on the tape: “healthy but discerning,” with “rational responses to data” — strong IPOs from strong companies, and dramatic punishment for negative updates. Graig Suvannavejh counts 19 Nasdaq issues year-to-date with the XBI still outperforming the broader market, but says the generalist investor is not fully back and “biotech is always a nervous type of a market.”
• Paul Matteis’s caveat: good vibes haven’t meant everyone made money — this is “an inefficient stock picker’s market,” not the usual everything-rally. Market-neutral funds that short liquid, catalyst-light names “to fund their longs” were challenged as heavily shorted stocks like Crinetics and Apogee were taken out. “Usually… it’s like everyone’s just making a ton of money. I don’t feel like that’s the case.”
• A Wall Street Journal piece Tess flagged — “Healthcare Investing Is Now an AI Short in Disguise” — reframes the flow-of-funds debate: healthcare isn’t cyclical like AI picks-and-shovels, and may be trading as an AI hedge. Forward growth rates have diverged so significantly that “healthcare looks cheap on a relative basis,” and “people are going to need medicines.”
• Two high-profile failures tested the market’s discipline: Tenex missed six-minute walk and KCCQ in HFpEF with a possible harm signal in healthier patients, and Sionna’s CF NBD1 corrector produced a 1 mmol sweat-chloride reduction versus a guided ≥10, sending the stock down ~90%. Paul’s confession: “I really thought this was going to work… this industry always really humbles you,” with a possible Trikafta drug-drug interaction as a potential explanation — but only 14 patients to interrogate.
• Jazz bought epilepsy-focused Actio for ~$1.3B ($820M upfront), extending a franchise built on the once-controversial GW Pharma deal, which Graig thought was around 2021 and perhaps ~$7B, whose Epidiolex he believes now does >$1B a year. Paul’s caution on targeted epilepsy plays like Actio’s KCNT1 program: “I love the therapeutic thesis, but the market thesis is often a little bit trickier to diligence” — he thinks Epidiolex is “probably less than $100 million” in Dravet itself.
• Psychedelics look close to a potential approval: MindMed’s Phase 3 LSD study in generalized anxiety succeeded with a very large effect size still intact 12 weeks after a single day-one dose, supporting quarterly-or-less dosing. Paul: “even as people are kinda jaded to it, [it] still sorta blows me away” — and the trial’s exclusion of comorbid depression addresses the FDA’s pseudospecificity question. With Lilly buying atai/Beckley’s 5-MeO-DMT, “it’s kinda hilarious to think about reticence around CBD five years ago.”
• Post-acquisition shoes kept dropping: seven deaths on Neurocrine’s Vykat XR were surfaced by the Foundation for Prader-Willi Research, and EMA pulled Amgen’s Tavneos after ChemoCentryx’s “incorrect and misleading pivotal data” — selective unblinding and re-adjudicating ~9 patients to rescue the p-value. Paul still sees Vykat as “a viable commercial product” (~$400M this year, hoped $500–600M) given prior noise was priced in; Chris Garabedian’s verdict on Tavneos: “it just doesn’t look good for our industry.”
• Belite Bio’s tinlarebant got a priority-review PDUFA of Feb 12, 2027 — potentially the first-ever Stargardt drug — just as Tarsus paid ~$800M for rival Alkeus, whose DRAGON Phase 3 won’t read out until 2H 2029 and could now face enrollment against an approved competitor. Separately, Tess’s bottom line from the Endpoints China debate: U.S.-competitiveness concerns are real, but a broad-based ban is “probably not the right legislative approach.”

Deep dive

1. Healthy but discerning — and not everyone is getting paid

• Tess’s framing of the public markets: the XBI is performing, IPOs are working — but only because “the companies that are going public are really strong companies with really strong data,” while negative updates get sold “pretty dramatically.” Her summary phrase: “healthy but discerning,” underwritten by “rational responses to data.” Chris’s gloss: this kind of tape favors decades-long specialty investors over tourists.
• Paul’s dissent from the good-vibes narrative: sentiment and fundraising look fine and nothing feels “super overheated,” yet “it hasn’t been a great year for everybody.” Funds with market-neutral or percent-short mandates habitually short liquid, catalyst-light names “to fund their longs” — and some of this year’s big takeouts involved names he viewed as heavily shorted, including Crinetics and Apogee, making that model more challenging. “Usually, when we’ve got an open IPO window, a big fundraising window… everyone’s just making a ton of money. I don’t feel like that’s the case.”
• Graig’s tally: 19 Nasdaq issues year-to-date by his count, biotech “the darling within healthcare,” XBI outperforming the broader market — but the generalist investor is not fully back, and “biotech is always a nervous type of a market.”

2. Healthcare as an AI short in disguise

• Tess loved the WSJ headline “Healthcare Investing Is Now an AI Short in Disguise.” The year-and-a-half flow-of-funds worry — is everyone selling Lilly to buy AI? — gets inverted: the AI winners are cyclical picks-and-shovels infrastructure names, healthcare is not cyclical, and healthcare stocks have done “a bit better when AI is doing poorly.” Her kicker: we still don’t know AI’s revenue curve, “and people are going to need medicines, right?”
• Chris’s historical contrast: unlike 25 years ago, when the dot-com and genomics bubbles were lumped together, today’s IPO market is mature — watch whether generalists rotate into biotech as tech gets “overheated and frothy.” Tess adds that forward growth rates have diverged significantly, so “healthcare looks cheap on a relative basis.”

3. Venture is thriving late-stage; failures feed a reverse-merger frenzy

• Chris’s week-in-venture: Boulevard Bio, a Deerfield spin-out company creation with substantial capital; Epic Bio’s $90M Series C for FSHD via epigenetic silencing of DUX4; and a Series B funding a Phase 3 for an allosteric AKT1/2 inhibitor in HHT. But Q2 seed and Series A counts fell — he’s watching for early-stage risk tolerance to return next year.
• The structural shift he flags: every failure of a single-product binary biotech now triggers “almost a frenzy of like, ‘Okay, are they gonna be the next reverse merger candidate?’” Public funds carrying “PTSD” from crossover rounds that trapped them when markets turned now prefer direct IPOs or PIPE reverse mergers over mezzanine rounds — two such deals were announced Monday, and Perceptive separately backed a newco to give new life to Ovid’s soticlestat.

4. Tenex and Sionna: the industry “always really humbles you”

• Tenex missed both the six-minute-walk primary and the KCCQ secondary in HFpEF — “notoriously very, very challenging” but with “enormous unmet need,” per Tess. The NT-proBNP reduction offers a mechanistic consolation, but a treatment-effect-by-severity pattern raised the uncomfortable question of whether the drug might be “kind of harmful” in healthier patients. Path forward: Type C meeting with FDA, EMA discussions, maybe enriching for severe patients — but “it certainly doesn’t seem like a straightforward path forward.”
• Sionna hurt more because Paul believed: “I really thought this was going to work… this industry always like really humbles you.” The clever design — an NBD1 corrector layered onto Trikafta inadequate responders, using sweat chloride, the field’s newborn-screening diagnostic and “pretty uncontroversial biomarker” — guided to ≥10 mmol; his team’s own PK/PD recreation suggested 10 could even be conservative. Result: 1 mmol, far from stat-sig, stock down 90%.
• The autopsy: a possible drug-drug interaction lowering Trikafta exposures “could easily erase a pharmacodynamic effect,” but with 14 patients they may never find “an unequivocal explanation” — and the CF HBE assay Sionna thought it had cracked à la Vertex “ultimately did not translate.” The lingering street debate stands: “Vertex has set such a high bar… how much more do we really need?”

5. Jazz–Actio: great therapeutic thesis, trickier market thesis

• Graig’s deal recap: ~$1.3B total ($820M upfront, $500M in regulatory/commercial milestones) to beef up an epilepsy franchise built on the GW Pharma acquisition, which he thought was around 2021 and perhaps ~$7B — “a very controversial drug at the time” that became a >$1B/year product in Epidiolex, with rare epilepsies offering better pricing.
• Paul’s two-sided take: targeted, genetically oriented neuro therapy is “theoretically very high interest” but brutal to execute in neuropsych or Alzheimer’s; epilepsy’s simpler biology is delivering “pretty great data” (Stoke and others). The rub is prevalence: from covering GW, his perception is Epidiolex is “probably less than $100 million” in Dravet, with most use in broad refractory epilepsy. On Actio’s KCNT1 program — they also have a SHANK3 program — “I love the therapeutic thesis, but the market thesis is often a little bit trickier to diligence.”

6. Financing off data — and RevMed partners with Zai Lab for China access

• Graig’s under-covered financings: AbCellera, a Canadian company listed on Nasdaq with an approximately $3.5B market cap, posted positive Phase 2 data in menopausal vasomotor symptoms and upsized a $175M follow-on to $200M — rare biotech innovation in women’s health since Veozah. Silence Therapeutics reported competitive Phase 2 siRNA data in polycythemia vera — a two-approved-drug market where Takeda/Protagonist have a Q3 PDUFA — and upsized $150M to $175M. Both, he says, “a good positive sign for the health in the biotech sector.”
• Tess on the RevMed deal with Zai Lab: a reversal of the usual China-to-U.S. flow — daraxonrasib, zoldonrasib, elironrasib and the G12V program supporting commercialization and development in Asia, plus a clinical collaboration around Boehringer Ingelheim’s PRMT5 and an EGFR/MET trispecific. No economics disclosed, which makes her wonder “which one of those came first, the PRMT5 or the China commercialization?”
• Her answer to Chris’s why-not-go-alone question: “China commercialization is really hard, right? Really hard” — NRDL pricing dynamics and the need for a large field force. RevMed could raise the money and hire the people “no problem,” but “do they necessarily wanna be focusing management time and attention there? Maybe not.”

7. MindMed’s LSD Phase 3, Cullinan’s lung-cancer win, MoonLake’s bimekizumab problem

• MindMed’s proprietary LSD formulation Phase 3 study in generalized anxiety succeeded with a very large effect size — still intact 12 weeks after day-one dosing in a placebo-controlled trial, supporting once-quarterly or less-frequent dosing. Paul: “even as people are kinda jaded to it, [it] still sorta blows me away.”
• The design detail he thinks is the real tell: MindMed went to lengths to enroll significant anxiety without comorbid depression — deliberately de-enriching the population — because FDA, having seen enough psychedelic depression data, wants to get at the question of pseudospecificity (as the psych division demands in schizophrenia cognition claims). With three positive large placebo-controlled studies (two anxiety, one depression) and one Phase 3 to go, “they could be looking at an approved product in the not so distant future.”
• The psychedelics arc in one exchange: GW reportedly had a thin buyer list over CBD stigma — “it’s kinda hilarious to think about reticence around CBD five years ago” now that Lilly, “the big dog,” is buying atai/Beckley’s 5-MeO-DMT. Graig: “it does take some time, but I think the time is now.”
• Elsewhere: Cullinan/Taiho posted a positive first-line Phase 3 in lung cancer on top of a filed second-line program (February PDUFA), in a space where J&J’s combo products are tracking above $1B this year. MoonLake’s IL-17 readout in biologic-naive psoriatic arthritis showed positive ACR50, ACR20, and PASI90 results — yet the stock fell, which Tess reads as the market asking “how is this gonna compete against bimekizumab?” Differentiation, not efficacy, is now the critical emphasis.

8. Shoes dropping post-M&A: Vykat deaths, Tavneos withdrawal

• The Vykat XR news came not from Neurocrine but from the Foundation for Prader-Willi Research, whose clinical recommendations for real-world use identified seven deaths that Graig did not think had previously been appreciated — he reads it as “a PSA,” not a pull-the-drug campaign, but alarms nonetheless.
• Paul’s stock math: Neurocrine paid a bit under $3B for an already-profitable Soleno, with reports of similar issues dating back about a year and a STAT News article by Adam Feuerstein also covering the matter, so expectations “aren’t huge” — ~$400M this year, hoped $500–600M. He concedes the efficacy data is “not overwhelmingly convincing” — the best study is a randomized-withdrawal trial, a design “better enriched for a larger effect size” — but the disease is severe enough that it stays “a viable commercial product… unless this is just the beginning of a lot more noise.”
• Amgen’s Tavneos was pulled by EMA after FDA urged U.S. withdrawal, which Amgen initially resisted. Per press reports Chris cites, ChemoCentryx provided “incorrect and misleading pivotal data”: selected team members were unblinded, saw the p-value missing, and re-adjudicated ~9 patients as responders to make it positive. “It just doesn’t look good for our industry.”
• On the broader shoe-drop pattern (GBT/Pfizer being another): Chris puts the onus on acquirer diligence — “short of anybody being fraudulent” — while Paul says “it’s a capitalist system”; these cases are “a black eye” but a “small minority.”

9. A Stargardt showdown, Axsome’s launch, and the China-ban debate

• Graig’s setup: Belite Bio’s tinlarebant NDA was accepted with priority review, breakthrough and orphan status — PDUFA February 12, 2027 — potentially the first-ever drug for Stargardt disease (~40,000 U.S. patients; the gene involved is described as too large for current gene-therapy approaches, “almost like a DMD story”). The prior week, Tarsus bought Belite’s biggest rival Alkeus for ~$800M — but Alkeus’s ongoing Phase 3, DRAGON, won’t read out until 2H 2029 and could now face enrollment against an approved competitor.
• Quick adds: Tarsus raised XDEMVY guidance to $685–705M; Axsome’s Auvelity launch in Alzheimer’s agitation is hard to read through IQVIA/Bloomberg given Medicare sampling, but new-to-brand scripts in the 65+ segment are up 126% — “so far so good.”
• The Endpoints debate — Peter Kolchinsky versus Ginkgo’s Jason Kelly, with Fiona Murray providing the cross-sector national-security lens as BIOSECURE’s reintroduction reignites the protectionism fight. Tess’s bottom line: concerns about U.S. biotech competitiveness are “reasonable and valid… let’s take that really seriously,” but a broad-based ban is “probably not the right legislative approach” — Peter kept pressing Jason on enforcement mechanics, including whether FDA could really block “Euro-washed” China-discovered drugs. Chris’s color: “He doesn’t suffer fools gladly,” yet Peter was “very restrained” — and letting Jason talk “kind of helped the argument he was trying to make.”