Summary

Biotech sentiment is back to “winter is coming” or “winter is here,” but Yaron Werber thinks the bottom may still be roughly three months away. His 15-year investor survey swung from record small-cap enthusiasm before the election to sharply weaker sentiment, while the XBI has delivered “exactly five years of nothing” as the S&P 500 doubled. Historically, similarly flat starts have usually led to flat-to-down years.
Financing conditions remain hostile even for credible companies with mature assets and enough IPO proceeds to reach catalysts. Septerna and Upstream Bio fell 50% and 54% respectively in 2025 without news, as investors penalized future capital needs, competition, and distant catalysts. Armistice Capital’s use of IOUs for redemptions added a separate liquidity warning: potentially highly profitable PIPE warrants can become difficult to liquidate when investors want cash.
A new cohort is reaching profitability because companies cannot keep raising money indefinitely, but earnings alone will not command biotech valuations. BeiGene, argenx, Legend, Ascendis, SpringWorks, and eventually Ultragenyx illustrate commercial maturation; Sam Fazeli’s qualification was blunt: nobody wants to own “a profitable biotech company whose profitability is going up at 10% a year.” The winners must still show a significant ramp or reinvest productively in R&D.
Moderna’s 2025 revenue range of $1.5 billion-$2.5 billion leaves its cancer vaccine as the pivotal trajectory-changing asset. The norovirus clinical hold, softer CMV timing, possible regulatory delay for the COVID-flu combination, and GSK’s description of RSV vaccines as a “two-player market” all make that readout more important. Bristol Myers Squibb’s Opdualag adjuvant melanoma failure removes one potential competitive bar but also warns that apparently plausible adjuvant trials can miss.
SpringWorks stood out as a plausible near-term M&A candidate, while differentiated obesity stories still need harder evidence. Werber expected a Merck KGaA transaction around the following week, citing SpringWorks’ rapid Ogsiveo launch and Gomekli’s broader label; Aardvark, by contrast, priced at $16 and immediately traded near $13.70-$13.80. Fazeli could not see an easy acquisition path until Aardvark produces “really hard, good-quality, replicable obesity data.”
FDA policy is a genuine two-sided trade: reform could accelerate development, but blunt staffing cuts could impair reviews for years. Steve Usdin argued that cutting personnel before determining what the agency needs is “kind of backwards,” especially when reviewers take about two years to become productive and most drug-review staffing is funded by industry user fees. The decisive contest is whether success means firing people or delivering more safe and effective medicines efficiently.
NIH presents an even sharper mix of legitimate reform targets and systemic research risk. Usdin agreed that NIH is too risk-averse, funds too much me-too work, and is not producing “$50 billion a year of science” from roughly $50 billion of spending; however, cuts could disrupt collaborations, SBIR funding, infectious-disease work, and the transfer of basic research into products. His one-year forecast was deliberately binary: the industry will likely feel “a lot better or a lot worse,” not merely unchanged.

Deep dive

1. Near-bottom sentiment is returning, but the historical bottoming process is incomplete

Chris Garabedian framed the malaise through Adam Feuerstein’s “Biotech Is in a Dark Place”: roughly 700 public biotechs, about 200 trading below cash, and a provocative proposal to delist three companies before permitting one new IPO. Ginkgo Bioworks co-founder Jason Kelly countered with Lilly and Novo’s value creation, though Chris treated those as exceptions rather than proof of broad health.
Werber’s “sentimentometer” surveys 150-180 investors quarterly, predominantly US sector specialists and more long-short than long-only. “History is a sage adviser, and it’s not always perfect, but it’s a lot more predictive than not”: across 24 years, only three began roughly flat, and those generally finished flat to down.
Sentiment moved from investors hiding in large caps to record small-cap enthusiasm before the election, then reversed violently. Werber now hears, “It’s never been this bad. It’s over. The VC model is broken” — rhetoric that usually signals proximity to a bottom, although outlier sentiment generally takes two quarters to bottom and he estimated perhaps another three months.
Fazeli’s simpler scoreboard was the XBI: “exactly five years of nothing,” versus a doubling of the S&P 500. He viewed the unexpectedly strong headline and core CPI as a January effect, but warned that tariffs and other inflationary pressures might still force an eventual rate increase.

2. Liquidity and post-IPO performance matter more than launch-day demand

Armistice Capital’s redemption IOUs appeared to reflect two portfolios inside one fund: liquid holdings and illiquid warrants accumulated through PIPEs. Werber noted that warrants can be “unbelievably profitable” when deals work, but cannot readily fund redemptions; Chris found the situation striking given approximately 250 public holdings and cautioned against extrapolating one fund into an industry trend.
IPO candidates have improved materially — later-stage programs, more data, stronger pedigrees, and differentiated mechanisms — but their aftermarket performance has been poor. Without company-specific news, Septerna had fallen 50% during 2025 and 60% from its high; Upstream Bio was down 54% for the year and 74% from its high.
Werber’s triumvirate of current investor fears is future financing, competitive exposure, and insufficient catalysts. Even well-funded companies get caught in a self-fulfilling spiral as falling valuations highlight the hundreds of millions they may need over five years: “The good news, there’s a great reception. The question is, what’s the staying power?”

3. Commercial maturation creates cash flow, not automatic growth premiums

Werber sees an “unbelievable amount of innovation” reaching commercial scale: BeiGene and argenx turning profitable in 2025; Legend and Ascendis around late 2025 into 2026; SpringWorks potentially within two years of launch; and Ultragenyx around late 2026-27 after four additional products. The forcing mechanism is straightforward: “Companies can’t continue to raise money forever anymore.”
Chris recalled the older playbook of delaying profitability to preserve a pipeline valuation, with Vertex continually funding R&D as the exemplar. Fazeli’s distinction was growth: argenx can justify ownership through a steep commercial ramp, whereas a biotech growing earnings only 10% annually offers neither classic biotech upside nor compelling reinvestment.
Ascendis’ Yorvipath was launching strongly in the US for hypoparathyroidism after European progress, while TransCon CNP was due for filing that quarter against BioMarin’s Voxzogo. Werber was waiting for AstraZeneca’s phase 3 eneboparatide results imminently but saw that daily injectable as perhaps 18 months behind and “a little bit of a me-too.”
Exelixis had a CABO franchise exceeding $2 billion globally, anticipated upside from neuroendocrine tumors, and generic protection through 2030. Yet its next cycle depends on zanzalintinib: early evidence had not shown ample differentiation from CABO, making acquisition unlikely and upcoming head-and-neck and colorectal data decisive.

4. Moderna’s cancer vaccine must offset weakening franchise optionality

Moderna maintained its recently issued 2025 guidance, but the $1.5 billion-$2.5 billion revenue range remained wide and consensus was still a little ahead of it. Fazeli said 2024 had once looked like the bottom, but “now 2025 could be the bottom,” with volatility amplified by a norovirus-vaccine clinical hold, CMV timing, and uncertainty around the COVID-flu combination’s regulatory path.
The cancer vaccine is the event Fazeli believes “is going to change the trajectory for this company.” Its controlled phase 2 melanoma trial was small, and the Keytruda control arm looked weaker than expected, leaving a nagging question over the larger adjuvant dataset anticipated around year-end or possibly 2026.
Bristol Myers Squibb’s Opdualag worked in metastatic melanoma but failed in adjuvant disease. Fazeli called that bad news for Bristol and said “a negative has been taken away” for Moderna because Opdualag might have raised the adjuvant bar; equally, a trial many expected to work had failed, sharpening uncertainty around Moderna’s own Merck-partnered bet.
GSK’s unusually explicit description of RSV vaccines as a “two-player market” appeared grounded in payer conversations and Moderna’s RSV sales. Fazeli did not read it as arrogance; he read it as another headache making a positive cancer-vaccine outcome still more important.

5. Obesity differentiation needs proof, while SpringWorks offered visible strategic fit

Fazeli contrasted Metsera’s comparatively conventional GLP-1, GIP, and amylin portfolio with Aardvark’s more nuanced entry through Prader-Willi syndrome and hypothalamic obesity. Aardvark raised roughly $94 million at $16, the bottom of its proposed range, then traded near $13.70-$13.80 immediately after listing.
Rhythm provided a roughly $2 billion valuation benchmark in Prader-Willi syndrome, although it was much further advanced and used a different mechanism. Aardvark’s ARD-101 plus a DPP-4 inhibitor had early evidence of enhancing GLP-1 efficacy, but Fazeli wanted a meaningful trial against increasingly demanding comparators such as tirzepatide and the coming “triple G.”
SpringWorks’ Ogsiveo dramatically exceeded first-year expectations: roughly $170 million in US sales versus consensus around $45 million-$50 million, with Europe expected later in 2025. Newly approved Gomekli offered adult and pediatric neurofibromatosis coverage and a cleaner label than AstraZeneca’s Koselugo.
Historically, potential buyers such as Ipsen, Jazz, Servier, and Exelixis were considered too small for a deal estimated at about $5 billion against roughly $1.5 billion in portfolio sales. Merck KGaA confirmed discussions, and Werber expected a deal before SpringWorks’ following Thursday earnings report. Its $60 billion market capitalization, $21 billion in sales, global reach, and neurology-oncology footprint made the fit credible. SpringWorks shares had risen from $32 after JPMorgan; Chris noted a recent multiyear peak above $60 and a 2021 peak of $90.

6. Below-cash companies can contain valuable science, but runway still rules

Xilio’s AbbVie discovery-option deal sent its shares up about 117%, yet the company remained below cash after including the new proceeds. Because Xilio would continue doing discovery work on masked T-cell engagers and had runway only through Q1 2026, Fazeli saw another capital raise approaching unless the agreement or new data restored investor demand.
The platform can mask CD3, the antigen-binding site, and even a costimulatory component, but Fazeli resisted declaring a renaissance. Longer follow-up from other masked engager programs must show that higher doses translate into higher efficacy: “There’s still some wood to be chopped here for the masking world.”
Chris added that his firm’s former investment, Landos, received a 200%-plus premium from AbbVie while trading at a negative enterprise value, supporting the view that AbbVie may buy undervalued public companies that fit its pipeline.
AnaptysBio’s PD-1 agonist in rheumatoid arthritis initially “scares the bejesus out of me,” Fazeli admitted, because it suppresses T-cell activity opposite to immuno-oncology. Yet 12-week efficacy looked broadly comparable to Rinvoq, helping shares move from roughly $12-$13 to above $20; unresolved week-12-to-week-14 behavior and Lilly’s discontinuation of a similar program left profile-versus-prioritization ambiguity.

7. FDA’s fate depends on whether reform or head-count reduction sets the metric

Usdin’s baseline was “Everything is uncertain.” FDA faced staffing cuts, one-hire-for-four-departures constraints, return-to-office pressure, low morale, and possible scientific-leadership removals; a 10% reduction concentrated in the Center for Tobacco Products might spare biopharma, whereas losing 10% of review staff could cause missed PDUFA deadlines and serious delays.
The most consequential leadership question after Marty Makary’s arrival, Usdin argued, would be Patrizia Cavazzoni’s replacement at CDER. That director exerts more day-to-day influence over issues vital to drug developers than the commissioner.
Fazeli asked whether FDA contained removable fat. Usdin’s answer preserved the distinction between possible inefficiency and reckless sequencing: firing 10%-20% first and analyzing needs later is “kind of backwards.” Reviewers may take two years to become productive, so an erroneous cut cannot be reversed when missed deadlines first appear.
Chris and Usdin identified the central contradiction: most drug-review staffing is paid through industry user fees, so cutting it does not necessarily save taxpayers money. DOGE may measure success by people fired, while Makary could measure it by delivering more safe and effective drugs efficiently; biotech requires credible regulation, not indiscriminate deregulation.

8. NIH reform has merit, but disruption could sever biotech’s research foundation

Usdin’s broader policy outlook included potential upside from pharmaceutical manufacturing onshoring, fixes to IRA Medicare negotiation, PBM reform, restoration of immediate R&D expensing, a return to traditional antitrust enforcement, and bipartisan legislation introduced to restore pediatric priority-review vouchers. Counterbalancing those possibilities was uncertainty over revival of most-favored-nation international reference pricing.
Usdin agreed with Jay Bhattacharya’s longstanding criticisms that NIH is too risk-averse, awards first grants too late, and funds excessive me-too research. “We’re certainly not getting $50 billion a year of science out of the roughly $50 billion a year that we’re paying for NIH,” he said.
The risks were more serious at NIH than FDA: Bhattacharya’s statements about COVID-19, which Usdin said were not based in science, raise concern about reversal of important infectious-disease work, while concern about pharmaceutical influence could disrupt an essential interface. Industry collaborations with NIH-funded researchers and SBIR funding for small biotechs could all be affected while the intended replacement mission remains unclear.
Werber’s closing question compressed the policy trade: better, unchanged, or more anxious in one year? Usdin rejected the middle — “either feeling a lot better or a lot worse” — because that is when staffing consequences, Makary’s reforms, CDER leadership, and user-fee reauthorization should reveal whether the system improved or was “blown up” into real operational problems.

DeepSeek, Kimi & MiniMax Attention Mechanisms, Paper by Paper Episode 130 - Feburary 7, 2025