All podcasts / Invest Like the Best / Summary

Alex Karnal - The Trillion-Dollar Health Revolution

2026-04-21 - 94 min - source - Read full transcript

Patrick O'Shaughnessy (host)Alex Karnal

Key insights

GLP-1 drugs are the first commercial proof that a much larger, trillion-dollar preventive health revolution is investable, not just a one-off obesity drug cycle.

Karnal argues GLP-1 revenue will exceed $100 billion a year, but the more important signal is that patients are voting with their feet for proactive, preventive medicine rather than waiting to treat disease after the fact. He frames this as the first proof point for a wider set of medicines (PCSK9 inhibitors, anti-amyloid drugs) that could cut a much larger share of the trillion-dollar annual cost of reactive healthcare.

preventive-medicine-economics

Patients are optimizing for tolerability and durability on GLP-1s, not maximum weight loss, which is the opposite of where Wall Street's attention goes.

Data Karnal's team collected shows patients want a dose that produces modest, sustainable weight loss with manageable side effects rather than the highest-efficacy dose. This matters for investing because it reframes what 'winning' the GLP-1 market means: staying-power and tolerability beat maximal potency.

glp1-and-metabolic-drugs

Price elasticity in the GLP-1 market is enormous, and the oral formulation of Wegovy at roughly $150-1,800/year is driving record adoption.

Compounded (non-approved) GLP-1s, priced at about half the branded cost, revealed that 15-20% of the market was being served entirely outside the traditional prescription and insurance system because branded pricing (over $4,500/month at points in 2025) was unaffordable. Since the oral Wegovy launch, weekly new-script volume moved from roughly 200,000 to 300,000 per week.

glp1-and-metabolic-drugs

PCSK9 inhibitors are close to a genuine 'free lunch' drug, unlike GLP-1s, which carry real tradeoffs.

GLP-1s cause meaningful nausea, vomiting, diarrhea, gallstones, pancreatitis, and muscle loss risk. PCSK9 inhibitors, by contrast, are modeled on people with a natural genetic mutation that eliminates PCSK9 protein production entirely - these people show an 88% reduction in cardiovascular disease risk over 15-year studies with no observed downside, and animal models suggest even pushing LDL toward zero causes no harm.

glp1-and-metabolic-drugs

Three structural barriers, not a lack of available medicines, are what keep people from the preventive health benefits already proven in trials: complexity, cost, and convenience.

Karnal describes his own two-year journey from a bad cholesterol test result to getting on a PCSK9 inhibitor as illustrative of system friction - repeated doctor visits, retested labs, and scheduling delays. He argues chronic preventive medicines are priced like acute treatments despite needing to be taken for decades, and that compliance collapses the more frequently and painfully a patient has to re-engage with the system.

preventive-medicine-economics

Cardiovascular risk reduction from GLP-1s appears independent of weight loss, pointing to a distinct biological mechanism beyond calorie reduction.

Novo Nordisk outcomes data showed north of a 20% reduction in heart attack/stroke risk from GLP-1 use that held even after controlling for weight loss, suggesting the drugs confer cardioprotective benefits through a separate pathway - which also implies that diet and willpower alone could not fully replicate the drug's benefit.

glp1-and-metabolic-drugs

AI-driven drug discovery was bottlenecked for years by bad training data, and the winners will be firms that can generate proprietary 'science tokens' through automated experimentation rather than just mining published literature.

Karnal notes a large fraction of published scientific literature does not replicate, so AI trained purely on existing papers risks garbage-in-garbage-out. He believes the firms that combine AI talent, capital, and a novel way to generate new experimental data (via fully automated, robotic wet labs like Lila Sciences) will build a durable moat and compound toward what he calls scientific superintelligence.

ai-in-drug-discovery

AI is already compressing molecule discovery timelines from roughly two years to about a month for known biological targets, though discovering genuinely novel targets is still early.

Karnal says the industry has not yet systematically discovered a target that didn't previously exist, but for known-but-hard-to-drug targets, AI systems can go from computational model to candidate molecule dramatically faster than legacy high-throughput screening approaches.

ai-in-drug-discovery

Sensitivity and specificity are the two metrics that separate useful cancer diagnostics from noise, and false positives from over-testing (e.g., full-body imaging) carry real psychological and downstream cost.

Karnal evaluates every new diagnostic on what fraction of true cancers it catches (sensitivity) versus what fraction of healthy people it incorrectly flags (specificity). He personally avoids broad imaging screens like Prenuvo because stacking many low-context measurements raises the odds of a disruptive false positive without a clear framework for acting on the result.

cancer-and-diagnostic-screening

Braidwell's investment process is built around three questions per company: will the science work, is the market opportunity real, and can capital be structured to produce an attractive risk-adjusted return.

The firm's daily morning meeting brings together scientists, biostatisticians, commercial experts, AI experts, and structured-finance specialists to jointly assess opportunities, reflecting Karnal's belief that no single type of expert can properly underwrite a biotech investment alone.

biotech-investing-philosophy

Karnal left Deerfield to found Braidwell after realizing the industry was optimizing for financial returns on drug approvals, not actual patient adherence or public-health impact.

A decade into his career, patient-journey data showed that most people prescribed chronic medicines stopped taking them within about a year. He describes this realization - after building his career and raising a family - as a personal turning point that made him ask whether backing 'invention' without pushing for 'impact' was the wrong goal, leading him to co-found Braidwell with Brian Kreider (former Bridgewater COO).

biotech-investing-philosophy

A grassroots 'citizen pharmacology' movement, mostly organized on Reddit around peptides, is running informal self-experimentation in parallel with the FDA's own push to move faster under new leadership.

Karnal frames the peptide and self-experimentation subculture as people unwilling to wait for the traditional RCT/FDA pathway, choosing to test hypotheses on themselves at higher personal risk in exchange for faster answers, occurring alongside a formal FDA effort (citing Commissioner Marty Makary) to use AI to speed document review and cut redundant study requirements.

preventive-medicine-economics

Companies

Braidwell - Life sciences investment firm Karnal co-founded with Brian Kreider after leaving Deerfield; runs a cross-functional team (scientists, biostatisticians, commercial experts, AI experts, structured finance) around a daily morning meeting
Deerfield Management - Where Karnal spent 15 years and helped build into a major source of life sciences capital, starting as a 24-year-old derivatives trader who pitched Jim Flynn on a lower-risk portfolio construction approach
Eli Lilly - Maker of Wegovy/Zepbound and anti-amyloid Alzheimer's drugs; pioneered Lilly Direct, a digital front end letting patients get GLP-1 prescriptions without traditional sales-rep-driven channels
Novo Nordisk - Maker of Ozempic/Wegovy; published cardiovascular outcomes data showing GLP-1s cut heart attack/stroke risk independent of weight loss
Amlin Pharmaceuticals - Karnal's first exposure to GLP-1 as an asset class in 2005, an early twice-daily injectable version of the hormone
Pfizer - Acquired a GLP-1 candidate positioned as a potential once-a-month dosing option
Amgen - Working on a once-a-month, potentially once-a-quarter, GLP-1 dosing regimen
Exact Sciences - Stool-based colorectal cancer screening test; first mover that Guardant Health iterated on
Guardant Health - Developed a blood-based colorectal cancer screening test after observing that stool-based test compliance was capped by sample-collection friction
Prenuvo - Full-body MRI imaging company; Karnal is cautious about it personally due to false-positive risk without clear context for results
Neko Health - Skin- and body-imaging screening company cited alongside Prenuvo as part of the emerging imaging-based screening category
Capstan Therapeutics - Developed an in-vivo (IV-administered) CAR-T approach that reprograms immune cells to attack cancer without extracting cells from the patient's body
Lila Sciences - AI-driven, fully automated robotic drug-discovery lab Karnal visited; can go from computational model to candidate molecule in about a month versus years historically
Biogen - Co-developer, with Eli Lilly, of anti-amyloid medicines that slow cognitive decline in later-stage Alzheimer's patients by roughly 30%
Hims and Hers - Cited as one of the channels through which compounded (non-FDA-approved) GLP-1s reached price-sensitive patients at roughly half the branded cost

Techniques and frameworks

Health stack - Karnal and O'Shaughnessy's framework borrowing the tech-stack metaphor for personal health: offense (nutrition, training, monitoring) and defense (five biological layers - lipids, cardiometabolic health, neurocognitive health, inflammation, blood pressure) that available medicines can already address
Right to win analysis - Braidwell's investment framework for evaluating whether a medicine will differentially win a market segment - applied to GLP-1s by breaking the obesity market into overweight, obese, and morbidly obese BMI bands
Sensitivity and specificity - Karnal's two core metrics for judging any new diagnostic test - sensitivity (percent of true cancers correctly flagged) and specificity (percent of true negatives correctly cleared)
Scientific method as iterative loop - Karnal's walkthrough of drug discovery: hypothesis, experimental design, execution, data review, confirm-or-reloop, moving from literature to in vitro to animal models to human trials

Summary

Alex Karnal, co-founder of the life sciences investment firm Braidwell and formerly a 15-year veteran of Deerfield Management, walks Patrick O'Shaughnessy through what he calls a "health stack": five biological layers - lipid optimization, cardiometabolic health, neurocognitive health, inflammation, and blood pressure - where medicines already exist to meaningfully extend healthy lifespan, but adoption lags far behind what's clinically possible. His central claim is that 2025 was the most exciting year of his 20-year career not because GLP-1 revenue is approaching $100 billion, but because GLP-1 adoption is the first commercial proof that people will proactively adopt preventive medicine at scale, which he believes sets up a much larger, trillion-dollar reduction in downstream reactive healthcare spending.

The conversation spends significant time on the mechanics and market dynamics of GLP-1 drugs (semaglutide, tirzepatide, and next-generation multi-agonists), including the discovery that patients optimize for tolerability and durability rather than maximum efficacy, and that price elasticity is dramatic - branded pricing near $4,500/month pushed 15-20% of demand into unregulated compounded alternatives, while oral Wegovy at a roughly $150/month, $1,800/year price point has nearly doubled weekly new-prescription volume. Karnal contrasts GLP-1s, which carry real tradeoffs (nausea, muscle loss, gallstones), with PCSK9 inhibitors for cardiovascular disease, which he calls close to a genuine "free lunch": people with a natural genetic mutation that eliminates the PCSK9 protein show an 88% reduction in cardiovascular disease risk over 15-year studies, with no apparent downside even as LDL is driven toward zero.

He identifies three structural barriers to broader adoption of medicines that already work - complexity, cost, and convenience - illustrating with his own two-year journey from an abnormal cholesterol test to actually getting on a PCSK9 inhibitor. On cancer and diagnostics, he lays out a sensitivity/specificity framework for judging new tests, praises the shift from stool-based (Exact Sciences) to blood-based (Guardant Health) colorectal cancer screening as reducing patient friction, and expresses personal caution about broad imaging screens like Prenuvo because stacking many low-context measurements raises the odds of a disruptive false positive.

On AI, Karnal argues the field was long constrained by low-quality training data - much published literature doesn't replicate - and that the firms positioned to win are those that can generate proprietary experimental data through automated, robotic wet labs (he cites a visit to Lila Sciences) rather than relying solely on published research. He says AI has already compressed molecule discovery for known targets from roughly two years to about a month, though discovering wholly new biological targets remains early-stage. He also discusses a parallel, informal "citizen pharmacology" movement - largely organized on Reddit around peptides like BPC-157 - where people self-experiment outside the FDA process to get faster answers, occurring alongside the FDA's own push (under Commissioner Marty Makary) to speed approvals using AI-assisted document review and reduced redundant study requirements.

The episode closes with Karnal's personal story: raised by an entrepreneurial father who never stuck with a business long enough to succeed, and a resilient, "we never quit" mother, he ended up on Merrill Lynch's derivatives desk assigned to biotech almost by accident, discovered a portfolio-construction insight that eventually got him hired at Deerfield at 24, and spent 15 years there before a realization - that most patients prescribed chronic medicines stop taking them within about a year - led him to question whether the industry was optimizing for financial returns on invention rather than actual patient impact. That realization, alongside co-founder Brian Kreider (former Bridgewater COO), led to founding Braidwell, whose investment process centers on three questions per opportunity: will the science work, is the market real, and can the investment be structured for an attractive return.

Notable Quotes

"What I'm excited about about the GLP-1 opportunity is we're seeing the first commercial proof that we're ready to head in that direction... a once in a lifetime trillion dollar revolution in all of public health." - Alex Karnal

"I think PCSK-9s are very much more that free lunch than GLP-1s. GLP-1s have real toxicity associated with them... But for PCSK-9, it's pretty much a free lunch." - Alex Karnal

"Why is it that I can go on my iPhone, I can go to Amazon, I click a button and the next day I can have toothpaste. But it takes me an incredible journey, hundreds of phone calls, multiple doctor visits, being pricked and prodded multiple times just to protect myself from having a heart attack. That doesn't make sense." - Alex Karnal

"We all know that medicine only works if we take it... yes, medicine only works if people take it, but they weren't." - Alex Karnal

"I think what you're basically describing on the AI side is our quest to try to get to a level of scientific superintelligence." - Alex Karnal