Adam May on $ABVX's blowout data and subsequent stock crash

Summary

• Abivax’s ($ABVX) phase-3 maintenance data didn’t just meet the blue-sky case — it exceeded a scenario nobody modeled. Clinical remission delta came in at 40% on the 50mg dose and 39% at 25mg versus Rinvoq’s 39%/30%, and on the GI doctors’ favored endpoint, endoscopic remission, obefazimod’s 38% delta more than doubled Rinvoq’s 18%. Adam May: “This outcome was not even listed on the scale for any of the scenario analyses of these data.”
• The stock spiked from ~$130 to $175–185 after hours on June 1, then crashed below $100 as the XBI fell as low as roughly $68 on a table showing seven cancer cases on the 50mg arm versus one on placebo — a reaction May calls “demonstrably false.” One case was colonic dysplasia (“by definition not cancer,” probably included via a coding error), four were non-melanoma skin cancers generally cured by a short local-anesthesia procedure, leaving one prostate and one breast case — “statistical noise without a doubt.”
• The rate math clears the drug: pooling phase 2 (seven years running, one cancer) with phase 3, and conservatively assigning the phase-2 case to 50mg, the incidence is 0.59 per 100 patient-years versus a ~0.5 background rate in UC patients. There’s no carcinogenic mechanism — obefazimod isn’t genotoxic, and severe infections were more common on placebo than on 50mg, so it’s “unequivocally” not immunosuppressive — and no clustering, since the two real cancers hit distinct organs.
• May pins the crash on management, not the molecule: they disclosed cancer cases in more detail than precedent required, “naively” assuming the market would read it as they did. After the stock cratered they pulled the Part 2 safety dataset — roughly 500–600 additional phase-3 patients, including several hundred on 50mg for a year — forward from October to before the end of June; about three new cancers would be consistent with background rates, while more could “spook the market again.”
• At $100 ($7B), May argues the market is pricing in a black-box warning — yet even the worst case (black-box, UC only) is a multibillion-dollar drug, given Rinvoq does roughly $4B peak in IBD with five black-box warnings plus lab monitoring. Blue sky — no black box and a mid-2027 Crohn’s hit — is $8–10B peak sales, “almost exceeds a pharmaceutical company’s ability to buy a company.”
• May sees a multiple-bidder possibility and believes several pharmaceutical companies could be interested. Walker gives the bullish ranges: roughly $16–18B fully diluted ($175–200/share) near term, potentially high hundreds or around $300 if the Crohn’s data arrive first, citing Cidara’s 110% takeout premium and Abivax’s CEO having sold CinCor six weeks after a bad readout.
• Nektar has drifted from a $92 secondary to the mid-50s on no news; May is “at a loss for words” but flags three overhangs: specialists sidelined until the Eli Lilly jury trial resolves in late summer or fall, buzzy competitor Q32 Bio’s open-label alopecia data (whose IL-7 targeting raises SCID-like safety concerns), and Sanofi not yet killing amlitelimab. Walker says Sanofi still projects amlitelimab as multibillion-dollar despite his view that it will receive a black-box warning. Asked to pick, May takes Abivax on “the degree of certainty that I have that the current price is wrong.”

Deep dive

1. The setup: efficacy was the question, safety was supposedly derisked

• Recap from the prior episode: induction data (8-week efficacy) had already hit — “cross-trial comparison caveat, probably the third most efficacious drug that had ever been studied in ulcerative colitis” — leaving the maintenance readout (44 more weeks) to answer durability and safety. May’s caveat stands out in hindsight: “even a year is a pretty short period of time to determine the safety of a drug.”
• Safety looked largely derisked going in because the unblinded Data Safety Monitoring Board had reviewed 80%, then 90% of the trial with no new signals — and as Walker put it, an unexpected DSMB halt is “every investor in a biotech’s nightmare.”
• The comp that frames everything is Rinvoq, the JAK inhibitor with a 39%/30% clinical remission delta but a black-box warning covering cancer, severe fatal infections, blood clots, heart attacks/strokes and MACE, plus lab-monitoring requirements — “it’s every way it can kill you except for like human causes,” Walker joked. Consensus for Abivax was a 20–30% delta; May had guessed 25%; a “three handle” was the blue-sky case.

2. A four-handle: efficacy beyond the wildest bull case

• The result: 40% clinical remission delta at 50mg and 39% at 25mg, versus Rinvoq’s 39% and 30% at the corresponding doses — “not something that people were thinking was feasible.”
• On endoscopic remission — the endpoint GI doctors favor most because a colonoscopy showing zero disease is objective, versus the “noisier” questionnaire-driven remission endpoint — Abivax’s 38% delta more than doubled Rinvoq’s 18%. “That result in particular was beyond the wildest bull-case dreams… not even listed on the scale for any of the scenario analyses.”
• Walker’s generalist framing of the stakes: uncontrolled UC can end in colectomy, and May didn’t sugarcoat it — a new hole in the abdomen, a colostomy pouch “for the rest of your life.” A drug that fully heals the colon “changes your life, lengthens your life.”

3. Then the crash: $185 to below $100 on seven cancer cases

• The 4:05pm June 1 release said “no new safety signals identified”; the stock unhalted from ~$130 into the $175–185 range, then bled to $160, $150, accelerated through $110 and broke triple digits. May initially saw the listed cases as passing the “sniff test” — “you see cases of cancer in large phase 3 studies all the time” — until messages arrived about seven cancers on the 50mg arm versus one on placebo.
• The selloff was violent enough to drag the whole sector: the XBI fell ~4–4.5% the next day and went as low as roughly $68 — “beyond an overshoot to the downside.”
• May’s diagnosis is that management “severely mishandled” the presentation: peers with UC drugs — even ones carrying cancer black boxes — don’t press-release cancer cases at that granularity. “They took for granted that the market would be able to rapidly interpret” what experts saw instantly: no cancer risk.

4. Seven becomes two: dysplasia, skin cancers, and a dermatologist’s read

• Case one was colonic dysplasia — “by definition not cancer,” a precancer common in inflamed colons — probably swept in by a coding error under a neoplasms category and then, worse, footnoted verbatim as “colon cancer. Completely wrong.”
• Four more were non-melanoma skin cancers, which May — a dermatologist — treats “10 times a day”: ~5 million US cases a year, 15–20x more common than the next cancer, and almost always cured by a roughly 15-minute procedure under local anesthesia. The FDA literally tracks drug cancer risk with NMSC as its own separate endpoint. Walker’s corroboration: his dad had one cut off the bottom of his nose last year — “not a big deal.”
• That leaves two “real” cancers — one prostate, one breast, the two most common types in the world — on an arm of almost 200. Walker’s framing: “It’s literally one” of each. May: “This is in my opinion statistical noise without a doubt,” and distilling seven to two up front “probably would have prevented this entire crash.”

5. No mechanism, no clustering, and the patient-years math the market skipped

• The FDA’s two carcinogenicity mechanisms both fail here: obefazimod was studied years ago and shown to be non-genotoxic, and severe infections were more common on placebo than on 50mg, so “unequivocally the drug is not immunosuppressive.” Nor is there clustering — carcinogenic drugs can drive one tumor type; breast and prostate are “completely distinct organs.”
• Comparing drug arm to placebo arm was itself an error — the trial is powered for remission delta, not rare events, and adverse-event capture was wildly asymmetric: ~80% of drug patients completed versus placebo patients dropping out en masse (Walker’s numbers: ~160 finished OB50 versus 66 placebo). If a dropout were diagnosed with lung cancer six months later, “we’ll never know.”
• The right comparator is background rate: ~0.5 cancers per 100 patient-years in UC patients. The largest safety database is counterintuitively the seven-year phase 2 study — hundreds of patient-years, one cancer. Pooled with phase 3, and conservatively assigning the phase-2 case to 50mg, the rate is 0.59 per 100 patient-years — “perfectly in range.” But “certainly people were not prepared to look at that in depth after hours on Monday night.”

6. Walker’s pushback: the market has had a week — why hasn’t it repriced?

• The bear framing Walker put squarely: this is “probably the most talked-about name in the biopharma sphere,” blanketed with sell-side research, banks downgraded after two days of digestion, and the stock sits at $100, not $130 — “is this really where you’re going to find edge?” May’s answer: “neither of us would be on this podcast right now… if we believed in efficient markets,” and “I am repeatedly amazed at the things that the market is able to ignore when it comes to deep analysis of data.”
• His receipts: before induction he published the math showing the trial was “as close as a guarantee as you can ever get to a 10x upside,” tens of thousands read it, and the stock traded down for 48 hours — then the data hit for roughly 1,000% upside. Walker’s fair counter: the stock is much bigger now, “and there are a lot more eyeballs on you, my friend.”
• A Jefferies note Walker quoted — “a good drug does not always equal a good stock,” and the JAK differentiation “narrowed” — drew May’s sharpest line: pre-data, merely matching Rinvoq was the blue-sky case, so “it feels like they’re being prisoners of the stock-price movement and not actually evaluating the fundamentals.”

7. The skew: even a black-box worst case is a multibillion-dollar drug

• The bet reduces to two variables: black box or not — a warnings-and-precautions note is “essentially the same as not having anything on the label at all” — and whether Crohn’s hits at the mid-2027 readout. Worst case — black box, UC only — Rinvoq proves the concept: worse efficacy, five fatal black-box warnings, lab monitoring, and still roughly $4B peak IBD sales. A $10B acquirer of a $2B-peak drug “at least is going to make your money back.”
• Blue sky — Crohn’s hit plus clean label — is $4–5B peak in UC plus $4–5B in Crohn’s: “an 8 to 10 billion drug… you’re getting in more of like a merger territory than a buyout territory.” And the UC efficacy magnitude “speaks to adjusting the probability of success upwards for Crohn’s as well.”
• At $100, May thinks “you’re pricing in a black-box warning and somebody could reasonably buy this for $130–150 a share” even assuming one. Walker’s structuring instinct: “This is a perfect one for CVRs” — on Crohn’s and on the label outcome.

8. Part two: 500–600 more patients of safety data within weeks — and the raise question

• Part two covers the ~550–600 of ~1,300 patients who started the trial outside the primary maintenance analysis — several hundred on 50mg for a year — and will matter primarily as a safety dataset. Originally slated for an October conference, it was pulled to before the end of June the day after the crash: “It does not seem like they were hiding that data. They just legitimately did not think people were going to be worried.”
• The bogey is roughly three cancers over one year in that population at background rates. Even three new cases would lower the current per-patient-year rate; fewer would lower it further, while more could “spook the market again.”
• Walker argued that reportable cancer cases would be sent to the FDA on a rolling basis — “if there was a cancer case in week one, that is mandated to be reported by week two” — and his “gun to my head” guess was that management probably knew the cancer counts. May said he could not confirm it, but also guessed they knew.
• On the $500M balance sheet (“nothing” for trials this size) and secondary fears: management had said it would raise after the maintenance data, so May assumed a raise if the stock recovered to around $150. A raise would not kill M&A; in the RNA precedent, the company did a secondary during takeover rumors and later sold at a bigger premium. A secondary may take M&A off the table for only a couple of weeks.

9. Nektar’s unexplained drift from $92 to the mid-50s

• The speakers recalled the post-April move differently — Walker described roughly $45 to $100, while May recalled the stock going from the mid-$70s to almost $110 — but agree that after the $92 raise, it has bled to the mid-50s on no updates. May: “I’m a little bit at a loss for words as to why this sell-off has been so sustained and severe,” especially with the XBI not weak, though Nektar “has always traded at a discount to what a fresh new IPO with a flashy management team would trade at” on the same asset.
• Three candidate overhangs: the Eli Lilly litigation, with resolution expected in late summer or early fall, sidelines biotech specialists because legal handicapping is “out of their mandate”; Q32 Bio’s upcoming open-label alopecia data — with a history of “hyping up data and then disappointing” and blinded data already worse than Nektar’s — plus its buzzy $55M late-May raise, with RA Capital both a major Nektar holder and a Q32 backer; and Sanofi not yet discontinuing amlitelimab despite the OX40L-class Kaposi sarcoma cases. Walker notes that Sanofi still projects amlitelimab as a multibillion-dollar drug despite his view that it will get a black-box warning in atopic dermatitis.
• May’s safety worry on Q32 is mechanistic: it targets IL-7, and humans with IL-7-receptor mutations develop SCID, fatal in most people — “I have concerns about the safety of that drug… but in the long run I’m not concerned about it as a competitor.” On Lilly, both agree the jury optics — “big bad Big Pharma stole this drug, messed it up, wouldn’t give it back” — make a settlement hard to imagine avoiding, yet “this stock is not trading as if they’re going to get anything out of that lawsuit.”

10. The verdict: Abivax, on certainty — and a multi-bidder endgame

• Asked which is more mispriced, May picks Abivax — defined not as most upside but “the degree of certainty that I have that the current price is wrong based on objective data.” Walker’s near-term range is roughly $16–18B fully diluted ($175–200/share, about 100% in six months); if they clear the safety signal and let the Crohn’s card flip, he sees high hundreds or possibly a $300 stock at an M&A price. These are blue-sky scenarios, but Walker sees downside as quite low since phase 3 is in and the worst case is still multibillion-dollar UC sales. Nektar could 3–5x longer-term, but Abivax’s margin of safety plus near-term M&A probability is “really hard to pass up.”
• On sentiment repair, May isn’t waiting on the tape: “ultimately this is a multiple-bidder scenario where the price is going to be determined by who has the best bid… market price fluctuations are probably noise.” He “fully trusts” pharmaceutical business-development and regulatory teams to run the same cancer-rate math he has. The comp both discussed was Cidara, bought at a 110% premium after a huge rally because multiple bidders showed interest.
• The color Walker loves: Abivax’s roughly 65-year-old CEO previously sold CinCor six weeks after a disappointing phase-3 readout, with a CVR and at a big premium. Walker said, “Lightning could strike twice”; May recalled that the premium was very large.