Psychiatry: Cobenfy Muted, ML-007 Binary, MDD Graveyard

Psychiatry is not neurology. The neurodegenerative shelf gets the Alzheimer’s headlines and the tau-versus-amyloid debates. Psychiatry gets higher placebo response rates, higher failure rates, and the first genuinely new mechanism in schizophrenia in fifty years. That mechanism is the muscarinic wave, and the question is whether it is a durable shift or a one-drug story. The answer, based on the latest data, is that the novel mechanisms are still unproven while a boring approved drug quietly compounds. Here is the ranking.

1. Bristol Myers Cobenfy: the first new schizophrenia mechanism in 50 years, and the launch is muted

Cobenfy (xanomeline-trospium) is the first antipsychotic with a novel mechanism approved since the 1950s, targeting M1/M4 muscarinic receptors instead of dopamine D2. Bristol Myers bought it in the $14 billion Karuna acquisition, and the launch has underwhelmed: $63 million in Q2 2026 revenue, missing consensus by about 4%. Run-rate that out and you get roughly $250 million annualized, a fraction of what a first-in-fifty-years antipsychotic was supposed to do. The problem is not efficacy; it is physician behavior. Chief commercial officer Adam Lenkowsky flagged two issues on the Q2 call: accelerating new patient starts and improving repeat prescribing in a disease where doctors have prescribed the same antipsychotics for decades. The Alzheimer’s psychosis expansion, which was supposed to be the next growth driver, was pushed to 2027 on slower enrollment. Cobenfy proved the market will re-rate a novel mechanism; it has not yet proved that mechanism can become a fast commercial ramp. Here is the read-through for the rest of the class: if the first-in-class with a $14 billion acquisition behind it cannot ramp past $250 million annualized in year two, the follow-ons like MapLight are climbing a steeper hill, because payers now have a benchmark for what the mechanism actually delivers in practice.

2. MapLight ML-007: the binary BID-versus-QD bet

MapLight’s ML-007 (betovumeline) is the direct follow-on to Cobenfy, using the same M1/M4 muscarinic approach with a different anticholinergic partner, fesoterodine instead of trospium, to target lower GI discontinuation. The Phase 2 ZEPHYR readout was a Rorschach test. The twice-daily dose hit the primary endpoint: a negative 4.5 point PANSS total score difference versus placebo at Week 5, effect size 0.37, p equal to 0.015. The once-daily dose missed. Investors focused on the miss, and MapLight (MPLT) crashed 73% in a day before rebounding to $11.36, a $530 million market cap. The bull case is the cognitive signal: in patients with baseline cognitive impairment, ML-007 showed a 0.51 effect size on the Cogstate battery independent of PANSS improvement, targeting a symptom domain no schizophrenia drug is approved to treat. The bear case is adherence: a drug that only works twice-daily may struggle against Cobenfy and its successors. ZEPHYR enrolled 307 patients randomized 1:1:1 across placebo, twice-daily, and once-daily arms in a five-week inpatient design, and safety was clean, with no serious drug-related adverse events and a 2% GI discontinuation rate. That GI number is the actual bull case against Cobenfy, whose registrational program carried a higher cholinergic discontinuation burden. This is a binary bet on whether MapLight’s Phase 3 design can salvage a once-daily dose or convince payers that twice-daily is enough.

3. MDD: Neumora navacaprant went 0-for-3, and that is the cautionary tale

The major depressive disorder lane just got a lesson in base rates. Neumora’s navacaprant, a kappa opioid receptor antagonist, failed all three of its Phase 3 KOASTAL trials and was discontinued, with 35% of staff laid off. KOASTAL-1 missed the MADRS primary endpoint, with a female-subgroup signal (negative 2.7 least-squares mean difference, p equal to 0.072) that did not survive. KOASTAL-2 was a dead heat (negative 0.3, p equal to 0.813). KOASTAL-3 had placebo outperform the drug (0.7, p equal to 0.480). The one flicker was anhedonia, where the Snaith-Hamilton Pleasure Scale separation in women was statistically significant (negative 2.3, p equal to 0.015), and 83.3% of completers elected to enroll in the long-term extension. None of it saved the program. Neumora now trades at $1.53 with a $290 million market cap, down roughly 95% from its 2025 peak. The lesson is the one from why Alzheimer’s drug trials fail: high placebo response is the killer, and MDD is where placebo response runs highest.

4. Axsome Auvelity: the boring winner nobody frames as a mood-disorder story

The durable winner on this shelf is Axsome’s Auvelity (dextromethorphan-bupropion), an approved MDD drug that works through glutamatergic NMDA modulation plus bupropion. Auvelity posted $180.3 million in Q2 2026 net sales, up 51% year over year, pushing Axsome’s total quarterly product revenue to $218 million, and management raised Auvelity’s peak sales projection to $8 billion after an Alzheimer’s agitation approval. Axsome trades at $205.40 with a $10.8 billion market cap. At $10.8 billion in market cap against roughly $870 million in annualized product revenue, Axsome trades near 12 times sales, a full price that assumes the $8 billion peak is reachable. The contrast with Cobenfy is the point: Cobenfy is the novel mechanism with a muted launch, while Auvelity is the novel-mechanism-turned-commodity that is already compounding revenue. The market is paying for Axsome’s execution, not its novelty.

5. Zuranolone and the frontier: a rolled-up asset and a 2029 binary

Sage’s zuranolone (Zurzuvae), the oral postpartum depression drug partnered with Biogen, has been rolled up by Supernus in an $8.50-per-share, $561 million acquisition, ending the standalone Sage story. Sage’s collapse is a cautionary tale in miniature: zuranolone won PPD approval in 2023 but failed to translate to the much larger MDD market, and a company that once traded above $50 a share sold itself for $8.50. The PPD-to-MDD translation question is exactly what every mood-disorder program faces. The PPD franchise is now a Supernus asset generating modest collaboration revenue; Sage’s 50% share was $23.2 million in Q2 2025. On the frontier, Lilly’s $2.8 billion AtaiBeckley acquisition puts BPL-003, an intranasal 5-MeO-DMT with Breakthrough Therapy designation, in Phase 3 for treatment-resistant depression with data not expected until 2029. That is a flag on the frontier, not a near-term catalyst.

The pattern

The pattern here repeats across the entire shelf. Mechanism novelty is necessary but not sufficient. The market paid up for Cobenfy’s new mechanism and then watched a muted launch. A novel mechanism did not stop navacaprant from going 0-for-3. Auvelity, the boring mechanism with good execution, compounds fastest. For psychiatry, the durable edge is not picking the newest molecule; it is picking the company that can actually sell a drug to psychiatrists, who are the most behaviorally entrenched prescribers in medicine.

The risks

The muscarinic wave could stall if MapLight’s Phase 3 fails to resolve the dosing question and Cobenfy’s ramp stays flat. MDD’s placebo response rate means any novel mechanism, psychedelic included, faces a coin flip at Phase 3. And Axsome’s $8 billion peak sales projection assumes the Alzheimer’s agitation and depression indications both scale, which is not guaranteed.

The verdict

Own Axsome for the compounding revenue story: Auvelity is growing 51% with an $8 billion peak sales target, and the stock is already proving execution. Trade MapLight only if you can size the binary properly, because a Phase 3 design that cannot save once-daily dosing would halve it. Watch Cobenfy as the mechanism bellwether, but do not expect the launch to re-rate Bristol Myers, where a $63 million quarter is a rounding error against a diversified oncology franchise. And treat the MDD novel-mechanism lane, navacaprant’s graveyard, with the base-rate skepticism it has earned. That leaves Auvelity as the only name on the shelf where the thesis is already visible in the income statement, and it is the one to own.