Orexin Agonists: $10B Wake-Up Call. Takeda Won Round One.

The narcolepsy field just got its first genuinely new mechanism in three decades, and the follow-on race is where the money is.

Takeda’s Orzeyful (oveporexton), an oral orexin receptor 2 agonist, won FDA approval on August 5, 2026 as the first treatment for narcolepsy type 1 that addresses the underlying orexin deficiency rather than masking symptoms. That approval validated the entire class and turned a niche orphan indication into a blueprint for a much bigger prize: narcolepsy type 2, idiopathic hypersomnia, and the excessive daytime sleepiness that rides along with sleep apnea.

The mechanism is the reason this is different. Orexin, also called hypocretin, is the brain’s master wakefulness switch, produced by a small population of neurons in the lateral hypothalamus. Narcolepsy type 1 patients have lost nearly all of those neurons, which is why they are not just sleepy but plagued by cataplexy, the sudden muscle weakness triggered by emotion. Traditional treatments, amphetamines and modafinil, force wakefulness with abuse liability and cardiovascular baggage. An OX2R agonist restores the natural signal. That is a $1B-plus story: a stimulant-free wakefulness drug. Here is the class, ranked by risk-reward, with prices from Polygon’s August 29 close.

Takeda Orzeyful: First Approval, Immaterial to the Stock

Takeda won the race, and the market shrugged. Orzeyful (oveporexton) was approved August 5, 2026 for narcolepsy type 1 in adults, several weeks ahead of its September 30 PDUFA target date, after the FDA accepted the NDA with priority review in February. In the FirstLight and RadiantLight Phase 3 trials, oveporexton met the primary endpoint with a mean sleep latency gain on the Maintenance of Wakefulness Test of 13.83 minutes at the 1 mg dose and 17.20 minutes at the 2 mg dose versus placebo (both p<0.001). Weekly cataplexy rate fell by a median of more than 80%, and cataplexy-free days rose from zero at baseline to 4 to 5 per week. The full data set is in Takeda’s World Sleep 2025 presentation. The oveporexton approval and pre-approval thesis is covered in detail in a prior piece.

The reason the stock did not move is arithmetic. Takeda is a $57.6 billion company. Even a successful narcolepsy franchise, which analysts peg in the $1 billion to $2 billion peak range, is under 4% of Takeda’s revenue base. The approval is a class validation event, not a Takeda stock event. Own Orzeyful as the proof the mechanism works; do not own TAK for the narcolepsy launch.

Axsome AXS-12: The Cataplexy Specialist

Axsome (AXSM) is the next FDA decision in the space, and it is a different drug with a different target. AXS-12 is reboxetine, a highly selective norepinephrine reuptake inhibitor, aimed specifically at cataplexy and excessive daytime sleepiness in narcolepsy. In the Phase 3 SYMPHONY trial, AXS-12 cut weekly cataplexy attacks 83% versus 66% for placebo (p=0.018), and 33% of patients achieved full cataplexy remission versus 9.5% on placebo. The confirmatory ENCORE trial also hit its primary endpoint. The NDA is filed with a PDUFA in May 2027.

At $199.02 and a $10.4 billion market cap, Axsome is not cheap, but it has a real commercial engine already, with Auvelity and Sunosi on the market. AXS-12 is not an orexin drug, which matters for positioning: oveporexton and the other OX2R agonists treat the sleepiness that defines narcolepsy, while AXS-12 competes most directly on cataplexy control. The two mechanisms may end up complementary rather than head-to-head. That is the bull case for AXSM into May 2027.

Alkermes Alixorexton: The Narcolepsy Type 2 Play

Alkermes (ALKS) owns the expansion data that matters in the class. Its alixorexton, formerly ALKS-2680, is an oral OX2R agonist, and the Phase 2 Vibrance-2 study was the first large randomized trial to show an OX2R agonist working in narcolepsy type 2, patients who do not have an orexin deficiency. Alixorexton met both primary endpoints at week 8, with the 18 mg dose achieving statistical significance on both the MWT (p=0.047) and the Epworth Sleepiness Scale (p=0.046), moving patients from severe sleepiness to normal or mild scores by week 13. The trial is registered as NCT06555783.

Why this matters: narcolepsy type 1 is roughly 25,000 to 35,000 U.S. patients. Narcolepsy type 2 and idiopathic hypersomnia are each comparable or larger, and together they are the expansion that turns the orexin class from an orphan market into a multi-billion-dollar one. Alkermes is running Phase 3 Brilliance studies in NT1 and NT2 and a Phase 2 in idiopathic hypersomnia. At $47.41 and a $7.95 billion market cap, ALKS is the pure play on the expansion thesis, and it already carries a profitable commercial business to fund the trials without a dilutive raise.

Centessa ORX750: The Best-in-Class Candidate

Centessa (CNTA) has aggressive claims and the data to back them so far. ORX750 is an oral OX2R agonist with a long half-life, and in the Phase 2a CRYSTAL-1 study it delivered a 1.5 mg dose that drove a greater than 20-minute gain in mean sleep latency on the MWT versus placebo (p=0.0026), cut the Epworth score to 5.1 versus 18.7 on placebo (p=0.0001), and produced an 87% relative reduction in weekly cataplexy attacks. Those numbers across narcolepsy type 1, type 2, and idiopathic hypersomnia are why Centessa calls it a potential best-in-class profile. The trial is registered as NCT06752668.

Centessa trades at $40.50 with a $6.26 billion market cap and $349 million in cash as of September 30, 2025, and it has said it will initiate a registrational program. The risk is execution: ORX750 is the furthest behind in the class, and Centessa will need to fund a large Phase 3 program, meaning dilution. The reward, if the registrational data matches the Phase 2a, is a best-in-class drug in the broadest indication set. This is the highest beta name on the list.

Apnimed AD109: The Sleep Apnea Adjacent Play

Apnimed (APMD) is not an orexin company, but it is the clearest read-through on where the wakefulness trade goes next. AD109 combines aroxybutynin and atomoxetine to target the neuromuscular cause of obstructive sleep apnea, and it has now completed two Phase 3 trials. In the second registrational study, AD109 reduced the apnea-hypopnea index 46.8% at week 26 versus 6.8% for placebo (p<0.001), with 22.9% of patients reaching complete disease control, defined as an AHI below 5. The pitch is the first oral pill for sleep apnea, a market where the standard of care is a CPAP machine that roughly half of patients abandon. Apnimed trades at $29.71 with a $1.18 billion market cap, post-IPO. The tie to the orexin story is the expansion: if oral drugs can keep airways open and, separately, keep the brain awake, the combined addressable market for stimulant-free wakefulness dwarfs narcolepsy alone.

The Verdict

The orexin class just cleared its biggest hurdle, and the trade now shifts from “will the mechanism work” to “who owns the expansion.” Takeda won round one, but at $57.6 billion the win does not trade. Axsome has the next FDA decision in May 2027, but it is a cataplexy drug, not an orexin drug.

Alkermes is the top pick. It is the only OX2R agonist with Phase 2 data in the narcolepsy type 2 and hypersomnia expansion. It already has a commercial business to fund Phase 3 without a dilutive raise, and at $7.95 billion it is not yet pricing in a $1B-plus franchise. Centessa is the higher-risk, higher-reward version of the same thesis: the highest single-dose efficacy in the class, but a funding gap to bridge. Apnimed is a separate bet on oral sleep medicine worth owning only if you believe the CPAP-avoidance market materializes. This whole shelf sits inside the 2026 neuroscience catalyst calendar, and the endpoints to watch are MWT sleep latency and the Epworth score. Position sizing should reflect the binary: 2% to 3% of a biotech allocation in Alkermes, 1% to 2% in Centessa, and treat Axsome and Apnimed as complementary, not core.