Compass Has Two Positive Phase 3 Trials. The FDA Has Final Guidance. So Why Does Psilocybin Still Feel Like a Gamble?

Robin Carhart-Harris chose his words carefully: “Hard to emphasize enough the challenge of treating anorexia. Was this our hardest trial? Maybe.” That sentence, from one of the most rigorous psychedelic researchers alive, landed the same week that Compass Pathways announced sustained six-month response data from its second Phase 3 COMP360 trial and the FDA published its final guidance for psychedelic clinical investigations. Three signals in a single news cycle, each coming from a different corner of the field, each pointing in the same direction. The convergence is hard to dismiss.

But convergence is not clarity. And the momentum visible on the surface of this moment conceals a set of structural problems that no single guidance document, however landmark, can resolve on its own.

The Numbers That Matter Most

Start with what Compass actually demonstrated. In a trial of nearly 600 patients with treatment-resistant depression, 39% of those receiving the active dose of COMP360 showed a meaningful clinical response after one to two doses, sustained at six months. Per Paul F. Austin‘s public post, that makes two positive Phase 3 trials in a row. Compass is now planning to complete its FDA filing in Q4 of this year, and with a priority review voucher, the standard twelve-month review window could compress to one or two months. A regulatory decision as early as late 2026 or early 2027 is plausible. If COMP360 clears, psilocybin would become the first classic psychedelic approved for medical use in the United States, which would simultaneously reschedule it off Schedule I.

That is a genuinely historic possibility. It is also where the comfortable narrative should end.

Chris Witowski, PhD, co-founder and CEO at Psilera, ran the arithmetic that most of the celebratory posts skipped. His analysis is the most operationally honest take in this week’s chatter. At roughly four dosing sessions per year, with six-hour monitored sessions per patient, psilocybin generates approximately 24 hours of clinical chair time per patient annually. Esketamine (Spravato), now running at a $2 billion-plus annual revenue rate, requires two-hour sessions dosed approximately weekly, producing close to 104 hours of monitored time per patient per year. The math cuts both ways: psychedelics are more time-efficient per session, but the infrastructure required to deliver even that reduced burden barely exists. Per Witowski, only 5% of treatment-resistant depression patients are currently on Spravato, despite years of commercial availability and no direct competitor. He does not expect that figure to exceed 10% even a decade from now, as infrastructure catches up. The bottleneck is not the molecule. The bottleneck is the clinic.

The Regulatory Bar Just Rose

The FDA’s final guidance, welcomed by Josh Hardman of Psychedelic Alpha as a signal that the agency is now engaging the field with genuine seriousness, is worth reading against what it changed from the draft. Hardman’s post identifies the most consequential shifts: the agency now describes a blinded twelve-month follow-up with prespecified retreatment criteria as “the most informative” trial design, a materially higher evidentiary standard than the draft implied. Safety packages must now include driving studies. The guidance explicitly closes the door on sponsors pleading ignorance about expected adverse events. These are not minor refinements. They reflect an agency that watched the MDMA-assisted therapy application collapse in 2024 and recalibrated accordingly.

The common assumption in this field is that FDA guidance signals an easier path to approval. The final psychedelics guidance signals a more demanding one.

Hardman also notes that of the 225 public comments submitted during the draft period, most remained unaddressed in the final document. Sponsors had pushed back on monitor-to-patient ratios, on the restriction against dosing session monitors participating in post-session therapy, and on the length of required follow-up periods. The agency held its position on all three. That matters operationally: a blinded twelve-month follow-up requirement, layered onto the inherent unblinding problem in psychedelic trials where most participants know whether they received an active dose, creates a methodological tension that no sponsor has yet resolved cleanly at scale. The September public hearing will open another comment window, but a public hearing is not an advisory committee. It will not produce binding guidance modifications.

Which leaves sponsors in a familiar position: substantial new clarity on what the FDA wants, and no established template for delivering it.

The Indication That Changes the Conversation

Carhart-Harris’s anorexia paper shifts the terrain in a different direction. Anorexia nervosa carries one of the highest mortality rates of any psychiatric disorder, as his post states plainly, and the treatment landscape is notoriously barren. A psychedelic intervention in this population is not an incremental line extension onto a crowded market. It is a probe into one of psychiatry’s most intractable failures. The fact that his team completed the trial at all, and that he described it as possibly the hardest work of his career, signals something about both the promise and the cost of expanding this field beyond depression.

Tadeusz Hawrot of PAREA pointed to the European College of Neuropsychopharmacology launching a free online course on psychedelic research, covering neurobiology, clinical protocols, trial design, and safety. The ECNP is not a fringe institution. Its decision to build a formal curriculum around this field is the kind of mainstreaming signal that precedes regulatory and reimbursement normalization in European markets. For global sponsors designing trials with eventual EMA submissions in mind, that institutional shift matters at least as much as the FDA’s final guidance.

Meanwhile, Definium Therapeutics released Phase 3 data on DT120, its sublingual LSD formulation, showing a clinically meaningful change on the MADRS scale at six weeks in major depressive disorder. Per Witowski’s analysis, the sublingual delivery format produced discharge-ready patients at five to six hours, attributed in part to LSD’s faster sublingual absorption reducing the tail of the experience. Two molecules, two delivery formats, two positive late-stage datasets inside the same month. The FDA is now looking at a field that is no longer asking for credibility. It is presenting evidence and waiting for a framework capable of processing it.

The practitioners, coaches, and facilitators whom Austin addresses in his post will not be wrong that approval creates demand. They will be wrong if they assume the system is ready to route that demand responsibly. Compass can file in Q4. The guidance is final. The September hearing is scheduled. None of that builds the clinical infrastructure, trains the monitors, resolves the unblinding confound, or answers the durability question that Witowski correctly identifies as the field’s largest unresolved variable. The molecules are ahead of the machinery. That gap is where the next phase of this story will be written, and sponsors who treat the Compass filing as a finish line are misreading which race they are in.

References

1. Robin Carhart-Harris, LinkedIn post on psilocybin-therapy for anorexia paper publication
2. Josh Hardman, Psychedelic Alpha, LinkedIn post on FDA final guidance for psychedelic clinical investigations
3. Paul F. Austin, LinkedIn post on Compass Pathways COMP360 six-month Phase 3 data
4. Chris Witowski, PhD, Psilera, LinkedIn post on psilocybin and LSD Phase 3 data and clinical infrastructure constraints
5. Tadeusz Hawrot, PAREA, LinkedIn post on ECNP psychedelic research online course launch