Big pharma has paid billions for psychedelic assets built on molecules nobody can own. This piece examines what the patents underneath those deals actually protect, and where each one is likely to break.
Warren Buffett's line about the tide going out and revealing who has been swimming naked applies neatly to psychedelic patents now that big money is coming in. Until recently, these patents were challenged mostly by nonprofits, through relatively inexpensive proceedings at patent offices, and rarely fought over in court. Full patent litigation costs millions of dollars, and no company spends millions attacking a patent that protects a drug that may never reach late-stage trials. That calculation has now changed. AbbVie, Lilly and Otsuka have paid, or committed to pay, sums reaching into the billions for psychedelic assets. Competitors, future acquirers and eventually generic manufacturers all now have a reason to look for the weakest claim in each portfolio.
The problem runs across the whole sector: almost none of these drugs are new molecules. 5-MeO-DMT, psilocybin, LSD and methylone are all decades-old chemistry, so nobody can patent the molecule itself. Companies patent everything but the molecule instead: a particular salt, a particular crystal form, a formulation, a dosing schedule, or the use of the drug for a specific disease. Each of these is narrower than a patent on the molecule, and each is easier to attack or work around.
1. AbbVie and Gilgamesh: When the Money Withdraws
Bretisilocin is a substituted tryptamine in the DMT family (chemically 5-fluoro-N-methyl-N-ethyltryptamine, or 5-F-MET). In August 2025, AbbVie agreed to acquire the program from Gilgamesh for up to $1.2 billion, counting the upfront payment and development milestones.
The patent risk comes from someone else's claims. Enveric Biosciences, a small biotech, holds U.S. Patent 12,138,276 on halogenated psilocybin derivatives, a category that includes fluorination, the same kind of modification behind bretisilocin. Enveric itself says the claims appear relevant to bretisilocin. Whether bretisilocin's specific structure actually falls inside them has never been tested.
Gilgamesh moved first. On August 12, 2025, thirteen days before the AbbVie deal was announced, it filed a post-grant review petition, a request that the patent office's internal review body, the Patent Trial and Appeal Board (PTAB), reconsider a recently issued patent. This is the standard play: challenge a blocking patent before spending big on the asset behind it. Enveric hired outside counsel in October and asked the board to decline the case.
Then AbbVie took over. The acquisition closed on October 17, 2025. Gilgamesh's other programs went into a separate spin-off, and the original entity, the one that had filed the petition, became a wholly owned AbbVie subsidiary. On November 20, AbbVie's lawyers asked to withdraw the petition, and the board granted the request. Enveric announced it in February 2026, leaving its patent unchallenged, and its CEO said the outcome showed the company's willingness to fully defend its intellectual property.
The withdrawal is a risk calculation. A withdrawn petition means no ruling either way, so the patent was never tested. Losing would have left AbbVie facing a patent that had survived a formal challenge, which is worse than the untested one Gilgamesh started with. Withdrawing keeps the options open: a license, an argument that bretisilocin falls outside the claims, or simply time, since approval is years away.
None of this makes bretisilocin a bad deal for AbbVie. It becomes one only if three things all go wrong: bretisilocin's structure falls inside Enveric's claims, Enveric refuses to license on workable terms, and AbbVie can't design around the patent. None of the three is established, and the first has never been tested. A small company holding a patent that touches a billion-dollar drug has strong reasons to negotiate rather than block. The likelier cost is a fee or royalty that trims AbbVie's return. The price tag is also softer than it sounds, since "up to $1.2 billion" includes milestone payments that are due only if development succeeds.
2. Lilly and AtaiBeckley: A Rival's Patents on both Sides of the Atlantic
Lilly's deal centres on BPL-003, a nasal spray of 5-MeO-DMT (also called mebufotenin). Because the molecule itself is decades old, BPL-003's protection rests on one specific salt of it, the benzoate. AtaiBeckley holds patents on that salt in the US, UK and Europe.
The problem is that a competitor, GH Research, got there first with a use patent, which is a different kind of patent. GH is developing its own 5-MeO-DMT product, GH001, an inhaled vapour that does not use the benzoate salt. So, instead of patenting a salt, GH patented a use: treating depression with 5-MeO-DMT. Here is the trick: a use patent covers the molecule whatever form it takes, so it can reach a rival's product even when that product is chemically different. This is the source of the conflict between GH Research and AtaiBeckley. A benzoate salt of 5-MeO-DMT given for depression is still 5-MeO-DMT given for depression, and Beckley's salt patent does not give it the right to do something GH's use patent covers.
GH sought this use patent in both Europe and the US, and the two patent offices answered differently. Before GH filed, surveys of people taking 5-MeO-DMT in ceremonial settings had already reported less depression afterwards. The US examiner, prompted by prior art that Porta Sophia submitted, treated those reports as making the idea obvious and refused the broad claim from GH. GH accepted a narrower US patent covering only inhaled use. The European Patent Office required proof that the effect was already established, not just hinted at, and so granted GH the broad version: any route, until at least 2040.
In Europe, there is a genuine thicket: two companies with overlapping patents on the same molecule, and BPL-003 falls inside GH's claim. Atai disclosed a patent dispute with GH over this European patent in January 2024. Porta Sophia and an anonymous party challenged the patent at the European Patent Office, and lost. This is where the new money matters. A nonprofit and an anonymous challenger could not break the patent. Lilly can afford to try, either in the courts or by paying GH for a license. Either way, the patent that once blocked a small biotech now stands in the way of a multibillion-dollar product.
In the US, there is no thicket for now. GH's patent covers inhalation and BPL-003 is a nasal spray, while Beckley's salt patent doesn't cover GH001. The products can coexist. Two things could change that. GH is still pursuing further US patents, some covering nasal delivery, and a new grant could reopen the conflict. More immediately, the pressure point flips: in the US, the patent most exposed is Lilly's own. Choosing a salt is a routine step in drug development, and US courts have struck down salt patents as obvious before. While BPL-003 belonged to a small company, nobody had much reason to test its salt patent. Owned by Lilly, it becomes worth attacking.
Lilly appears to have priced this in. Up to $1 billion of the deal is paid only if later milestones are met, including US approval and the Drug Enforcement Administration moving the drugs to a less restrictive schedule.
3. Compass: Approval Turns the Thicket into a Target
Compass hasn't been acquired, but approval will do for its patents what the acquisitions did for the others: attach real revenue to them, and with it a reason to attack. Compass has built the sector's most developed thicket around psilocybin, a molecule it cannot own. It holds roughly ten granted patents across the US, UK, Germany and Hong Kong, covering "Polymorph A" (the specific crystal form in its drug COMP360), a related hydrate form, formulations, and uses for depression and anxiety. Many molecules can crystallise in several different arrangements, called polymorphs, and psilocybin can also be made in a non-crystalline form.
The thicket has already been tested once. In 2021, a nonprofit called Freedom to Operate (not to be confused with the legal concept of the same name) asked the US patent appeals board to review two of Compass's Polymorph A patents. In 2022, the board refused to put them on trial. But it read Compass's claims very narrowly: they cover only psilocybin whose x-ray signature matches Polymorph A's exactly, a measurement called x-ray powder diffraction. The challenger argued this leaves other crystal forms free for anyone to use. Some patent lawyers saw the ruling as a clean win for Compass instead.
The real test runs on a predictable schedule. Because psilocybin has never been approved as a drug, COMP360 should receive five years of protection as a new chemical entity, during which the FDA will not accept a generic version regardless of the patents. A generic company can file at year four, however, if it declares that Compass's patents are invalid or don't cover its product. That declaration automatically triggers a lawsuit. If COMP360 is approved in late 2026 or early 2027, the first challenge could come around 2030 or 2031, and how hard it is fought will depend on how much COMP360 is selling by then.
This isn't hypothetical: Johnson & Johnson's Spravato, a ketamine-derived nasal spray approved in 2019, followed the same script, with generic manufacturers filing to enter as soon as the law allowed and Janssen suing four of them for patent infringement in 2023. Some of those fights have already settled, one is still headed to trial, and Compass should expect the same mix once its own five years run out.
The likeliest generic strategy is to walk around the thicket rather than cut through it, by using a different crystal form of psilocybin. That strategy is more powerful than it first looks, because Compass's patents are not independent layers: its thicket has many branches, but they grow from a single root, Polymorph A. Its claims on formulations, doses and uses are tied to the same crystal. A generic that avoids Polymorph A may avoid most of the thicket at once.
A second psilocybin program adds pressure from the opposite direction. In April 2026, the nonprofit Usona Institute received one of three psychedelic priority vouchers the FDA issued that month, for psilocybin in major depression. Usona's psilocybin is not yet before the FDA for approval, so the competitive pressure is not here just yet, but it's coming. Usona publishes its research openly, and everything it publishes adds to the body of earlier public work (in patent law, "prior art") that challengers can cite against Compass's later claims. Big money gives challengers a motive. Open science hands them ammunition.
4. Definium: A Thin Patent, An Exclusive Platform
Definium hasn't been acquired, and no rival has emerged with overlapping claims against it. Its LSD tablet patents have never been challenged. What it has is three positive Phase 3 readouts in 2026, a pre-NDA meeting with the FDA scheduled for the fourth quarter, an NDA filing planned for the first half of 2027, and investor materials projecting roughly $2 billion in annual revenue for every one percent of its target population treated. A patent portfolio nobody had reason to test is about to sit underneath a potential blockbuster.
That portfolio is thin, and young. Definium's first US patent on its dissolving LSD tablet claims priority to August 2021 and was granted in July 2024. It is also very specific. It claims a freeze-dried tablet made from a solution containing maltodextrin at 2 to 10 percent, hydroxypropyl methylcellulose as a binder at 1 to 5 percent, mannitol or trehalose at 2 to 10 percent, and LSD tartrate below 1 percent, dissolving on the tongue within 60 seconds. Two features make it fragile.
The first is how easily it can be designed around. The claim uses the phrase "consists essentially of," which in patent language locks it to that recipe: any ingredient that meaningfully changes the product falls outside the claim. Definium's later patents widen the list of usable LSD salts, from hydrochloride to mesylate, but they do not widen the recipe. Maltodextrin as the matrix former and hydroxypropyl methylcellulose as the binder appear in every claim in the family. A competitor that builds its tablet on gelatin, the classic base for freeze-dried tablets, falls outside all of them.
The second is obviousness. Freeze-dried tablets that dissolve on the tongue have existed for decades, with Catalent's Zydis technology the standard example. Putting a known drug into a well-established tablet format is a textbook target for the argument that the invention was obvious.
Definium's strongest protection may not be its patents at all. It holds exclusive rights to Zydis for every salt and crystal form of LSD, a contract no patent challenge can touch. A rival could design around the claims and still be unable to use the platform itself. That is a useful qualification to everything else in this piece: money will test these patents, but a patent is not always where the moat is.
There is no thicket around DT120 yet. Definium has few direct competitors and is well ahead of them, so nobody has had reason to build one. Approval will change that. We can expect a rush of filings on dosing regimens, methods for specific conditions, delivery formats, and LSD prodrugs (compounds the body converts into LSD) and close chemical relatives. The thicket forms after the money arrives, not before, and most of what grows in it will be weak. Prodrugs such as 1P-LSD were sold openly on grey markets for years, and that public record counts against anyone trying to patent them now. Patent law asks whether something was available to the public, not whether it was legal, so even illicit sales can count, provided a challenger can produce dated proof of what was sold and when. Grey-market vendors left that trail in archived websites and published chemical analyses. Street dealers did not.
5. Otsuka and Transcend: Many Claims, No Anchor
Otsuka paid $700 million at closing for Transcend, with up to $525 million more tied to future sales, for a total of up to $1.225 billion. What it bought is a company founded in 2021 whose lead drug is a molecule first made thirty years ago and long since in the public domain. So why so much money invested here? Here is the commercial logic: TSND-201 (methylone), a member of the MDMA family, acts on the transporters that recycle serotonin, dopamine and noradrenaline, but not on the serotonin receptor behind classic psychedelic effects, which may make its effects milder and its trials easier to keep blind than the MDMA trials that ran into trouble at the FDA. It carries Breakthrough Therapy designation and one of the agency's psychedelic priority vouchers. The patent position underneath all that is the weakest in this piece.
The question starts with Alexander Shulgin. Peyton Jacob and Shulgin first synthesised methylone and patented it in 1996 (WO1996/039133) as a potential antidepressant and treatment for Parkinson's disease. The patent expired long ago, so the molecule belongs to everyone. Worse for Otsuka, the expired patent is a published document linking methylone to psychiatric treatment, which is exactly the kind of prior art a challenger cites. Transcend's protection therefore rests on narrower claims covering post-traumatic stress disorder, dosing, or formulation. Its own materials describe TSND-201 as a proprietary formulation of methylone.
The obviousness argument writes itself. Combine Shulgin's description of methylone as an antidepressant with the large body of research on MDMA for post-traumatic stress disorder, and a court could well find it obvious to try a close chemical relative of MDMA for the same condition. Transcend wasn't the only one seeing the opening. Terran Biosciences began filing preliminary methylone applications in 2021, the year Transcend was founded.
This is a third kind of structure. Compass built a deep thicket with a single root. DT120 has no thicket yet. Methylone has neither: no strong central claim, and no single owner. The molecule is public, Shulgin's expired patent already ties it to psychiatric treatment, and what remains are narrow claims on formulation and indication, filed by several companies at roughly the same time. In the other cases there is one patent the rest depends on. Here there is nothing at the centre to depend on.
The wider family of methylenedioxy compounds, which includes MDMA, MDA and methylone, faces the same problem for the same reason. Because the parent molecules are all old and public, companies are patenting mirror-image versions instead. Many of these drugs exist in two forms that mirror each other (enantiomers), and the standard drug is a 50/50 mix of both. R-MDMA, one form alone, is being developed by AtaiBeckley (EMP-01, now Lilly's) and by Definium (DT402), while Arcadia is working on AM-1002, a version with a different ratio of the two forms. Isolating one form from a known mixture is usually judged obvious unless the patent holder can show it performs in ways nobody predicted. The result is a field filling with thin, overlapping claims that will hold up only where clinical data backs them. TSND-201 is one of those claims. Its formulation differs from plain methylone the way R-MDMA differs from plain MDMA: a narrow variation on public chemistry, defensible only if the clinic proves it behaves differently.
Otsuka seems to understand this. More than a third of what it agreed to pay depends on the product actually selling. That structure prices risk of every kind, but patent risk is part of it. Transcend is also developing methylone prodrugs, compounds the body converts into methylone, which it classifies as genuinely new chemical entities. That is the tell. The surest way out of a weak patent position built on an old molecule is to stop relying on the old molecule.