Jan-Christian Ehlers

your capital is at risk - relax

[Jan-Christian Ehlers @JanChristan]

Selling my stake in $ATAI 339% up (after 5 Years).

I initiated positions in $XE $GFS and $SIVE.

holding $MSFT.

[Doug Drysdale @insidepharma]

Roughly 1.1 billion people live with migraine. For most it's an occasional bad day, but for some patients it takes more than half the month, even after the older preventives and the newer CGRP drugs. Many of these patients keep having frequent attacks after several preventives and need new options.

Here's why I think BETR-001 deserves a serious test in those patients.

First, the targets have a track record. BETR-001 is an agonist at 5-HT1B, 1D and 1F, the receptors triptans and lasmiditan already act on. We know that biology works for migraine. We aren't betting on an unproven pathway.

Second, it adds something those drugs don't. It's also an agonist at 5-HT2A, the receptor linked to neural plasticity as seen with psychedelics.

Third, the practical profile is designed for real life. It's an oral capsule, taken at home. Preclinical work shows it doesn't agonize 5-HT2B, the receptor that has caused cardiac safety trouble for some other drugs, and it isn't hallucinogenic.

Fourth, there's an early human signal. A 2010 pilot by Karst and colleagues gave a racemic version of this compound to patients with cluster headache.

We're targeting the IND for Q1 2027, and then into the clinic.

[Jan-Christian Ehlers @JanChristan]

As ATAI was my highest conviction bet ever and I think there a presents that just keep on giving I thought I might as well share my thesis for $BETR.CN 🧵

2026 is the year psychedelic medicine became big pharma business. With huge progress in the space surfacing to markets and bigger investors. The next gen of that medicine may not need a psychedelic experience at all.
Here is where the sector stands, why neuroplasticity is the next frontier, and why $BETR.CN could be one of its most interesting companies too look at.

Start with the deals. $LLY closed its acquisition of AtaiBeckley in September (~US$2.8B upfront plus a CVR worth up to ~US$1B). Otsuka agreed to pay US$700M upfront for Transcend. $ABBV bought bretisilocin from Gilgamesh in 2025 for up to US$1.2B.

Then the data: $DFTX (formerly MindMed) reported three positive Phase 3 trials of its LSD tablet in 2026, in depression and anxiety, and plans to file in H1 2027. $CMPS has two positive Phase 3s for psilocybin in TRD and expects to complete its NDA in Q4 2026 ans is most likely to see FDA approval soon.

$HELP (formerly Cybin) says Phase 3 topline data for HLP003 in depression are on track for Q4 2026. Washington is helping: an executive order in April and FDA priority vouchers for Compass, Usona and Otsuka/Transcend. Q2 2026 funding topped US$1B, a record.

But most of these drugs come with a trip. Patients sit through supervised sessions of 6-8 hours in specialized clinics (sometimes less, e.g. BPL-003 was bought for pobably that reason). In Definium’s Panorama trial, participants were monitored for at least 8 hours. That works, but it limits how many patients a clinic can treat.

So the big question for the next wave: can you keep the benefit without the trip while continuing your life while being on medication. The leading theory is neuroplasticity. Since 2018 labs have shown that psychedelics make neurons grow new branches and spines, rewiring thought to drive lasting effects.

The classic route runs through the serotonin 5-HT2A receptor, the same one that triggers the trip. But in 2023 a Nature Neuroscience paper from Eero Castren’s lab found that LSD and psilocin also bind TrkB, the BDNF receptor, with ~1,000x the affinity of other antidepressants. So What If one could make use of that?

Now Smart money is moving and one of the purest listed plays on this idea is $BETR.CN.
Its drug BETR-001 is a form of 2-bromo-LSD (BOL-148): LSD’s scaffold, but only partial 5-HT2A activation (~60% vs ~92% for LSD), no head-twitch in mice, and it blocks 5-HT2B. Blocking 5-HT2B matters: that receptor is linked to the heart-valve damage seen with older ergot drugs.

BetterLife now leads with headache, a smart first test, since non-hallucinogenic ergot-type drugs have long been used there. Chronic cluster headache, the “suicide headache”, has no approved preventive in the US, and CGRP antibodies failed in chronic cluster trials.

Next is chronic migraine, where $ABBV two oral CGRP drugs sold US$2.3B in 2025. And the idea came from patients: the "Clusterbusters" community self-treated with psilocybin and LSD, and in a 2006 Harvard/McLean survey 22 of 53 reported benefit at sub-hallucinogenic doses.

Researchers then tried BOL-148 itself. In an open-label 2010 case series at Hannover Medical School (Harvard/McLean co-author), 5 patients who had failed verapamil got 3 doses. 3 of 5 went into remission for over a month; one chronic patient was still improved after 9 months.

In September BetterLife closed a US$100M (C$138.75M) offering at C$0.25, led by Deep Track Capital with RA Capital (9.0% stake), Perceptive, Marshall Wace and Coastlands. This month it added a Scientific Advisory Board co-chaired by two of headache medicine’s biggest names.

Peter Goadsby (King’s College London) shared the 2021 Brain Prize for the CGRP research behind modern migraine drugs. Richard Lipton runs the Montefiore Headache Center and is a past president of the American Headache Society. Both joined in October.

Executive Chairman Doug Drysdale (@insidepharma) is the former CEO of Cybin, CMO Dr. Mahan Chehrenama worked on Amgen’s Aimovig program, and Head of Clinical Operations Sheila Leitzinger, who joined on Oct 15, has run first-in-human studies over nearly 30 years.

Catalysts:
Dec 9 shareholder vote on Nasdaq listing requirements, the IND targeted for Q1 2027, Ceruvia’s Phase 1 data on the same molecule in Q2 2027, efficacy data from 2028. And for the sector, Helus’s Phase 3 data and Compass’s NDA are both due in Q4 2026.
Next comes the FDA IND in Q1 2027, then Phase 1A in healthy volunteers and Phase 1B in cluster headache and migraine in parallel, then Phase 2, on a budget of about C$105M funded by the raise. 2-bromo-LSD is not a scheduled drug in the US or Canada. And $BETR.CN doesn't need LSD or any scheduled prodrug for production.

Risks:
BETR-001 is not yet officaly in humans, the 2010 data are small and open-label, the IND slipped for years before the money came, and the CEO was fined C$50K in 2025 for late insider reports. The 1957 molecule isn’t patentable; Ceruvia develops the same drug.

Plus a legal dispute: director Steven Sangha lent C$2M in June, convertible at C$0.06 per share plus warrant. After the C$0.25 raise the company repaid him in cash; he returned the C$2M and is suing to convert into 33.3M shares + 33.3M warrants. Worst case: 4-7.5% dilution.

Thesis:
psychedelics have shown they can win Phase 3 trials and billion-dollar deals. The next leg is plasticity without the trip, and BetterLife is a funded, single-asset bet on exactly that, with about two-thirds of its value in cash and lots of start participants in the space lined up behind it.
Not investment advice.

[Jan-Christian Ehlers @JanChristan]

@HenryCryptoG hey, none you have to go through IBKR. They plan to list on NASDAQ soon.

[Jan-Christian Ehlers @JanChristan]

Closed
$SHMD
$AMBA
For profit

Increased
$BETR.CN

Also doubled Down on $XE

[Elon Musk @elonmusk]

Since I’ve been asked a lot:

Buy stock in several companies that make products & services that *you* believe in.

Only sell if you think their products & services are trending worse. Don’t panic when the market does.

This will serve you well in the long-term.

[Jan-Christian Ehlers @JanChristan]

@aleabitoreddit @EscapeTheLab

[Directionally Correct @directioncorrec]

Josh brolin is long semiconductors and Ben affleck is trying to cosplay as an autistic tech founder what stage of the cycle are we in

[Jan-Christian Ehlers @JanChristan]

@HedgeyeComm lol you fool

[tetsuo @tetsuoai]

The first 40 seconds of WANTED.

[Ren @ren_stocks]

I had the chance to sit and dig deep into Bank of America's latest AI data center forecast, and it's the most bullish table I've seen this year.

BofA sees AI data center systems going from ~$280B in 2025 to ~$2.2T by 2030.

That's roughly 8x in five years and a 40% CAGR from 2026. By then, AI makes up 86% of all data center spend, up from 55% last year, while non-AI data centers grow just 7% a year.

BofA has the memory market going from $110B in 2025 to ~$494B in 2026, taking memory from 22% of data center spend to 59% in a single year. DRAM alone passes $1T by 2030.

Interconnect is the fastest-growing part of the stack.

AI networking grows 43% a year to ~$394B, AI switching 51%, and optical reaches $111B. Co-packaged optics goes from basically zero to $14B by 2030.

AI CPUs for agentic workloads grow from $18.5B to $180B, a 44% CAGR.

Every AI layer on this table compounds three to four times faster.

I’m positioning myself to capture most of the up and coming upside.

You simple are not bullish enough for the AI trade.

[Drey @thedreydossier]

pov: the other woman is an autonomous drone.

[Jan-Christian Ehlers @JanChristan]

If anybody reading this wants to talk algo trading in crypto. Especially Infra and execution DM me.
I’ll keep going anyways lol 😆

[Jan-Christian Ehlers @JanChristan]

@karlbooklover You might try and use Hoff-Lead-Lag-Transformation as a Signal. Very interesting model technique.
Can I send you a DM? :)
Best!

[Jan-Christian Ehlers @JanChristan]

This and suffering a drawdown 😝

[Jan-Christian Ehlers @JanChristan]

@DualAcies That’s probably what makes you money in the end :)

[Jan-Christian Ehlers @JanChristan]

Bought $CCXI

Wow seem like I am an disciple of the guy with white hair. (lolz)

Really do have my own thesis.

AI = Prometheus

Fund him!

[Jan-Christian Ehlers @JanChristan]

@chooi_jeq Chill tf out; touch some grass 😂

[Elon Musk @elonmusk]

@Geiger_Capital Instagram is for girls

[▲ Marktechpost AI @Marktechpost]

NVIDIA just released Kumo Tabular, an open tabular foundation model that predicts new rows in a single forward pass. It ranks first on TabArena with an Elo of 1950, and NVIDIA reports it runs 17x faster than LimiX-2.

The setup will feel familiar if you know TabPFN or TabICL. You pass labeled rows as context, and the model predicts the labels of new rows. There is no training, no tuning, and no feature engineering.

Under the hood, it uses column, row, and in-context attention. 4 [CLS] tokens compress each row, so cost stops depending on column count. Query rows attend only to the context, so its keys and values are computed once and reused. A length-aware attention temperature keeps attention sharp as tables grow.

It comes in Small, Medium, and Large sizes, from about 28M to 215M parameters, and was pretrained only on synthetic tables. NVIDIA reports it also places first on BeyondArena, TALENT, and ScoringBench. It reads numerical and categorical columns natively, and text or timestamps need preprocessing.

The weights are released under OpenMDW-1.1, which allows commercial use. It runs through NVIDIA's new GPU-native structured-data-models (SDM) library......

Full analysis: https://t.co/hChqR3sBhm

Model: https://t.co/XJrkb6N8xl

GitHub: https://t.co/kHcuYh2qXm

Technical details: https://t.co/ihCl5WXayq

@NVIDIAAI @nvidia @NVIDIAAIDev