HELPCybin Inc.

HELP — Cybin Inc. / Helus Pharma Inc. — dossier 2026-09-23

Slot names-ai-capex-5, cloud research routine, unattended. Model-drafted, not human-verified. HELP is a Tier 0 promotion (queue.json, priority 3, 6-K filed 2026-09-23) worked in an ai-capex slot; it is a biotech_binary name and its cluster stages wait for cluster-biotech-binary. Stages closed here: S1, S2 (incl. capital allocation), S3 (gate-minimum, incl. earnings quality), S5, S7 valuation table, S8, S9.

Names. The registry id and Nasdaq ticker are HELP. The SEC filer is still CYBIN INC.; the business name Helus Pharma was adopted 2026-01-05 and the legal name changes to Helus Pharma Inc. effective 2026-10-01. The programs were renamed in the same exercise: CYB003 → HLP003, CYB004 → HLP004, CYB005 → HLP005. All amounts are US$ thousands as filed, under IFRS, with a 31 March fiscal year end.

Every figure carries the document it was read from; quotes were located by exact substring search in a text extraction of the filing at the stated URL — never re-typed.

Sources

Not read (gate-minimum, §3): the Annual Information Form (Ex-99.1, 1.4 MB) except by reference through the MD&A's own risk-factor section; the FY2025 40-F; the Q2 and Q3 FY2026 interims; the management information circular; the Form 3s and Form 4s themselves. There are no earnings-call transcripts — ROIC.ai lists none for HELP and Equibles was not spent on it; a Canadian MJDS clinical-stage issuer that furnishes conference-fireside announcements rather than hosting quarterly calls gives S3's transcript step nothing to read, and S8 is built on written milestone disclosure instead. That is a real narrowing of this name's evidence base and it is why no dodge log exists.

S1 — Source

watchlist (owner, Apple Stocks, 2026-09-20), added 2026-09-23, coverage sec_fpi, cik 1833141, cluster biotech_binary. Tier 0 promotion: queue.json promoted HELP on 2026-09-23 at priority 3 on the 6-K filed that day (QUEUE-2026-09-23). That 6-K is the AGM result and the name change — see S5. The watchlist note says "files 40-F and 6-K (Canadian MJDS), no 10-K/10-Q/Form 4"; the Form 4 part is now wrong, see S5.

S2 — Kill test

There is no revenue, no product approved anywhere, and no margin to trend. The kill test for a clinical-stage issuer is cash against burn against the readout date.

(MDA-FY26 selected annual and quarterly information; MDA-Q1-FY27.)

❗ The burn nearly doubled in FY2026 — net loss $81,607 → $147,998 — as the Phase 3 program went into the field. Q1 FY2027 alone lost $47,819 and consumed $37,070 of cash.

Runway. $166,365 of cash at 2026-06-30 against a ~$37,070 quarterly operating burn is roughly 4.5 quarters, to around mid-2027 on the current rate — and the rate is rising, not falling, because EMBRACE and EXTEND are still enrolling. Working capital was $175,127 against current liabilities of $17,568. Against that sits $82,713 of further study commitments the company expects to pay within 24 months and up to $9,500 of contingent milestone consideration under the Mindset licence, payable in cash or shares at the company's discretion (MDA-Q1-FY27, Contractual Obligations).

No going-concern qualification. The financial statements carry none; what the MD&A says instead is the standard early-stage formula, and it is worth quoting because it is the whole capital structure in one sentence: "The Company has negative cash flow from operating activities and has historically incurred net losses… The Company will be required to raise additional funds through the issuance of additional equity securities, through loan financing, or other means" (MDA-Q1-FY27).

Survives or dies: the company is funded past the event that decides it. APPROACH topline efficacy data is expected in Q4 2026 — inside three months of this dossier — with approximately $23,686 of remaining spend to that milestone as at 2026-03-31 out of a ~$49,857 total (MDA-FY26). It does not need to raise before the readout. It will need to raise after it, on terms the readout sets, and that is the correct way to think about the name: cash is not the risk, the MADRS curve is.

Capital allocation and returns

There is no ROIC to compute — invested capital is a clinical programme and NOPAT is negative by construction. What can be scored is the price at which capital has been raised and whether the company spent it where it said it would.

(MDA-Q1-FY27.)

⚠️ Three raises at successively lower prices — $7.83, then $6.51, then $4.85 — while the weighted average share count rose 134% year on year (22,401,197 → 52,318,151). That is the arithmetic of a clinical-stage issuer financing a Phase 3 into a falling tape.

The counterpoint, and it is a real one: the stock is now $13.71 (PX-2026-09-22), and the option grant of 2026-08-19 was struck at $11.78 (FORMD-2026-09-01). Whatever the market thought in June, it has re-rated roughly threefold since — which makes the unused $100,000 ATM the single most consequential live instrument on this balance sheet. It is authorised to 2027-10-17, and at $13.71 it would raise the same money for a third of the dilution the June offering cost.

Where the money went. The October 2025 prospectus supplement's use-of-proceeds table, restated in the FY2026 MD&A against actual spend as at 2026-03-31 (MDA-FY26, "Update on Use of Proceeds"):

Verdict: ⚠️ neutral — spent as stated, raised at falling prices. Every completed line came in exactly on plan and no line was re-budgeted. Note the flip side of that tidiness: the "revised estimated use of proceeds" column is identical to the original in all nine lines, which is either genuine budget discipline six months into a multinational Phase 3 or a table that has not been re-underwritten. A second detail argues for reading it carefully — the MD&A elsewhere says "The Company spent approximately $12,766 to initiate enrollment in the second Phase 3 study, EMBRACE" while this table books the same milestone at $1,196, because the table covers only the October 2025 proceeds. Two numbers for one milestone in one document, on different bases, both correct.

S3 — Filings, gate-minimum

Read: the FY2026 40-F wrapper and its MD&A (Ex-99.3) in the relevant sections — programs, use of proceeds, liquidity, risk factors; the CEO and CFO certifications; the Q1 FY2027 MD&A; the AGM 6-K of 2026-09-23 and its press release; the two Schedule 13Ds of 2026-09-09; the Form D of 2026-09-01. The full AIF (Ex-99.1) was not read; the MD&A's own risk-factor section was.

The three biggest risks management itself discloses

From MDA-FY26, Risk Factors, in the company's own words:

1. The compounds may simply not work, and the odds are known. Under Early Stage of the Industry and Product Development: "Many prescription drug product candidates never reach the stage of clinical testing and even those that do have only a small chance of successfully completing clinical development and gaining regulatory approval." And, on the specific risk of this moment: "Many companies in the pharmaceutical and biotechnology industries have suffered significant setbacks in later-stage clinical trials after achieving positive results in early-stage development, and the Company cannot be certain that it will not face similar setbacks."
2. Announced timings are estimates, not commitments. Under Achieving Publicly Announced Milestones: "the actual timing of such events may differ from what has been publicly disclosed… Any variation in the timing of previously announced milestones could have a material adverse effect on the Company's business plan, financial condition or operating results and the trading price of the Common Shares." This is the risk factor that governs the Q4 2026 readout.
3. The whole commercial case depends on drug scheduling changing. Under Regulatory Risks and Uncertainties: HLP003 and HLP004 are "classified as Schedule I drugs under the CSA" in the United States, and "The success of the Company's business is dependent on the reform of controlled substances laws pertaining to HLP003 and HLP004. If controlled substances laws are not favourably reformed in the United States, the UK, the EU, and other global jurisdictions, the commercial opportunity that the Company is pursuing may be highly limited."

A fourth, worth naming because it is where the money goes: "The Company may never have a product that is commercially successful… The level of acceptance the Company ultimately achieves may be affected by negative public perceptions and historic media coverage of serotonergic agonist substances."

The clinical facts that decide the name

HLP003 (deuterated psilocin) for adjunctive MDD — the PARADIGM Phase 3 program, initiated 2024-11-13 (MDA-FY26):

• APPROACH — n=220, randomised 1:1 to 16 mg HLP003 or inactive placebo, two doses three weeks apart, moderate-to-severe MDD (MADRS≥24) inadequately responding on a stable antidepressant. Primary endpoint: change in MADRS from baseline at six weeks. ~45 US sites. Dosing underway; topline efficacy expected Q4 2026.
• EMBRACE — n=330, 1:1:1 to 16 mg, 8 mg or inactive placebo, same population and endpoint, ~60 sites with minimal overlap. MHRA approval 2025-07-17, EU approval via the Irish Medicines Board as reference member state 2025-08-07 (Ireland, Poland, Greece), Australia 2025-08-26.
• EXTEND — rollover, up to n=550, long-term extension with optional additional 16 mg doses.
• Blinding measure disclosed: "Across all three studies, raters will be remote, independent, and blinded with no information on the dose received or the participant's dosing experience. Effects during the dosing session will be firewalled to ensure that the study team stays blinded."

❗ The read-across from the company's own Phase 2 is the most important analytical fact available on this name, and it cuts both ways. On 2026-03-05 the company reported topline results from the Phase 2 HLP004 study in GAD, whose design used "two low-dose control administrations of sub-therapeutic doses of HLP004" rather than inactive placebo. The result (MDA-FY26):

"Participants randomized to both 20 mg and 2mg dosing arms experienced meaningful subjective effects and showed clinically significant responses over Standard of Care, with 59% meeting the criteria for response and 32% for remission in the 20mg arm and a 30% responder and remitter rate in the 2mg arm at week 6."

A sub-therapeutic control arm delivering roughly half the active arm's response is the signature problem of this whole therapeutic class: participants know whether they were dosed, and expectancy does work that the statistics cannot separate from pharmacology. APPROACH and EMBRACE use inactive placebo, which makes the unblinding gap larger, not smaller — the 90-minute acute effect the company advertises as a commercial advantage ("acute drug effects lasting approximately 90 minutes and discharge readiness within approximately three hours for 100% of participants") is the same property that tells every participant which arm they are in. Against that, the headline efficacy was strong on its own terms: "Patients that received 20mg HLP004 adjunctive to Standard of Care therapy achieved mean reduction of 10.4-points (p<0.0001) in the HAM-A from baseline at six weeks", with "no drug-related serious adverse events or suicidality-related safety signals" and, at six months, "67% responders and 39% remitters" in the pooled population. Note that the pooled six-month figure mixes the arms — a durability number without a comparator.

Earnings quality

There are no earnings. The applicable checks, and what they return:

1. Non-GAAP bridge: none. The company reports IFRS net loss and does not present adjusted earnings. A clean result, and it is worth saying so explicitly rather than skipping it.
2. Share-based compensation was $11,574 in Q1 FY2027 alone against a $47,819 net loss — 24% of the quarterly loss, and roughly a third of the cash operating burn is non-cash comp. On 2026-08-19 the company granted options exercisable at $11.78 until August 2036, vesting in eight quarterly tranches beginning 2026-08-19, to a list of fourteen officers and directors, with maximum potential proceeds of $261,987 on exercise (FORMD-2026-09-01). ⚠️ Flag 1 — the grant is dated weeks before a pivotal readout, on a ten-year term, with vesting starting immediately.
3. Cash conversion: the Q1 FY2027 reconciliation is unusually clean and the MD&A states it plainly: "cash used in operating activities was $37,070 driven by a net loss for the period of $47,819 and an increase in working capital of $875, partially offset by the following non-cash items: share-based compensation of $11,574, foreign translation loss of $19 and depreciation and amortization of $31." Net loss to cash burn is explained almost entirely by SBC. No flag.
4. Working capital / receivables: no revenue, no receivables. n/a.
5. Costs moved onto the balance sheet: no capitalised development cost is evident; R&D is expensed. No flag.
6. Sector equivalent — the milestone-spend disclosure. The Canadian use-of-proceeds regime (§2 above) is the closest thing this issuer has to an earnings-quality artefact, and it holds up: planned equals actual on every completed line. ⚠️ Flag 2 is the counterpart: the "revised estimated use of proceeds" column has not moved a dollar from the original in any of nine lines, six months into a three-study Phase 3 program — a table that never changes is a table nobody is re-underwriting, and the same MD&A carries two different figures for the EMBRACE-enrolment milestone on two different bases.

EQ: 2 flags — SBC is 24% of the quarterly net loss, with a 10-year option grant at $11.78 struck weeks before the pivotal readout; the use-of-proceeds "revised estimate" is unchanged from plan in all nine lines. No non-GAAP measures are presented at all, which is a clean result.

Auditor: Zeifmans LLP, reappointed at the 2026 AGM on 99.207% of votes cast (6K-AGM-2026-09-23). A small Toronto firm auditing a company running a three-study multinational Phase 3 with $166m of cash. Not a finding on its own; noted, as the NBIS/Reanda case in this same slot shows why.

Transcripts: none exist to read. ROIC.ai lists no earnings calls for HELP. The company's 6-K stream in September 2026 is almost entirely conference appearances — Cantor (2026-09-03), H.C. Wainwright (2026-09-09), Morgan Stanley (2026-09-10), Psych Congress (2026-09-21), TD Cowen (2026-09-22) — announced but not transcribed into any source reachable here. No dodge log can be built for this name from this environment, and that is a gap rather than a clean bill.

S5 — Ownership

❗ The filing record changed character in September 2026, and it looks like a change of reporting regime. Across the whole EDGAR history for cik 1833141: the first-ever Form S-8 was filed 2026-09-01; the first-ever Form 3s on 2026-09-09 (three of them, plus one on 2026-09-17); the first-ever Form 4s on 2026-09-09 (eight since); and the first-ever Schedule 13Ds the same day. Every one of those is US domestic-issuer machinery: a foreign private issuer uses F-forms rather than S-8, and its officers, directors and principal shareholders are exempt from Section 16 — as the company's own Q1 FY2027 MD&A still says: "the Company's officers, directors, and principal shareholders are exempt from the reporting and short-swing profit recovery provisions of Section 16 of the Exchange Act" (MDA-Q1-FY27, "Foreign Private Issuer" Status). The same MD&A carries a risk factor headed "The Company May Lose 'Foreign Private Issuer' Status in the Future".

The inference — that HELP has lost or is losing FPI status and has begun reporting as a domestic issuer — is the natural reading of S-8 + Form 3 + Form 4 + Schedule 13D all appearing for the first time within nine days, but no filing read in this pass states it, and the company continued to furnish 6-Ks through 2026-09-23. This is the single most important thing to confirm on this name and it is flagged as an inference, not a fact. It matters practically: a domestic issuer files 10-K/10-Q instead of 40-F/6-K, loses MJDS, loses the Nasdaq foreign-issuer governance exemptions, and its insiders' trades become visible within two business days instead of on Canadian timelines.

The two new 13D filers are the co-founders, and they are insiders (13D-SO, 13D-GLAVINE, both event date 2026-09-01):

Both are listed as Executive Officer and Director on the Form D of 2026-09-01 (FORMD-2026-09-01). Schedule 13D — not 13G — is the control-intent form.

❗ The company is run by an interim CEO into its pivotal readout. The FY2026 annual report's CEO certification, dated 2026-06-29, is signed "/s/ Eric So — Eric So — Interim Chief Executive Officer (Principal Executive Officer)" (CERT-CEO-FY26); the CFO certification is signed by Greg Cavers (CERT-CFO-FY26). The Form D's list of executive officers and directors — Eric So, Michael Halstead, Greg Cavers, Paul Glavine, Gabriel Fahel, Alex Nivorozhkin, Aaron Bartlone, George Tziras, Amir Inamdar as officers; Theresa Firestone, Mark Lawson, Grant Froese, Eric Hoskins, Freda Lewis-Hall as directors — contains no permanent chief executive. A co-founder is running the company on an interim basis, holding 5.8% of it, three months before the data that decides it.

13F register at 2026-06-30 (13F-Q2-3S, 72 filers down from 82, 26,604,749 shares of roughly 73.1m outstanding; dollar figures are at the 2026-06-30 price, not today's):

This is the opposite of the NBIS and DOCN registers in the same slot. The buyers are healthcare specialists — Venrock, Adage, Rosalind, Sphera, Sio, Boxer, Logos, DAFNA — not dealers, and they added into the June offering at $4.85. OrbiMed trimmed. Dedicated biotech money positioning ahead of a binary readout is exactly what one would expect to see here, and it is present.

Shares outstanding. The AGM press release states that 46,452,598 shares were represented, "representing 63.58% of Helus's total issued and outstanding common shares as at the record date" (6K-AGM-2026-09-23), which implies ~73.1 million shares outstanding. That figure is arithmetic on the company's own two numbers, not a stated share count, and is used as such below.

Shelf / ATM / buyback. The 2026 ATM Program authorises up to $100,000 and runs to 2027-10-17; nothing had been sold under it as at 2026-06-30 (MDA-Q1-FY27). There is no buyback and there will not be one. A $100m unused ATM over a ~$1.0bn market cap is the single largest overhang on this name, and at today's price it is also management's cheapest financing.

S7 — valuation table (the readout), and its assumptions

A revenue-and-margin out-year model is the wrong instrument here and is not attempted. HELP has no revenue, no approved product, and one event in Q4 2026 that determines whether the equity is worth its cash or a multiple of it. What follows is a scenario table around that event, with no probability attached to any column — assigning a probability of success would require a base rate for adjunctive-MDD Phase 3 trials in this class that this pass has no sourced basis for, and inventing one would be the exact failure mode the gate rules exist to prevent.

At $13.71 (PX-2026-09-22) on ~73.1m shares, market value is ~$1.00bn. Cash was $166,365 at 2026-06-30; assume roughly $115m–$130m at the readout after two more quarters of burn. Enterprise value is therefore ~$0.87bn — the market is paying about $870m for a Phase 3 asset that has not read out. ADV is ~$33m (2.42m shares × $13.71), the thinnest of the three names in this slot by an order of magnitude.

Assumptions: the bear column values the residual as ~$120m of net cash plus a modest option value on HLP004 (whose Phase 2 read positively) and the IP estate, with limited dilution because a failed readout closes the ATM; the base and bull columns assume the unused $100m ATM is drawn into strength (hence 85m shares) and apply a 2.4×/4.6× re-rating of today's enterprise value, which is a judgement about how these assets have historically re-priced on a first positive pivotal, not an observation. The spread is ~23×, which is what a binary is. A second pivotal (EMBRACE) and the EXTEND long-term data still sit between the base case and an NDA.

What would have to be true for $13.71 to be right: the market is currently pricing something close to even odds of the base case, net of the cash. Given the Phase 2 control-arm result described in S3 — a sub-therapeutic arm at 30% response against 59% for the active dose — that is not an obviously wrong price, and it is not an obviously right one either. The honest statement is that this name is not valued, it is handicapped, and the handicapping is the owner's judgement, not the routine's.

No buy price and no position size (owner decision 2026-09-23).

S8 — Management credibility

Seven rows. All the promises are dated operational and spending commitments — there is no financial guidance, no revenue to guide, and no earnings call on which to make one.

An eighth, a resequencing that was disclosed rather than buried and is therefore context rather than a failure: "The Company previously mentioned that it intended to file a clinical trial application with the UK MHRA related to the HLP003 Program upon completion of its pre-clinical studies and CMC development. The Company had then decided that it would first proceed in the U.S. and would subsequently file a clinical trial application with the MHRA" — and the MHRA approval duly arrived on 2025-07-17 (MDA-FY26).

Promises belong to people, and the people have changed. ❗ The FY2026 annual report is certified by an Interim Chief Executive Officer (CERT-CEO-FY26), and the 2026-09-01 Form D's list of officers and directors contains no permanent CEO (FORMD-2026-09-01). The record above was largely set under a management that is no longer wholly in place; a co-founder holding 5.8% is running the company into the readout on an interim basis. Under the skill's rule this is close to a new team — thin record case. It does not quite become one, because the CFO who signs the numbers (Greg Cavers) is unchanged and the milestone-spend record is the company's, not an individual's.

Dodge log: none, and not because there are no dodges — because there are no transcripts. ROIC lists no earnings calls for HELP and the company furnishes conference appearances rather than hosting them. S8 here is built entirely on written milestone and use-of-proceeds disclosure.

Verdict: ✅ reliable (partial) — four met, one withdrawn, two open, on process rather than outcomes. Everything the company budgeted for, it spent roughly as stated and delivered on time, including a Phase 2 readout that arrived with positive topline data. The one withdrawal — the HLP005 candidate, slipped roughly a year — is the preclinical programme nobody is paying for, and it was disclosed, in a footnote. The qualifier is the whole caveat: these are commitments to do things, not to achieve them, the record has no downside quarter in it, there is no live Q&A anywhere in the evidence base, and the chief executive who made most of it has been replaced by an interim. The first genuine test of this management's credibility is three months away and it is not a promise at all — it is a p-value.

S9 — Monitoring

• Next event: ❗ APPROACH topline efficacy data, Q4 2026 — the readout that decides the equity, inside three months. Before it: the HLP004 next-study design, committed "by the end of Q3 2026", i.e. 2026-09-30, one week from this dossier — a free, dated, near-term test of whether this management hits its own dates.
• Next results filing: Q2 FY2027 (quarter ending 2026-09-30), expected mid-November 2026 by cadence (Q1 FY2027 was furnished 2026-08-14; FY2026 on 2026-06-29). Derived, not sourced.
• Filing feed: the Tier 0 EDGAR sweep covers cik 1833141 — and this name is the argument for reading 6-K exhibits: HELP furnishes roughly one 6-K a week, most of them conference announcements, and the material ones (results, AGM, financings) are indistinguishable from the noise at the wrapper level. If the FPI inference in S5 is right, the feed will change shape without notice: 10-Q and 10-K replacing 6-K and 40-F, and Form 4s giving two-business-day insider visibility that has never existed for this name before. Watch for the first 10-Q.
• Leading indicator between quarters: ❗ the unused $100,000 2026 ATM Program (effective to 2027-10-17). Drawing it at $13.71 rather than at June's $4.85 would be straightforwardly good capital allocation; not drawing it before a binary readout, with roughly 4.5 quarters of cash, is a statement of confidence that is worth reading as one either way. Also: the newly available Form 4 stream from the two co-founder 5.8% holders and the fourteen option grantees at $11.78.
• The trigger that converts this name to a buy is not a valuation level: it is APPROACH hitting its primary endpoint with an effect size that survives the unblinding critique — a MADRS separation at six weeks that is large relative to the placebo arm's own movement, in a trial where the control is inactive and the drug produces 90 minutes of unmistakable acute effect. The company's own Phase 2 in GAD is the reference case to judge it against, and there the sub-therapeutic control arm reached 30% response against 59% active. A positive APPROACH whose placebo arm behaves like that 2 mg arm is not a de-risking event; a positive APPROACH whose placebo arm behaves like a placebo is the whole thesis.
• Kill criteria (specific, testable):
• APPROACH misses its primary endpoint, or hits on a margin the placebo arm's response makes uninterpretable. This is the kill criterion; everything else is a detail.
• The Q4 2026 readout slips. The company's own risk factor warns that announced timings move, and the HLP005 candidate has already slipped a year. A slip announced without a new date, weeks before the guided window, is the shape to watch for.
• An equity raise before the readout. With ~4.5 quarters of cash and $23,686 of remaining milestone spend, there is no need to raise before Q4 2026. Raising anyway — at any price — would say something about the sponsor's own read that the disclosure does not.
• The interim CEO becomes permanent without a search being disclosed, or a second senior departure follows. The company is in the most consequential quarter of its life without a permanent chief executive.
• The name change lands and the readout does not. The legal name changes to Helus Pharma Inc. on 2026-10-01 to reflect "the Company's anticipated transformation from a clinical stage discovery and development company to a commercial-ready pharmaceutical company" — a commercial identity adopted before the data that would justify it. That is not a kill on its own. It is the sentence to re-read if APPROACH misses.