Thesis

Genelux (GNLX, $2.54) is a $100M micro-cap biotech with a single asset — Olvi-Vec, an oncolytic virus that resensitizes tumors to platinum chemotherapy — and a Phase 3 readout in H2 2026 that will send the stock to $15-20 or to zero. The market is pricing Phase 3 success at approximately 7%. The oncology Phase 3 base rate is 58%. Our estimate, after adjusting for the indication's failure history and the small single-arm Phase 2, is 30-55%.

The edge survives aggressive haircuts. Even at 30% P(success), EV is $6.65 — 162% above current. Breakeven is 6.5%. The question is whether there's a structural reason the market is right to price below base rate, or whether it's systematic neglect of a micro-cap in a stigmatized drug class.

The trade: equity at $2.54, lottery-ticket sizing (1-2% max), positioned before three sequential catalysts that stack favorably over the next nine months.

The 7% Problem

We can back out what the market thinks from the current stock price.

Using our scenario framework — bull $18 on Phase 3 success, base $5 on mixed signal, bear $0.50 on failure — with 20% base case probability, the market-implied probability of Phase 3 success is approximately 7%.

Seven percent. For a trial with FDA Fast Track, JAMA-published Phase 2 data, and explicit FDA guidance that PFS alone could support traditional approval.

The oncology Phase 3 base rate — the average across ALL Phase 3 trials, including the garbage — is 58%. The PRROC-specific rate is lower, given the graveyard (JAVELIN, NINJA, CORAIL, FORWARD I). But 7% is below even the most bearish reasonable estimate.

Four reasons the market might be right:

1. Pre-readout death. Cash runs out before 127 PFS events accrue. But $33.1M pro forma covers through Q1 2027, and readout is guided for H2 2026.

2. Selection bias in N=24. Single-arm, no control. Lurbinectedin showed 30% ORR in Phase 2 PRROC and collapsed to 14% in Phase 3 (CORAIL). Average Phase 2-to-3 ORR shrinkage is -4.2%, but range is -46.6% to +47.3%.

3. Class stigma. The oncolytic virus class has one approved drug (T-VEC) that was a commercial flop. Replimune's RP1 got a CRL in July 2025. Investors pattern-match "oncolytic virus" to "doesn't work."

4. Institutional abandonment. Searching 5,618 earnings transcripts for "Genelux" or "Olvi-Vec" returned zero results. Nobody with real money discusses this stock.

These are real discounts. They are not reasons for a 93% discount to base rate. To believe 7%, you have to believe Olvi-Vec is less likely to succeed than the average Phase 3 drug, despite dramatically better Phase 2 data and unusually specific FDA engagement.

The honest calibration: 30-55% probability of Phase 3 success. Every estimate in that range produces strongly positive EV.

Why We Think 30-55%

The Phase 2 data, published in JAMA Oncology (May 2023), tested Olvi-Vec in 24 patients with platinum-resistant/refractory ovarian cancer:

The effect sizes are large enough to survive significant Phase 3 shrinkage. The ORR could drop by 40% (from 54% to 32%) and still match the best Phase 3 result ever seen in PRROC — mirvetuximab's 42% in the biomarker-selected MIRASOL trial. The PFS could halve from 11.0 to 5.5 months and remain clinically meaningful against a historical benchmark under 4 months.

The mechanism matters for why this isn't just another checkpoint inhibitor entering PRROC's graveyard. JAVELIN, NINJA — the checkpoint monotherapy failures — failed because checkpoint inhibitors don't work as single agents in this setting. Olvi-Vec does something different: it primes tumors to respond to platinum chemotherapy again. Patients who stopped responding to platinum start responding. That's not incremental — it's a mechanism shift, and it's why the FDA granted Fast Track and wrote, in March 2025, that "a clinically meaningful PFS advantage... could potentially support traditional approval."

The FDA also recommended a pre-BLA meeting after topline data. For a 25-employee, $100M company. That's the FDA leaning in, not hedging.

What makes us cautious: N=24, single-arm, no control. PRROC has buried multiple Phase 3 programs that looked promising in Phase 2. The dose-response signal from the China SCLC trial (67% ORR in highest dose cohort) and early NSCLC data (60% DCR in 5 patients) suggest the mechanism is real — but that's hypothesis-generating, not confirmatory. Peer review in JAMA Oncology adds credibility. It doesn't add patients to the sample size.

Why the Market Thinks 7%

Everything around the science is on fire. The 10-K, filed March 19, 2026, reads like a distress filing wrapped around a promising drug.

Going concern — second consecutive year. Management: "These factors raise substantial doubt about our ability to continue as a going concern." Cash at year-end $14.6M. Pro forma (after a January 2026 raise at $3.00/share): $33.1M. Runway "into Q1 2027." Burn accelerating — $25.3M in 2025, up 19% from 2024.

Serial dilution at declining prices. Three offerings in 18 months: $4.00 (May 2024) → $3.50 (March 2025) → $3.00 (January 2026). Each cheaper. 41.6% dilutive overhang from options, warrants, and RSUs. Stock options repriced twice — from $9-10.50 to $6.00 in 2022, then from >$5.00 to $3.33 in September 2025. You don't reprice options twice unless the stock has been underwater for years.

Insiders selling, not buying. CEO sold 17,182 shares in November 2025. Multiple officers selling. Sales are small (sell-to-cover on vestings), but zero open-market buying at 52-week lows. If management believed the readout was going to hit, at least one of them would put personal money in at $2.54. Nobody has.

Partner stress. Chinese partner Newsoara deferred reimbursement obligations in September 2025, citing need to complete its next financing round. Deadline: December 31, 2026. Newsoara signed a new $135M deal with vTv Therapeutics in February 2026, so it's alive — but it's prioritizing new deals over existing GNLX obligations. GNLX is fronting costs for Newsoara's trial while itself on going concern.

The market sees a company in survival mode and prices accordingly. Every piece of this is public, priced, and real. The question is whether it justifies a 93% discount to the Phase 3 base rate.

Three Catalysts, Stacking

The position doesn't sit dead for nine months.

1. Replimune RP1 PDUFA — April 10 (3 weeks)

Different oncolytic virus (HSV-1, melanoma). Got a CRL in July 2025 on single-arm trial design. Resubmitted October 2025. We put 50% on approval.

Approved → oncolytic virus class validated. GNLX catches +15-25% sentiment tailwind. Rejected → GNLX takes -10-15% guilt-by-association, but Olvi-Vec's randomized Phase 3 avoids RP1's specific failure mode. Net asymmetric: approval helps more than rejection hurts.

2. Relacorilant PDUFA — July 11

Corcept's relacorilant for PRROC. ROSELLA Phase 3 met both PFS (HR 0.70) and OS (HR 0.65). Lancet published. We give it 65%.

Approved → validates PRROC as a treatable indication. Different mechanism (cortisol modulation vs platinum resensitization). Market creation > market competition at this stage. Net positive for GNLX.

3. Phase 3 OnPrime readout — H2 2026

127 PFS events needed. If positive: $15-20+. If negative: $0.50 or below. No middle ground.

The catalysts stack: class validation (April), indication validation (July), company-specific readout (H2). Each positive outcome incrementally reprices GNLX before the main event.

The Trade

Instrument: Common equity. Options chain is dead — only two expirations (Apr 17, Jul 17), 15 total OI in July, no January 2027 available. Can't position through the readout with options.

Entry: $2.54. RSI 28 (oversold), 8% of 52-week range. 7.6:1 risk/reward. Breakeven P(success) is 6.5%. Enter before RP1 PDUFA (April 10) to capture class upside on the first catalyst.

Size: 1-2% of portfolio. Full Kelly says 49% — obviously insane for a binary with uncertain probabilities and going concern status. Size for surviving the zero.

Factor profile: 88% idiosyncratic (iev regress: R² = 11.7%, β = -0.58 to SPY, +1.02 to XBI). Edge% = ≈72%, concentrated entirely in the Phase 3 probability gap.

Exit triggers:

• Phase 3 positive → hold through BLA filing, reassess at $15-20
• Phase 3 negative → immediate exit, expect $0.30-0.70
• Stock runs to $5+ pre-readout → trim 50% to lock in gains
• Cash below $10M with no financing → exit regardless

What Would Kill It

1. Phase 3 enrollment slippage. Readout pushes to 2027, cash cliff becomes binding. Desperation offering at $1-2 before data arrives. Event count (gap-zseccg) is the single most important unknown — we don't know how close they are to 127 PFS events.

2. RP1 rejection + GNLX timeline miss. Double negative: class rejection AND delayed readout could push below the $1.99 52-week low and make the going concern real.

3. Insider selling accelerates. Current pattern is small sell-to-cover. If CEO starts selling larger blocks, management sees something in the interim data we don't.

4. DSMB stops the trial. No disclosure on whether an interim futility analysis has occurred. In oncology, no news from the DSMB is generally good news — they stop failing trials. But absence of evidence is not evidence of absence.

What I Don't Know

1. Phase 3 event count. The gap. "H2 2026" could mean July or December. If events are at 60, H2 2026 is aspirational. If at 110, imminent. Company hasn't disclosed.

2. Management unscripted commentary. No earnings calls found. We have the 10-K narrative but no Q&A, no conference presentations, no unscripted tone. Transparency gap.

3. DSMB interim analysis. Has there been one? Was it clean? No disclosure.

4. Commercial viability post-approval. PRROC is crowding — at least 7 major pharma programs plus approved Elahere and filing relacorilant. Olvi-Vec's mechanism is differentiated, but commercial execution by 25 people against big pharma sales forces is a real question the $18 bull target may not fully account for.

Scenarios

Conviction

The market is wrong at 7%. Not because I have special insight into the Phase 3 data — nobody outside the DSMB does — but because 7% is below the base rate for any oncology Phase 3 trial, and Olvi-Vec's Phase 2 data and FDA engagement place it meaningfully above average, not below.

The structural reasons for the mispricing are identifiable: micro-cap invisibility, oncolytic virus class stigma, going concern scaring away capital, serial dilution signaling capitulation. Real reasons for a discount. Not reasons for pricing below the success rate of the average bad drug in an average hard indication.

The probability range is 30-55%. Size at 1-2% — small enough that total loss is manageable, large enough that a 5-7x matters. The catalysts stack favorably (RP1 April 10 → relacorilant July 11 → Phase 3 readout H2 2026).

This is a good bet at a bad company. Size accordingly.

Evidence