GNLX — Genelux Corporation

Company context · prepared 2026-08-11 · company sweep 2026-08-11 · USD · framework v5.6.1

How to read this. Written for a reader who has not studied pharmacology or finance. General terms are defined once in framework/04-glossary.md. Tags: [VERIFIED — source] / [UNVERIFIED] / [WEB ESTIMATE — source, date].

Program analyses for this company are the sub-folders listed in C.2.

C.0 Company-tier coverage — CLEARED

Verdict: CLEARED. No mandatory row reads NOT CALLED and none reads BLOCKED.

C.1 What the company is

Genelux Corporation is a clinical-stage cancer-drug company based in Westlake Village, California, incorporated in 2001 and listed on the Nasdaq Capital Market. It has no product revenue and has never sold a medicine [VERIFIED — BPIQ fetch_company_info 2026-08-11; 2025 Form 10-K reporting revenue of $0.01M against a net loss of $32.1M, TradingView summary 2026-03-24].

It is a one-molecule company, and that is stronger than “single-asset”. The pipeline feed returns three rows and all three are the same drug, Olvi-Vec (olvimulogene nanivacirepvec), in three different cancers [VERIFIED — BPIQ fetch_company_drugs 2026-08-11]. A company with several assets can survive one failure. A company with one molecule in three indications cannot diversify away a result that speaks to the molecule itself.

The lead programme is Olvi-Vec in platinum-resistant or platinum-refractory ovarian cancer, in a randomised Phase 3 trial called OnPrime (also GOG-3076) run with the GOG Foundation, a co-operative research group in gynaecologic cancer [VERIFIED — ClinicalTrials.gov NCT05281471, read 2026-08-11]. The company also holds a technology platform it calls CHOICE and a library of engineered vaccinia viruses, and has out-licensed Olvi-Vec in China, Taiwan, Hong Kong and Macau to Newsoara BioPharma [VERIFIED — BPIQ fetch_company_info, company description].

The share price this document is written against is $2.72, and its date is 2026-08-10. The connector carries no price date, so the date was established rather than assumed — see C.4.

C.2 Pipeline — every program

Three rows are listed because three rows were returned. No row is marked Failed or Discontinued. The drug name is effectively identical on all three — one row carries a trailing space — so the integer bpiq_drug_id is the only thing that separates them. Programs are keyed on that integer throughout, never on the name.

All three rows carry has_catalyst: true and all three carry the same placeholder date, 2026-12-31. None of the three discloses a day. Two of them disclose only a calendar year. The consequence for the lead programme’s own attribution is worked through in olvi-vec-platinum-resistant-ovarian/README.md under Attribution.

C.3 Financial position

Second-quarter 2026 results had not been published when this sweep ran, and that is a finding rather than a gap in the sweep. Two independent sources agree: the EDGAR filing index for CIK 0001231457 shows the newest periodic filing is the Q1 2026 10-Q of 2026-05-07, with nothing since 2026-06-22 but Form 4s, and the BPIQ press feed, updated 2026-08-10, carries no results release [VERIFIED — EDGAR submissions and BPIQ fetch_company_press_releases, both read 2026-08-11]. The 2025 equivalent came on 2025-08-07 [VERIFIED — press feed], and a Reuters earnings preview dated 2026-08-02 expected a loss of 20 cents a share [WEB ESTIMATE — TradingView/Reuters, 2026-08-02]. The statutory deadline for a non-accelerated filer is 45 days after quarter end, i.e. 2026-08-14. So the filing is late against the company’s own precedent but not yet late against the rule. The consequence for this document is concrete: every cash, burn and runway figure above is a 2026-03-31 figure, and the refresh that would replace it is days away.

The company is not funded past its own catalyst with any margin. The readout window judged in the program document runs from 2026-10-01 to 2027-04-30. On the connector’s burn figure roughly $8.0M would remain at 2026-12-31; on the net-loss route, nothing would [UNVERIFIED — modelled from the verified cash and burn figures; assumes a flat burn and no financing]. A company reading out a pivotal trial with one to two quarters of cash left has almost no ability to wait for a better price before raising, whichever way the result goes.

Dilution history. Every raise in the 24 months to 2026-08-11, plus the standing facility.

The at-the-market facility was essentially undrawn through 2026-05-03, and that is new information. This was the single largest hole in the previous version of this document, and the EDGAR share count closes most of it. The filed share count was 44,805,811 at 2026-03-16 (the 2025 Form 10-K cover) and 44,840,416 at 2026-05-03 (the Q1 2026 10-Q cover) — a rise of 34,605 shares across the seven weeks that span the facility’s own establishment on 2026-03-19 [VERIFIED — EDGAR XBRL companyconcept, dei:EntityCommonStockSharesOutstanding, read 2026-08-11]. At $2.72 that is about $94,000 of stock, which is restricted-unit vesting, not an equity programme being drawn. The same series shows 38,139,144 shares at 2025-12-31 rising to 44,840,416 at 2026-03-31, a difference of 6,701,272 against a January offering of 6,666,667 shares — so the January raise and routine vesting account for the entire increase [VERIFIED — EDGAR XBRL companyconcept, us-gaap:CommonStockSharesOutstanding].

What that does and does not establish. It establishes that management did not sell stock into a $2.90–$3.00 market in the six weeks after opening the facility. It does not cover 2026-05-03 to today: the next cover-date figure arrives with the overdue Q2 10-Q, and the stock has traded between $2.51 and $3.08 in the interval. The reading that fits both facts is that the full $100M remains available to be issued into a positive readout, which is a larger overhang than a partly-drawn facility would be, not a smaller one.

Is a raise before the next catalyst likely? The mechanism is in place, uncapped, and now confirmed unused. A $100M facility against a market capitalisation of about $122M is an authorisation to issue roughly four-fifths of the company. The company has raised equity in each of the last two Januaries or Marches, and on the twice-repeated pattern in C.7 it sells stock within 72 hours of good news [UNVERIFIED — judgement, built on the verified facility, cash, burn and share count above].

Both recent placements are under water. The January 2026 offering priced at $3.00 and the March 2025 offering at $3.50, against a spot of $2.72. Every share sold in either deal is at a loss [VERIFIED — arithmetic on the two verified placement prices].

C.4 Valuation frame

The price cache stops five trading days short of this sweep, and that is disclosed rather than papered over. data/prices/GNLX.json was last fetched 2026-08-06 and its final bar is 2026-08-05, close $2.51. The 52-week range above is unaffected — neither the high nor the low falls in the missing window — but the cache cannot yet confirm the $2.72 spot, which rests on BPIQ alone. Refreshing the cache is scripts/fetch-prices.mjs’s job and sits outside this analysis [VERIFIED — data/prices/GNLX.json, fetched field].

There has been no run-up. The stock is sitting near its floor going into its own pivotal readout, and it made a new recent low six days ago. It is 6.9% of the way up its 52-week range and only 19% above the low. On 2026-08-05 the stock fell from $2.82 to $2.51, about −11%, on 949,100 shares — roughly three times its average daily volume — and no company announcement appears in the press feed for that day or the days around it [VERIFIED — data/prices/GNLX.json; BPIQ press feed]. A third-party item dated 2026-08-06 independently describes the stock as sitting “near US$2.51” [WEB ESTIMATE — Sahm Capital, 2026-08-06]. The move is real; its cause is not in any source this sweep could read, and it is flagged in C.8 rather than explained.

At about $96M of enterprise value the market is paying roughly two-thirds of a single year of the base-case peak sales modelled in the program document, and roughly one-sixth of the low end of the published analyst price targets. That is not a price that has priced in success. Whether it has priced in failure is the question the program document takes up.

C.5 Ownership and flow

The short position is small in percentage terms and large in days-to-cover terms, and the second number is the one that matters here. 3.1 million shares short is under 7% of the company, which on its own would be unremarkable. But at an average daily volume of 174,004 shares it takes 17.9 trading days to buy that position back [VERIFIED — FINRA, 2026-07-15 settlement]. The short position also grew 3.5% over the preceding fortnight, from 3,007,000 shares at the 2026-06-30 settlement, having fallen from 3,183,715 the fortnight before that. This is a thinly traded stock in which a forced repurchase would have to be spread over weeks.

No dedicated fund is positioned either way. Zero 13F-reporting hedge funds hold this stock. There is no fast money to sell a disappointment and none to add on a win, so the reaction will be driven by a retail-dominated register and by whoever the company sells stock to afterwards.

Insider flow: there is not one open-market purchase in the entire feed. Across 250 transactions back to 2023-09-11, no officer or director bought a single share on the open market at any price [VERIFIED — BPIQ fetch_company_insider_transactions, 250 rows read in full]. The six acquisitions that carry a non-zero price are all option exercises at a fixed $6.00 strike in September and October 2023. Every other acquisition in the file is a grant recorded at a price of $0.00. This remains true through 2026-08-04, the newest transaction in the file.

The founder sold out. Aladar Szalay, recorded as a 10% owner, disposed of 1,116,788 shares for roughly $10.2M across 52 transactions between 2023-09-12 and 2024-09-13, and made no acquisition of any kind in the file. The selling ran from $27.16 down to $1.98 — it continued all the way through a 93% decline rather than stopping when the price fell [VERIFIED — same feed].

Directors have been selling into the readout. John Thomas sold 10,000 shares at $5.00 on 2025-12-01, 10,000 at $2.90 on 2026-03-02 and 10,000 at $2.98 on 2026-06-01; John Smither sold 12,000 at $2.91 on 2026-07-01. The press feed records these as 10b5-1 plan sales, which means they were scheduled in advance and are weak evidence of intent, but they are still net supply from the board in the months before a pivotal result [VERIFIED — BPIQ insider transactions; Stock Titan Form 144 and Form 4 items].

New since the last sweep: three tax-withholding sales on 2026-08-04. Chief executive Thomas Zindrick sold 3,023 shares at $2.9459, chief technical officer Joseph Cappello 653 at $2.9467 and chief scientific officer Yong Yu 653 at $2.9434 [VERIFIED — BPIQ insider transactions]. These are routine share-withholding to cover tax on vesting units, they total under $13,000, and they are not evidence of intent in either direction. They are recorded because the feed records them.

Management economics were reset immediately before the readout, and the previous version of this document under-counted the block. On 2026-06-16 the company granted restricted stock units covering 1,651,259 shares in one day — 730,601 to the chief executive, 278,906 to the chief financial officer, 233,957 to the general counsel, 174,841 to the chief technical officer, 119,318 to the chief scientific officer and 28,409 to each of four directors (John Thomas, James Tyree, John Smither and Mary Mirabelli) — plus options over a further 126,264 shares to those same four directors. That block is 3.68% of the shares outstanding, awarded five months before a binary event, at a share price near the 52-week low [VERIFIED — BPIQ insider transactions, all nine Form 4 rows summed individually]. The 2026-08-05 version of this document recorded 1,622,850 units, which counted three directors rather than four; the corrected figure is 28,409 shares higher. A separate inducement grant of 275,000 options went to the incoming chief medical officer on 2026-01-02.

C.6 Options chain and its readability

Plain takeaway: this chain cannot price the event, cannot bracket it, and no number taken from it should be used for anything. It is less readable than at the last sweep, not more. Only three strikes exist in the entire chain — $2.50, $5.00 and $7.50 — on a $2.72 share, so there is no at-the-money contract at any expiry. Total open interest across every strike and every expiry is 527 contracts, and the whole chain traded zero contracts on the sweep date.

The single largest position anywhere in the chain is 401 contracts of open interest on the 2026-09-18 $5.00 call, a strike 84% above spot, on zero volume — up by one contract since the last sweep. It looks like one institutional position or a data artifact and is not resolvable from this feed. It is flagged in C.8 and not interpreted.

C.7 Reaction to past catalysts

Five rows were returned but only four distinct dates: the 2026-01-05 lung release is recorded twice, once for small cell and once for non-small cell lung cancer, with byte-identical price fields. No new catalyst row has appeared since the last sweep.

Not one of these four is a randomised efficacy readout, and both data days carry a financing. The two clean moves are a regulatory designation (+14%) and a first-patient-dosed milestone (+12%). The two data days are each entangled with an equity offering priced within seventy-two hours. C.7 therefore establishes that this stock pays low-double-digit percentages for good news of secondary importance, gives almost no guide to the magnitude of a pivotal move, and establishes a strong, twice-repeated pattern: Genelux sells stock immediately after it releases good data.

C.8 Company data-quality flags

• max_52_week_position reproduces the documented previous-close bug to five decimal places again. The connector reports 0.05925; substituting the previous close of $2.66 recovers exactly that figure, while the last price of $2.72 gives 0.06886. C.4 derives the row from the price cache instead, per the current spec, so the bug is designed around rather than corrected for.
• market_cap carries the same previous-close bug, and this sweep proves it exactly. $119,275,512 ÷ $2.66 = 44,840,418 shares, against an EDGAR-filed count of 44,840,416 — a match to two shares. Dividing by the last price of $2.72 would imply 43,851,291 shares, which no filing supports. The filed count is used throughout C.4.
• The connector’s enterprise value cannot be reconciled to its own market capitalisation and cash, and the residual is unstable. EV − (market cap − cash) is $2,494,994 today and was $1,149,787 on the 2026-08-05 sweep, with no financing, filing or disclosure in between that could change a debt balance. Whatever the field is doing, it is not “market cap minus cash plus debt” on the same inputs. No debt figure is inferred from it anywhere in this document, which is a change from the previous version, where the residual was reported as approximately $1.2M of implied debt.
• The price feed carries no date, and the date used here is derived rather than supplied. See C.4: two BPIQ sweeps checked against the committed price cache establish that last_price is the previous session’s close. This is worth restating because it means every BPIQ price in this repo is a close, never an intraday quote.
• The committed price cache ends 2026-08-05, five trading days before this sweep. The 52-week range is unaffected, but the $2.72 spot rests on BPIQ alone and is not yet cross-checkable.
• An 11% single-day fall on 2026-08-05 has no cause in any source this sweep could read. The stock went from $2.82 to $2.51 on roughly three times average volume with no company announcement in the press feed. Recorded, not explained.
• The 2026-01-05 catalyst row is duplicated and its price fields are misattributed. Two rows (ids 6398 and 6409) carry the same date, the same source URL and byte-identical price fields for two different indications. The recorded −24.838% intraday move is almost certainly the 2026-01-08 offering rather than the 2026-01-05 data, on the press-feed evidence in C.7. Flagged, and the row is not used as an efficacy-reaction precedent.
• open_price_gap_percent disagrees in sign with intra_day_price_change_percent on two of four catalyst rows. This is the documented unreliability of that field. Neither figure is used alone.
• Options implied volatility now shows both artifacts on all four expiries. Exactly 0.01488 (the floor) on three $2.50 calls, each with a delta of exactly 1.0 and a gamma of exactly 0, and exactly 9.99512 (the ceiling) on the fourth. All four treated as missing. At the last sweep one contract still produced a real number; none does now.
• An unexplained block of option open interest persists. 401 contracts sit on the 2026-09-18 $5.00 call, an 84%-out-of-the-money strike, against 126 contracts across the rest of the chain, on zero volume. Flagged, not interpreted.
• The burn figure disagrees with the income statement by 47%. The connector’s monthly burn of $2.020M against a Q1 2026 net loss implying $2.967M. Share-based compensation is the likely reconciling item, and this company grants heavily — a 1.65M-unit block in June 2026 alone. C.3 carries both routes and neither is presented as the answer.
• Cash, burn and enterprise value are four and a half months old, and the refresh is overdue against the company’s own precedent. finance_updated_at is 2026-03-31. Q2 2026 results had not been published on either EDGAR or the press feed when this sweep ran; the 2025 equivalent came 2025-08-07 and the statutory deadline is 2026-08-14. This is the thing most likely to date this document, and it is likely to do so within days.
• The previous version of this document over-stated one insider figure. The 2026-06-16 grant block is 1,651,259 restricted stock units, not 1,622,850: the earlier total counted 28,409 shares for three directors where the feed shows four. Corrected in C.5.
• No earnings-call transcript exists for this ticker through the connector. Two further attempts this sweep returned no content. Management commentary throughout is taken from the quarterly press releases instead, which are shorter and carry no analyst questions.
• Drug names in the pipeline feed are dirty. The same molecule appears as "Olvi-Vec (Olvimulogene nanivacirepvec) + platinum-doublet + bevacizumab", "Olvi-Vec (Olvimulogene nanivacirepvec) " with a trailing space, and "Olvi-Vec (Olvimulogene nanivacirepvec)". indications_text does not separate them either — row 17406 holds “Cancer, Platinum-resistant ovarian cancer, Ovarian cancer” in one field. Programs are keyed on the integer bpiq_drug_id throughout.
• All three pipeline rows carry the same synthesized catalyst date, 2026-12-31. One derives from “H2 2026” and two from “2026”. None discloses a day. Rule 23 applies to all three, and the consequence for clustering is that no conflict between them can ever be confirmed on placeholder dates alone.