Percheron's A$2.3 Million Funds the Trial, Not the Company

Percheron Therapeutics, an Australian-listed biotech, just sold roughly one new share for every three already out there, for A$2.3 million (about US$1.5 million). The money is meant to fully fund a clinical trial of its lead cancer drug, HMBD-002, in acute myeloid leukemia — a study designed and run by Vanderbilt, paid for by Percheron. The shares went for A$0.005 each. Twenty-two months earlier the same company sold shares for A$0.08. The whole business is now worth about A$7.6 million.

There are two ways to read that headline. One: a nearly broke micro-cap found a cheap, credible route to a clinical data point. The other: it sold a chunk of itself to buy an answer it may not survive to see. Both readings are true. Which one you weigh more is the entire investment case, so it is worth understanding what the money actually buys.

The Vanderbilt arrangement is called an investigator-sponsored trial, and that label does real work. Vanderbilt designs the study, runs it, and handles the FDA. Percheron writes a financial grant, supplies the drug and technical support, and keeps full access to the data. Up to 38 patients with high-risk acute myeloid leukemia or myelodysplastic syndrome get HMBD-002 added to azacitidine and venetoclax, the standard treatment backbone, with recruitment targeted for the fourth quarter of this year.

This is the biotech version of doing things that don't scale. A company-run phase 2 to ask the same question could cost tens of millions and take years. An academic trial costs a fraction of that, because the university already owns the investigators, the ethics approval, the clinical sites, and the patients. For a company of Percheron's size, spending scarce money on the cheapest possible credible experiment is the right instinct — the same logic that says test the smallest version of the product before building the big one.

The first thing you learn about Percheron is what it stopped paying for. The company was formerly Antisense Therapeutics, built around an antisense drug, avicursen, for Duchenne muscular dystrophy. In December 2024 that program's phase 2b trial failed and was terminated. Research spending fell from A$10.8 million in fiscal 2025 to A$2.1 million in fiscal 2026. The rebuilt company is one licensed asset, one clinical bet, and an insistence on spending almost nothing until the bet is tested.

The question is whether the experiment can answer what it asks. The evidence behind an AML bet is thin, and it is worth stating precisely. HMBD-002 is a monoclonal antibody against VISTA, an immune checkpoint that is highly expressed in AML and has been linked to relapse and treatment resistance. That rationale is real. But the clinical evidence comes from a phase 1 study in 48 patients with advanced solid tumors — pancreatic, colorectal, breast, lung, liposarcoma, and head and neck cancers — designed for safety, not efficacy. The best story it produced was one triple-negative breast cancer patient whose tumor shrank 27 percent after five weeks, just short of a partial response, plus several patients with prolonged stable disease. No objective responses were confirmed. The safety side was genuinely good — well tolerated up to 1,400 milligrams a week, low rates of serious treatment-related side effects, no cytokine release syndrome.

AML is not a soft landing for that evidence. It is the most common acute leukemia in adults, with roughly 20,000 new U.S. cases a year, and the most crowded corner of oncology. Azacitidine plus venetoclax is already the standard for older or unfit patients, and every competitor with a plausible molecule is testing a combination on top of it. A 38-patient academic study is a hypothesis-generating data point, not a registration path. If blocking VISTA shows a signal in this hard-to-treat population, a partner might appear. If it doesn't, nothing happens. That is the realistic shape of the bet.

There is also the matter of who gets the upside first. Percheron did not invent this drug. It licensed HMBD-002 from Hummingbird Bioscience in June 2025 for an upfront of US$3 million, milestone payments that can reach US$290 million, and royalties. The structure means Hummingbird holds the first claim on any success — the milestones and royalties come off the top before Percheron's shareholders get much. And the invoices are real and near: the license carries roughly A$1.4 million in near-term commitments to Hummingbird, half already paid in July and the rest due by the end of September.

That context explains the placement's most flattered number, too. The new shares were priced at A$0.005, described as an 8.9 percent premium to the stock's 15-day average. A premium is easier to claim when there is no real price discovery left — the stock trades so thinly that some days only a few thousand shares change hands, so the average it beats was already near zero. The raise is structured in three tranches, with up to 230 million free attached options exercisable at A$0.01, and part of it needs shareholder approval at an October meeting. The one detail that reads like conviction rather than survival is the third tranche: up to 22 million shares reserved for Dr. Michael Baker and related parties. Baker was appointed CEO in August, and he is buying in at the same price the placement sells at — the price anyone can pay today. That is the strongest signal in the announcement, and it is still a weak one.

Now the arithmetic that matters. Percheron ended June with A$4.05 million in cash and a quarterly operating outflow of a bit over A$1 million. Its full-year net loss was A$4.8 million, down from A$14.9 million the year before but still about double what this placement adds to the bank. One research note published a couple of weeks before the raise said new capital or a partnership would be needed by early 2027. This placement pushes that horizon out by a few quarters. "Fully fund the Vanderbilt trial" is true, as far as these things go. "Fully fund the company" is a different sentence.

Here is a model that fits. You are exchanging equity worth about a third of the company for one specific, cheap experiment: does blocking VISTA add anything, in 38 patients, to a standard AML regimen? Buying an experiment cheaply is sensible. The mistake is treating "fully funded" as "company funded." So watch two things on a schedule. First, recruitment: the trial is supposed to start enrolling in the fourth quarter, with a fresh batch of drug substance the company manufactured in June and says it will release in September — a real, dated commitment to check. Second, the cash count at every quarterly report, and whether a partner appears before the money runs out. The discipline here is genuine. The runway is the enemy. In a company whose only asset was licensed, whose licensor takes the first cut, and whose only currency is equity, the winning path was always going to be a partner. This raise keeps that path open for a while. It does not make it shorter.