The Firmus float died in a private room. That is the part worth fixing

A quarter off, in a day, before a share traded

Firmus Technologies went to market at $11 a share, targeting a $44 billion valuation and what would have been the largest Australian float since Telstra in 1997. The day before it withdrew its listing application, reports had the company and its advisers weighing a cut to $8.25 and a smaller raise. That is roughly a quarter of the asking price, reconsidered in public reporting rather than in a lodged document, and then abandoned entirely.

Nothing about that is illegal and little about it is unusual. It is simply how large Australian floats are sold. A company lodges a prospectus, then its bankers spend a fortnight asking large fund managers what they will actually pay. If the answers are bad enough, the price falls or the deal dies. Here the deal died, and the reported reasons were concrete: the company’s debt, the absence of detail, and a broader nervousness about artificial intelligence infrastructure being valued ahead of being built. Firmus itself cited market volatility and said the offer would not appropriately reflect its long-term growth prospects. It will now raise privately.

The part that should bother you

If institutional investors walked away citing a lack of detail, the obvious question is what the prospectus said. A prospectus is meant to be the document that lets a stranger assess the business. When professional investors need a roadshow to work out the debt stack and how much of the planned capacity is contracted rather than aspirational — and the item notes most of the Tasmanian centres are not built yet — the public document was not carrying its weight.

That matters because the prospectus is the only thing a retail buyer ever sees. Institutions get the document and the meeting. Everyone else gets the document and the price. When the meeting is where the real information flows, the retail offer is not a smaller version of the institutional offer. It is a different, worse product sold at the same price.

The defence is that retail never got hurt here, because the deal never happened. True, and lucky. But the filter worked because this was the biggest float in a generation, with the scrutiny that attracts. Scale the same structure down to a $300 million listing that nobody writes about, and the same information asymmetry produces a completed deal instead of a withdrawn one.

What a fix looks like

The fix is narrow and cheap. Where a bookbuild changes the economics of an offer — the price range, the size, the forecast, the capital structure — the issuer should lodge the changed material publicly before the book closes, not explain it afterwards in a statement about market conditions. Australia already has the machinery for this in supplementary and replacement prospectuses. What is missing is an expectation that it gets used in real time, and a regulator willing to say that selective disclosure to a bookbuild is still selective disclosure.

ASX could help by publishing a simple annual tally: listing applications lodged, repriced, withdrawn, and at what stage. Right now the only way to learn that a float moved 25 per cent in private is to read the business pages the morning before it collapses.

The case against this argument

The strongest objection is that the bookbuild is a feature, not a bug. The whole point of asking institutions privately is that they will say candidly what they will not say publicly, and that candour is what produces an honest price. Force every revision into a lodged document and you slow the process, raise the legal cost of floating, and push more companies to do exactly what Firmus has now said it will do — skip the public market and raise privately, where there is no prospectus at all and no retail investor to protect.

That last risk is real, and it is the quiet cost of this week. Firmus has not disappeared; it has moved to a venue with thinner disclosure. The same Australian capital pools that declined at $11 in daylight may well meet the company in a private round where the terms are known to a handful of parties and the eventual holders are ordinary savers through their retirement funds, two steps removed and unable to read anything.

But that argument cuts against the status quo, not for it. If the public route is already so forgiving that institutions must extract the basics in a meeting, its main advantage over the private route — a document a stranger can rely on — has been given away for nothing. Making the prospectus actually carry the information is what justifies the cost of listing in the first place.

What this episode actually proved

Investors can still say no to a very large number attached to a very new industry. That is worth something, and it is the healthiest thing in this story. But a market that only functions when the deal is big enough to be watched is not a market that functions. The filter here was journalism and a few dozen fund managers, not disclosure. Those are not substitutes.