Australian AI cloud firm Firmus has backed out of a planned IPO, blaming volatile market conditions.

The company, which has signed deals to provide data center space for Meta and OpenAI, was due to IPO on October 23, but has now changed its plans.

It had been hoped the listing would raise $5 billion, making it one of the largest in Australian Stock Exchange history. However, as reported by DCD on Wednesday, the company had reduced its initial share price from AU$11 ($7.65) to AU$9 ($6.26) due to lack of interest from investors at home and abroad.

Firmus IPO not in firm's "best interests"

In a statement shared with CNBC, Firmus said the terms of the IPO did not match the strength of the company’s proposition and its outlook for growth.

“The board therefore concluded that proceeding with the offer was not in the best interests of the company and its shareholders," the Firmus statement said. "Firmus will now pursue capital from the private markets and consider alternative public and private market options.”

It caps a bad week for Firmus, after it was reported it had pulled the plug on its collaboration with CDC Data Centres, dubbed Project Southgate, that could have seen 1.6GW of capacity built in Australia. Only 43MW was delivered under the agreement, though Firmus is seemingly still going ahead with the build-out, but without CDC.

In August, the company raised $2bn from backers including Nvidia and Blackstone, following a previous $505m equity investment round in April. The August round gave Firmus a post-money valuation of more than $10.5 billion.

It has signed agreements with Meta and OpenAI to provide data center capacity in Indonesia and Malaysia.

Are the public markets shying away from backing AI data center firms?

News of the IPO being canceled will call into question the appetite of the public markets to bet on the future growth of neocloud companies like Firmus, which are planning massive AI data center build-outs but have little track record of being able to deliver infrastructure at scale.

It may also set nerves jangling at other AI cloud firms, such UK-based Nscale and US firm Lambda, which are planning listings of their own in the coming months.

Neil Osnato, founder of infrastructure advisory firm Persistence Analytics Group, said he does not believe the problems experienced by Firmus are "evidence that AI infrastructure demand is unreal," but that "capital markets are beginning to distinguish more sharply between future demand narratives and demonstrated execution."

Osnato said: "The key distinction is that market enthusiasm, customer agreements, announced capacity and delivered operating infrastructure are not the same evidentiary state.

"In Firmus’s case, the contrast is especially visible: 1.6GW was contemplated under Project Southgate, while only roughly 43MW was actually delivered under that collaboration."

He added that, in future, companies like Firmus will have to provide evidence they can deliver on their commitments, not just raise funding and attract customers, if they are to appeal to investors on the public markets.

"That distinction becomes particularly important in neocloud financing because extremely large capital commitments can be made while delivery capability, customer concentration, power availability, construction execution and future utilization remain dependent on one another," Osnato said.