Australian AI data center operator Firmus has withdrawn its planned $5 billion initial public offering, citing market volatility and unfavorable conditions after determining that the proposed terms did not adequately capture the strength of its business and long-term growth prospects.
The decision puts on hold one of Australia’s largest prospective share offerings and highlights the challenge of translating investor enthusiasm for artificial intelligence infrastructure into public-market valuations that companies are willing to accept. Firmus, which is backed by Nvidia, had reportedly planned to price its shares at A$11 apiece, valuing the company at approximately $30.6 billion.
In an emailed statement to CNBC, the company said its board concluded that proceeding with the offering would not serve the interests of the business or its shareholders.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
“The board therefore concluded that proceeding with the offer was not in the best interests of the company and its shareholders,” Firmus said.
“Firmus will now pursue capital from the private markets and consider alternative public and private market options,” it added.
The withdrawal marks a shift in Firmus’s funding strategy rather than an abandonment of its expansion plans. The company intends to continue seeking capital privately while keeping other financing routes open, including a potential future return to public markets.
However, the gap between Firmus’s reported IPO valuation and its previous funding valuation raises questions about how investors are pricing the enormous capital requirements and long-term commercial potential of AI infrastructure businesses.
Firmus had planned to raise $5 billion through the offering, which would have made it the second-largest new share sale in Australia’s history. The proposed valuation of $30.6 billion would have represented a substantial increase from the company’s valuation of more than $10.5 billion following a major funding round announced in August.
The decision to withdraw suggests that the proposed offering did not deliver terms the board considered acceptable, although the company did not disclose specific investor objections, subscription levels, or the precise market factors that led to the decision.
Firmus’s abandoned IPO comes after a rapid expansion in its fundraising and valuation over the past year. In August, the company announced a $2 billion funding round backed by Nvidia, Coatue Management, Blackstone and Jane Street. That investment brought its total equity raised over the preceding year to more than $3 billion and its valuation to above $10.5 billion.
The proposed IPO would have tested whether public-market investors were prepared to value the company at a substantially higher level, reflecting expectations for continued growth in AI computing demand.
AI data centers have become central to the technology industry’s investment cycle as companies compete to secure the computing capacity needed to develop and run increasingly sophisticated models. The facilities require significant spending on specialized processors, networking equipment, power infrastructure and cooling systems, often before the full commercial returns from their capacity become visible.
For operators such as Firmus, the opportunity lies in supplying the computing resources that AI developers and technology companies need without requiring those customers to build every facility themselves. Yet the scale of the opportunity also comes with financial challenges. Expanding capacity requires sustained access to capital, while the economics of each project depend on demand, utilization, operating costs, and the ability to secure long-term customers.
A public listing would have provided Firmus with a potentially substantial pool of equity capital and a publicly traded valuation. But it would also have exposed the company to the demands of investors assessing its growth assumptions, funding requirements and future profitability against prevailing market conditions.
Firmus said market volatility and the terms available for the proposed offering were central to its decision. Its statement that the offer did not adequately reflect the business’s strength and long-term outlook indicates that the board believed the proposed valuation or other terms failed to capture what it considered the company’s full potential.
The withdrawal also shows that strong private-market fundraising does not automatically translate into an acceptable public listing. Private investors may be willing to commit capital based on long-term growth expectations, while an IPO must attract sufficient demand at a price that satisfies both the company and investors buying into the business.
The difference is relevant for AI infrastructure firms, whose valuations are tied to expectations of future computing demand and whose expansion can require large amounts of upfront capital. When market conditions become less favorable, companies may choose to delay listings rather than accept terms they consider too low.
Firmus’s next financing decisions will be important in determining how quickly it can expand while preserving the valuation expectations established during its private fundraising.
Meta Agreements Strengthen Firmus’s AI Infrastructure Ambitions
Firmus’s growth strategy is supported by its relationships with major technology companies and its focus on AI computing infrastructure in Southeast Asia.
Last month, the company announced agreements with Meta to provide GPU computing capacity at its AI data centers in the region. The facilities are built on Nvidia’s DSX platform and are intended to support Meta’s AI research, model development and training.
The agreements give Firmus a connection to one of the world’s largest AI developers at a time when technology companies are competing to secure the computing resources needed to advance their models. They also position the Australian operator within Nvidia’s broader ecosystem, where its hardware and infrastructure platforms underpin a growing range of AI deployments.
The Meta agreements are expected to help Firmus demonstrate the commercial relevance of its data centre investments by linking its infrastructure to the computing needs of a major customer. Demand from large technology companies is important in a capital-intensive business because contracted computing requirements can help operators plan capacity and finance expansion.
However, the announcement does not disclose the financial value of the Meta agreements, the amount of computing capacity involved, or the revenue Firmus expects to generate from them. Their precise contribution to the company’s future earnings and valuation therefore cannot be established from the disclosed information.
Nvidia’s involvement is another significant part of Firmus’s positioning. Beyond its backing in the August funding round, Nvidia’s DSX platform connects the company’s infrastructure strategy to the technology supplier at the center of the AI computing boom. That relationship can support Firmus’s efforts to build facilities suited to the demands of modern AI workloads, although it does not remove the commercial risks associated with financing and operating those facilities.
Thus, the immediate question is: can Firmus secure enough private capital to sustain its expansion without the proceeds of the planned IPO? The company has said it will pursue private-market funding and consider alternative public and private financing options, leaving open the possibility of another listing when conditions improve.
Its withdrawal does not establish that investor demand for AI infrastructure has weakened across the market. Firmus specifically cited market volatility and the terms of its proposed offering, rather than announcing a change in its business strategy or disclosing a deterioration in customer demand.
Nevertheless, the decision demonstrates that the AI infrastructure boom is not insulating every company from valuation discipline. Even businesses backed by major investors and connected to prominent AI customers must balance ambitious growth plans against the cost of capital and the price investors are willing to pay.