Key Points
- 01Firmus cancels ASX listing plan valued at A$43.7–44.0 billion
- 02IPO had aimed to raise about A$7–8 billion at A$11 per share
- 03Only around 46MW of 912MW signed capacity is currently running
- 04Board will seek private funding after scrapping the float
Firmus shelves high-profile ASX listing
Firmus Technologies has withdrawn its planned initial public offering and pulled its application to list on the Australian Securities Exchange, abandoning a large-scale float that had drawn significant market attention. The proposed listing implied a market valuation of about A$43.7–44.0 billion, positioning the deal among the largest prospective offerings in the local market.
The company’s bookbuild had marketed shares at A$11 each, targeting proceeds of approximately A$7–8 billion. With the application now withdrawn, those plans have been set aside and the shares will not be offered to public investors on the Australian exchange under the previously announced structure.
Scale of planned raise and business build-out
The size of the proposed fundraise reflected the scale of Firmus’s expansion plans in data-centre and AI infrastructure. The business had signed capacity of about 912 megawatts but only around 46 megawatts were running, meaning the vast majority of its promised capacity remained under development.
Reporting described roughly 97% of the company’s data-centre capacity as not yet built. This gap between contracted capacity and operating assets was a central feature of how investors assessed the profile of the planned IPO, with much of the growth story still dependent on future construction and ramp-up.
Investor concerns over leverage and risk
Analysts and bankers cited projections that Firmus could reach about US$30 billion of debt by 2028 as they evaluated the listing. This large expected leverage, combined with limited current operating capacity and revenue, was reported as a key factor weighing on institutional demand for the offer.
The combination of substantial planned borrowing and a capital-intensive build-out highlighted the execution risk embedded in the business model. These factors contributed to cautious positioning among potential buyers during the bookbuild process, despite the growth potential implied by the company’s signed capacity.
Role of major backers and path forward
Firmus counts Nvidia (NVDA), Blackstone (BX) and Coatue among its major backers, with Nvidia (NVDA) both an investor and a supplier of chips to the company. The presence of these high-profile supporters did not ultimately secure sufficient investor conviction to carry the transaction to market at the proposed terms.
In its statement on the withdrawal, Firmus said the board had concluded that proceeding with the offer would not be in the best interests of the company and its shareholders. The company said it would now pursue private-market capital and explore alternative public or private options for funding its expansion plans.
Implications for AI infrastructure deals in Asia
The collapse of such a large, AI-focused infrastructure IPO underscores the level of scrutiny investors are applying to early-stage capacity build-outs and highly leveraged financing structures. For other AI and data-centre operators in the region considering public listings, the Firmus outcome highlights the need to align offer terms and business risk with institutional appetite in current market conditions.
Key Takeaways
- 01The withdrawn Firmus IPO shows that even large, growth-focused offerings with high-profile backers can fail if execution risk and leverage projections concern investors.
- 02A wide gap between contracted capacity and operating assets was central to how the market viewed Firmus, signalling the importance of demonstrated utilisation for capital-intensive AI infrastructure plays.
- 03Projected debt levels around US$30 billion by 2028 became a focal point for risk assessment, illustrating how balance-sheet outlooks can shape demand for new listings.
- 04By pivoting to private-market funding, Firmus highlights an alternative path for large AI infrastructure projects when public valuations and structures do not meet internal expectations.