NextFin News - Some potential investors in Firmus Grid Ltd.'s blockbuster initial public offering are turning cautious as the Australian data centre operator approaches its stock market debut, concerned that existing shareholders could flood the market soon after the listing, according to people familiar with the matter. The worry over a potential stock overhang is compounding unease about what some investors view as an aggressive pricing strategy for a deal that values the company at about A$43.7 billion ($30.3 billion).
The hesitancy is striking because, on paper, demand for the offering is strong. Indicative orders from potential investors have already exceeded the size of the deal, and the company has moved to allocate roughly half of the IPO shares to existing shareholders - a structure that lets anchor backers such as Nvidia and Blackstone increase their stakes. The tension between those two facts is the story: the same investors who are being asked to anchor today's offering are the ones who could become tomorrow's sellers, and new money is starting to price that in before a single share has traded.
The Deal, and the Doubt
Firmus is preparing for what would be the second-largest initial public offering in Australian history, trailing only Telstra's A$10 billion listing in 1997. The company is seeking to raise as much as US$5.5 billion, including an over-allotment option, with Bank of America, JPMorgan, Morgan Stanley and Morgans leading the offering. Shares are priced at A$11 each, and trading on the Australian Securities Exchange is due to begin on October 23.
The valuation has become the flashpoint. At A$11 a share, Firmus carries an equity value of about A$43.7 billion - nearly triple the roughly US$10.5 billion post-money mark the company achieved in an early-August fundraising round. That leap, accomplished in less than two months, is what some investors call aggressive and what others call the price of admission to the AI infrastructure trade.
The overhang concern adds a second layer. An IPO allocation tilted toward existing holders means the tradable float on day one is thinner than the headline raise suggests. In the short run, that can amplify early strength if passive funds and momentum traders compete for a limited supply. But it also concentrates the eventual exit: the shareholders who take up their allocation now can supply the market later, and the prospect of that supply is already changing how new investors underwrite the risk.
Firmus declined to comment on the matter. The company has said it is seeking to capitalize on artificial intelligence demand, and it has inked large deals with high-profile clients as it builds out capacity across the Asia-Pacific region.
Why Overhang Matters More Than Headline Demand
The central question of this IPO is not whether Firmus can sell its shares. It can - the book is reportedly oversubscribed. The question is who ends up holding the stock on day one, and what they do next.
When about half of an IPO is allocated to existing investors, two mechanics kick in. First, a thin float can create a short-term squeeze: fewer shares in genuinely new hands means index funds and momentum buyers can push the price up even when long-term conviction is mixed. This is why landmark listings often pop on debut regardless of underlying debate. Second, and more important for the wary investors here, the allocation structure concentrates future supply. The anchor holders of today are the potential sellers of tomorrow, and a market that anticipates that supply discounts it upfront.
This is the classic IPO overhang problem, sharpened by the valuation trajectory. Firmus' equity value has nearly tripled since early August. Investors who bought in at the lower mark are sitting on substantial paper gains. The incentive to monetize - or at least hedge - once public liquidity arrives is real, and new investors know they are being asked to underwrite both the company's future and the earlier investors' exit.
The mechanics are unforgiving. A shareholder who entered at the August round is sitting on roughly a threefold gain. Even a partial exit after listing locks in a return that the public buyer, entering at A$11, can only hope to earn. That asymmetry is why overhang concerns tend to surface most sharply in deals where the private-to-public jump has been steep: the earlier the entry, the stronger the temptation to sell, and the more the public price must discount the eventual supply.
"Our process prevents us from effectively buying into, kind of, the hopes and dreams,"
said Kirit Hara, portfolio manager at Merlon Capital Partners.
"We're anchored towards what's actually on the table. And what's on the table in terms of operating and development, we're struggling to get to that kind of A$40-odd billion dollar market cap."
Hara also flagged the company's debt pile - approximately US$30 billion - as a concern given rising costs and interest rates. That debt, added to the equity value, implies an enterprise value approaching A$60 billion.
Pricing, Debt, and the Execution Gap
The overhang does not sit alone. It compounds a deeper debate about whether the valuation is achievable at all.
Firmus began in 2019 as a Bitcoin mining operation in Tasmania and has pivoted to AI data centre infrastructure. It currently operates facilities in Melbourne and Singapore, and it estimates that five planned data centres across Asia-Pacific, together with those existing sites, will generate about US$5 billion in annual earnings within five years. That projection is the bridge between today's price and a defensible valuation - and it is also the part of the story that requires the most faith.
"We are struggling with the fundamental arithmetic, and that there is an enormous amount of execution which is required to justify the valuation,"
said Romano Sala Tenna, portfolio manager at Katana Asset Management, who said he remained undecided on whether to participate. The company did not meet his firm's investment criteria on cash flow and earnings, he said, but could still be attractive as a trade if passive fund buying creates strong demand after listing.
That distinction - a business to own versus a trade to take - captures the split in the investor base. One camp is underwriting a five-year build-out story. The other is underwriting a shorter-term flow dynamic: index inclusion, passive mandates, and the momentum that comes with a landmark listing. The overhang worry is, at its core, a bet that the second camp will not be large enough, or persistent enough, to absorb the supply from the first.
The arithmetic behind the skepticism is simple to state and hard to execute. A US$30 billion debt stack sitting on top of a US$5 billion annual earnings target - itself five years away and dependent on five sites that do not yet exist - leaves little room for cost inflation, construction delays, or a softer-than-expected pricing environment for data centre capacity. Morningstar strategist Lochlan Halloway put the debt at around six times forecast earnings and said Firmus showed hallmarks of a "boom phase," pointing to the "wild increase in valuation in such a short period of time."
Not everyone is sitting out. Sydney's Blackwattle Investment Partners said it would not bid for the stock at all.
"Because we don't have enough confidence in the delivery of the future projects, we can't get any confidence on the valuation that underpins that,"
said portfolio manager Joseph Koh. He said his firm prefers exposure to the AI theme through companies like Goodman Group, which he argued has a better track record of delivering data centres on time and on budget.
"There are so many unknowns,"
Koh said.
"But that doesn't mean we would look to short it either."
Cyclical Boom or Structural Shift? Both Are Present
The Firmus debate is a cleaner version of the question running through global markets right now: is the AI infrastructure build-out a structural shift in capital allocation, or a cyclical boom that will mean-revert?
The structural case is straightforward. Hyperscale cloud demand, sovereign AI ambitions, and the power-and-cooling constraints of next-generation chips have created a multi-year capacity shortage. Firmus is not selling a concept alone: it has operating facilities in Melbourne and Singapore, a project on Indonesia's Batam Island being developed with DayOne Data Centers with Nvidia's support, and reported compute agreements with large clients. If data centre capacity remains scarce for years, today's valuation looks like an entry price rather than a top.
The cyclical case is equally forceful. Data centre construction is capital-intensive and slow, but it is not insulated from the cycle. When capacity finally arrives, utilization rates fall, pricing power erodes, and the earnings that justified peak valuations fail to materialize. Firmus' own five-year earnings target - US$5 billion annually - assumes five unbuilt sites come online on time and on budget, in a market where cost inflation in power, chips, and construction has been the norm rather than the exception.
The honest read is that both forces are at work. Demand for AI compute is structural; the timing and pricing of the capacity response is cyclical. Firmus is priced as if the structural leg dominates. The overhang concern is the market's way of saying the cyclical leg has been underweighted.
One wrinkle matters here. According to an investor presentation reviewed ahead of the offering, the shares of Firmus' founders would be locked up after listing under escrow arrangements, with only 10% released after one year and a further 39.9% after two years. That is a tight leash on the insiders who know the business best. But the overhang worry is not primarily about the founders. It is about the broader set of pre-IPO holders - the venture funds, corporate backers, and crossover investors who bought in during the valuation run-up and may face fewer restrictions. The escrow on founder shares limits one source of supply; it does not eliminate the concern that other early holders will look to take profits once trading begins.
What Comes Next, and What Would Prove It Wrong
For investors, the implications split by time horizon.
In the short term, the allocation structure - half the deal to existing holders, strong indicative demand, a landmark listing - argues for a firm debut and possible early strength, particularly if passive funds are forced buyers. Overhang fears tend to be priced in before they are realized, and a thin float can amplify moves in both directions. The first-day move will tell you more about flow mechanics than about the business.
Over the medium term, the stock will be judged on delivery: whether the Batam project and the other planned sites reach operation on schedule, whether reported customer commitments convert into revenue, and whether the debt stack - roughly six times forecast earnings - remains serviceable as financing costs evolve. This is where the overhang becomes real rather than anticipated. If execution slips, the same shareholders who were welcomed as anchor investors become the supply wall that new holders feared.
Longer term, the question is whether Firmus becomes a permanent piece of Asia-Pacific AI infrastructure or a boom-phase valuation that the cycle corrects. The structural demand for data centre capacity is not in doubt; the question is who captures the returns, and at what price.
Base case: the stock lists firmly, trades on the back of passive and momentum flows, and then converges toward fundamentals as project milestones land or slip. Upside case: capacity stays scarce, the five planned sites deliver, and the A$43.7 billion equity value proves to have been a reasonable entry point. Downside case: construction delays or cost overruns force a reassessment, and the overhang that investors now fear becomes the selling pressure they experience.
The signal to watch is not the first-day pop. It is whether the pre-IPO shareholders who were not bound by the founder escrow are net buyers or net sellers in the months after listing - and whether the company can show, project by project, that the US$5 billion earnings target is a plan rather than a pitch. Miss two consecutive project milestones, and the structural-demand thesis loses its near-term support.
The broader market takeaway extends beyond one Australian float. Australia's primary market has been quiet - year-to-date IPO proceeds sit just above US$1 billion after topping US$2 billion in each of the prior two years. A successful Firmus debut would reopen the door for other issuers waiting in the wings. A troubled one, particularly if the overhang materializes into sustained selling, would close it again. The company's backers are not just selling shares; they are testing whether public markets will price a two-year-old data centre operator like a hyperscale cloud giant.
Firmus is not being priced on the data centres it has built, but on the ones it has promised - and the investors now hesitating are betting that promises, unlike capacity, do not come with a lock-up.
更多独家洞察尽在 nextfin.ai.
