Firmus IPO: The Bull and Bear Case Ahead of Its ASX Listing

Firmus is building the infrastructure that powers the AI boom.

Unlike a traditional data-centre operator that primarily provides space, power and cooling, Firmus designs and operates integrated "AI factories", combining large amounts of power capacity with Nvidia GPUs and high-performance computing infrastructure.

In simple terms, Firmus takes megawatts of electricity and turns them into AI compute, selling that capacity to some of the world's largest technology companies.

The company has announced Meta and OpenAI as customers, while Nvidia is also a major strategic partner and shareholder. Firmus says it now has more than 900MW of contracted capacity across its customer portfolio, spanning seven AI factories across Australia, Singapore, Indonesia and Malaysia. Two sites are currently operational and five are under development.

Firmus has priced its IPO at A$11 per share, implying an equity valuation of around A$43.7 billion, or approximately US$30 billion.

The base IPO is seeking to raise around US$5 billion, potentially increasing to US$5.5 billion if the over-allotment option is exercised. That would make Firmus the second-largest Australian-listed IPO on record, behind Telstra's 1997 float.

Trading is currently expected to begin on the ASX on 23 October.

The institutional bookbuild is the first major test. Firmus has reportedly already received indications of interest exceeding the shares available, with the bookbuild close brought forward to 8 October.

That makes the eventual level of oversubscription and allocation particularly important for understanding where the Firmus share price could open.

Size alone makes Firmus significant, but this is more than simply a large Australian IPO.

Firmus gives Australian investors exposure to one of the dominant structural themes in global markets: the extraordinary capital expenditure required to build the infrastructure behind artificial intelligence.

There is also no obvious large-cap ASX equivalent offering exactly the same combination of data-centre infrastructure, GPUs and direct AI compute capacity. Investors looking for comparable businesses are more likely to turn to overseas companies such as CoreWeave and Nebius.

That makes establishing a fair valuation considerably more difficult.

And that is where things get interesting.

The A$43.7 billion IPO valuation represents a dramatic increase in Firmus' value.

The company was valued at more than US$10.5 billion following an August capital raising. At the A$11 IPO price, Firmus is being valued at around US$30.6 billion, almost three times that level in less than two months.

For the bulls, that increase reflects the extraordinary growth in contracted capacity, new customer agreements and the strategic value of AI infrastructure.

For the bears, investors are being asked to pay today for a very large amount of capacity, earnings and cash flow that still need to be delivered.

This could be the central argument that drives the Firmus share price well beyond listing day.

Firmus has more than 900MW of contracted capacity, which provides considerable visibility around future demand.

The challenge is converting those contracts into operational AI factories, revenue, earnings and ultimately cash flow.

Only a relatively small proportion of Firmus' contracted capacity is currently operational, with five of its seven AI factory sites still under development.

That creates significant execution risk.

Building AI infrastructure at this scale requires power, land, GPUs, financing, construction capability and customers to come together on schedule.

Delays or cost overruns could therefore have an outsized impact on future returns.

AI factories are enormously capital intensive.

Firmus needs significant capital to build facilities and purchase the GPUs required to deliver its contracted compute capacity. That means the company is likely to remain highly dependent on equity and debt markets as it expands.

This makes the marginal cost of capital particularly important.

As Firmus grows, investors may increasingly focus not simply on revenue or EBITDA growth, but on the return generated from each new AI factory relative to the cost of financing it.

If credit spreads fall and funding remains readily available, the economics of future projects become more attractive.

If borrowing costs rise, credit conditions tighten or leverage becomes excessive, those economics move in the opposite direction.

For a capital-intensive AI infrastructure company, the credit market could therefore become almost as important to the equity story as demand for AI itself.

CoreWeave provides a useful offshore case study.

The point isn't that Firmus will trade exactly like CoreWeave. Rather, it demonstrates how quickly investors can shift their focus from extraordinary AI growth to debt, capital expenditure, interest expense and free cash flow.

CoreWeave provides a useful offshore case study.

The point isn't that Firmus will trade exactly like CoreWeave. Rather, it demonstrates how quickly investors can shift their focus from extraordinary AI growth to debt, capital expenditure, interest expense and free cash flow.

The stronger the demand and execution, and the lower the cost of funding, the stronger the potential economics.

Another fascinating feature of the IPO is the proposed allocation.

Firmus reportedly plans to allocate around half of the IPO shares to selected existing strategic and financial investors, potentially allowing major existing shareholders to increase their exposure. At the same time, indications of interest have reportedly exceeded the shares available.

That combination could matter for the opening trade.

If the institutional book is heavily oversubscribed and new investors receive smaller allocations than they want, they may need to buy additional shares once Firmus begins trading.

That secondary-market demand could support a premium to the A$11 IPO price.

But there is another side to that equation.

Some existing investors entered Firmus at valuations significantly below the IPO valuation and may therefore be sitting on substantial paper gains.

Final escrow arrangements will be critical.

Market reports have suggested that some pre-IPO investors could potentially sell up to 20% of their existing holdings from the first day of trading, although the final arrangements have yet to be confirmed. One estimate suggested that this could theoretically make as much as US$5 billion of legacy stock available.

That does not mean all US$5 billion will be sold.

But it illustrates the potential scale of the supply overhang and creates an interesting tension around the opening trade.

Underallocated investors could be trying to buy.

Existing investors could be looking to monetise some of their gains.

The interaction between those two groups could become a major driver of the Firmus share price.

There is a credible argument for Firmus trading above A$11.

Demand for the IPO has reportedly exceeded available shares, while allocating around half the offer to existing investors could leave some new institutions underallocated. Firmus also carries the validation of heavyweight investors including Nvidia, Blackstone and Coatue, while Meta and OpenAI provide credibility on the customer side.

More fundamentally, global investment in AI infrastructure remains enormous. If Firmus can successfully convert its contracted capacity into operational AI factories while maintaining attractive project economics, the earnings opportunity could be significant.

There is also scarcity value. Australian investors have few direct ways to obtain this type of listed AI infrastructure exposure.

A strong opening could therefore attract institutional investors, momentum traders and retail investors who missed or received limited IPO allocations.

The bear case starts with valuation.

At approximately A$43.7 billion, investors are being asked to place a substantial value on earnings and capacity that largely sit in the future.

Execution risk is therefore significant. Five of the company's seven AI factory sites remain under development, and bringing hundreds of megawatts of capacity online is an enormous operational challenge.

Funding is another risk. AI infrastructure requires huge capital investment, potentially leaving the equity sensitive to credit spreads, refinancing conditions and interest rates.

Technology also moves incredibly quickly. GPUs are expensive assets that depreciate rapidly, while each new generation can potentially alter the economics of existing hardware.

Finally, there is the potential seller overhang. If pre-IPO shareholders use the listing to realise gains, secondary-market supply could challenge demand from new investors.

The bull case requires strong execution.

The bear case doesn't necessarily require Firmus to fail. At this valuation, disappointment may be enough.

The first major catalyst is the institutional bookbuild. Investors will want to know not simply whether it is oversubscribed, but how heavily.

The next major information event will be the prospectus, providing greater visibility around Firmus' financials, forecasts, debt structure, capital requirements and escrow arrangements.

Then comes the listing itself.

The A$11 IPO price provides the market with a clear reference point. The opening auction will provide the first genuine public-market verdict on whether that valuation is too low, too high or roughly right.

But the debate won't end with the first trade.

Where Firmus closes on day one, trades one week later and ultimately sits one month after listing may tell us far more about how public-market investors value the company.

Given the extraordinary interest in the Firmus IPO and sharply divided views around its valuation, execution risk and potential supply overhang, Pepperstone will launch a grey market ahead of the ASX listing.

The grey market price will reflect where aggregated supply and demand expects Firmus shares to open on their first day of trading.

With the IPO priced at A$11, traders can take a tactical view on whether, and to what extent, Firmus could open at a premium or discount to the offer price. Pricing will evolve as order flow, expectations and new information enter the market.

Shorting the physical stock could initially be constrained by limited stock borrow. Pepperstone's grey market provides a tradeable way to express both bullish and bearish views ahead of the listing, followed by the Firmus equity CFD once trading begins.

For traders, investors and media, the grey market can therefore provide a live gauge of expectations around one of Australia's most closely watched IPOs.

The central Firmus story is compelling.

A roughly A$43.7 billion valuation. One of Australia's biggest IPOs. Nvidia backing. Meta and OpenAI as customers. More than 900MW of contracted capacity. Huge future capital requirements. Significant execution risk. Potential seller overhang. And bulls and bears with dramatically different views on fair value.

That is exactly why Firmus could become one of the most closely watched Australian equity events in years.

It is a fascinating IPO from both a strategic and tactical perspective, and where the Firmus share price trades on the opening print, at the end of day one, one week later and one month later will continue to divide market opinion.

Pepperstone's grey market, followed by the Firmus equity CFD once the stock lists, gives traders the opportunity to express that view in either direction and over the timeframe of their choosing.

Whether Firmus ultimately trades at a premium or discount to its A$11 IPO price, the debate around fair value could make the price discovery fascinating to watch.