The artificial intelligence (AI) boom has attracted billions of dollars from investors hoping to benefit from the rapid growth of data centres and computing infrastructure. But a major IPO setback in Australia suggests that enthusiasm for AI companies may no longer be enough to win investor support.
Firmus Technologies, an Australian data centre developer backed by Nvidia and Blackstone, withdrew its planned initial public offering (IPO) on October 9, 2026. The company had aimed for a valuation of around A$44 billion (approximately USD $31 billion), which would have made it one of Australia’s largest potential stock market listings. According to Reuters, the company abandoned the offering after investors raised concerns about its valuation and business prospects.
The decision highlights a growing challenge for AI companies looking to raise money. Investors who once rushed to secure a stake in promising AI businesses are becoming more careful about how much they are willing to pay and whether companies can deliver on their growth plans.
Why Did Firmus Technologies Cancel Its $31 Billion IPO?
Firmus had planned to offer shares at A$11 each, which would have valued the company at nearly A$44 billion. However, the proposed valuation proved difficult to justify to prospective investors.
According to Reuters, the target valuation was about three times the value established during a funding round in August 2026. It was also around 23 times the company’s valuation a year earlier. Such a rapid increase raised questions about whether the business could support the price investors were being asked to pay.
Market volatility also played a role in the decision. Firmus said that recent market conditions meant the terms of the offering would not adequately reflect the company’s business strength and long-term growth outlook.
However, wider market uncertainty was not the only concern. Investors were also assessing the company’s operating capacity, its plans for expansion and the risks involved in building large-scale data centre infrastructure.
Firmus said it would explore other ways to raise money through public and private markets rather than proceed with the planned listing.
ALSO READ: AI Safety and Bubble Risks Take Center Stage at Singapore Conferences
Investors Are Becoming More Careful About AI Companies
For much of the AI boom, investors have been eager to back companies that could benefit from rising demand for computing power. The rapid expansion of AI models and services has increased the need for data centres, advanced chips and cloud infrastructure.
This enthusiasm has helped companies attract large investments based on expectations of future growth. In some cases, investors have been willing to accept high valuations because they fear missing out on the next major AI business.
Firmus’s failed IPO suggests that this enthusiasm has limits. Investors still see opportunities in AI infrastructure, but they also want stronger evidence that a company can turn its plans into operating capacity and future revenue.
The change matters because data centre projects require significant spending on land, power, cooling systems and computing equipment. Building these facilities takes time, and delays can affect when a company starts generating revenue.
Investors must therefore consider more than the expected growth of the AI industry. They also need to assess how quickly an individual business can build infrastructure, attract customers and generate returns.
Firmus Faces Questions Over Its Data Centre Expansion
Firmus’s business plans have also drawn attention to the gap between its current operations and its long-term ambitions.
The company aims to build data centre capacity to support growing demand from major AI and cloud computing customers. However, Reuters reported that Firmus had built only 42 megawatts of capacity against a target of one gigawatt, or 1,000 megawatts.
That gap is important because data centre developers need to turn construction plans into working facilities before they can fully benefit from rising demand for AI computing.
The company has also shifted its strategy towards expansion in Asian markets, including Southeast Asia. Changes in business plans can create additional uncertainty for investors, particularly when a company is seeking a very high valuation.
Firmus’s relationship with data centre partner CDC Data Centres also became a concern, adding to questions about its plans and ability to execute its strategy. These issues do not necessarily mean the company cannot succeed. However, they help explain why investors may have been unwilling to accept its proposed valuation.
AI Infrastructure Companies Face Growing Investor Pressure
Firmus’s IPO setback comes at a time when investors are reassessing the cost and potential returns of the AI boom. Building the infrastructure needed for AI requires substantial investment. Companies must spend heavily before they can generate revenue from new facilities, and the returns depend on customer demand, electricity availability and the speed of construction.
Investors are also watching how established AI companies manage their spending. If major customers slow their expansion plans or take longer to commit to new infrastructure, data centre developers could face additional financial pressure.
Firmus’s withdrawal does not prove that demand for AI infrastructure is declining. Instead, it highlights the difference between believing in the long-term growth of AI and agreeing to pay a particular price for an individual company.
The distinction could become increasingly important as more AI-related businesses consider going public.
ALSO READ: Nvidia and Broadcom Face Less Risk From US AI Data Center Power Shortage
What Firmus’s Failed IPO Means for the AI Boom
The cancellation is a warning sign for companies hoping to use the AI boom to secure high valuations on the stock market. Investors may continue to support businesses with strong customer relationships, clear revenue prospects and the ability to deliver projects on schedule.
But companies with ambitious expansion plans may face tougher questions if their current operations do not match their long-term targets. The setback could also influence how future AI infrastructure IPOs are priced. Companies may need to offer more realistic valuations and provide clearer evidence of their financial performance to attract buyers.
For now, Firmus plans to explore alternative fundraising options. Whether it eventually returns to the public market will depend on its business progress, financing needs and investor appetite.
The wider message is that enthusiasm for AI remains a powerful force in financial markets, but it does not guarantee that every AI-related investment will succeed. As investors become more selective, companies may need to prove their business value rather than rely on the promise of future AI growth alone.