Why Firmus's $5 Billion IPO Failed and Why Other Listings Stalled
$5 billion was the size of the Firmus float that failed in Australia in October. The company withdrew its planned stock market listing, citing market volatility. Reuters Investing.com
In Australia, the collapse was described as a blow to Australia's shrinking share market. One account described the failure as a domino to fall in the collapse of the AI trade. Investing.com AFR
Firmus did not stand alone. A string of high-profile IPOs, or first-time sales of shares to the public, were abandoned, delayed or reworked in 2026. Oura postponed its IPO and blamed uncertainty in the IPO market. Bamboo Insurance and Holtec were among failed high-profile listings in the US. OpenAI said it will not proceed with an IPO in 2026. Reuters Fortune FT Reuters
Secondary markets moved first. AI-related stocks tumbled across Wall Street, Europe and Asia. AI-linked Asian stocks slumped after top AI lab CEOs called for slowing down technology development. Reuters
That fall was concentrated. Forty-one AI-related stocks accounted for 75% of S&P 500 returns and 80% of S&P 500 earnings growth. The S&P 500 is the main index for large US stocks. Reuters
The pipeline numbers show the funding size. Anthropic's IPO could value the company at more than $2 trillion. It lost $42 billion in 2025 and spent $7.33 billion on compute, the computing power used to build AI. Reuters
The broader context here is new listings trying to price long, costly buildouts while gains rest on few names. When 41 stocks drive three-quarters of returns and four-fifths of earnings growth, broad demand that normally takes up large floats is thin. Volatility then hits bookbuilding, the search for buyers, price cuts and withdrawal risk. The window shut fast.
Looking at what this means for equity capital markets, the market for raising money by selling shares, the order matters. The pullbacks by Oura, Firmus and others, plus OpenAI stepping back from a 2026 IPO, point to price discovery failing rather than delay for better prices. With large compute spend against current losses, public investors are asked to fund multi-year plans at prices tied to a narrow earnings base. That math moves fast from Sydney to New York when AI-linked shares move together, with little room for single-company stories until breadth improves or deal terms change.
