AI sector faces valuation reset as investors reassess tech valuations

Market observers are increasingly suggesting that a reassessment of the potential value of artificial intelligence businesses is under way, following recent volatility in technology-heavy indices. The sentiment emerged after the Nasdaq 100, which tracks the 100 largest non-financial stocks on the American exchange, fell by 5.5% last week. This decline triggered widespread concern that the current AI boom may be unsustainable, although tech stocks subsequently recovered during Wall Street trading on Monday night. The recent market wobble has prompted a broader examination of whether the estimated one trillion US dollars in annual AI investment is built on a solid foundation or is at risk of collapse.

The initial signs of market unease may have originated from the failure of Firmus, an Australian datacentre company, to list on the Australian Securities Exchange. The multi-billion dollar initial public offering had previously been on track to value the firm higher than Telstra, despite the company not yet having delivered on its promised earnings. This outcome served as a wake-up call for investors in Australia and globally to reconsider the valuations of similar businesses. Gerard Minack of Minack Advisors stated that investors are fundamentally overestimating the returns that AI providers will be able to deliver. While he expressed growing confidence that returns on AI investment will disappoint, he noted that he lacks the confidence to call the precise top of the AI market.

Despite the recent pullback, the Nasdaq 100 remains worth nearly three times more than at the start of 2023. Within the S&P 500, technology giants including Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla now account for 35% of the index, a significant increase from 12% a decade ago. The total market capitalisation of these so-called magnificent seven companies has risen by more than 50% over the past two years. Raisah Rasid, a Singapore-based global market strategist at JP Morgan Asset Management, described the recent sell-off as a valuation reset rather than a fundamental questioning of AI efficacy. She characterised the trend as a long-term structural transformation, distinguishing it from the dotcom bubble.

Rasid identified two potentially more meaningful threats to the sector. The first is the possibility that an open source AI model, potentially from China, could undermine the economics of AI and the hundreds of billions of dollars in associated investment. The second risk is the potential for popular opposition to gain critical mass, driven by anger over datacentre developments or concerns that AI will do more harm to workers and society than good. The imminent quarterly earnings season in the United States will provide the latest verdict on whether AI companies justify their high valuations, although Minack cautioned that sharemarkets can continue to climb well after warning signs over earnings emerge.

Sally Auld, chief economist at NAB, agreed that the recent pullback appears to be investors taking a breath rather than the entire trade running out of momentum. She noted that historically, every major new technology, from railroads to the internet, has been accompanied by overbuilt capacity. This suggests investors may be adopting a more balanced approach, indicating a reassessment under way rather than a total disaster. The recent market action may also reflect growing mindfulness of the risks associated with tech companies’ increasing reliance on debt. Major American hyperscalers, including Amazon, Microsoft, Google and Apple, have increased borrowing to fund spending on datacentres and AI infrastructure.

Minack stated that these companies spent 425 billion US dollars in AI-related capital expenditure in 2025 and are on track to invest 800 billion US dollars this year, with over 1.1 trillion US dollars expected in the following year. According to the International Monetary Fund, corporate bonds issued by these companies now account for nearly 10% of the global investment-grade credit market, up from 0.5% at the end of 2024. This rapid financialisation of the AI story helped explain why the IMF recently placed the bursting of the AI investment bubble at the top of a trio of risks threatening the global economy, alongside rising borrowing costs and the threat of major cyber-attacks.

Another area of concern is the growing circularity of funding arrangements between companies, such as manufacturers agreeing to help pay for datacentre construction in exchange for the centre owner purchasing their microchips. Auld noted that these circular financing deals are only as strong as their weakest link, meaning that if demand for AI is insufficient, neither the datacentre nor the chips may be needed. In Australia, where money is pouring into datacentres to power computing needs, CBA economists have estimated there will be 150 billion US dollars in datacentre investment by 2030 and 260 billion US dollars by the mid-2030s. While this investment is meaningful, Auld stated it is not a major part of Australia’s economy, suggesting the country would not be at the epicentre of any AI bust.