The world is holding its breath as it waits to find out whether this year’s estimated $US1tn investment in artificial intelligence is resting on a castle made of sand.
Last week, the Nasdaq 100 – an index of the 100 biggest non-financial stocks on the American tech-heavy Nasdaq exchange – dropped by 5.5%, triggering the usual angst that the AI boom is about to go bust.
The recent rot may have started with the failure of Firmus, an Australian datacentre company, to list on the ASX.
The multi-billion dollar initial public offering was, at one stage, on track to make Firmus worth more than Telstra – even before having delivered on its promised earnings.
That proved to be a fever dream of its founders and a wake-up call to investors in Australia and around the world to reconsider the potential value of similar businesses.
If this is the beginning of the end of the AI boom, it is only the very beginning, and tech stocks bounced back in Wall Street trade on Monday night.
Gerard Minack, of Minack Advisors, said his view was that investors were fundamentally overestimating the returns that AI providers will be able to deliver.
“While I am becoming more confident that returns on AI investment will disappoint, I have no confidence calling the precise top of the AI market,” Minack said.
Despite the recent wobbles, the Nasdaq 100 is still worth nearly three times more than at the start of 2023.
On the S&P 500, meanwhile, tech titans Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla now account for 35% of the American sharemarket index, up from 12% a decade ago.
The total market capitalisation of the so called “magnificent seven” companies has increased by more than 50% in the past two years.
Raisah Rasid, a Singapore-based global market strategist at JP Morgan Asset Management, said what she was seeing in the recent sell-off was “a valuation reset, rather than the market questioning the efficacy of AI itself”.
“This is still a long-term, structural transformation trend,” Rasid said. “This is very different from the dotcom bubble.”
Rasid sees two potentially more meaningful threats. First, that an open source AI model, potentially from China, could undermine the economics of AI and the hundreds of billions of dollars of investment – a concern shared by Minack.
The second, said Rasid, is the risk that popular opposition gains a critical mass: for example, anger about datacentre developments or that AI will do more harm to workers and society than good.
The imminent quarterly earnings season in the US will deliver the latest verdict on whether the AI companies are worth their extremely high valuations. But Minack said sharemarkets can continue to climb well after warning signs over earnings emerge.
Sally Auld, NAB’s chief economist, agreed that last week’s pullback in valuations looked more like investors taking a breath rather than the entire trade running out of puff altogether.
after newsletter promotion
“When you look back in time, every big new technology, from railroads to the internet, we always overbuild capacity. Why do we think this one will be any different?”
“It suggests investors may be taking a more balanced approach. It doesn’t mean the whole thing is a disaster; there’s a reassessment under way.”
The recent market action may also be evidence that investors are becoming more mindful of the risks associated with tech companies’ growing reliance on debt.
Companies like Amazon, Microsoft, Google and Apple have gone on a borrowing binge to help pay for spending on datacentres and associated AI infrastructure. These American “hyperscalers” spent $US425bn in AI-related capital expenditure in 2025, Minack said, and are on track to invest $US800bn this year and over $US1.1tn in the next.
Corporate bonds issued by these companies now account for nearly 10% of the global investment-grade credit market, from 0.5% at the end of 2024, according to the International Monetary Fund.
The fast-growing financialisation of the AI story helps explain why the IMF last week put the bursting of the AI investment bubble at the top of a trio of risks that could threaten the global economy, alongside rapidly rising borrowing costs and the threat of a major cyber-attack.
Another cause for concern in the AI market is the growing “circularity” of funding arrangements between companies – where, for example, a manufacturer agrees to help pay to build a datacentre in return for the centre owner buying its microchips.
These circular financing deals are only as strong as their weakest link.
“If the demand for AI is not there, then not only does the datacentre not get built, but also the chips don’t need to be made,” Auld said. “But that’s far more of an issue for the US than in Australia.”
In Australia, money is pouring into datacentres to power the computing needs of the vaunted AI revolution. CBA economists have estimated there will be $150bn in datacentre investment in this country by 2030, and $260bn by the mid-2030s, creating echoes of the mining investment boom of the 2010s.
The investment is meaningful, but is not a “major part” of Australia’s economy, Auld said.
“We would not be at the epicentre of any AI bust.”