Why the AI boom could leave investors licking their wounds
With 30 per cent of the U.S. S&P 500 index trading at more than 10 times sales, and 13 per cent of the S&P 500 by index weight trading at 20 times price-to-sales (P/S) (a level that now surpasses the tech boom of 1999-2000), we’ll take a break this week from commenting further on market valuations and leave it to you to decide whether it’s wise to be fully invested in equities, or diversified.
This article was first published in The Australian on 21 November 2025.
Perhaps more interesting is the idea that a single sector is fuelling the U.S. economy; capital spending on artificial intelligence (AI) infrastructure, which is offsetting weakness in other sectors. The AI boom, of course, is driving the stock market. And for that reason the U.S. is in something of a dilemma. If the government tries to generate more balanced or stable growth, it could trigger a stock market correction which, of course, would lead to an economic downturn.
Further risks to economic growth and stability are posed by data centre developers reportedly approaching multiple energy utilities to secure power purchase agreements for the same project. And this behaviour could adversely affect the economy and the stock market.
Let me explain.
According to reports, in a fiercely competitive arms race, AI data centre developers are overstating energy needs and keeping projects alive even after they become unviable. If reports of dubious forecasts caused by double, triple and quadruple-counting are true, consumers could end up paying for “stranded and underutilised” power plants and transmission lines.
It’s simple: the “power panic” pressure on utilities could lead to the construction of unnecessary or underutilised infrastructure, resulting in higher bills. Utilities across the U.S. are already reportedly seeking rate increases in response to new data centre demand, but individuals, homes and businesses could end up burdened with extra costs from overbuilt, unnecessary or underutilised infrastructure.
And what of growth estimates baked into AI-related share prices based on chip demand? What if that chip demand is lower because many data centre proposals are unviable and, consequently, fewer will be built than the most optimistic estimates initially implied?
Last month, AEP Ohio, a subsidiary of American Electric Power, cut its pending projects list by nearly a third, forced to focus on developers with the financial strength to complete a data centre. Similarly, California utility PG&E revised down its data centre pipeline by 400 megawatts – equivalent to about 25 data centres.
As well as energy grid, energy pricing and economic instability, there’s the instability that comes from what’s known as “hoarding”. Remember just after Covid, when everyone was looking for computer chips? I remember writing that cars purchased in 2020-2021 would have a tough time being resold in the years to come because many lacked the chips needed to operate all the luxury features expected in modern cars.
That same dynamic could be happening in the chip market again. Enterprise consumers – in this case, the AI boom participants – exaggerate their eagerness to purchase scarce goods (by forecasting unrealistic data centre builds), aiming to secure a share of the limited supply of chips. And because it is incentivised to paint the appearance of strong growth, the supplier of the goods ramps up output to meet this apparent surge. But then, for the reasons noted above, the enthusiasm and demand vaporise (it was largely illusory to begin with), growth estimates are revised down, and share price multiples are instantly unjustifiable.
An argument that top-line growth is pursued at all costs can be made by examining the receivables item on the balance sheets of artificial intelligence (AI) chip suppliers. It’s not unusual for receivables to rise at the pace of revenue. That’s because some proportion of customers inevitably will be sold goods on credit. However, what should raise questions is a rise in receivables that far outpaces revenue growth.
According to London-based MacroStrategy’s Julien Garran, Andrew Lees and James Ferguson, “over the past 30 months, Nvidia’s receivables have risen from 56 per cent of quarterly revenues up to 85 per cent.
“[Our] understanding is that the increase is largely deals whereby Nvidia sells chipsets to data centres for ‘future compute’.”
“Jensen Huang has said the compute will be for Nvidia’s service for robot and driverless car developers training on synthetic data, but it is unclear how much revenue Nvidia is receiving from this source, as they don’t split it out, despite the fact that it is a multibillion-dollar investment.”
One of the big risks investors must contend with from time to time is presented when an investment theme is treated as structural. Historically, this is a regular occurrence, and it happens when a new technology is lauded as humanity-altering.
Chip, server, data centre, and energy demand are all assumed to follow a steep, structural, north-easterly straight line, thanks to AI. But the reality is that consumer or customer demand is usually cyclical, not structural. And when a structural theme meets a cyclical commerciality, investors tend to do rather worse than they had hoped.
I think the description of the AI boom rendering the U.S. a “one-legged” economy could be apt. If the boom slows, which is likely due to the cyclical nature of demand or because energy and chip hoarding prove that demand estimates were grossly exaggerated to begin with, then not only will investors be licking their wounds, but so too will the U.S. economy.
This article was first published in The Australian on 21 November 2025.
Roger Montgomery is the Founder and Chairman of Montgomery Investment Management. Roger has over three decades of experience in funds management and related activities, including equities analysis, equity and derivatives strategy, trading and stockbroking. Prior to establishing Montgomery, Roger held positions at Ord Minnett Jardine Fleming, BT (Australia) Limited and Merrill Lynch.
He is also author of best-selling investment guide-book for the stock market, Value.able – how to value the best stocks and buy them for less than they are worth.
Roger appears regularly on television and radio, and in the press, including ABC radio and TV, The Australian and Ausbiz. View upcoming media appearances.
This post was contributed by a representative of Montgomery Investment Management Pty Limited (AFSL No. 354564). The principal purpose of this post is to provide factual information and not provide financial product advice. Additionally, the information provided is not intended to provide any recommendation or opinion about any financial product. Any commentary and statements of opinion however may contain general advice only that is prepared without taking into account your personal objectives, financial circumstances or needs. Because of this, before acting on any of the information provided, you should always consider its appropriateness in light of your personal objectives, financial circumstances and needs and should consider seeking independent advice from a financial advisor if necessary before making any decisions. This post specifically excludes personal advice.
Buy and hope.