Revisit Edition-ROI-McKeefer: Wall Street's bet on artificial intelligence is now too big to bear: McKeefer

Monday’s column article is being redistributed to more subscribers, without any modifications. The views expressed in this article are solely those of the author, who is a Reuters columnist.

Jamie McGeever

Reuters, Orlando, Florida, October 12 - The bar for the relentlessly rising AI-driven stock market is staggeringly high. As the seemingly invincible tech giants start to show cracks, more and more signs are pointing to the reality that AI risk is now far more than just a “harbinger of trouble.”

Although the S&P 500 Index .SPX and the Nasdaq Index .IXIC reached record highs last week, a series of negative headlines related to artificial intelligence—from delayed IPOs in Australia, to reports that OpenAI, the developer of ChatGPT, saw revenues (link) miss earlier expectations—suggest the market may soon face some tough obstacles.

In the coming weeks, financial results and forecasts from major AI companies might ease these worries. But the market's expectations are sky-high—extraordinarily so. According to Tajinder Dhillon, Head of Earnings and Equity Research at London Stock Exchange Group (LSEG), analysts expect S&P 500 component companies to grow Q3 earnings over 30% from last year, two-thirds of which is due to technology and AI. Some analysts say as much as 80% of overall S&P 500 earnings growth could come from this sector.

Investors are right to remain cautious.

Last Friday, Firmus (link), an Australian data center operator backed by Nvidia NVDA.O, put its $5 billion IPO on hold. The deal would have been the country’s second largest IPO ever, and the fourth largest globally this year.

This is clearly a negative turn, but even good news out of AI companies may struggle to lift market confidence. On Thursday, Samsung (link) 005930.KS, the world’s largest memory chip maker, said it expects to post around $80 billion in Q3 profits—a record for a technology company’s quarterly results.

On the same day, Taiwan’s TSMC (link)2330.TW, the world's largest contract chipmaker, reported record Q3 revenues of $46.7 billion, beating forecasts and up 50% year-on-year. Yet, both companies' share prices fell.

Of course, isolated share price swings and specific company announcements do not necessarily reflect the overall direction of this multi-trillion-dollar industry.

But these headlines are part of an ongoing wave of less-than-encouraging news. Don't forget, there was also news last week that OpenAI’s annualized revenues in September (link) were about $50 billion—$20 billion below prior estimates.

All this points to the likelihood that this earnings season will see the entire AI sector undergo exceptionally intense scrutiny.

The Highest Bar

The scale of investments pouring into the AI revolution—and the capital needed to achieve profitability—are both staggering.

According to Columbia University professor Stijn Van Nieuwerburgh (link), the total cost of funding AI buildout from last year to 2032 is expected to hit $10.3 trillion, averaging 3.63% of US gross domestic product (GDP) per year.

Assuming a conservative 10% rate of return—which is below the 15%-20% AI investment target cited by many analysts—Van Nieuwerburgh estimates that by 2032, “the annual revenue needed to recoup investment, replenish depreciated capital, and compensate investors' opportunity cost of capital” is about $3.55 trillion.

This figure is more than 9% of GDP, roughly equal to Americans’ total spending on motor vehicles, transportation, household goods, apparel, and energy. Van Nieuwerburgh notes that, considering OpenAI and Anthropic currently generate about $100 billion per year, total revenues would need to grow over 80% per year to reach the required $3.55 trillion by 2032.

“If the economic lifespan of IT equipment shortens, the revenue requirement will increase significantly,” he says.

The Greatest Resource Misallocation in History?

Until now, profits generated by AI and tech firms have mostly lived up to Wall Street’s ever-rising expectations. Can they keep up that momentum?

Michael Cembalest from JP Morgan Private Bank notes that since ChatGPT’s launch in November 2022, the 41 AI-related stocks he tracks were responsible for 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending growth. For so few companies, this is an incredibly heavy burden.

In the coming weeks, the outlook provided by the CEOs and CFOs of these 41 companies could signal whether a market correction is near—and how severe it might be.

AI is undoubtedly the largest investment boom in US history. But will it ultimately be the biggest capital misallocation ever?

(The views expressed in this article are solely those of the author, a Reuters columnist.)

Enjoyed this column? Follow Reuters "Open Interest" (and "ROI"), your essential new source for global financial commentary. Follow ROI on LinkedIn and X.

You can also listen to the daily podcast "Morning Bid" on Apple, Spotify, or the Reuters app. Subscribe to hear Reuters journalists discuss the day’s big stories in markets and finance, seven days a week.

(To make Reuters coverage accessible to non-native English speakers, Reuters automatically translates its reports into several other languages. Because automated translation may contain errors or lack necessary context, Reuters does not guarantee the accuracy of the translated text, which is provided for reader convenience only. Reuters accepts no responsibility for any harm or loss arising from the use of automated translation.)