Fears of a bursting of the AI investment bubble, which have been increasingly voiced for some time, are now manifesting themselves both on the stock market and in investment decisions.
AI and tech stocks took a hit on Wall Street this week when the private capital group Blue Owl announced it would not be going ahead with a $10 billion deal to build a data processing centre for the tech firm Oracle in Saline Township, Michigan.
Blue Owl has been the primary backer for Oracle, funding a $15 billion site in Abilene, Texas and an $18 billion project in New Mexico.
The Larry Ellison owned company has been engaged in a race to try to catch up with Google, Amazon and Microsoft to capitalise on AI and has been making large investments funded in the main by debt.
Oracle’s debt has increased substantially this year, rising from $92.6 billion at the end of May to around $111 billion today, an increase of 44 percent compared to the year before. And it is set to go even higher with Morgan Stanley predicting it will reach $290 billion by 2028.
Oracle shares tumbled on the news and are now down 46 percent since they reached their peak in early September. But Oracle is not the only company to be caught in the slide. The high-tech companies Broadcom and CoreWeave have experienced significant falls. In the case of Coreweave, this amounts to a 65 percent decline, with its share falling from a high of $186 earlier this year to $64 in a situation which has been described as “getting worse by the day.”
The Blue Owl decision to pull out of the latest Oracle project reverberated through the market because it was taken as a sign that “they’re not so bullish as [some investors] are” on the AI boom, according to one analyst who spoke to the Financial Times (FT).
Speaking to the Wall Street Journal, Rishi Jaluria, an analyst at RBC Capital Markets, said: “This is all compounding on itself. People are viewing Oracle as barometer right now and saying ‘what does this mean for chips, or power?’ There’s a lot of downstream impact.”
Steve Wyatt, the chief investment strategist at BOK Financial, raised broader issues in his remarks to the newspaper.
“I think it’s going to be a big question for the market: How patient are we going to be for those companies to get past the enthusiasm for the buildout of AI to the timeframe when we start expecting a return?”
A recent comment piece on Bloomberg by Brad Stone began by citing a warning issued two months before the Wall Street crash of 1929 which warned that “sooner or later a crash is coming and it may be terrific.” There was an initial dip on the warning but then the markets continued to power ahead because of optimism of new mass markets opened up by autos and the radio.
The article noted that tech companies were on pace to spend around $1.6 trillion annually on data centres by 2030 under conditions where the profit-making capacities of AI at this stage “remain entirely hypothetical.”
According to Gil Luria managing director at investment firm DA Davidson, whose remarks were cited: “When we have entities building tens of billions worth of data centres based on borrowed money without real customers, that is when I start worrying.”
The Bloomberg writer said he had no strong opinion on whether there was an AI bubble but then went on to make an important broader point.
“If you define a speculative bubble as any phenomenon where the worth of a certain asset rises unsustainably beyond a definable fundamental value, then bubbles are pretty much everywhere you look. And they seem to be inflating and deflating in lockstep.”
Crypto is a case in point—an asset which has no intrinsic value—but whose market value rose $636 billion from the start of the year to October before losing most of that in the two months to December. Then there is the $3 trillion market in loans by private credit.
In a recent email, cited in the Bloomberg piece, Carlota Perez, a British Venezuelan economic historian and analyst, who has studied the development of capitalist economy for decades, said the innovation in high tech had created a casino economy that was overleveraged, fragile and prone to bubbles.
“If AI and crypto were to crash,” she wrote, “they are likely to trigger a global collapse of unimaginable proportions.”
The AI boom is not only impacting the financial markets, sending Wall Street to record highs this year, but resting on an increasingly narrow group of companies it has played a major role in the growth of the real economy.
It is estimated that data centre investments have accounted for 80 percent of the increase in US private sector demand for the first half of the year. Some estimates put it even higher at 92 percent. Overall, AI-related capital expenditures make up around 5 percent of total US GDP. If this dried up for any reason or were significantly reduced the US economy would fall rapidly into recession.
There are any number of triggers that have the potential set off a collapse. These include the sudden withdrawal of financial backing from the private credit market, as seen in the case of Blue Owl and Oracle, or doubts about what revenue the massive AI data centres will generate when they come online.
In addition, there is an existential threat to the present investment model. It is based on the belief that the construction of giant centres under the control of the so-called hyperscalers will place them in control of the market, enabling then to extract super-profits because of their monopolistic position.
As was noted in a recent comment piece published in the FT: “Current AI valuations assume massive durable moats. Investors have priced in the assumption that only a few companies can build frontier AI models, allowing them to extract monopoly rents.
“But if open-source models can match the performance of closed models at a fraction of the cost, that assumption collapses.”
The entry of the open-source DeepSeek model back in January, which sent a shock wave through the markets, may have been only the first of such developments, the like of which are to be found throughout the economic history of capitalism, which have the potential to bring down the financial edifice built on AI.