The Canada Nation

Your Trusted news Source

The tech-AI bubble is deflating. Should governments ride to the rescue if it bursts?

The tech-AI bubble is deflating. Should governments ride to the rescue if it bursts?



Open this photo in gallery:

People work on the floor at the New York Stock Exchange on Tuesday. There are signs the AI peak has been reached and that trouble lies ahead, Eric Reguly writes.Seth Wenig/The Associated Press

Sergio Ermotti, CEO of Switzerland’s UBS, one of Europe’s largest banks, will not go so far as to say that the burgeoning AI industry has achieved bubble status. But he comes close.

During a talk early this week at the Lugano Global Forum, just ahead of the blockbuster Anthropic public-share sale, he said that the tech and AI markets “are still pricing perfection” in spite of their decline from recent peaks. He added that the “economy has been sustained by the AI cycle.”

The implication: If AI comes crashing down, the damage could be widespread. There were no government leaders in the room; if there were, they might have asked themselves a few tough questions. If the collapse of AI reverberates through the markets and the economy, triggering a financial crisis, are we going to come to the rescue again? Would we have the moral right and the financial firepower to do so?

Major stock indexes plunge as Anthropic fails to rekindle AI trade

The questions seem premature, given that AI has been a godsend to the markets and the economy, especially the U.S. one, and hold enormous potential to create value for years to come. In spite of its rapid progress, you could argue the industry is still in its infancy. Anthropic did not exist six years ago. On Thursday, the day of its initial public offering, it ended the trading session with a (fully diluted) market value of US$1.8-trillion.

Still, there are signs that the AI peak has already been reached and that trouble lies ahead.

Shares of Meta META-Q, owner of Facebook, are down by almost a quarter from their recent peak of US$780. Alphabet GOOGL-Q, owner of Google, is trading near its 52-week low. Nvidia NVDA-Q hit a year-low on Friday, taking its loss to 22 per cent since this time last year. Anthropic ANT-Q on Friday lost much of the gains it made on its first day of trading the day before. Apple AAPL-Q is among the few tech giants holding up well.

There is no one reason for the general sell-off. Nvidia’s fall was triggered by the sex-spy scandal that cost Jensen Huang his job as CEO – he resigned on Nov. 12. The sweeping Democratic victory in the U.S. midterm elections raised the prospect of AI regulation and a clampdown on the mushroom-like growth of AI data centres. Either scenario would slow industry growth.

Many investors fear that the sheer scale of the AI investments cannot be sustained. The Belgian-American economist Stijn Van Nieuwerburgh recently estimated that the spending on all things AI will average an extraordinary 3.6 per cent of U.S. GDP between 2025 and 2032, which is well more than that of the late-19th-century American rail boom. The stock markets appear to be wildly unbalanced and set for a correction – or worse. At last count, the top 10 tech companies most heavily geared to AI accounted for almost 40 per cent of the value of the S&P 500 index.

Decoder: Has Canada’s stock market outperformance run its course?

Surely, something has to give. The question is whether the grind-down will last weeks or months, or longer, or transform suddenly into outright rout. Neither can be ruled out. Bubbles form and tend to get punctured. Investors of a certain vintage will remember the Japanese asset-price bubble of the late 1980s; the dot-com bubble in 2000; and the mid-2000s housing and credit bubble, sponsored by the subprime mortgage crisis.

Which brings us to the possible government and central banks’ responses, or lack thereof, to a tech-AI meltdown.

In previous crises, governments and central banks moved with alacrity to minimize the damage. During the COVID-19 pandemic, governments used massive fiscal stimulus, including direct cash transfers to hundreds of millions of people, to help replace lost income. During the 2008 financial crisis and the subsequent Eurozone sovereign debt debacle, interest rates dropped and central banks launched quantitative easing (QE), essentially the printing of money to buy government bonds in an effort to drive down long-term interest rates and improve liquidity.

For the most part, the formula worked – but at a price. Governments’ debt loads soared and there is a strong argument to be made that QE helped to stoke inflation.

A tech meltdown may not come with copycat responses. With budget deficits and debt loads on the upswing, governments don’t have the fiscal wiggle room they had a decade ago, and central banks would be reluctant to get involved unless rattled consumers slow spending to the point a recession becomes imminent.

The moral argument for bailing out the markets and the economy is dubious. In the United States, certainly, and to a lesser degree elsewhere, the AI rush evolved into national greed projects. Energy- and water-gobbling data centres were built with abandon. Regulations were scant, despite warnings from the AI companies themselves that unchecked AI could destroy humanity. Donald Trump in particular resisted U.S. regulations, arguing they would hand the competitive advantage to the Chinese AI industry. No one wanted to interfere in the market euphoria.

Having, in effect, encouraged a bubble, governments – and investors, banks and underwriters – will have to live with the consequences. They hope that the bubble merely deflates slowly, as it appears to be doing now, and not burst, triggering a financial crisis that they had a hand in creating.