Will AI crash the world economy?

Artificial superintelligence could one day destroy humanity. But the AI investment boom could turn to bust and cause a global economic earthquake.

A full-blown economic crisis seems unlikely given the stories about how quickly AI is revolutionising entire industries, professions, and jobs. Every week we hear about new companies, new deals and new investments, many involving staggering sums of money.

AI tech investment has been powering the growth of the US economy; it would likely be in recession without it. But dig deeper, and the warning signs are there.

On Thursday, sales of US tech stocks increased as the market had its worst day in a month. What is happening? The hype about the potential of artificial intelligence sparked a frenzy of investment that powered an upward spiral in the stock prices of the top tech firms. Share prices are overvalued relative to likely future profits from AI.

Most analysts expected a major market correction very soon, which means investors selling shares and share prices falling. Whether that correction turns into a crash depends on the level of panic and uncertainty. It’s a confidence game, which is fine as long as there is confidence that AI will deliver.

Big tech revenues are enormous. In the last year, Meta earned approximately $134.9bn, Alphabet $350bn, Microsoft $236bn and Amazon $574bn. With vast piles of cash reserves, there is no danger they will go bust. Their AI bets are big, but they are essentially betting with other people’s money by issuing debt bonds.

Borrowing is a better option than using cash reserves because debt is tax-deductible, which means paying less tax on profits. The Financial Times estimated that Google, Amazon, Microsoft and Meta will spend more than $400bn on data centres in 2026 after spending $350bn this year.

Debt bonds from private creditors and banks are paying, so that means deep and long-term exposure. As the costs spiral and the big AI profits stretch further into the future, investors are increasingly nervous and have started selling some of this debt.

What about all those big money deals? Even they look less than impressive when you examine the details. Microsoft’s multi-billion-dollar investment in OpenAI involved taking a stake in the company and getting a share of future profits. Nvidia invested $100bn in OpenAI, and in turn, OpenAI is purchasing Nvidia’s Graphics Processing Units.

Amazon is a major investor in Anthropic. This week, Japanese tech investor SoftBank revealed it sold off $5.8bn of shares in Nvidia. If profit projections shrink or fade in the future, the entire interconnected structure will collapse.

Some UK companies are reporting significant productivity gains and increased revenue after using AI models, but more broadly, experts report a lack of targeted AI use and a human skills gap, muting more significant gains. Despite that, service sector companies are betting big on AI bots and reducing thousands of junior human jobs.

There are now a multitude of new small businesses offering an innovative angle on existing services, leveraging the magic sauce of AI to generate higher profits through productivity and efficiency gains. Some may succeed, but many resemble the pre-dot.com bust proliferation of ventures with weak business models that attracted investors because they had a website. A major market decline will hit confidence in all AI companies and could wipe out many of these companies, burning investors and banks.

The truth is, relative to the huge investment and gigantic debt, AI isn’t making that much money. The Economist estimates that revenue from AI is just $50bn a year, and an MIT survey found that 95 per cent of businesses’ AI pilots failed to make a profit.

The entire AI boom is based on a fixed idea that might not materialise; that future profits will be massive because the technology is so transformative. But what if it fails to deliver?

The journalist and author Andrew Ross Sorkin, who has written books about the global crashes of 2008 and 1929, recently said, “We will have a crash; I just can’t tell you when, and I can’t tell you how deep.”

US tech companies are in an intense AI arms race and the US is locked in an aggressive AI arms race with China.

They are all spending vast sums of money, building more data centres that need ever more energy infrastructure. But arms races never end well, and a country usually collapses.

____________________

James Sorene is a commentator and writer.

LBC Opinion provides a platform for diverse opinions on current affairs and matters of public interest.

The views expressed are those of the authors and do not necessarily reflect the official LBC position.

To contact us email opinion@lbc.co.uk