LEILA FADEL, HOST:
Thursday was a bad day on Wall Street. The NASDAQ fell 2.6%, the largest one-day decline since June. The NASDAQ is full of tech stocks - companies that make AI tools, companies that make the chips that power AI and companies that use a lot of AI. Enthusiasm around AI drove their prices up, so why are they now falling? For answers, we've called David Wessel. He's director of the Hutchins Center at the Brookings Institution. Good morning, David, and welcome back.
DAVID WESSEL: Thank you. Good morning, Leila.
FADEL: So what happened yesterday?
WESSEL: The apparent trigger was word from Meta - that's the parent of Facebook - that it'll need to raise hundreds of billions of dollars over the next few days to finance its bet on AI. That spooked the market, pushed Meta's shares down 7%. That erases $125 billion from their market value. And what's going on is it highlighted building concerns that - about the amount of debt that the hyperscalers and other tech companies are raising to make these big bets on AI.
And their bet is that AI will prove to be not only a historic advance in technology on the order of say, electricity, but also that it will produce huge profits. And that's the problem. There's some doubts about how certain those profits will be. So it wasn't just Meta yesterday. Shares of almost all the tech companies with big stakes in AI fell sharply - Nvidia, Oracle, Alphabet, Amazon, SpaceX. There's an index of companies that make semiconductors, the chips that power all these computers. It lost nearly 7% yesterday, although it's still 10% higher than it was six months ago.
FADEL: Is this a sign of an AI bubble?
WESSEL: It's really too early to make that call. You know, the shares could very well bounce back today. But the size of the bets that these companies are making are enormous. Goldman Sachs puts AI-related investment around the world at about a trillion dollars this year, more than half of that in the U.S. And a Columbia University professor, Stijn Van Nieuwerburgh, predicts that AI will - investments over the next decade measured against the share of the - as a share of GDP, the size of the economy, are three times what we spent to build the internet highway - internet - the - sorry, the interstate highway system.
And AI companies are increasingly financing all this investment with debt, borrowing money. And that's the problem. People are worried they might not make enough money to pay off this debt. You know, recall in the 19th century that a lot of people lost money on railroad bonds, but the nation's still gotten benefited from railroads, and we could be doing that again.
FADEL: The NASDAQ dropped a lot yesterday, but it wasn't just tech stocks that fell, right?
WESSEL: No. Moderna stock fell yesterday because there was some data on a cancer drug trial that was disappointing. But the broader stock market indexes like the S&P 500 also fell yesterday. There's a continuing anxiety about the Middle East. Oil prices are up again, and investors are growing increasingly nervous about France, which has a big debt burden and a political system unable to address it. I mean, the only silver lining yesterday was that interest rates came down, and that helps - may make it easier for other parts of the economy, which have not been doing as well as AI, from suffering.
FADEL: And really quickly, what will you be watching today and in the days ahead?
WESSEL: I'm watching to see if there's some ripple effect from this, whether companies begin to pull back on their AI investments. According to the Financial Times, Treasury Secretary Scott Bessent told a closed-door meeting of investors on Wednesday that he doesn't see a systemic problem with AI. But I recall that's what his predecessor said about the subprime housing market, and we know how that turned out.
FADEL: David Wessel of the Brookings Institution. Thank you.
(SOUNDBITE OF MUSIC) Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
