A MARTÍNEZ, HOST:
Tech companies are pouring hundreds of billions of dollars into AI data centers, and the big question is whether all that money represents a financial bubble. NPR's Bobby Allyn has been watching. So, Bobby, bubble or no bubble?
BOBBY ALLYN, BYLINE: That's the question, isn't it? And it really depends on whom you ask, right? So to tech executives and investors, who of course, have the most to gain by keeping this momentum going, they say, no, absolutely not. No bubble to see here. But to financial analysts, economists and other experts I've talked to, A, the situation is quite alarming.
Let's just begin with the money, right? Amazon, Google, Meta and Microsoft are set to spend $400 billion just this year on AI, and that's mostly on data centers. OpenAI says it wants to spend more than $1 trillion on data centers in the coming years. But the rub is that most AI chatbots - so ChatGPT and the rest - are not making money. One analysis from Menlo Ventures found that just 3% of customers are paying for AI. And OpenAI, the lead AI company, doesn't expect to be profitable for another five years. So some see this gusher of money as a huge overinvestment. Then there is the technology itself, which has basically plateaued, according to Paul Kedrosky. He's a venture capitalist who researches AI at MIT.
PAUL KEDROSKY: The technology is very useful, but the pace at which it's improving has more or less ground to a halt. So the notion that the revolution continues with the same drum music playing for the next five years is sadly mistaken.
MARTÍNEZ: OK. Then how do tech companies defend their spending?
ALLYN: Yeah. Well, Silicon Valley boosters say AI is a once-in-a-generation technology. They compare it to the early internet or even the railway system. And they say all this infrastructure is needed to stay globally competitive, especially with China. White House AI czar and venture capitalist David Sacks wrote on X recently that slowing down AI spending would even risk a recession, writing, quote, "we can't afford to go backwards." But if you put a magnifying glass, A, on the details of some of this spending, there is a lot of reason to have doubt.
MARTÍNEZ: Ooh. So what sort of doubts? I mean, what about the spending is raising questions?
ALLYN: Yeah. Mainly two things. The first is something known as circular spending. We're seeing a lot of this. One company will give another company billions of dollars to buy their own products, usually semiconductor chips. And this kind of investing has happened for a while, but we've never seen it happen at the scale of hundreds of billions of dollars. The fear is that those kinds of deals are basically artificially inflating what the demand actually is for AI. The second issue are what's known as special-purpose vehicles. Big tech companies are teaming up with Wall Street giants to build these data centers, and the debt they are using is being kept off the balance sheets of the tech companies. Analyst Gil Luria of the D.A. Davidson investment firm told me the industry is now up to $100 billion of debt for building AI data centers.
GIL LURIA: At this point, if the market for artificial intelligence was even to just steady in its growth, then pretty quickly, we will have overbuilt capacity. And that debt will be worthless, and those financial institutions will lose money.
ALLYN: Yeah. And Luria says it's a situation that wouldn't just be a problem for investors but would hit all of us, right? The stock market could plunge, and that could pack a punch for pensions, retirement funds, the job market. Borrowing could get harder and more expensive. So if this does happen, A, the economic impacts would be widely felt.
MARTÍNEZ: That's NPR's Bobby Allyn. Thanks a lot, Bobby.
ALLYN: Thank you, A. 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.
