How the AI Bubble is Different

**Brian Lehrer** (0:10)
Brian Lehrer on WNYC. Now we'll talk about a potential AI bubble. Sam Altman, who leads OpenAI, Sam Altman of all people, OpenAI, the company behind ChatGPT, believes that an AI bubble is upon us, and it's not just him. The IMF, the International Monetary Fund, is calling it a risk to financial stability.
There's little doubt that AI has been fueling the stock market. Boom, the value of AI-linked companies is up an estimated $27 trillion in the last three years, according to a published stat I read, more than a third of the value of the entire US stock market. But here's the thing, so far, revenue and actual profitability have not paid off from these valuations. In the words of Annie Lowrey, who writes about politics and the economy for the Atlantic and is about to join us, market watchers are starting to get worried. They see financiers bulldozing giant piles of money to private AI startups with no realistic path to profitability. Tech companies reliant on other tech companies for revenue growth and non-tech businesses without a lot to show for their AI investments.
And he argues that this is a different kind of bubble. So what if it pops or when it pops depending on your outlook? Annie Lowrey is a staff writer at The Atlantic, author of Give People Money, and has a new book coming out called The Time Tax, How the Government Wastes Our Time and How to Fix It. She joins me now to talk about her latest Atlantic piece, The AI Bubble is No Ordinary Bubble. Annie, always great to have you. Welcome back to WNYC.

**Annie Lowrey** (1:51)
Thank you for having me.

**Brian Lehrer** (1:52)
And, listeners, you're invited in on this one. Do you work in tech or on Wall Street or in an industry that's been touched by the AI buildout in some way that's relevant to this discussion? 212-433-WNYC, 212-433-9692.
Has your job changed because of AI, for better or worse? Are you watching your retirement account and wondering how exposed you are to all this, maybe trying to manage that? 212-433-WNYC, 433-9692, call or text. What do you mean by no ordinary bubble, Annie?

**Annie Lowrey** (2:33)
It's an unusual bubble for a couple of reasons. But first of all, I think it makes sense to think of it as two bubbles. So there's a bubble maybe having to do with the capital expenditure, which is the cost of the AI buildout.
Unlike a lot of tech products, AI uses a lot of real resources, especially chips. So in the United States, we're expecting roughly $2.7 trillion to be spent on data centers and AI infrastructure through 2030, which is obviously just a couple of years from now. There's also an equity bubble, perhaps, a valuations bubble, which is something we're maybe more familiar with.
What's unusual about this, I think, is that your average person has nothing to do with this. They're not investing in these companies. This is not driven by active trade trading. This is not something like Bitcoin, where your cousin is getting in on it, or the housing bubble, where a huge number of Americans were involved in taking out mortgages on speculative properties. The other unusual thing is that it's happening not when credit is super cheap. This is not a zero interest rate environment. In fact, relative to where it's been, credit is kind of expensive. So these big companies are financing that build out in part with their own real profitability. That's the real money that they're making, but in part with corporate debt and private deals made through non-banks, but sometimes called the shadow banking system. So those are really unusual facets.

**Brian Lehrer** (4:06)
How convinced are you that it is a bubble? You mentioned the mortgage bubble. You know, most people listening now can remember how devastating and long-term the effects were when that bubble burst in 2008

**Annie Lowrey** (4:21)
Yeah. And, you know, I think that the fact that most Americans are not involved, right? You and I could not right now decide to go invest in Anthropic and OpenAI, right? Those are not public companies. We'd have no way to do it, right? They are not taking our money.
So that could shield people from this in some ways. The other thing is that big tech companies right now are very, very profitable. They're very credit worthy. And that's real, right? Just Google, just Google by itself made $160 billion in profit last year, not in revenue, in profit. So I think that that adds this little bit of the buffer. But I think about that number, $2.7 trillion being spent on this. That would mean that, you know, these businesses would need a $2.7 trillion plus interest return just to break even. That's a lot of money. That's a lot of money. And it would need to happen kind of quickly. We don't have a lot of visibility into the structure of these deals, right? When are things going to need to be paid back? Some of it's very long range, but some of it might be shorter range. Some of it might be very leveraged in ways that we don't understand. That's the stuff that I really worry about.