The AI bubble, explained

CFOs are under pressure from their bosses to budget for AI and weave it into every workflow, or at the very least, to show some sort of justification for the added spend. But as AI seems to be injected into every aspect of life—from the mundane of recipe hunting to the more complex identification of business opportunities—chatter of an “AI bubble” is happening all over.

As IT Brew recently reported, Alphabet’s CEO Sundar Pichai told investors during the company’s latest earnings call that AI helped drive a $100 billion quarter: “Five years ago, our quarterly revenue was at $50 billion. Our revenue number has doubled since then, and we are firmly in the generative AI era.”

But this week, the Wall Street Journal reported that investors, “already nervous about the sky-high valuations of AI businesses, have taken note of the weakness in the bond market.” The analysis came on the heels of reporting that raised questions about Meta’s use of “aggressive” structured financing, as it uses debt to fund a gigantic $27 billion data center for AI in Louisiana.

We asked Aswath Damodaran, a professor of finance at the Stern School of Business at New York University, to explain what’s causing the bubble, how it’s affecting retail investment, how to know when it might burst, and what consequences CFOs need to be aware of if it does.

This interview has been edited and condensed for clarity.

What’s driving the AI bubble?

Two things—one is greed and the other is FOMO. If you’re a portfolio manager, your clients are asking: “Where is your AI investment?” and you’re looking for something to show...I have a feeling that a lot of the pricing is being driven by investors wanting something with AI in its name in their portfolios. There’s no serious thinking here of what the business is going to look like, what the revenues and profits are. They’re just buying a name because their clients want something that they can point to and say, “that’s my AI investment.” So the demand is there because there’s a lot of money chasing after these relatively few AI names.

When will we know if the AI bubble is going to burst?

The more institutional investors get excited about something, the more likely the bubble is going to burst. We talk about retail investors being sheep. Institutional investors are the biggest sheep of them all because they have to live on mood and momentum. They’re so caught up in having to convince their clients that they have to have an AI name.

If the momentum shifts and goes the other direction, the same forces that cause AI names to get pushed up will get pushed down. They’re all running for the exit at the same time. The more excited institution investors get, with little substance to back themselves up, the more I worry about this being getting pretty close to a tipping point.

Is there a bubble brewing on the AI infrastructure side?

The infrastructure has some substance…The architecture bubble is going to be that there’s revenues, earnings, cash flows from architecture, but you’re building in a presumption that these revenues can grow 30%, 40%, 50%. Who’s going to be spending hundreds of billions of dollars in architecture if there’s nothing coming out from that architecture? I think that there’s an architecture bubble, but that bubble is easier to explain away because you have revenues and earnings.

I don’t even know what [OpenAI’s] revenue model is going to be. It can’t be just a subscription that they charge you and me. There’s not enough of us to justify [it]. It’s got to be that they eventually end up becoming licensed in products and services. They collect a licensing fee on tens of millions of products. I don’t think OpenAI itself is aware of what that endgame is. So I think we might be getting a little ahead of ourselves on those companies in terms of the value and price we attach.

What’s going on with the mergers and acquisitions market when it comes to AI?

Good sense has never driven the M&A market. It’s never been about value. I do a session called “Acquires Anonymous: Seven Steps to Sobriety” because it’s an addiction to companies. Acquiring is an easy, quick way to get in. There will be big companies that feel they’re being left behind…Acquisitions have never been about getting something at a decent price. It becomes almost defensive, which is you’re doing it because if you don’t do it, your competitor will, and you feel left out.

What happens when the companies investing in AI have such deep pockets?

They have cash until they get punished…They have to be extremely sensitive to how much shareholders trust them with the cash, but for the moment they’ve been given slack. But it’s like a rope that you’ve been given extra room. I mean, you can hang yourself as a company.

What will it mean for CFOs who have been implementing AI tools when the bubble does burst?

Your technology might disappear. ChatGPT is going nowhere; even if OpenAI crashes and burns, somebody will buy ChatGPT. Your basic products and services will survive. The question is: Who’s owning them?...I’ll be careful about not replacing people with cheap technology right now, because that cheap technology might quadruple in price. Two years from now, the people are gone. You can’t get them back.