What happens when the AI bubble pops?

Now every tech company seems to have its own AI: Google Gemini. ChatGPT by OpenAI. MetaAI. Spending on AI is Record highs reachedPower supply a big boom for the stock market. Even those White House wants to join in the fun.

So are we in an AI bubble – an over-investment period that is bound to peter out? Yes, argues Paul Kedrosky, partner at SK Ventures and fellow at MIT's Initiative on the Digital Economy. But not everyone's bubble thinks We are in. “AI is obviously an extremely important technology,” Kedrosky said Explained today Co-host Noel King. So what then?

It is The money flows into the AI ​​infrastructure like data centers, which worries Kedrosky: “We are spending so much money on the underlying infrastructure for AI without likely being able to recoup the majority of those costs, and there is a significant likelihood that most of these assets will become worthless due to the rate at which they lose value.”

What happens if the bubble bursts? And can past bubbles tell us anything about the future?

Below is an excerpt from their conversation, edited for length and clarity. There's a lot more throughout the episode, so check it out Explained today Wherever you get podcasts, including Apple Podcasts, PandoraAnd Spotify.

How much money goes into these data centers?

It will be on the Order of trillions Now. Future data center spending is projected to exceed $2 trillion. But a growing portion of the money being spent on all of these things that allow us to proliferate AI, like electricity, comes from debt. And debt comes with obligations. You can't just walk away. That makes this moment even more dangerous.

If AI is so important, why doesn't it make sense for trillions of dollars to flow into it? Isn't that what we should do?

We should be. But the problem, of course, is that there is this idea of ​​a so-called rational bubble. Everyone thinks they're doing the right thing, but when you add up everyone's “right” thing, you end up with a tremendous amount of waste.

It's no different than going back to the railway bubbles of the 19th century in Britain and the US. There were simply too many tracks, too many enthusiastic railway builders building tracks almost next to each other in the same locations. And that resulted in an incredible amount of waste. But it also led to corporate bankruptcies and various market crises in the USA and repeatedly in Great Britain in the 19th century. It's not as simple as saying, “Well, this is important, so we should build it and not worry about what it costs and not worry about the consequences.”

If so many smart people think we're in a bubble, why is so much money still flowing into data centers and other AI infrastructure?

I'm not convinced that many people think it's a bubble. When I talk to technology people, the most common answer I get is that not only is it not a bubble, but that it is probably the most important technology of our lifetime. We have the opportunity to build a superintelligence, a godlike intelligence, on top of all these chips and buildings and this AI electricity thing that we're creating. And saying we should slow down at this point is just a big mistake, according to the tech community. But there are people outside of technology who say, “Oh, that's an incredible amount of spending.” The Bank of England it said. Other people warn about this, but not within technology.

The United States, and humanity at large, has had no shortage of bubbles throughout history. You mentioned the railways; Take us through some famous American bubbles.

The railroad is probably one of the most important in the United States and that, in turn, inspired enthusiasm for the idea. The same thing happened in the 1920s during electrification. In the 1920s, the percentage of rural areas that had access to electricity was in the single digits. [to] by the end of the decade it was more or less ubiquitous. Everyone had access to electricity. But at the same time, it led to a proliferation of utility companies, of companies doing all sorts of questionable things and overspending. One could argue that electrification and the excitement surrounding it sparked the stock market rise of the 1920s, which led to the Crash of 1929 and helped trigger the Great Depression.

The telecom and dot-com bubbles are pretty familiar to people, but the biggest historical analogue to what's happening now is railroads and electrification. Just like we don't need two routes to Philadelphia, we probably don't need the same number of companies delivering these so-called large language models, these AI models that people use. These will of course shrink.

How destructive are bubbles and what do they destroy?

They all cause immense damage. It's about how big the bubble is and where the damage goes.

So if you're just holding an index fund and you think you're very conservative, you're just diving into AI. If everything is reversed, so it goes 20 or 30 percent the other way, you will be much poorer than before. This will change your spending. And that has an impact on recessions.

Isn't it always the case that the bubble bursts and then something is left behind, maybe not something beautiful, but something doable?

This is some kind of chatter from the technology community. But the reality is that almost every financial and technological revolution has caused great damage and it can take decades to get back to where we were before. And as they say in economics: In the long run it may work, but in the long run we are all dead.