Is AI a bubble that’s about to pop?

0:00This is The Guardian.

0:10Today, from tulips to microchips, is the AI bubble about to burst?

0:25Right, I'm going to take you back to 1637. When a bunch of, let's call them floral enthusiasts, gathered in Harlem a day's walk west of Amsterdam with their eyes on a very expensive prize. Tulip bulbs. Across the country, some of these bulbs were selling for well over£700 ,000 in today's money. So when the auctioneer stood up and began the bidding process, he figured he was on to a pretty big payday.

1:00But he was met with silence. He lowers his prices. Once. Twice. Still no buyers. He lowers them again. And again. And again. Within days, tulip bowl prices plummeted. The bubble had burst. The get-rich-quick greed that eclipsed common sense had turned to panic and collapsed the market. But did we learn the lesson?

1:37In the 1840s, an explosion of interest in railways drove UK stocks to an all-time high.

1:47Then that bubble burst.

1:532000. Hundreds of companies were fighting over dot-com domains, convinced the internet would make them a fortune. Nearly every day, another dot-com fails. Online grocer Webvan will file for chapter... The bubble burst.

2:092008. An abundance of cheap debt had people borrowing at extreme levels. There could be a moment where the market says, you know what, we don't like it this way. And when those moments happen, if enough of that happens at one time, things start to break. And yes, the bubble burst.

2:30And now, trillions of dollars is being pumped into the AI revolution. It's an extraordinary moment, even by Silicon Valley standards. Clearly a lot of excess investment. No company is going to be immune. And it's fair to ask. Are we going to have an AI bubble? The obvious question is around the whole of this country and the whole of the world right now is, is it a bubble? Lots of activity happening in the AI space. Can't stop, won't stop. A lot of bubble allegations too. From The Guardian, I'm Noshin Iqbal. Today in Focus, bubble or boom? Is AI at risk of collapsing the economy?

3:13Blake Montgomery, you're the tech editor for The Guardian US. Welcome back to Today in Focus. Now, we've all been witness to the huge rise of AI in the last three years since OpenAI launched ChatGPT. Can you give us a sense of just how important AI has become to the US economy? Enormously. It's a huge concentration of money in one singular industry. So the magnificent seven, which are the biggest tech companies in the U.S., some of the biggest publicly traded companies in the entire world, Alphabet, Amazon, Apple, Meta, Microsoft, NVIDIA, and Tesla. Those seven companies make up one third of the S &P 500, which is an index of the 500 biggest stocks in the United States stock market.

4:02All of those companies, to a greater and lesser degree, are very involved in the AI boom. And for seven of them to have a third of that entire value makes for this enormously top-heavy financial landscape. And if any one of those companies falters in any significant way, it sends shockwaves through the entire US economy. I guess all this doesn't sound wholly unfamiliar territory when it comes to boosterism around big tech. But, you know, there is this fear that we may be on the precipice of this big AI bubble and that the whole thing is about to just pop and crash. Where did that nervousness first come from?

4:41There are some evaluations that say like the entirety of the growth of the U.S. economy for recent quarters can be attributed to spending on AI infrastructure, namely like the construction of these enormous and expensive data centers. Normally, the economy is driven by a variation of factors, including consumer spending being the most prominent. But spending on AI infrastructure and data centers and chips and power acquisition and land and water rights and construction, spending on all of that is so huge that it outpaced consumer spending by some estimates in recent financial quarters, which is a shocking figure.

5:25It's just so much money and speaks to the concentration of wealth, not only in like the Magnificent Seven at the top of the stock market, but also in like the Silicon Valley giants. They are, it feels like in the U.S. economy, they are the only ones who are doing like super, super well. This image and this one and this one and these, they all have something in common. They were all made by machines, generative artificial intelligence software to be exact. So billions of people are encountering generative AI every day. 800 billion people are using ChatGPT. Even more are using generative AI in forms they may not recognize.

6:04So Google has an AI overviews feature in Google Search. That's the most visited website on Earth. But will generative AI make enough money to justify the amount of investment that has been just poured into it like a waterfall? That is where the fears of the bubble come from. and are not unjustified, I don't think. I think it's right to say we are nervous that the US economy seems to be in the hands of seven companies, basically. Well, there has been this investment frenzy in recent years, and lots of excitable chat about the boon to the economy that AI will eventually provide. But do companies like OpenAI actually make any money?

6:47OpenAI is operating at pretty big losses and burning just enormous amounts of money. And the amount of money that OpenAI would have to make to generate the returns on investment that are expected from this dump truck of cash that has been pulled up to its store is enormous. The amount of money is enormous. OpenAI declaring code red as the AI race intensifies. Projections are that OpenAI can't even turn a profit until probably 2030. They're talking about spending more than a trillion dollars on infrastructure. So how they're going to get there is not really clear. Whether we burn$500 million a year or$5 billion or$50 billion a year, I don't care.

7:30I genuinely don't. As long as we can, I think, stay on a trajectory where eventually we create way more value for society than that. And as long as we can figure out a way to pay the bills, like we're making AGI. It's going to be expensive. It's totally worth it. Side note, to a layperson, where has all that money come from? Who's betting that hard and that high? The primary backer of OpenAI is Microsoft. They have poured the most money in. Every venture capitalist firm in San Francisco and California has invested. And also where OpenAI gets the revenue that it makes. It's not operating at a profit, but it makes money based on deals with other companies and its enormous base of users.

8:14ChatGPT has 800 million users every month. And that's not nothing. That's enormous. That's a huge user base. It's also how big MySpace was. And MySpace went away. What? Who? Right. MySpace at its peak had 800 million users, just like ChatGPT. So there is precedent for something that large to disappear. I guess the thing that worries people the most is the circular nature of it, and that AI companies and chip manufacturers, namely NVIDIA, they're all investing in each other. Can you explain why that's a significant problem? What happens is that NVIDIA, which makes the kind of chips that are seen as the cutting edge and the best for the creation of artificial intelligence products by virtue of their processing power and the software used to integrate them, everyone's their chips.

9:04There's a backlog on orders of years and years. They just can't make them fast enough. So everyone wants NVIDIA's chips. NVIDIA is worth a ton of money. NVIDIA is investing in the companies that buy from it. So creating this flow of cash that is, like you said, a circle. So it's promised to invest$100 billion in OpenAI, for example. And OpenAI, big customer of NVIDIA's. We are following the U.S. push to lead the AI race on planet Earth. NVIDIA announces it will invest up to$100 billion into OpenAI. They'll supply those high-performance chips and fund some new data centers. This announcement comes shortly after NVIDIA made a$5 billion investment in Intel.

9:45So open AI and AMD announcing a five year AI data center deal. Open AI has signed a$300 billion deal with Oracle for cloud computing power. Meta is paying nearly$15 billion for a scale AI stake. What the problem might be is that if one of these companies falters and if NVIDIA say like can't give that$100 billion to open AI, open AI may not have enough money to buy from NVIDIA at that point. It's a house of cards, right? Right. That circle starts to crack and the cars on top of that circular road, which is the US economy, might fall off. Yikes.

10:35Let's look at the evidence, I guess. I mean, what's been happening in the markets in recent weeks? What story are they telling? So in recent months, there has been like market nervousness around this concentration, these circular deals, which has led to this kind of skittish tech sell off, which is not so much borne out in the financial results of all of these companies. actually. Their earnings reports, by and large, were quite good. But there is still this Wall Street nervousness and in a stock market that is so rife with speculation, which is like what we're talking about, the speculative nature of like, AI is bringing about this change.

11:11We want to get in on it really early. We're pouring a ton of money in. It's not investing in a real thing that's happening just yet. Everyone thinks it's right around the corner. So there has been this nervous tech sell-off, which has affected other sectors of tech, actually. Crypto is falling off a cliff, an economy that is based heavily on speculation. And now it's really flopping to the great chagrin of the Trump family, who has lost, they've lost at this point, like$1 billion in paper value of their crypto ventures. So sad, I know. The Trump-branded meme coin has lost roughly a quarter of its value since August.

11:53And shares of Trump's social media company, which began hoarding Bitcoin this year, are hovering near record lows. This sell-off is part of a larger wipeout. One analyst that we talked to about NVIDIA earnings said the spending is real. The amount of money being spent on AI infrastructure is real. Like, whether OpenAI will become the world's biggest company that is part of every layer of government and our personal lives and business, that's a different question. But the real winners so far have been the people who sell shovels to the gold miners, like NVIDIA and all the construction firms that are building all this stuff.

12:34But there have been chunks taken out of the market value of these giant companies because of a growing chorus of investors who say this is frothy and will come to nothing. Michael Burry, who you probably know from The Big Short, is now calling AI a bubble. A recent filing shows that at the end of September, he was betting against two of the biggest AI winners, NVIDIA and Palantir. How are tech leaders handling these AI bubble fears? I mean, Jensen Wang of NVIDIA addressed it directly in their most recent earnings call. So he said, quote, there's been a lot of talk of an AI bubble. From our vantage point, we see something very different.

13:17and NVIDIA is unlike any other accelerator. We excel at every phase of AI. I don't believe we're in an AI bubble, and the reason for that is we're going through a natural transition from an old computing model based on general-purpose computing to accelerated computing. One thing that I thought was pretty interesting was that the Guardian's headline about NVIDIA's most recent earnings used the phrase, like, fears of an AI bubble, and talked about this market sell-off that we've been talking about. Our headline about NVIDIA last quarter said amid fears of an AI bubble. Like the fears are persistent, but it hasn't happened in the past six months and the company has continued to deliver stellar results.

13:56The chip maker, NVIDIA, has become the first ever company to reach a market value of$5 trillion, reflecting investors' confidence in the future of artificial intelligence. Blake, how do you think an AI bubble bursting would compare to the dot-com bubble of 2000 or other asset bubbles before it? Tulips, subprime mortgages, take your pick. It would be pretty severe. The worth of all of these companies and the fact that they command a third of the Standard & Poor 500 means that any kind of correction, downturn would be quite financially painful in the United States. And I think across the globe, even though they are mostly American companies.

14:40That's where these fears of the AI bubble are coming from. It's less about the technology than the financial concentration. So the former IMF chief, Gita Gopinath, estimated that a correction or a crash could wipe out $20 trillion from the U.S. economy. And serious financial institutions are sounding the alarm and warning. The Bank of England said this could wipe out tens of billions of dollars from the U.K. economy. And financial regulators are sounding the alarm, trying to take protective measures or let people know this is precipitous and has a lot of the hallmarks of, like you said, the subprime mortgage prices in 2008, the dot-com bubble.

15:25I don't know if anyone's likening AI to tulips, but it looms large as the original crash. It is close to the gold rush of California. Like it has created in a very parallel way a migration to California, to the American West, and this build out in formerly desolate areas like the Nevada desert, no water, enormous complexes of data centers. It's having effects on the real world in ways that are quite surprising.

16:05Blake, one could argue that this is a normal part of the process, that whenever a major new technology is rolled out, be it railroads or the internet, that there are bumps along the way, but ultimately things level out. Is this any different? I do think it's different in its scale. The amount of investment is just an order of magnitude larger than a lot of the technological bubbles we've seen before. Or Pets.com did not command a third of the S &P 500, for example. Dot-com bubble had a really big effect on the U.S. economy. But if AI were to crash completely based on some sort of house of cards math, it might be more akin to the subprime mortgage crisis in its magnitude.

16:54So that is what is frightening about it, is the scale. And if all of this investment from these companies in this infrastructure stopped, it would mean that the U.S. economy and increasingly the global economy would see quite a lot less growth. And what about the impact, do you think, on ordinary people who don't work in tech or invest in it? So someone who's not invested in a total stock market index may not feel the immediate effects of a crash. they would, I think, because this would have this sort of like meteor hits the dinosaurs effect of chilling the entire global economy. I think people would feel it like a stock market downturn causes companies to have less like cash flow.

17:46And so invest less and hire fewer people. So I think even if someone believes themselves to be isolated from the boom and not involving in it, I mean, people choose not to use generative AI for a whole host of reasons. A personal isolation from it may not insulate them from a potential financial crash.

18:11Coming up, can we just turn it all off and on again?

18:24understandably the last few years of being pretty turbulent there's been a lot of excitement about AI there's been a lot of fear you know this small matter of that it might you know eventually end humanity and all that do you think in that light the bubble popping could be a good thing I don't want to wish like financial harm on the globe. I think like there are great ethical and environmental fears over AI. Like, does it pose a grave risk to the livelihood of artists who made their living creating the kind of stuff that AI can create with a stamp of a finger? Yes, it does. does it use enormous amounts of resources like electricity and water also yes it may threaten people's like very jobs like white collar jobs that we kind of thought were insulated from technological threats because of human knowledge and aptitude and for those reasons like there are quite a lot of people who say we need to get rid of this not use it so i think to that contingent And those people would say like, yes, get rid of it, burn it down.

19:33I don't think that's very likely. I don't think anyone is going to flip the switch off and say no. I think the technology is probably here to stay in some form. And also when there's so much money to be made, no one is going to say, yeah, I think we're okay. Blake, when it comes to why we're here in the first place and why AI is scaling and growing at such speed, are there any other reasons at play beyond the sheer opportunity to do so? There is a belief in kind of the government sector that AI can cut bureaucracy, make democracy more participatory and streamline government operations and be extremely useful in national security contexts.

20:21Palantir, which is part of this AI cohort of companies that we're talking about, is deeply enmeshed in U.S. immigration enforcement, for example. And there is this enormous fear at the kind of macro government level that if one country gets very, very far ahead in AI, they will command a huge lead and wield lots of power over everyone else. So there is this giant race at this point between the United States and China to be the leader in artificial intelligence and with a likening to the race for nuclear power. Blake, finally, do you think that this is a bubble that is about to pop? I don't think so.

21:10I think the money is real. I think there may be market corrections and bubble doomsayers will say this the bubble has popped and bubble yay sayers will say no it's just minor in comparison to ai's major gains uh i don't think it's out of the realm of outcomes that there is some sort of enormous financial crash related to ai but so far the steam and the momentum behind it is so strong that at least as a journalist, we have to treat it as it is now, which is this enormous force financially and technologically. It is like the great majority of the topic of the stories that we work on, like on the Guardian tech desk.

22:00And so it just has all the momentum in the world, but you know, so did the housing sector. So did myspace so did the dot coms so i don't think so i don't think it's a bubble yet i'm also not a wall street analyst so who's to say smarter financial people than me are doom saying but smarter people are also saying we're all going to be rich forever have you invested i wish i wouldn't be doing this are you kidding where are you putting your money um well i promise you. If it does burst, you will be the first person I'm asking back on. Blake, thank you so much for your time. Thank you for having me.

22:51That was Blake Montgomery, tech editor for The Guardian US. My thanks to him. You can read lots more about AI, the boom, the bubble, all at theguardian.com. But that's it for today This episode was presented by me, Noshin Iqbal It was produced by Eli Block and George McDonagh Sound design is by Joel Cox The executive producer was Huma Khalili We'll be back again this afternoon with the latest which you can find wherever you watch or listen to your podcasts

23:30This is The Guardian again.