You've probably noticed that many people are saying we're in an AI bubble. "OpenAI's Sam Altman sees AI bubble forming," CNBC reports. "Fears of an AI bubble are growing," NBC News states. "AI Is the Bubble to Burst Them All," Wired announces. "Mark Zuckerberg says a ‘collapse’ is ‘definitely a possibility,’" Fortune declares.
You've probably also noticed that almost everyone who uses the term “AI bubble” starts to sound very boring and confusing immediately afterward. This is because the people who bring up the AI bubble tend to be market analysts, and most market analysts would rather have potatoes growing inside their bodies than talk like a normal person for four seconds.
So what is an AI bubble? Are we in one? How scared should you be? What happens if the artificial intelligence industry blows a trillion-dollar hole in the non-artificial economy? What happens if it doesn't? Let's walk through this thing together.
1. Are we in an AI bubble?
Yes.
2. How can you be so sure?
Because the gigantic numbers floating around the AI economy—deals worth trillions, data centers costing tens of billions, 75 percent of gains in the S&P 500 centered on AI stocks—are almost entirely driven by hype, and the hype has come unglued from reality. The hype around AI insists that it's a world-transforming technology that will revolutionize every aspect of human society. The reality, which we'll explore in more detail in a second, is that AI companies are burning through staggering amounts of money (and fossil fuels) with no clear path to profitability, that the companies themselves aren't super clear about what their products are for, and that many of those products have failed to perform in the applications they've been assigned to (AI search engines return inaccurate information, AI teachers impair learning, AI therapists make mental health worse). Worse yet for the industry, the biggest players are increasingly tied up in time-bomb financial deals that look disastrous for their futures in any but the rosiest of best-case scenarios.
But because the hype around AI is so pervasive—who doesn't want to own a chunk of the technology that will determine every aspect of humanity's future?—investors have continued plowing money into the field despite the warning signs, sending stock prices into orbit. As of last week, Nvidia, the leading AI chipmaker, is the first company in history whose stock is priced at over $5 trillion in total—meaning this one individual company accounts for roughly 8 percent of the S&P 500.
That's a bubble.
Plain English Tackles the AI Bubble
3. What is a bubble?
It's what happens when the price people are willing to pay for a thing exceeds the value of the thing by some drastically untenable amount. It's hard to give a more technical definition, because determining the correct value of anything is weirdly tricky beyond a certain point. Markets, for all that people pretend otherwise, are extremely susceptible to vibes.
Past bubbles have inflated around everything from tulips to Japanese real estate to the original wave of dot-com stocks in the early 2000s, but let's run with a simpler example. Say there's a new hype sneaker. It costs $18 to manufacture, it retails for $100, and it's going for $1,800 on StockX. What's its correct value? Everyone can kinda sense that $1,800 is too much to pay for it, but the price keeps going up, so people keep snapping up more pairs, often hoarding multiples in the hope of flipping them for a profit. The price hits $2,000. Then, suddenly, a guy gets nervous and sells 10 pairs for a little less than the current max. People notice this and start selling off their own pairs for whatever they can get before the price craters. By the end of the next day, the bubble's popped. The price is down to $150. People who held their pairs take a massive loss; people who flip sneakers full time have spent a fortune and can't sell their current pairs for enough to refresh their inventory.
Why did the price spike? Because people convinced themselves it made sense to spend that much. Why did the price collapse? Because the people who collectively propped up the delusion that the price made sense got spooked and decided to get out with what they could before everyone else decided it didn't.
4. What does the susceptibility of markets to vibes mean for the AI bubble in 2025?
It means that the AI industry's most important product at this moment is not a chatbot or a video generator; it's the story the AI industry is telling about itself. There's a saying stock market people pass around: "The market can remain irrational longer than you can remain solvent." The market is definitely irrational right now; the goal of the AI industry is to keep it irrational until the industry can figure out a way around the mess it's created for itself.
How bad is the mess? OpenAI's ChatGPT, by far the most successful generative AI product to date, requires so much expensive computing power to run that it loses money almost every time you use it. (For the first half of 2025, OpenAI took in $4.3 billion in income and still reported a net loss of $13.5 billion.) A recent Deutsche Bank report says that the AI boom is "unsustainable." Bain & Co.'s annual global technology report says the AI industry will need $2 trillion in annual revenue by 2030 to continue at its current rate. (That number will be, um, hard to come by.) A Bank of England report warns that the market could experience a "sharp correction" due to the overvaluation of AI companies.
Companies with no clear plan for making money from AI have poured cash into the construction of data centers so massive that they're straining the electrical grid and driving up the price of electricity for the rest of us. They've poured cash into what's called "compute"—processing power, basically, used to train AI models and run search queries—at a rate they can't keep up unless the entire world economy reorganizes itself around their products by, like, Tuesday. All that tech has a shelf life, just as your iPhone does; the industry will have to keep spending ruinous sums every few years to replace it. And to keep the ball in the air, the companies have made a number of increasingly bizarre, shady, and circular deals with each other: Nvidia, for instance, has invested $100 billion in OpenAI, which OpenAI must then turn around and spend on Nvidia products. If I give your lemonade stand $10 so you can buy my $10 lemons, we can't tell our investors (your mom) that we've boosted the lemonade economy by $20. But that's the kind of reasoning that fuels more and more of Silicon Valley's grandiose claims about itself.
The thing is, though, the price of a stock is not determined by the soundness of the company. It's determined by the willingness of investors to pay a certain price. And there are ways to convince investors to pay high prices for shares even when it means driving through an endless series of flashing red lights. If you can, say, leverage three decades of mythmaking about the transformative impact of the tech industry by convincing a compliant press to spread the word that your new product is poised to make every sci-fi dream a reality—cure cancer, bring humanity to other planets, create flying cars—then the sheer size of the vision you're peddling might convince investors to buy in, even if doing so seems reckless.
Tech investors, especially the superrich ones, are always in the market for stories scaled to their egos, and their egos are getting more messianic all the time. And in the era of crypto and Robinhood, regular idiots like us increasingly see the stock market as a casino. Pushing some chips onto the new, hot, famous, sexy thing seems way more exciting than researching the fundamentals of some well-run company no one's talking about. Plus, everyone else is doing it. OpenAI is currently targeting a $1 trillion IPO. The basis of that number is not—like, extremely not—the solid state of the business. It's the story the company is telling.
Talking a big game has thus become the essential and necessary activity of AI companies. (This includes acknowledging the possibility of a bubble as a form of panic mitigation: That way, if the bubble does pop, the wise, patient tech CEOs can say they expected it all along.) AI storytelling is an amalgam of several different narratives, including:
- Inevitability: AI is the future; its eventual supremacy is both imminent and certain, and therefore anyone who doesn't want to be left behind had better embrace the technology. See Jensen Huang, the CEO of Nvidia, insisting earlier this year that every job in the world will be impacted by AI "immediately."
- Functionality: AI performs miracles, and the AI products that have been released to the public wildly outperform the products they aim to replace. To believe this requires us to ignore the evidence obtained with our own eyes and ears, which tells us in many cases that the products barely work at all, but it's the premise of every TV ad you watch out of the corner of your eye during a sports telecast.
- Grandiosity: The world will never be the same; AI will change everything. This is the biggest and most important story AI companies tell, and as with the other two narratives, big tech seems determined to repeat it so insistently that we come to believe it without looking for any evidence that it's true.
As far as I can make out, the scheme is essentially: Keep the ship floating for as long as possible, keep inhaling as much capital as possible, and maybe the tech will get somewhere that justifies the absurd valuations, or maybe we'll worm our way so far into the government that it'll have to bail us out, or maybe some other paradigm-altering development will fall from the sky. And the way to keep the ship floating is to keep peddling the vision and to seem more confident that the dream is inevitable the less it appears to be coming true.
5. Are there any indications of how well this is working?
Yes. Last week, as Nvidia was crossing the $5 trillion threshold—making the chip manufacturer worth approximately as much as the gross domestic products of France and South Korea combined—several of the so-called Magnificent Seven tech companies released their earnings reports. It was a sensitive moment for big tech. All year, murmurs of a bubble had been growing louder. Deutsche Bank called summer 2025 "the summer AI turned ugly." A series of bad earnings reports could have triggered a major sell-off. The air over Northern California could have echoed with a hollow popping sound.
Instead, the earnings reports were … OK? Alphabet, Google's parent company, notched its first $100 billion quarter. Microsoft beat expectations. Meta's stock slid after it announced a one-time tax charge, but not drastically. No bubbles were seriously threatened.
If you looked a little more closely, though, there were further signs of strain in the industry. Microsoft's stock slid 5 percent immediately after its earnings call, which analysts interpreted as a sign that the market is mildly freaked out about how much money the company is plowing into AI infrastructure. ("The vibe is cautious," TheTradable reported.) The companies were a little vague about how much revenue they're actually bringing in from AI, or how they expect to make money from it. Zuckerberg, defending the vast sums Meta is pouring into its data centers, unwittingly summed up the uncertain mood when he said that this kind of infrastructure spending "is very likely to be a profitable thing over … over some period."
6. What happens if the AI bubble pops? How bad would the damage be to the U.S. economy?
Pretty bad, man! On paper, at least, the ungodly sums AI companies are spending on their infrastructure are propping up the American economy almost single-handedly. In the first half of 2025, spending on data centers accounted for most of the economic growth in the U.S., outpacing all consumer spending combined. (You'll see this reported in ways that make it sound as if the economy is being bolstered by the money AI companies are making, but no, it's the money they're spending that's relevant.)
On the other hand, "on paper" may not be the best way to understand the state of the U.S. economy right now. It's been widely observed that many regular people feel the economy is bad, while economists, when they look at the numbers, say it's relatively strong. This disparity is probably in part because, on paper, Meta dumping a thousand trillion dollars into a Death Star–sized “superintelligence lab” with 11 employees in rural Idaho boosts the numbers, but in reality, Zuckerberg's real estate binges have a limited effect on conditions outside the semi-closed loop of the AI industry. (This may be especially true now that private equity funds are getting involved in data center financing.) Groceries are getting more expensive. Utilities are getting more expensive. Thanks to the Trump tariffs, almost everything we buy is getting more expensive. None of that is offset by AI investment, and if AI investment stopped tomorrow, the immediate impact on many people's daily lives might be less than you'd expect.
That also applies to, as you might guess, non-AI businesses, many of which are also struggling. Many have scrambled to incorporate AI into their operations, drawn by the promise of a technology that would cut labor costs and revolutionize productivity. Yet according to an MIT study released in August, 95 percent of businesses that have deployed generative AI have gotten no value from it. It's not clear that the AI bubble bursting would have a devastating effect on them either, at least directly.
Indirectly, of course, things look quite different. A sharp contraction in tech would put tens of thousands of people out of work, vaporize trillions of investment dollars, torpedo retirement and education funds, obliterate life savings, and ruin lives. It would be a violent shock to a system that's already under strain. Of course, one of the reasons the economy is under strain is the massive concentration of wealth in the hands of a small number of oligarchs, and those same oligarchs would like to use AI to replace the human labor force—likely putting tens of thousands of people out of work, vaporizing non-oligarchic wealth, obliterating savings, etc. It may be better for the world for the bubble to pop, whatever damage is done.
For now, though, we are living in the storytelling economy. As long as the shared fantasy of the AI industry holds Silicon Valley and its investors in thrall, the reckoning can be deferred. And in the meantime, who knows? Maybe the bubble will never pop. Maybe the people who've sold the fantasy can figure out how to keep us believing long enough for someone to create it, or at least to dream up a new one and hope we don't notice the difference.
It's bitterly funny, though, that up until now, some of the biggest victims of AI companies have been the artists, writers, actors, and filmmakers whose creative work the companies want to cannibalize and replace. The industry has shown nothing but contempt for narrative art that speaks to the human imagination—the very thing that's keeping it alive.
