Thanks to artificial intelligence, the American stock market is booming. Tech giants are borrowing billions to attract AI talent, buy chips and hardware and build data centers. And market watchers are starting to worry. They see financiers dumping huge piles of money on private AI startups with no realistic path to profitability, tech companies that rely on other tech companies for revenue growth, and non-tech companies that don’t have much to show for their AI investments. The value of AI-related companies has increased by $27 trillion over the past three years – a staggering amount that is now equivalent to 36 percent of the value of the entire U.S. stock market. Although future income could While these valuations are justified, as Dominic Wilson and Vickie Chang of Goldman Sachs argued in a note to clients, earnings expectations require Panglossian optimism.
No less than Sam Altman argues that we are in an AI bubble. The International Monetary Fund calls it a significant risk to financial stability and warns of what could happen if it bursts: lower investment, tighter credit, lower consumption, disrupted trade flows.
That’s pretty much what happens when a bubble bursts, as a Dutch tulip obsessive in 1637 or a Bitcoin evangelist in 2011, 2013, 2014, 2018, or 2022 might have told you. But the AI bubble is no ordinary bubble. It’s fueled by hyper-rich corporations, not investors sitting at the kitchen table. They blow it when loans are pretty expensive and not dirt cheap. This could make the bubble less fragile and more durable than in the past. But it doesn’t make it any less painful if it bursts.
The dot-com bubble of the late 1990s and the real estate bubble of the late 1990s were remarkably broad-based compared to today’s AI bubble. Uncle Ted got a huge desktop computer, opened a new type of e-trade account, and started day trading stocks in Apple and Pets.com. The share of American households owning stocks increased by 13 percentage points from 1995 to 2001. During this time, more than 2,752 companies went public (far more than the 730 operating companies that went public in the last six years). Half a decade later, Aunt Linda bought a condo with no down payment, resold it, and mortgaged, sight unseen, three new investment properties in the Phoenix suburbs. The home ownership rate rose 5 percentage points as the housing bubble inflated; 40 percent of mortgages issued at its peak were for investment or vacation properties.
In both cases, ordinary people bet their savings on a seemingly successful bet: the Internet changes everything, real estate prices never go down. In both cases, cheap credit fueled irrational exuberance. Low interest rates allow venture capitalists to fund nonsensical web companies and banks to give junk loans to borrowers with bad credit. In both cases, rising interest rates caused the bubble to burst.
Today, Uncle Ted and Aunt Linda aren’t really involved in the AI frenzy. The share of Americans who own stocks has remained stable. Household debt has increased, but it has fallen relative to disposable income and GDP. Everyone seems to know someone who was personally hurt by the dot-com collapse and housing crisis. How many people today know someone who is all about OpenAI and Anthropic? (The companies aren’t public, after all.) How many people even know someone whose livelihood is directly affected by the AI frenzy?
The AI bubble’s insularity isn’t the only thing that makes it unusual, and we should probably think of it as two overlapping bubbles rather than just one. AI leads to huge capital expenditure – physical infrastructure, software development. And it leads to a huge increase in company valuations.
Digital innovations typically do not require a lot of labor or heavy equipment. The companies that offer them tend to be capital-light, like marketers and insurers, rather than capital-intensive, like airlines and hotel chains. But AI is different. A startup may need thousands of times as much computing power to train an algorithm as it would take to develop a traditional software product. To get that energy, Silicon Valley is building 1,500 data centers across the United States and counting, and is buying massive amounts of semiconductor chips. (The Wall Street Journal calls these chips “the most important market of the 21st century.”) Amazon, Microsoft, Alphabet and Meta alone are spending more than $700 billion on expansion this year. Investments in AI infrastructure are currently responsible for essentially all of America’s GDP growth. Without it, in other words, we could find ourselves in a recession.
The boom increases the value of farmland, drives up the cost of construction jobs and pours huge sums into water and electric utilities. Still, tech companies are the main beneficiaries: Nvidia sells chips to Meta; Amazon sells cloud computing capacity to OpenAI. This increase in sales and excitement about how much money can be made when AI starts increasing productivity and profits is driving valuations higher. The Magnificent Seven – Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla – now account for a third of the value of the Standard & Poor’s 500. OpenAI is worth more than Eli Lilly, JPMorgan Chase, Visa, Costco, Exxon Mobil, Wells Fargo, CVS Health, McDonald’s and Boeing.
To justify these valuations, technology companies need to generate huge revenues and profits. According to calculations by PitchBook’s Harrison Rolfes, OpenAI needs to generate around $100 billion in free cash flow by 2030. Analysts expect the company to lose $10 billion to $30 billion this year. If that’s the case – or if more communities ban data centers, or if non-tech companies become reluctant to buy AI software, or if China develops AI models that don’t require as much computing power – we could be in for a massive correction. The AI economy is a trillion-dollar environment of buying and selling, investments and equity holdings, all happening between San Francisco and San Jose. Big Tech provides money to AI startups to buy cloud services from Big Tech, which uses the revenue to run new AI models… you get my point. What happens to one company could happen to everyone.
Tech companies will also need to start generating even larger revenues and profits to pay off their debts. In the early days of AI, venture capitalists, wealthy individuals and large technology companies invested cash in the new technology. But the expansion proved so costly that Silicon Valley switched to corporate bonds and private loans, i.e. loans made by companies other than banks. Even though the AI boom occurred at a time when interest rates are quite high, there is still a lot of leverage associated with it. The deals are complicated, opaque and structured in such a way that they are invisible on traditional balance sheets, making the “ultimate risk allocation less transparent,” noted Stijn Van Nieuwerburgh of Columbia Business School. Lenders are already feeling sick — “buy-side indigestion,” as Morningstar describes the market’s recent reluctance to meet Silicon Valley demands.
When the bubble bursts, Uncle Ted and Aunt Linda will be affected, even if they were never the ones who fueled the madness. Your retirement plans and pensions are invested in Silicon Valley stocks that could fall; Your small businesses rely on access to credit, which may be limited. But hey, there’s always a chance that the AI will steal their job before all that happens.
https://www.theatlantic.com/ideas/2026/07/ai-economy-stock-market/688004/
