The AI Bubble Is Real. The Question Is How It Ends.
Every generation gets its bubble.
Every generation gets its bubble. The 1990s had dot-com. The 2000s had housing. Both ended in recession, one manageable, one catastrophic. Now, with AI stock prices defying gravity and Silicon Valley awash in capital, a growing number of economists are asking whether we're living through the next one and, more urgently, whether anyone in power is paying attention.
The numbers are hard to ignore. Since early 2023, Nvidia's stock has multiplied more than tenfold, driven almost entirely by insatiable demand for the chips that power AI systems. Microsoft and Google's parent company, Alphabet, have more than doubled and tripled in value, respectively. The broader S&P 500, by comparison, grew by less than twice its value over the same period, and that index already includes these AI giants. Strip them out, and the gap between AI-related stocks and the rest of the economy becomes even more striking.
This is not normal growth. This is a bet on the future, placed at enormous scale, by investors who are convinced or at least hopeful that AI will be the most transformative technology since the internet. Whether that conviction is justified is a separate question. What matters economically is that when the price of an asset races ahead of its underlying value, you have a bubble. And this one has been inflating for years.
Why Bubbles Form in the First Place
A stock's price has two layers: what it's actually worth based on future earnings, and the inflated premium driven by belief, excitement, or fear of missing out. When that premium grows large enough and detaches sufficiently from reality, you have a bubble.
Bubbles are also not random. They tend to emerge when there is more money looking for a home than there are good places to put it. Low interest rates over the past decade made safe investments, such as government bonds and savings accounts, essentially pointless. That money had to go somewhere. AI, arriving at exactly the right moment with exactly the right story, became the destination.
Not every investor is a true believer, either. Some genuinely think AI will reshape civilisation and that the companies building it will generate almost unlimited profits. Others are more sceptical, viewing the technology as sophisticated software rather than a genuine leap forward. And then there's a third group, perhaps the most telling, who know the bubble exists but keep investing anyway, betting they'll be smart enough to get out before it pops.
It was this exact logic that led Citigroup's CEO to say, just before the 2008 crash: as long as the music is playing, you have to get up and dance. He stayed on the dance floor too long. Most people who think they'll time the exit perfectly do.
Three Ways This Could End
The good outcome is that the bubble, whatever its excesses, ends up funding something genuinely transformative. The dot-com era burned through enormous amounts of capital on bad ideas. There is no kinder way to describe a company that tried to sell pet food online and went bankrupt within a year of its IPO. But the same era also bankrolled Google, Amazon, and the infrastructure that made the modern internet possible.
AI investment, even if overheated, could follow a similar path. The technology is real. The applications are expanding. If the flood of capital produces breakthroughs that genuinely lift productivity, improve healthcare, accelerate scientific research, and create new industries, then the bubble will have served a purpose even as it deflates. This scenario requires governments to move quickly on regulation, and it assumes AI displaces jobs slowly enough that economies can absorb the disruption. Neither assumption is guaranteed, but neither is impossible.
The bad outcome is a market correction sharp enough to tip into recession. Bubbles always burst; eventually, the only question is what triggers the sell-off. When enough investors conclude that multiple AI companies are simultaneously overvalued, the rush to exit becomes self-reinforcing. Prices fall, confidence drops, spending contracts, and the broader economy slows.
The closest parallel here is 2001, when the dot-com bubble burst, and the United States entered a recession that lasted less than a year. It was painful, but recoverable. An AI correction could follow a similar arc except for one important difference. Far more ordinary households now have money tied up in the stock market, either directly or through pension funds and mutual funds, than they did twenty years ago. A sharp correction in AI stocks would not stay in Silicon Valley. It would show up in retirement accounts and household savings across the country.
The ugly outcome is the one nobody wants to say out loud, but economists are starting to. It draws uncomfortable comparisons to 2008, not because AI stocks are mortgages, but because of what is happening beneath the surface of the balance sheets.
Major tech companies are currently financing the construction of vast new data centres almost entirely through debt. Some of that infrastructure is being built to serve demand that does not yet exist, on the assumption that demand will materialise before the bills come due. If it doesn't, if AI adoption plateaus, if the revenue projections prove wildly optimistic, if enough large companies find themselves unable to service their debt, nobody knows exactly how the financial system responds.
The 2008 crisis was not caused by falling house prices alone. It was caused by the way those mortgages had been sliced, packaged, and sold throughout the global financial system, so that when one part failed, everything connected to it failed too. AI financing is not structured the same way. But the honest answer is that we don't yet fully understand the exposure and that uncertainty, in a crisis, is often as dangerous as the problem itself.
The bubble will not stay inflated forever. Asset prices cannot permanently outrun the earnings that are supposed to justify them. At some point, triggered by a disappointing earnings report, a regulatory shock, a broader economic slowdown, or simply the moment when enough sophisticated investors decide to take their profits, the correction will come.
Central banks could, in theory, raise interest rates to deflate the bubble gradually before it grows any larger. The more controlled the deflation, the less damage it causes. But central banks are already navigating inflation, employment, and geopolitical uncertainty. Adding asset price management to that list is easier said than done.
The only real variables left are the timing, the trigger, and how much of the fallout lands on people who never bought a single share of Nvidia in their lives and who will have no idea why their retirement savings suddenly look different.
About the Creator
Mark Lim
Hi I am mark an automotive student and a car, tech and food enthusiast ! Im gonna try and post daily & hope you enjoy what I write and do share my page with people you know. I would gladly appreciate it! Cheers
Enjoyed the story? Support the Creator.
Subscribe for free to receive all their stories in your feed.
Comments
There are no comments for this story
Be the first to respond and start the conversation.