AI bubble now 17 times bigger than dot-com boom and four times larger than subprime crisis

Despite mounting warnings about the massive sums inflating the so-called “AI bubble,” the boom shows no signs of bursting — in fact, it keeps expanding. The scale of the surge has now eclipsed all previous market bubbles, from the dot-com era to the 2008 financial crisis.

Julien Garran, a research analyst at MacroStrategy Partnership, estimates that the current AI bubble is roughly 17 times larger than the dot-com bubble of the early 2000s and more than four times the size of the subprime mortgage bubble that triggered the global financial meltdown in 2008. Unlike the internet frenzy two decades ago, which had limited impact on GDP, Garran warns that today’s AI-driven speculation is deeply embedded in overall economic growth — making the stakes much higher.

Before the 2008 crash, investors helped inflate a fragile property market built on risky, overleveraged mortgages. Garran argues that AI has become a similar kind of bubble — an industry propped up by hype rather than proven value. “You can’t create an app with meaningful commercial value,” he told MarketWatch. “It’s either generic, like video games that don’t sell, regurgitated public-domain content, or caught up in copyright issues.”

Meanwhile, the costs of developing AI systems are soaring as their performance gains diminish. Garran noted that some startups are already facing backlash, citing one New York company whose subway ads have been defaced with graffiti. “To know whether we’ve hit a wall, watch the LLM developers,” he said. “If they release a model that costs ten times more, uses twenty times more compute, and isn’t much better, that’s when we’ll know.”

Even without a full collapse, Garran cautions that economic growth is already slowing, and the tech sector’s momentum could reverse as quickly as it did in the early 2000s. Given the size of the AI bubble, he suggests the best time for it to burst “was yesterday — and the second best time is now.”