The AI investment frenzy just got its most damning assessment yet. Using a historical framework that's accurately predicted every major tech bubble, researchers are sounding maximum alarm bells about artificial intelligence - giving it a perfect 8 out of 8 on their bubble scale. With Nvidia now worth $5 trillion and Big Tech pledging $400 billion more, the question isn't whether we're in a bubble anymore.
Nvidia just became the world's first $5 trillion company - that's 2.5 times Canada's entire economy. Meanwhile, 95% of businesses deploying AI say they're seeing zero returns. If this doesn't scream bubble, what does?
Brian Merchant, author of "Blood in the Machine" and WIRED contributor, decided to cut through the vibes-based bubble chatter with actual science. He turned to a 2019 academic framework by scholars Brent Goldfarb and David Kirsch that's successfully identified every major tech bubble in history - from the dot-com crash to the aviation boom of the 1920s.
The verdict? AI scores a perfect 8 out of 8 on their bubble scale. "We've got the maximum level of bubble alert here," Merchant tells WIRED's Uncanny Valley podcast.
The framework looks at four key factors, and AI checks every box with flying colors. First is uncertainty about business models - just like electricity in the 1800s, everyone knows AI is powerful, but nobody's figured out how to reliably make money from it. When OpenAI CEO Sam Altman was asked about their business plan years ago, he literally said they'd "build AGI and ask it" how to make money. Talk about uncertainty.
Second comes "pure play" companies - businesses that live or die entirely on the bubble technology. Nvidia has become the poster child here, transforming from a gaming chip company into an AI infrastructure juggernaut. "It's the classic case of selling shovels during the gold rush," Merchant explains. If AI tanks, so does Nvidia - and that company now represents 8% of the entire stock market.
The third factor is novice investor access, and boy do we have that. Anyone with a Robinhood app can pour money into Nvidia. But here's the kicker - according to Goldfarb, AI is so loaded with uncertainty that even sophisticated investors become novices. "Nobody knows what this future is going to be," Merchant notes.
The final ingredient is coordinated belief among investors about the technology's potential. AI's narrative is literally "it will do everything" - cure cancer, fight climate change, automate all jobs, babysit kids. "It's the story to end all stories," Merchant says. There's something for every investor to believe in.
But the circular money flows are getting weird. OpenAI cut deals with chipmaker AMD, meaning retail investors in AMD now have exposure to AI's fate. Data center real estate plays are buried in pension portfolios. "We have a growing number of things that expose regular retail investors to an AI bust," Merchant warns.
The scale dwarfs previous bubbles. Google, Meta, Microsoft and Amazon are already on track to spend $400 billion on AI this year - and they're promising even more for 2025. Yet that MIT study showing 95% of AI-investing companies see no returns suggests a massive gap between investment and utility.
Historically, bubbles don't destroy the underlying technology - the internet survived the dot-com crash, and we all work for websites now. But the economic carnage can be brutal. The dot-com crash looks quaint compared to what AI's scale could unleash.
"This could be a bubble to end all bubbles," Merchant argues. The infrastructure built during previous crashes - like fiber optic cables - stayed useful. But AI chips get upgraded constantly, making today's hardware potentially worthless in a decade.
There's one wild card: government intervention. The current administration took a 10% stake in Intel and clearly loves AI. If the bubble starts popping, we might see unprecedented federal bailouts of AI companies - something that's never happened during previous tech crashes.
Even Nvidia CEO Jensen Huang insists we're not in a bubble, calling it a "natural transition" to accelerated computing. But when your company represents 8% of the stock market based on selling shovels for digital gold that might not exist, natural feels like the wrong word.
The framework doesn't predict when bubbles burst, just that they will. With all four indicators maxed out and investment still accelerating, we're in uncharted territory. The question isn't whether AI will survive - it's whether the economy can handle what happens when the music stops.
The academic framework is clear - AI exhibits all four classic bubble indicators at maximum intensity. With trillion-dollar valuations built on uncertain returns and coordinated beliefs about unlimited potential, we're witnessing what could be the biggest tech bubble in history. The technology will likely survive whatever comes next, but the economic fallout from unwinding $400+ billion in annual investments could make the dot-com crash look like a minor correction. Whether government intervention changes this equation remains the biggest unknown in an already uncertain equation.
