Here’s a fun paradox: when tech stocks get *cheaper*, that’s sometimes when you should get *more* nervous.
Tom Essaye, who runs Sevons Report Research, just dropped a note that’s basically the financial equivalent of “we need to talk.” His argument? The fact that AI stocks are trading at surprisingly low valuations right now isn’t a bargain—it’s a warning sign that investors are quietly freaking out about whether the whole data center boom is actually worth it.
Think about how stock valuations normally work. Growth stocks command premium prices because everyone’s betting on massive future earnings. So when those same stocks suddenly look cheap? That means the market is basically saying, “Yeah, we don’t think those earnings are actually coming.”
Essaye walked through four examples that tell the story. Nvidia’s sitting at a 21x forward PE ratio (remember, the S&P 500 is at 21.5). Broadcom’s at 24x. But then you’ve got Micron at 10x and SanDisk at 14x—despite SanDisk being up nearly 4,500% in the last year. That’s not a discount. That’s the market pricing in serious doubt.
Here’s the nightmare scenario Essaye laid out: imagine Google decides building 10 new data centers is too expensive and the ROI isn’t there. Suddenly, Nvidia, Micron, Broadcom, and SanDisk don’t get those massive chip orders. No chips needed means no sales. No sales means the whole thesis collapses.
And it’s not theoretical. Oracle stock has tanked about 25% since early June because investors watched the company throw billions at AI infrastructure and thought, “Uh, is this actually going to pay off?” That kind of investor hesitation is contagious.
Here’s where it gets spicy: Essaye drew a direct line to the dot-com bubble. Not because he’s saying we’re definitely heading for a crash, but because the *pattern* is eerily similar. Back in 2000, everyone was obsessed with internet connectivity. The problem? It wasn’t nearly as profitable as people assumed. So the buildout stopped. Companies that were supposed to be goldmines became graveyards.
The parallel is uncomfortable: what if AI adoption and effectiveness just… don’t live up to the hype? What if companies realize they’re spending billions on infrastructure for returns that don’t materialize? That’s when the music stops.
Now, Essaye isn’t exactly sounding the alarm that the market’s about to crater tomorrow. This fear has been floating around for months without materializing. But he’s basically saying: watch those valuations. When growth stocks get cheap, it’s not always a gift. Sometimes it’s the market whispering that it’s lost faith.
The AI boom has been the story of 2026. But stories can have plot twists. And right now, the market’s pricing in the possibility of a pretty big one.
