Remember 1999? When everyone and their cousin’s dog was buying tech stocks and nobody asked pesky questions like “but what does it actually do?” Well, buckle up—BCA Research thinks we’re about to relive that fever dream, except this time it’s all about AI.
Here’s the setup: BCA Research is predicting what they’re calling a “melt-up”—basically a 30% stock rally that could send the S&P 500 past 9,200. Sounds great, right? The catch? It looks suspiciously like the dot-com boom, which, as you might recall, didn’t end with champagne and yacht parties for everyone.
The firm’s thesis is that the AI trade is hitting its late-stage growth phase. You know, that moment when things get *really* exciting right before they potentially get *really* messy. They’re watching four key signals, and honestly, some of them are pretty wild.
**First, everyone’s suddenly buying AI.** The Ramp AI Index—which tracks how many US businesses have paid subscriptions to AI tools—just hit 50% for the first time ever in March. That’s not gradual adoption; that’s a stampede. When half of American businesses are suddenly paying for something, you know the hype cycle is in full swing.
**Second, GPUs are expensive as hell.** Nvidia’s B200 cloud GPU was averaging $5.09 per hour in March, up 13% from the month before. These chips are the pickaxes in the AI gold rush, and everyone’s scrambling to buy them. When the infrastructure costs are skyrocketing, it usually means people are betting big.
**Third, tech giants are throwing absurd amounts of money at this.** Amazon, Google, Meta, and Microsoft are collectively spending an estimated $587 billion on capital expenditures this year. That’s not “we’re interested in AI”—that’s “we’re mortgaging the future on this.” The risk, as BCA notes, is that we end up with an AI-powered economy that doesn’t actually need trillions in data center investments. Oops.
**Fourth—and this is the spicy part—financial risk is rising.** Credit spreads for tech debt are widening, which is investor-speak for “we’re getting nervous.” BCA called this “the most worrying set of indicators,” which is basically them saying, “Yeah, this could get ugly.”
Here’s the thing: BCA isn’t saying AI is a dud. They’re saying we might be building way more infrastructure than we actually need. It’s like everyone showing up to a party with enough snacks to feed a small nation, then realizing only 20 people are coming.
The real question isn’t whether AI will transform the economy—it probably will. The question is whether we’re about to pay $587 billion to build something that could’ve been built for $200 billion. And whether the stock market’s going to party like it’s 1999 before figuring that out.
So yeah, a 30% rally is possible. But remember: in 1999, the party was *amazing*. The hangover, though? That lasted years.
