So Jim Chanos—you know, the guy who basically called “BS” on Enron before it became a dumpster fire—just shared some spicy takes on AI investing that might make you rethink your portfolio.
Here’s the deal: While everyone’s been losing their minds over data centers (because AI needs somewhere to live, right?), Chanos is basically saying “hold up, you’re looking at this all wrong.”
His hot take? Stop throwing money at the companies building AI data centers and start investing in the ones actually building the AI models. Why? Because as he puts it so elegantly: “The former are technology companies, the latter are REITs.”
Translation: Data center companies might look all techy and futuristic, but economically speaking, they’re basically fancy real estate plays. And we all know how that story usually ends.
This whole conversation started when Nvidia dropped $2 billion on CoreWeave, an AI infrastructure company that’s been the poster child for the data center boom. But Chanos isn’t buying the hype. He’s out here asking the uncomfortable questions like, “Does anyone still bother to check these wildly bullish claims with, you know, their actual financial statements?”
Ouch. But also… fair point?
Here’s where it gets interesting: CoreWeave reported better-than-expected revenue last quarter, which sounds great until you realize they also scaled back their guidance and revealed some “temporary delays” with a data center partner. Classic case of “the numbers look good until you read the fine print.”
Chanos did some quick math and pointed out that even with generous 10-year depreciation on their GPUs, CoreWeave would still be reporting losses. That’s like buying a Ferrari and realizing you can’t afford the insurance.
But this isn’t just about CoreWeave. Chanos has been side-eyeing the entire data center boom for a while now. His concern? All this massive spending on AI infrastructure might hit a wall when companies start asking the dreaded question: “What’s our actual return on investment here?”
Think about it: Everyone’s been throwing money at AI like it’s 1999 and the internet just got invented. But what happens when the CFOs start demanding to see some actual profits from all those shiny new data centers?
Chanos’s advice is refreshingly simple: If you want to play the AI game, bet on the companies creating the actual AI magic—the models, the algorithms, the stuff that makes ChatGPT go “brrr.” These are the real tech companies with actual intellectual property and competitive moats.
The data center companies? They’re essentially landlords with really expensive tenants. And we all know landlords don’t get rich from the property—they get rich from the rent. But what happens when the tenants can’t pay?
So next time someone tries to sell you on the “AI infrastructure revolution,” maybe ask yourself: Am I investing in the future of technology, or am I just buying really expensive real estate with a tech sticker on it?
Jim Chanos has been right about these things before. Just saying.