Remember when everyone was freaking out about “circularity” in AI deals last month? Well, buckle up buttercup, because Wall Street has found a new word to lose sleep over: depreciation.
Yeah, I know. It sounds about as exciting as watching paint dry on a rainy Tuesday. But this boring accounting term is currently giving AI stocks the kind of beating usually reserved for crypto during a bear market.
Here’s the deal: All those shiny, expensive GPUs and semiconductor chips that tech companies have been buying like they’re Pokemon cards? Some very smart (and very pessimistic) people think these gadgets are going to lose their value way faster than anyone expected. And when that happens, it’s going to hit company earnings like a freight train carrying bad news.
The poster children for this doom-and-gloom scenario? None other than Michael Burry (yes, the guy who predicted the housing crash and got a movie made about him) and Jim Chanos (another legendary short seller who’s basically the Grim Reaper of overvalued stocks).
Burry’s doing some back-of-the-napkin math that would make your accountant cry. He thinks Big Tech is underestimating depreciation by a whopping $176 billion between 2026-2028. His reasoning? These chips probably have a lifespan of 2-3 years, not the 6 years companies are banking on. That’s like buying a car thinking it’ll last a decade, only to find out it turns into a pumpkin after two years.
Peter Berezin from BCA Research dropped some numbers that’ll make you spit out your coffee: if tech companies hold $2.5 trillion in AI assets by 2030 (which seems likely at this rate), and those assets depreciate at 20% annually, we’re looking at $500 billion in yearly depreciation expenses. For context, that’s more than all these companies’ combined profits for 2025.
Kai Wu from Sparkline Capital painted an even grimmer picture, suggesting depreciation could jump from $150 billion to $400 billion annually in just five years. He compared current AI spending to historical booms like railroads and the internet, and guess what? When you adjust for how quickly AI chips become obsolete, today’s AI buildout makes those look like child’s play.
Now, before you start panic-selling your tech stocks, remember that not everyone’s buying into this doomsday scenario. Bernstein analyst Stacy Rasgon thinks GPUs can actually run profitably for about 6 years, and that most companies’ depreciation accounting is reasonable.
But here’s the thing about markets: sometimes it doesn’t matter if the fear is rational or not. The mere possibility that these expensive toys might turn into expensive paperweights faster than expected is enough to send investors running for the hills. The Nasdaq 100 is down 6.3% over recent weeks, and tech stocks are getting hammered harder than a nail in a construction zone.
So what’s the takeaway? In the world of AI investing, yesterday’s revolutionary breakthrough could be tomorrow’s museum piece. And in a game where the rules change faster than TikTok trends, maybe it’s worth asking: are we building the future, or just really expensive digital sandcastles?