The AI Bubble Debate: What the Critics Get Wrong

Garrison Lovely opens his new book Obsolete with a paradox: the amount of money flowing into artificial intelligence is "truly world historic," but the most prominent arguments that it's all a bubble rest on shakier ground than most people realize. Lovely, a freelance journalist who spent years reporting on the AI industry, doesn't dismiss the possibility of a correction. Instead, he argues that the popular bubble narrative—the one you hear from commentators like Ed Zitron and Cory Doctorow—misses what's actually happening and why it matters.

The Nvidia Problem and Circular Financing

The most compelling piece of the bubble case centers on Nvidia. At the time of recording, the chipmaker was valued at roughly $5.3 trillion—"the first five trillion dollar company," Lovely notes, and "two and a half times bigger than the biggest monopolies in the gilded age, even adjusted for GDP." Nvidia designs the best AI chips in the world with enormous margins, and its customers are almost entirely the hyperscalers: Google, Amazon, Meta, and Microsoft. If any of them decide to stop building data centers, Nvidia's revenue collapses, and so does its valuation.