I’ve lived through a number of hypercharged investing sprees in my day, some of which have
proved prescient, and others decidedly less so—the first internet surge, the rise of the alts, fracking, the most recent mobile-fueled Silicon Valley venture bubble, and the streaming wars, to name just a handful. But I’ve never seen anything like the unprecedented capex emanating from our economy’s so-called hyperscalers—Meta, Google, Amazon, and Microsoft, etcetera—in pursuit of artificial intelligence hegemony.
The Four, which earn between $50 billion (Meta) and $175 billion (Amazon) per quarter, are on pace to spend a staggering $360 billion on A.I. infrastructure this year as they buy G.P.U.s, build out extensive data centers, and manage the costs and hazards of operating these extraordinary cloud computing mini cities. Their logic is simple, if somewhat imprecise: A.I. and superintelligence will become the transformative technology of our time, of course,
and while its impact will be vast and ubiquitous, the platform shift won’t confer a rose to every suitor.
The Four, along with Apple, have been the most prosperous companies of the internet age—and, you could argue, of all time—with attendant fortress balance sheets. They may have created their fortunes in software, products, and logistics, but they aren’t going to cede their advantages to upstarts
like OpenAI and Anthropic—both of which are spending like mad, too. Meanwhile, equity analysts are rewarding these companies for using their billions to moat their traditional businesses while investing in their futures; Morgan Stanley analysts predicted in a recent note that an additional $3 trillion could be invested in data centers through 2028. Also, none of their A.I. businesses are yet profitable. Heavy cake.
In a pair of extraordinary pieces this week, my newest partner, Ian Krietzberg, assessed the less appreciated economic underbelly of the A.I. revolution: We’ve gone full throttle into a Wild West—a true gold rush. In A.I. Bubble Bust Theories, Ian looks for clues regarding whether these extraordinary
investments are actually trickling down into their operating plans, and what a correction might look like—and which companies and sectors it could impact. And in a subsequent piece, Slaughterhouse GPT-5, he offered an acute look at the industry’s most famous insurgent—and how it’s managing the comedown from its latest hype cycle.
But
if you only have time to read one piece this weekend, I’d turn your attention to my partner Bill Cohan’s latest masterstroke, Jim Chanos in the Metaverse. Bill phoned up the legendary short seller on a late summer pleasure cruise and engaged him and his newest fixation: Meta’s depreciation schedule for, you guessed it, the
company’s massive program of capital spending. Chanos’s insights are provocative and compelling, and merely the latest indication that we’re in the prehistory of this extraordinary saga—one of the true stories of our time, and especially what you should expect to read about voraciously at Puck.