Michael Varley, of “The Big Short” fame, said some of the world’s biggest AI companies are exaggerating the lifespan of their Nvidia chips to inflate short-term profits. Now, one hedge fund boss has warned that these chips could become obsolete.
“The depreciation schedule is probably going to be much faster, and in some ways I hope so,” Greg Jensen, co-lead investor at Ray Dalio’s Bridgewater Associates, said on the “In Good Company” podcast this week.
Jensen explained that there is a “competition for resources” for AI as companies compete for scarce land, energy, microchips and scientists, and that technology company leaders hope that AI itself will be useful.
“One of the things they have to do is find ways to make chips more efficient, more energy efficient, and they’re going to use AI to do those things,” he said.
“AI will generate better ways to do this,” Jensen predicted, adding that “some of the advances in the science of depreciating current assets will come from those assets themselves.”
Puts, trades, and ecosystems
Barry shot to fame after his huge bet on the US housing bubble was immortalized in the book The Big Short and the film adaptation starring actor Christian Bale as Barry.
He reappeared on X in late October after more than two years of silence. Since then, he has sounded the alarm on the AI bubble, closed his hedge fund to outside funding, launched Substack to share research results, and disclosed that he owns bearish put options on Nvidia and another AI darling, Palantir.
Burley faulted the AI giant for stretching depreciation from about three years to more than six years, noting that current-generation chips are likely to lose value faster because Nvidia is releasing new chips faster and faster.
“Hyperscalers are investing hundreds of billions of dollars in graphics chips while accelerating planned obsolescence, thereby systematically extending the useful life of chips and servers for depreciation purposes,” he wrote on Substack this week.
The investor also criticized the sprawling web of “give-and-take deals” between AI companies.
Jensen said these are not the product of “normal bubble dynamics” (where companies squeeze their financials to justify high valuations) and that given the huge demand for their chips, “NVIDIA can make as much money as they want.”
Instead, NVIDIA is bent on building its own ecosystem of buyers who won’t develop their own chips to prevent Alphabet from owning the entire AI “stack,” Jensen said.
“They’re like the Standard Oil of the Gilded Age, trying to monopolize things,” Jensen said of Nvidia. He added, “Everyone has to be restrained as to who I team up with and where I get my chips and power. If I don’t, I’m going to die.”
Jensen also said the AI investment boom is not a typical capital cycle, as executives such as Elon Musk and Sam Altman believe they are in a race to develop the best minds and are willing to spend whatever it takes to win.
