Michael Burley recently sounded the alarm on the AI bubble on the Substack exchange, calling out market darlings and warning that a long recession is on the way.
The investor of “The Big Short” fame took aim at the tech boom in a written discussion with Anthropic co-founder Jack Clark and podcaster Dwarkesh Patel.
Escalator that goes on forever
Burley warned that so-called hyperscalers such as Microsoft and Alphabet are wasting huge sums of money on quickly obsolete microchips and data centers to power AI tools such as chatbots, which are becoming commoditized.
He drew similarities to Hochschild-Korn, a department store in Baltimore that was briefly owned by Warren Buffett in the 1960s.
“When the escalator was installed at the department store across the street, he had to do it too,” Barry wrote. “In the end, neither benefited from that expensive project. There was no durable margin improvement or cost improvement. And both were in exactly the same position. This is how most AI implementations play out.”
When one local department store upgraded its windows or implemented a new checkout system, “others were forced to follow suit,” writes author Alice Schroeder in The Snowball: Warren Budget and the Business of Life.
She added that Buffett and his late business partner, Charlie Munger, viewed the situation as “standing on tiptoe in a parade.” “Once someone did it, everyone had to do it.”
Barry wrote on Substack: “This is why trillions of dollars in spending without a clear path to use by the real economy is so concerning. Most companies do not benefit, because their competitors would benefit as much, and neither would have a competitive advantage.”
Given how the previous capex boom has played out, Barry writes that we are now “past the stage where stock prices reward investors for further build-out and entering a period where true costs and revenue shortfalls begin to become apparent.”
Looking ahead, he predicted that tech industry employment would be “lower or not growing as much because we think we’re heading into a very long recession.”
“Poster Children”
Barry shot to fame after his contrarian bet on the mid-2000s housing bubble was chronicled in Michael Lewis’ book The Big Short. In the film version, he was played by actor Christian Bale.
He transitioned from running a hedge fund to writing for Substack late last year. Some of his early posts took aim at skyrocketing AI stocks, warning that they were overvalued and destined to crash.
“I think the market misunderstands the two companies that are the biggest names in AI, Nvidia and Palantir,” he told Clark and Patel.
The chipmaker and data analytics specialist are “the two luckiest companies” because their products happened to be AI-friendly from the start, he added.
“NVIDIA is holding the fort as a power-hungry, dirty solution until a competitor comes along with a completely different approach,” Burry wrote.
He also said that Palantir CEO Alex Karp’s criticism of him for shorting the stock shows that Palantir is “not a confident CEO.”
“He’s marketing as hard as he can to keep this going, but it’s going to fail,” Burley added.
Palantir and NVIDIA did not respond to requests for comment from Business Insider.
Surprise and plumber
The value investor shared three things that surprised him about the AI boom. The first, he said, is that Google “failed and created an opportunity for a competitor that was far less advantageous.” “Google is catching up to startups in AI, which is amazing.”
Second, OpenAI’s chatbot ChatGPT has “launched a multi-trillion dollar infrastructure race.”
“It’s like someone built a robot prototype and every company in the world started investing in a robotic future,” he wrote.
Third, Nvidia continued to dominate, as it expected more power-efficient chips would now be commonplace.
Speaking about the real-world impact of AI, Barry questioned the idea that a career in trading is an “AI-free option.”
“If I was middle-class and asked for an $800 plumber or electrician, I might use Claude,” he said, referring to Anthropic’s AI chatbot.
“I like being able to take a photo and figure out everything I need to do to fix it,” he added.
Burley also raised the prospect that AI chatbots would “make people stupid,” such as doctors who use chatbots so much that they begin to forget medical knowledge.
