Michael Burry, known for predicting the 2008 housing meltdown, now focuses on artificial intelligence. He deregistered his hedge fund but remains active in investing. Phil Clifton’s research supports skepticism on the economics behind AI infrastructure buildout not justifying costs.

AI usage is widespread, but demand economics are small. OpenAI’s revenue is set to surpass $20 billion this year, but it’s tiny compared to the massive AI build-out. Clifton questions if the sums spent on infrastructure are justified by generative AI services.

Scion draws parallels between the AI infrastructure race and the early-2000s telecom boom. Heavy investment could outpace future demand, leading to potential economic issues. Some Big Tech companies are already hesitating on AI infrastructure commitments.

Nvidia has benefited greatly from AI spending, but Scion doubts the economic returns on that investment. Tech giants’ lengthened server lifespans may conflict with Nvidia’s product cycles, raising questions about the company’s defense of its hardware efficiency.

Michael Burry launches a Substack newsletter to lay out his bearish thesis on AI. The debate continues on whether generative AI is a bubble. Burry positions himself cautiously in the fast-moving AI story, challenging the economic viability of AI infrastructure.

Read more at CNBC: An inside look at his analysis showing AI is a bubble