The "Big Short" investor accuses major tech firms of inflating earnings by extending the useful life of expensive AI chips to lower depreciation expenses.
Michael Burry, the contrarian investor who famously bet against subprime mortgages before the 2008 financial crisis, is raising a new alarm about the artificial intelligence boom. This follows news that the investor has bet against the AI boom.
In a post on X this week, Burry alleged that major technology companies operating massive AI infrastructure are using aggressive accounting practices to pad their reported profits. He says, "Understating depreciation by extending useful life of assets artificially boosts earnings -one of the more common frauds of the modern era." Specifically, he claims they are understating depreciation expenses by estimating that expensive semiconductor chips and servers will last longer than is realistic, given the rapid pace of technological change.
Burry's accusation centers on a common but consequential accounting maneuver. When companies purchase large assets like computer chips, they spread the cost over multiple years based on how quickly those assets lose value. By extending the estimated useful life of equipment, companies can lower their yearly depreciation expense and boost reported earnings.
And as CNBC reports, Burry estimates this practice could understate depreciation by roughly 176 billion dollars between 2026 and 2028 across the industry. He specifically mentions Oracle (ORCL) and Meta (META), suggesting their profits could be overstated by as much as 27 percent and 21 percent respectively by 2028.
The accusation has sparked considerable discussion among observers and commenters online. Many in the community have expressed skepticism about the sustainability of current AI spending levels, with some noting the irony of companies making massive capital investments while claiming their equipment will remain useful far longer than typical product cycles suggest.
The challenge for regulators and investors is that companies do have legitimate leeway under accounting standards to estimate asset lifespans. Proving that a company has deliberately extended these estimates to manipulate earnings would require demonstrating intent, which is difficult. Burry's claims remain unconfirmed by independent analysis, and the companies involved have not yet responded substantively to the allegations.
What is clear is that the rapid expansion of AI infrastructure spending has created a new frontier for accounting scrutiny, and questions about whether current valuations reflect realistic assumptions about future returns are becoming harder to ignore. And this happens with agenic operating systems facing challenges winning trust from some users.
Burry has promised more detailed analysis on November 25, suggesting this conversation is far from over. Whether his warnings prove prescient or premature, they highlight a broader tension in today's market: the gap between the enthusiasm surrounding artificial intelligence and the financial reality of sustaining massive infrastructure investments in a field where technological obsolescence happens at breathtaking speed.
