Steve Eisman, a prominent investor known for his role in the 2008 financial crisis, has publicly pushed back against fellow “Big Short” investor Michael Burry’s concerns regarding the accounting practices of major artificial intelligence companies. During a discussion on the New Money podcast, Eisman challenged Burry’s thesis that hyperscalers are overstating earnings by underreporting depreciation on AI hardware.
Eisman argued that the current demand for chips, including older models, is so high that Burry’s concerns are not significant enough to alter the broader outcome of the AI trade. He stated, “Where I think he is wrong, for the moment, is that there is such demand for chips right now that there is still huge demand for the older chips.” Eisman further characterized Burry’s argument as “too academic,” suggesting that while the issue is real, it is not a primary driver of the sector’s performance.
The disagreement centers on how companies like Nvidia Corp. (NASDAQ: NVDA) and others manage the depreciation of their A100, H100, and B200 chips. Burry has previously accused hyperscalers of using accounting practices to inflate earnings, comparing Nvidia’s defense of its depreciation schedules to a market bubble. In response, Eisman noted that if AI growth drives strong performance for firms like Anthropic and OpenAI, changes to depreciation schedules would be relatively insignificant.
Despite challenging Burry’s specific depreciation argument, Eisman remains cautious about the broader AI trade. He previously reduced his exposure to the sector, partially hedging four AI positions. Eisman also highlighted specific risks within the trade, estimating that OpenAI and Anthropic account for approximately 70% of AI-related revenue at Microsoft Corp. (NASDAQ: MSFT), Amazon.com Inc. (NASDAQ: AMZN), Alphabet Inc. (NASDAQ: GOOGL) and Oracle. He warned that difficulties at either AI lab could have significant implications for the cloud giants.
