Michael Burry, the acclaimed investor celebrated for his prescient stance during the 2008 financial crisis and immortalized in the book and film ‘The Big Short,’ has once again taken a defiant position against prevailing market sentiment. In a detailed post on his recently launched Substack, Burry intensified his critical commentary on Nvidia and other leading artificial intelligence firms, revealing that he has taken explicit short positions against both Nvidia and Palantir—a bold move intended to underline his skepticism about the current AI investment frenzy.
In his Substack entry, provocatively titled ‘Unicorns and Cockroaches: Blessed Fraud,’ Burry expressed astonishment at the tone and content of Nvidia’s recent internal memo addressed to Wall Street analysts. According to Burry, the memo appeared to be a defensive reaction to claims he never actually made, creating arguments he described as “straw men” designed to misrepresent his original observations. He professed genuine disbelief that such reasoning could emerge from what is presently the world’s most valuable publicly traded company, suggesting that the memo “almost reads like a hoax.” To him, the document exemplified a kind of rhetorical deflection, substituting caricature for substantive engagement.
Burry elaborated on his frustration by clarifying that his criticism had been misconstrued. Since Nvidia functions primarily as a designer of semiconductors rather than an industrial-level manufacturer, he had never alleged that the firm was artificially extending the depreciation period for its property, plant, and equipment. The technical distinction is crucial, as Nvidia’s business model is based on design innovation rather than capital-intensive production facilities. “No one cares about Nvidia’s own depreciation,” he remarked bluntly, emphasizing that such a claim would make little financial or operational sense in context. He then metaphorically punctuated his point with the phrase, “One straw man burnt.”
Turning to another of Nvidia’s arguments, Burry dismissed the company’s insistence that continued use of older-generation chips invalidates his perspective. His real concern, he explained, is forward-looking—that the next iterations of these chips could rapidly lose their technological relevance between 2026 and 2028 as newer architectures emerge. This potential obsolescence, he warned, poses a serious issue for investors assessing long-term valuations within the increasingly competitive AI sector. “I am looking forward because I see problems that are relevant to investors today,” he wrote, effectively burning what he called “a second straw man.” In his view, Nvidia’s response to his critique was not only inadequate but “disingenuous on its face, and disappointing,” signaling a defensive corporate posture rather than constructive dialogue.
The investor further disclosed that, consistent with his critical observations, he continues to hold put options—bearish financial positions—on both Nvidia and Palantir, two of the market’s most prominent AI-related equities. These positions reflect his expectation that their valuations could decline from what he perceives as unsustainably euphoric levels. While he intends to write more about these trades in the future, his revelation already underscores a deep conviction that current AI stock pricing is vulnerable to correction. When approached for comment on Burry’s latest remarks, Nvidia did not issue an immediate response.
Central to Burry’s broader argument is his ongoing concern over how artificial intelligence companies are managing their depreciation accounting—specifically, how rapidly or slowly they record the decline in their asset values over time. This technical yet crucial area of financial reporting can significantly influence both near-term profit figures and long-term balance sheet integrity. Companies can, for example, boost their short-term profitability and inflate their asset valuations by extending the projected useful life of expensive technical infrastructure—such as data center servers or GPUs—from three years to five or six. However, Burry warned that such accounting practices, while superficially beneficial, may produce severe financial repercussions later, as overstated asset values are eventually subjected to steep writedowns.
To illustrate his point, Burry referenced a recent public statement by Microsoft CEO Satya Nadella, who disclosed that Microsoft had deliberately slowed down its data center expansion earlier in the year. Nadella explained that his caution stemmed from concerns over building infrastructure optimized for one generation of AI chips, which could soon be rendered inefficient by the differing energy and cooling requirements of next-generation models. Burry interpreted Nadella’s comments as tacit confirmation that the pace of technological change in AI hardware is dramatically accelerating, further magnifying the depreciation risk.
Expanding on this theme, Burry accused major cloud-service providers—often called “hyperscalers”—of methodically extending the official useful lives of their chips and servers for depreciation purposes, even as they collectively inject hundreds of billions of dollars into hardware that is, paradoxically, designed to become obsolete more quickly. He described the situation as a structural contradiction within the sector: intense spending justified by optimism in AI advancement, coupled with accounting adjustments that obscure the true cost of this aggressive capital allocation. Reflecting on the scale of response his commentary has generated, Burry admitted, “I have been drawn into something much bigger than me,” acknowledging that the pushback from Nvidia and the broader market reaction highlight the sensitivity surrounding his claims.
Meanwhile, Nvidia’s stock price has reflected a degree of volatility, dropping approximately fourteen percent from its high point on November 3. This decline underscores growing investor anxiety that the AI sector may be overextended, with valuations detached from fundamental earnings potential. The market’s unease mirrors Burry’s long-standing warnings that periods of technological euphoria often precede painful corrections.
Burry’s reputation as a contrarian thinker stems from his famous and lucrative short against the U.S. housing bubble nearly two decades ago—a trade that became emblematic of prescient skepticism amid widespread mania. Since then, he has repeatedly cautioned investors about cyclical excesses and the dangers of herd behavior in financial markets. Following a two-year hiatus from social media, he returned to the platform X (formerly Twitter) in late October, once again sharpening his focus on what he perceives as unsustainable overvaluation—this time within the realm of artificial intelligence. Through both his public remarks and his portfolio positioning, he has made it clear that he views the current AI boom as a speculative bubble reminiscent of previous periods of irrational exuberance.
His investment management firm, Scion Asset Management, disclosed on November 3 that it had established bearish positions in the form of put options against both Nvidia and Palantir as of the end of September. The notional combined value of these positions reached roughly $1.1 billion, though Burry clarified in his most recent post that the actual premium outlay required for each was closer to $10 million. The leverage inherent in options allows such positions to express large notional exposures with limited upfront capital—emphasizing the scale of his conviction without committing an excessive share of his assets.
By unveiling these moves and articulating his rationale at length, Michael Burry has once again positioned himself as one of the market’s most outspoken skeptics. Whether his criticisms will prove prophetic or premature remains to be seen, but his engagement ensures that the debate over AI’s real versus perceived value is far from settled.
Sourse: https://www.businessinsider.com/big-short-michael-burry-substack-nvidia-memo-depreciation-ai-bubble-2025-11
