The debate between "The Big Short vs NVIDIA" continues! "The Big Short" refutes NVIDIA's response point by point: Can't believe this comes from the world's most valuable company

The battle between the "big short" and NVIDIA is escalating. Michael Burry, the prototype of the movie "The Big Short," has once again directed his criticism at NVIDIA, calling an internal memo distributed to Wall Street analysts in response to his critiques "disappointing" and filled with "one straw man argument after another."

On Tuesday, November 26, Burry wrote on his newly launched Substack blog that he "couldn't believe" these responses came from the world's most valuable publicly traded company, describing the document as "reading almost like a scam."

This latest statement continues Burry's recent series of warnings about the AI industry bubble. Previously, he expressed his views on social media platform X regarding AI stocks being in a bubble. The direct consequence of this round of confrontation is that the market's scrutiny of AI-related companies has become increasingly harsh, with NVIDIA's stock price falling about 14% from its peak on November 3, indicating a subtle shift in investor sentiment.

Burry revealed in his article that he is currently "still holding put options against Palantir and NVIDIA." This confirms that he has not closed his previous short positions, directly signaling his bearish outlook on the prospects of these two AI star companies to the market.

"Straw man" rebuttal?

Burry's core rebuttal point is that he believes NVIDIA's memo misrepresents his criticisms and attacks points he never raised.

First, Burry points out that he has never questioned NVIDIA's fixed asset (PP&E) depreciation policy. He emphasizes that NVIDIA is primarily a chip design company with minimal capital expenditure, so "no one cares about NVIDIA's own depreciation." He considers NVIDIA's defense on this issue to be the first "burned straw man."

Second, regarding NVIDIA's claim that its older generation chips are still in use, Burry is equally unconvinced. He clarifies that his focus is on the future, specifically that new chips may become functionally obsolete between 2026 and 2028.

"I look forward because I see issues relevant to today's investors," he wrote, calling this the "second burned straw man." Burry concludes that NVIDIA's rebuttal is "superficially insincere and disappointing."

Core of the debate: depreciation, bubbles, and future impairment risks

The core of Burry's argument this time directly points to the accounting practices of artificial intelligence companies, especially the depreciation policies of their clients—those "hyperscalers" that spend huge sums on AI chips.

He explained that these companies are systematically extending the "useful life" of chips and servers in order to spread depreciation costs over time in accounting. For example, extending the depreciation period from three years to five or six years can beautify profits and increase the book value of assets in the short term.

However, Burry warned that the risk of this practice lies in the rapid technological iteration of AI chips. He cited a recent interview with Microsoft CEO Satya Nadella as evidence, where Nadella stated that he slowed down the construction of data centers earlier this year due to concerns about overbuilding infrastructure for a generation of AI chips, as the next generation of chips will have different power and cooling requirements.

Burry believes that if the "planned obsolescence" of chips accelerates, companies that have extended their depreciation periods will face significant asset write-down risks in the future.

Details of Short Positions and Market Background

Burry not only reiterated his bearish stance but also disclosed more details about his positions. According to documents submitted by his asset management company Scion Asset Management on November 3, the company held put options on NVIDIA and Palantir at the end of September.

In his latest article, Burry clarified that although the nominal value of these positions is as high as $1.1 billion, their actual cost is "only about $10 million each." This detail provides investors with a more accurate perspective on the true scale of his bets.

The backdrop of this debate is that market enthusiasm for artificial intelligence has pushed the stock prices of related companies to historic highs this year. However, since early November, as the earnings season unfolds and warnings from well-known investors like Burry emerge, market concerns have begun to surface. Besides NVIDIA, Palantir's stock price has also fallen 20% from recent highs, although its year-to-date gains remain impressive. Burry's continued commentary undoubtedly adds more uncertainty to the debate about whether AI is a revolution or a bubble.

NVIDIA's Internal Memo

In the face of stock price pressure and market doubts, NVIDIA previously took a series of rare defensive communications.

Last weekend, NVIDIA distributed a seven-page memo to Wall Street analysts aimed at "refuting" point by point the accusations of accounting fraud, circular financing, and AI bubble raised by critics, including the "big short" Burry. The memo clarified:

• Accounting Practices: NVIDIA emphasized that its business is economically sound, its reports are complete and transparent, and there is no comparability with historical accounting fraud cases

• Equipment Depreciation: In response to accusations regarding the lifespan of AI chips being only 2-3 years, NVIDIA stated that its customers typically set the depreciation period for GPUs at 4 to 6 years, which aligns with the actual usage lifespan of the equipment.

• Revolving Financing: Regarding this accusation, NVIDIA clarified that its strategic investments in the third quarter amounted to only $3.7 billion, a small portion of its revenue, and the claim of revolving financing is "baseless."

Additionally, an article from Wall Street Insight reported that NVIDIA publicly declared on the social platform X on November 25: "NVIDIA currently leads the industry by a generation—we are the only platform capable of running all AI models and is versatile across various computing scenarios." The company emphasized that its GPUs provide "higher performance, versatility, and interchangeability" compared to ASIC chips like Google's TPU