Big Short Michael Burry doubles down on shorting Nvidia and Tesla: Korean chips are "the beginning of the end of AI"

Michael Burry, the real-life inspiration for "The Big Short" and founder of Scion, has recently increased his short positions on the AI sector, revealing five short targets: Tesla (entry average price $416.22), Caterpillar ($1,060.98), Applied Materials ($729.40), NVIDIA ($198.09), and the semiconductor ETF SOXX ($642.80). The trigger was South Korea's announcement of a massive chip investment plan, which Burry directly called "the beginning of the end" of the AI frenzy. He also pointed out that the Philadelphia Semiconductor Index is currently more than 65% above its 200-day moving average, a deviation last seen at the peak of the 2000 dot-com bubble. (Background: Samsung and SK Hynix invest 2,000 trillion won to build wafer fabs! Betting big on memory for the next 10 years, Gwangju becomes the biggest winner) (Context: The Magnificent Seven start underperforming the market! Report: AI capital expenditures are eating into the cash flow of the Mag 7) Table of Contents Toggle

• "The Beginning of the End": South Korea's Chip Plan Gave Him the Entry Signal
• SOX Deviates 65% from Moving Average, a Level Last Seen in 2000
• $176 Billion in Depreciation Understated Key Summary
• Burry reveals entry average prices for five short targets: NVIDIA $198.09, Tesla $416.22, SOXX $642.80, Applied Materials $729.40, Caterpillar $1,060.98, betting on a broad downturn in AI infrastructure theme
• On the day South Korea announced its massive chip investment plan, Burry called it "the beginning of the end"; Caterpillar is his self-proclaimed first short, as its stock surged 86% in the first half of 2026 and its price-to-sales ratio hit a 30-year high
• Philadelphia Semiconductor Index deviates 65% from its 200-day moving average; Burry notes the last such deviation was during the 2000 dot-com bubble; he also estimates hyperscalers will understate about $176 billion in depreciation expenses from 2026 to 2028, inflating book profits On the day South Korea announced its massive chip investment plan, AI-related stocks rallied across the board, but Michael Burry chose to go against the tide and increase his short positions. The founder of Scion Asset Management, who famously profited from accurately predicting the 2008 subprime mortgage crisis and was portrayed in the film "The Big Short," recently disclosed his latest short positions, spanning Tesla, Caterpillar, Applied Materials, NVIDIA, and the iShares Semiconductor ETF (SOXX) that tracks U.S. semiconductor stocks. This is not just betting against a single company but a comprehensive bet on a downturn in the entire AI infrastructure theme.

Wrong direction. This is like Nordstrom’s opening up Nordstrom’s Rack. But Neiman Marcus opens up Neiman Marcus Rack, Bloomingdale’s opens up Bloomingdale’s Rack, Dillard’s opened up Dillard’s rack. In the end all it proves is there are too many clothes for sale. Meta pops 9%… — Cassandra Unchained (@michaeljburry) July 2, 2026

"The Beginning of the End": South Korea's Chip Plan Gave Him the Entry Signal

Burry directly cited the news from South Korea as his entry signal in his post, writing: "The proximate cause of today’s rally is big spending announced out of Korea. Well, I see that as the beginning of the end. It is only a matter of time now." The market interpreted South Korea's spending as a positive catalyst for accelerating the AI supply chain, but Burry read it as a signal of overheating. This is Burry's typical contrarian approach: at the peak of market euphoria, he bets on a valuation reckoning. Burry also disclosed the specific entry average prices for the five targets: Tesla $416.22, Caterpillar $1,060.98, Applied Materials $729.40, NVIDIA $198.09, and SOXX $642.80. Regarding Caterpillar, Burry claimed this was his "first" short on the industrial veteran, with a straightforward reason: Caterpillar's stock surged 86% in the first half of 2026, and its price-to-sales ratio (P/S ratio, market cap divided by annual revenue, measuring the premium investors pay for each dollar of revenue) rose to a nearly 30-year high, far exceeding fundamental support.

SOX Deviates 65% from Moving Average, a Level Last Seen in 2000

The Philadelphia Semiconductor Index (SOX, commonly known as the "Philly Semi," a key benchmark for the temperature of U.S. semiconductor stocks) is currently about 65% above its 200-day moving average. A moving average can be understood as the average stock price over a certain period; the larger the deviation, the more the current stock price has departed from its long-term trajectory. Burry directly compares this 65% figure: the last time such a deviation occurred was at the peak of the 2000 dot-com bubble, just before the Nasdaq crash. Burry established his short positions in NVIDIA and Palantir as early as November 2025, using put options (derivatives that bet on a decline in the underlying asset; the deeper the decline, the greater the profit). These include put options on 1 million shares of NVIDIA with a strike price of $110, expiring in 2027, and two batches of Palantir: put options expiring at the end of 2026 with a strike price of $100, and put options expiring mid-2027 with a strike price of $50. These strike prices are significantly below current market levels, indicating he is betting on a deep, long-term correction rather than short-term volatility.

$176 Billion in Depreciation Understated

Burry's other weapon against the AI frenzy targets the financial reports of hyperscalers (broadly referring to giant cloud platforms like AWS, Microsoft Azure, Google Cloud). He accuses these companies of systematically extending the depreciation lives of AI chips, understating annual amortization expenses. In simple terms, if a batch of GPUs is normally depreciated over five years, changing it to eight years reduces the annual book cost, making current profits look "better." Burry estimates that from 2026 to 2028, the entire AI industry will understate approximately $176 billion in depreciation expenses, meaning a significant portion of book profits is built on accounting assumptions. When part of the "return" on capital expenditure is manufactured by understating depreciation, the profit growth figures investors see are subject to recalculation. Scion Asset Management filed to deregister with the SEC at the end of November 2025, so Burry is no longer bound by formal fund structure regulations. He now directly discloses his holdings and views on X and his Substack "Cassandra Unchained." Without the obligation of quarterly filings, he speaks faster and more directly, and his target list has grown longer. Frequently Asked Questions Which AI-related stocks did Michael Burry short, and what are the specific entry average prices? Burry revealed five targets and their average prices: Tesla $416.22, Caterpillar $1,060.98, Applied Materials $729.40, NVIDIA $198.09, and semiconductor ETF SOXX $642.80, spanning technology stocks, industrial stocks, and a basket of semiconductor stocks, collectively betting on a broad downturn in the AI infrastructure theme. What does Burry mean by hyperscaler depreciation "padding"? Burry accuses large cloud platforms such as AWS, Microsoft Azure, and Google Cloud of extending the depreciation lives of AI chips, reducing annual amortization expenses to make book profits look better. He estimates that from 2026 to 2028, the entire industry will understate about $176 billion in depreciation, meaning that part of the AI profit growth seen by the market is built on accounting assumptions. This article is for reference only and does not constitute investment advice. Markets are highly volatile; please conduct your own research and carefully assess risks before investing.