Michael Burry, a big short seller, has broken down his short selling accuracy in the past two years: non AI and Chinese concept short sellers have shorted all 8 times, while AI short sellers have a winning rate of only 33% BlockBeats News: On July 3rd, Michael Burry, the prototype of the "big bear", announced today that he was targeting Micron Technology for a short of $1051.87 on Wednesday, sparking market attention once again. According to data compiled by BlockBeats based on SEC 13F publicly available position directions, Michael Burry has hit 9 out of 11 put options in the past two years, with an overall win rate of approximately 82%. But upon closer inspection, there is extreme internal differentiation: a total of 8 short positions were made by non AI and Chinese concept traders, all of which were successful. BABA, JD, PDD, and others contributed the vast majority of the winning percentage. On the other hand, AI related short positions have only a 1/3 hit rate, with Nvidia, the most popular short seller, failing twice: its NVDA rose 46% during Q1 2025 and another 4% in Q3, becoming the core source of all failed cases. The PLTR short selling disclosed in Q3 2025 only fell by 2.9%, barely hitting, and the AI related short selling win rate was only 33%. The reason why the market generally feels that Michael Burry is "not accurate recently" is because the most discussed behavior is precisely AI short selling, as well as the newly added TSLA, AMAT, CAT, SOXX, MU short positions in July 2026. However, the medium probability and non AI positions that increase their overall winning rate have extremely low publicity and are easily overlooked in the public impression. In addition, the 13F document does not disclose the option premium, exercise price, and expiration date, so hitting the direction does not necessarily mean actual profit. However, the data clearly reveals that Michael Burry, the "big bear," has a much stronger ability to short on medium and traditional targets than on the AI track. Michael Burry has recently further expanded his short positions in the AI sector, including targets such as Tesla, Applied Materials, Caterpillar, and semiconductor ETF SOXX. Burry stated that South Korea's recent announcement of investment plans to build a large-scale chip industry cluster is the "beginning of the end" of the AI boom, and questioned whether the continued expansion of capital expenditures could bring corresponding returns. Burry continues to maintain a bearish stance on Nvidia and is betting on SOXX's decline through put options. He had previously predicted that the shares of Nvidia and Palantir Technologies would have a significant correction in the next few years, believing that there was a foam risk in the current valuation of the AI sector.