Stanley Druckenmiller's latest holdings: selling MU and AVGO, re betting on GOOGL and AMZN. Is he in the next stage of trading AI? First, let's talk about Druckenmiller's background One of the most legendary macro traders on Wall Street, with deep connections in the US political and economic circles. Kevin Warsh is his apprentice, and Scott Bessent is also an important disciple of him. Druckenmiller tends to trade in the short to medium term and macroeconomic trends, and will quickly adjust positions based on economic cycles, liquidity, and market changes. So you can refer to his judgment on macro cycles and the direction of capital rotation. I recently looked at Druckenmiller's latest disclosed holdings and found that the adjustment range is actually quite large. First of all, let me emphasize the time: the deadline for this 13F is June 30, 2026, so all the actions reflected in it are completed before the end of the second quarter. It has been over a month now, and his position is likely to have changed again. 13F is suitable for studying its configuration direction and investment logic, and cannot directly copy assignments. He cashed out a portion of the semiconductors that had risen significantly earlier. MU and AVGO have both been cleared. But that doesn't mean he's bearish on semiconductors. TSM has become the second largest holding, accounting for approximately 5.4%, while STM has risen to the third largest holding, accounting for approximately 4.4%, and AMD has also been newly established. So this can be considered as a rotation inside his semiconductor. MU and AVGO, which have already risen significantly, will be cashed out first, and then the funds will look for better directions for risk/reward. Another change is that he has started to invest money in Mega cap AI/Cloud. GOOGL is an important new position in Q2, and AMZN has also been significantly added back. At the same time, the upward elasticity of technology stocks has been retained through META Call and TSLA Call, and PANW has been newly established. In Q1, he had just withdrawn from GOOGL and significantly reduced his holdings in AMZN, but in Q2, he immediately repurchased it. AVGO was bought in Q1 and sold in Q2. This actually fits Druckenmiller's style very well. He won't keep holding onto something just because he was bullish on it a few months ago. If the logic and risk/reward change, the position will also change accordingly. I think there is also a layer of AI Trade's rotation logic behind this. In the past few years, GPU was the first to make money, and later it gradually spread to Networking/Optical, HBM/Memory, Server, and Data Center. GOOGL, AMZN, and other Hyperscalers are the ones who spend money in the first half, constantly buying GPUs, building data centers, and improving CapEx. Now these infrastructures are gradually being put into operation, and the next stage of the market is more focused on the AI CapEx spent in the past few years. When will it truly become Cloud Revenue, AI Revenue, and FCF? If this profit transmission begins to materialize, the cloud giants that used to be responsible for paying the bill will gradually enter the harvest period. Of course, he didn't bet all his money on AI either. NTRA still holds the largest position in the gap, accounting for approximately 16% of the public portfolio, and EWZ Call has also entered the top five, indicating that he is also working on opportunities in healthcare and global macro. So overall, I think Druckenmiller's strategy is to cash in a portion of the already soaring AI Hardware while continuing to retain semiconductor positions, and then advance some of the funds to areas where AI may begin to cash in profits in the next stage. Finally, it should be emphasized that these are actions taken before June 30th, not his current real-time positions. Druckenmiller has been adjusting its position very quickly, and after more than a month, its current position is likely to change again.