Billionaire investor Stanley Druckenmiller spent the spring loading up on tech and semiconductor stocks. Then the market turned on them.
The legendary investor's family office, Duquesne, added to several big AI-related positions during the April-to-June quarter. The timing looks rough now, because many of those same stocks tumbled in July as investors started questioning whether the AI spending boom could keep going.
Druckenmiller's Duquesne Family Office raised its Amazon stake to roughly $129 million during the second quarter. The firm also opened a $120 million position in Alphabet, the parent company of Google, after dumping that stock in the first three months of the year.
The chip bets were even bigger. Duquesne boosted its Taiwan Semiconductor Manufacturing Co. stake by 19%, bringing it to $282 million. It also increased its STMicroelectronics position by nearly 19%, to $232 million. Those two semiconductor companies ended the quarter as Duquesne's second- and third-largest common-stock holdings.
The firm added smaller new positions in Advanced Micro Devices and Palo Alto Networks as well. Natera, a genetic-testing company, remained the biggest holding, valued at more than $800 million on June 30.
If you are done chasing hot stocks, grab the free Always Be Buying eBook to build wealth steadily.
Druckenmiller is one of Wall Street's most closely watched investors. He made his name working alongside George Soros at Quantum Fund, where he helped orchestrate the celebrated 1992 trade that shorted the British pound and reportedly returned $10 billion. Afterward, he led the $12 billion investing operation Duquesne Capital Management, then closed the fund in 2010 and converted it into a family office.
The stock buys came after a strong spring rally in tech and chip shares, fueled by excitement over AI spending. By the time the quarterly report was made public on Aug. 17, 2026, the picture had changed.
Many of those same stocks fell hard in July as investors worried that prices had gotten too high and questioned whether the AI infrastructure boom could keep delivering. The disclosure is dated June 30, so there is no way to know whether Druckenmiller held through the slump, sold into it, or bought even more after the drop.
That is the tricky thing about these quarterly reports. They show a snapshot of what a fund owned at one moment, not what it owns now. By the time you see the numbers, weeks have passed and the market may have moved on.
There is a useful lesson here for regular investors, even if you are working with a lot less than $800 million. One of the best stock pickers of his generation bought into the AI trade at what looked like a great moment, and the market immediately turned against him.
That does not mean Druckenmiller was wrong. It means timing is hard for everyone, even the pros. The AI spending boom could easily continue for years, and these same stocks might look cheap a year from now.
Or the pullback could keep going. Nobody knows yet.
What the report does show is conviction. Druckenmiller did not nibble around the edges. He made big, clear bets on the idea that AI infrastructure companies will keep winning. Whether those bets pay off is a question the next few quarters will answer.
For your own portfolio, the takeaway is simpler. If you own tech stocks, expect bumps like July's. If you are thinking about buying the dip, remember that even the smartest money in the room cannot tell you when the bottom is in.
When market timing fails, the free Always Be Buying eBook shows a calmer path to growing your money.