Billionaire investor Stanley Druckenmiller just revealed his latest AI bets, and Nvidia is still not on the list. His first quarter 2025 filings show fresh positions in Broadcom, Micron Technology and Intel, signaling a clear shift toward the next wave of artificial intelligence infrastructure.
Why Druckenmiller Walked Away From Nvidia
The story starts in 2023, when Nvidia sat as the single largest holding in Druckenmiller’s portfolio at the Duquesne Family Office. The legendary investor, known for delivering 30% average annual returns over three decades without a single losing year at Duquesne Capital Management, began trimming his Nvidia stake late that year.
By 2024, he had sold every last share.
In a Bloomberg interview that fall, Druckenmiller called the early exit one of the biggest mistakes of his career. He hinted he would buy back in if the valuation came down. Nvidia’s forward price-to-earnings ratio has indeed dropped since then, yet the billionaire has not pulled the trigger.
His Q1 2025 SEC filing confirms it: zero Nvidia shares purchased. Instead, he deployed capital into three companies he apparently sees as better positioned for what comes next in AI.
The Three AI Stocks Druckenmiller Just Bought
According to the 13F filing made public this month, Druckenmiller opened or expanded positions in three semiconductor names during the first quarter:
- Broadcom (AVGO): 195,955 shares purchased, now 1.8% of his portfolio
- Intel (INTC): 411,400 shares purchased, now 0.5% of his portfolio
- Micron Technology (MU): 23,400 shares purchased, now 0.2% of his portfolio
Taiwan Semiconductor Manufacturing also remains his third largest holding at nearly 5% of the portfolio. Together, these positions paint a picture of someone betting heavily on the broader AI supply chain rather than a single dominant chipmaker.
What Makes Broadcom His Biggest New Bet
Broadcom stands out as the largest fresh purchase by portfolio weight. The company has carved a unique space in AI by designing custom chips tailored to specific workloads for hyperscale cloud customers like Google and Meta.
While Nvidia sells general purpose GPUs that work across many applications, Broadcom builds application specific integrated circuits (ASICs) that deliver maximum performance for narrowly defined tasks. This approach has fueled massive revenue growth in its AI segment.
Broadcom’s AI revenue grew over 220% year over year in its most recent fiscal quarter. The company now commands a market capitalization near $2 trillion, reflecting Wall Street’s confidence in the custom silicon opportunity.
For Druckenmiller, this bet suggests he believes the AI chip market is big enough for more than one winner, and that customization will matter more as AI models become specialized.
Intel and Micron Fill Critical Gaps in the AI Stack
The Intel position is perhaps the most contrarian move. Intel has struggled to keep pace during the AI boom, losing ground to Nvidia in data center computing. But the company is staging a quiet comeback.
Intel recently partnered with Nvidia to pair Intel CPUs with Nvidia GPU chiplets in personal computing systems. AI workloads still require strong central processors for tasks like data preprocessing and inference at the edge. Intel dominates that market.
The company is also investing billions in new fabrication plants across the United States, positioning itself as a domestic manufacturing alternative at a time when supply chain security matters more than ever.
Micron tells a different but equally compelling story. AI models are hungry for memory. Training large language models and running inference at scale requires enormous amounts of high bandwidth memory (HBM), and Micron is one of only three companies in the world that can make it.
Micron posted record revenue, record gross margin and record earnings per share in its latest quarter. Management expects more records in the current period. The company’s biggest problem right now is not demand. It is keeping up with orders.
Should Investors Follow This Playbook
Druckenmiller’s track record demands attention. But copying a billionaire’s trades comes with important caveats.
First, timing matters. These 13F filings reflect positions as of March 31, meaning Druckenmiller could have already adjusted his holdings since then. Second, valuation tells a nuanced story right now:
| Stock | Forward P/E Ratio | 5-Year Return |
|---|---|---|
| Nvidia (NVDA) | ~26x | Over 1,500% |
| Broadcom (AVGO) | ~32x | Over 400% |
| Micron (MU) | ~15x | Over 60% |
| Intel (INTC) | N/A (rebuilding earnings) | Negative |
Nvidia actually looks cheaper than Broadcom on a forward earnings basis today. Value focused investors might find that surprising.
However, Druckenmiller has never been a pure value investor. He looks for inflection points. His bets on Broadcom, Micron and Intel suggest he sees these companies entering a new growth phase as AI spending broadens beyond just GPU purchases into custom chips, memory and traditional processors.
The AI boom is clearly not a one stock story anymore. Druckenmiller’s latest moves remind us that the biggest opportunities often sit just outside the spotlight, in the companies building the roads rather than driving the fastest car. Whether you follow his lead or stick with Nvidia, the underlying message is the same: AI infrastructure spending is accelerating, and the winners will be those who supply what the industry cannot function without. Share your thoughts on which AI stock you are betting on this year.
