In the third quarter of 2025, Thiel Macro LLC, the billionaire’s hedge fund, sold all 537,742 of its Nvidia shares (well over $100 million), according to regulatory filings. The shift marked one of Thiel’s most significant repositionings of the year: The stake in Nvidia had accounted for roughly 40% of the fund’s entire portfolio.

Making money moves

The move didn’t happen in a vacuum. Alongside the Nvidia sale, the fund took up new positions in Apple and Microsoft. Taken together, the trades cut Thiel Macro’s total U.S. equity exposure by more than half, from $212 million down to $74 million. Nvidia may be building AI’s future, but the platforms integrating AI with consumer products could prove the more durable bet. Companies like Nvidia have soared in value in the first few years of the AI boom, benefiting from a mass infrastructure build-out. But a different business model might eventually win out over the long term.

Nvidia has been the reigning beneficiary of the AI boom. Its chips power virtually every major model being trained today, and its stock reflects that status, when it recently became the world’s most valuable company after its market cap hit $5 trillion in October. But Nvidia’s over-performance hasn’t insulated it from growing fears of an AI bubble, particularly after it became clear that the company was involved in multiple circular deals with other big names in AI. 

These arrangements mostly look like this: Nvidia makes a capital investment in an AI startup or cloud provider, in return for commitments from that company to purchase Nvidia’s chips, meaning Nvidia is effectively paying itself to produce chips. The prevalence of these deals—which involve companies including OpenAI and CoreWeave—has raised even more questions as to how much higher the chipmaker’s stock value could soar.

Thiel’s move might have just been to divest himself from Nvidia at a huge profit, but the PayPal and Palantir cofounder has been more measured in his AI takes than most of his Silicon Valley peers. 

“It’s more than a nothing burger, and it’s less than the total transformation of our society,” he said in an interview with the New York Times last year. In that interview, he compared the rise of AI to the advent of the internet in the late 1990s. 

What remained after his repositioning tells its own story. Equally revealing is what Thiel bought when he dumped Nvidia. His Apple and Microsoft buys combined were worth around $45 million, making up the majority of his portfolio, and many commentators online read his repositioning as a cautious bet that AI really is a bubble. Even Thiel himself has called AI “extremely bubbly,” and warned that the technology could go the way of the dotcom crash that many companies were unable to escape from, although the investor has tended to describe AI as ultimately valuable and potentially transformative. 

Apple and Microsoft are not AI underdogs—they are the companies most likely to turn AI from a capital expenditure into a consumer product. Microsoft’s deep integration of AI across Office, Azure, and Copilot has made it the most direct corporate beneficiary of the OpenAI relationship. Apple, meanwhile, is sitting on over 2 billion active devices and a services business that just posted record quarterly revenue—a distribution moat that no chipmaker can replicate.

These companies also have the benefit of being more diversified than Nvidia. While the chipmaker has gone all in on AI, and most of its revenue comes from sales to a small number of large tech companies, Microsoft and Apple both enjoy multifaceted businesses that happen to include AI. Both will certainly be impacted if the AI trade fails to pay off, but their core franchises give them more ways to grow than a chipmaker whose fortunes are tied so tightly to demand for AI compute.