When Peter Thiel moves money around, people pay attention. And his latest moves are raising eyebrows across Wall Street. The billionaire venture capitalist and PayPal co-founder has completely dumped his Nvidia Corporation (NVDA) stake and dramatically reduced his Tesla Inc. (TSLA) holdings, according to a recent Form 13F filing from his Thiel Macro fund.

A Complete Reversal on Nvidia

The numbers tell a striking story. Thiel offloaded approximately 537,742 Nvidia shares during the July-September quarter. By September 30, he held exactly zero shares in the company that's been the poster child for the AI revolution. For someone who spent much of 2024 talking up Nvidia's dominance in AI chips, it's quite the about-face.

But wait, there's more. Thiel also slashed his Tesla position from 272,613 shares down to just 65,000 shares—a roughly 76% reduction. He also completely exited his 208,747-share stake in Vistra Energy Corp (VST), a utility company that had been riding the AI data center power demand wave.

The Mag 7 Shuffle

Here's where it gets interesting. Thiel didn't just take his chips off the table—he moved them to different parts of the casino. He significantly bulked up on two other "Magnificent 7" stocks, purchasing 79,181 shares of Apple Inc. (AAPL) and 49,000 shares of Microsoft Corporation (MSFT).

So what's the logic? If you're worried about an AI bubble, why rotate from Nvidia into Microsoft, which has poured billions into OpenAI and is itself deeply tied to AI infrastructure? Perhaps Thiel sees Apple and Microsoft as having more diversified revenue streams and less extreme valuations relative to their AI exposure. Or maybe he's simply taking profits after Nvidia's monster run.

SoftBank Did the Same Thing

Thiel isn't alone in his Nvidia exit. SoftBank completely dumped its Nvidia position in October, selling all 32.1 million shares for roughly $5.83 billion. This came despite SoftBank CEO Masayoshi Son publicly declaring earlier that month that Nvidia was "undervalued."

Son dismissed concerns about an AI bubble entirely, painting a massive vision for artificial superintelligence that would require 400 gigawatts of data-center power and up to 200 million AI chips—representing about $9 trillion in cumulative investment. He argued that even this astronomical level of spending would be "reasonable," and possibly "too small." Yet somehow, SoftBank still sold every single share. Make of that what you will.

What Changed Thiel's Mind?

The timing is particularly curious given Thiel's previous public statements about Nvidia. Back in August, he argued that the real focus in AI should be on Nvidia's chip dominance rather than on model developers like Meta Platforms (META), OpenAI, or xAI. He noted that Nvidia was capturing most of the industry's profits while others were losing money, thanks partly to a generational talent shift that left the semiconductor sector undervalued.

But even then, Thiel had expressed some reservations. In July 2024, he called it "very strange" that 80-85% of all profits in AI were flowing to a single company—Nvidia. He found it odd that most of the money was being made in the hardware layer, a part of the tech stack that Silicon Valley wasn't deeply familiar with anymore.

Perhaps those concerns eventually outweighed his bullish thesis. When one company is capturing that much of an industry's profits, it creates an uncomfortable concentration of risk. And when that company's valuation reflects sky-high expectations for continued dominance, the margin for error gets pretty thin.

The Valuation Picture

According to market data, Nvidia ranks in the 98th percentile for growth but only the 3rd percentile for value—a profile that screams "priced for perfection." For a value-conscious investor like Thiel, that might be reason enough to take profits and rotate into companies with more reasonable valuations, even if their growth potential is somewhat lower.

Whether this marks the beginning of a broader rotation out of AI high-fliers remains to be seen. But when prominent tech investors start shifting billions from the AI trade's biggest winner into steadier blue chips, it's worth paying attention. Sometimes the smartest move is knowing when to leave the party.