This OpenAI Researcher-Turned-Hedge Fund Manager Is Long Intel and Short Nvidia, TSMC, and Broadcom. Is a Changing of the Guard on the Horizon?

It's 13F season, the time of the quarter when major hedge funds disclose their buys and sells from the prior quarter.

With so much uncertainty around the longevity of the artificial intelligence (AI) boom, both to the upside and downside, investors are no doubt searching for clues as to what the so-called "smart money" is thinking about leading technology companies.

While many investors follow investing legends, such as Warren Buffett, Stan Druckenmiller, Bill Ackman, Phillipe Laffont, and others, interested investors should especially follow one new hedge fund founded just last year, run by a 23-year-old former OpenAI researcher. Though this manager is a big believer in the artificial intelligence boom -- or at least was previously -- his third-quarter trades may shock fellow AI bulls.

Image source: Getty Images.

Leopold Aschenbrenner's Situational Awareness

In 2024, then-22-year-old Leopold Aschenbrenner founded a new hedge fund called Situational Awareness. The term "situational awareness" originates from the AI world, where an AI model becomes aware of its own circumstances -- a potential security risk much like the computer HAL from 2001: A Space Odyssey.

Some might question following the bets of a 23-year-old or how he was able to raise hundreds of millions of dollars to start a hedge fund. Then again, Aschenbrenner is no ordinary 20-something. He graduated as valedictorian from Columbia University at the age of 19, with a dual major in economics and mathematics -- an age when most of us were just attending our first frat party as freshmen.

In 2023, he joined OpenAI as part of the Superalignment team tasked with cutting-edge research pursuing artificial general intelligence (AGI). Aschenbrenner was known to push buttons at OpenAI and was fired in 2024 over an allegation that he leaked sensitive information to third parties. Aschenbrenner denied this, claiming that he had shared hypothetical scenarios related to safety and AGI with outside parties as part of a brainstorming session, a type of activity common in Silicon Valley. His Superalignment team was later dissolved when other members, such as famed research Ilya Sutskever, left to found their own AI companies.

After his firing, Aschenbrenner published a 165-page essay called "Situational Awareness" in which he anticipated a fairly rapid path to AGI starting in 2027, when he predicts AI models will attain the ability to conduct their own research. Aschenbrenner also warned about the inevitable safety and geopolitical risks that come with the emergence of AGI.

At the same time, he also launched a hedge fund, specifically to capitalize on AGI trends in the market.

Q3 moves: Short Nvidia, TSMC, and Broadcom?

Situational Awareness's buys and sells in the third quarter were extremely surprising in light of what Aschenbrenner seems to believe about AI progress -- or at least what he believed as recently as 2024.

In his 2024 essay, Aschenbrenner anticipated AI training clusters would continue to grow exponentially in size in the years ahead, assuming technological progress proceeds at the current rate. That would see AI clusters going from 100,000 H100-equivalent clusters fed by 100 megawatts of power in 2024, costing billions, to a whopping 100 million H100-equivalent clusters powered by 100 gigawatts of power by 2030, costing as much as $1 trillion apiece. Yes, that's "trillion" with a "T."

Thus, it was a huge surprise to see Aschenbrenner actually go short the foundational leaders in chipmaking today -- Nvidia (NASDAQ: NVDA), the leader in AI graphics processing units (GPUs), Broadcom (NASDAQ: AVGO), which helps large cloud companies make their custom ASICs and is a leader in networking chips, and Taiwan Semiconductor (NYSE: TSM), the world's dominant semiconductor foundry for leading-edge chips.

While 13F filings don't require the disclosure of short positions, they do show whether a fund is long put options, which is another way of being short a stock. In the third quarter, Situational Awareness (SA) bought put options on all three stocks, with the notional value of the puts on Nvidia equating to 6.95% of the portfolio, Broadcom puts equating to 1.77%, and Taiwan Semiconductor (TSMC) puts at 1.76%. While not SA's largest positions, they are still notable, with the Nvidia put position being the sixth-largest overall in the portfolio as of the end of the third quarter.

It's also notable that SA turned bearish on Broadcom. In fact, in the second quarter, Broadcom was SA's third-largest long position, at 11.7% of the portfolio. So, Aschenbrenner flipped from bullish to bearish on Broadcom in a very short period of time.

SA maintains its leveraged bet on Intel

Meanwhile, SA maintained its largest long position from the second quarter, comprised of Intel (NASDAQ: INTC) call options, with a notional amount accounting for 16.41% of the SA portfolio. This was a position that Aschenbrenner first took in the first quarter of this year, with call options essentially making this a leveraged long bet on Intel stock.

Being long Intel and short Nvidia, Taiwan Semiconductor, and Broadcom is a very out-of-consensus position. So, what could be the explanation?

One reason could be Lip-Bu Tan becoming CEO of Intel on March 12. Tan has a long track record of success in his career in the semiconductor industry, so the Intel position could be a bet on Tan's ability to lead Intel to a turnaround.

Another explanation could be Aschenbrenner's belief in Intel's new 18A node being successful. This node was the endpoint of former CEO Pat Gelsinger's "five nodes in four years" plan he initiated in 2021 and is just entering high-volume manufacturing. It was also the node at which Intel was supposed to catch up with or surpass TSMC in process technology. So, Aschenbrenner may think 18A will work, potentially upending the current paradigm of TSMC's dominance.

Finally, the bets could be mere valuation calls, with Intel having a tiny market cap compared with the others. Nvidia's, TSMC's, and Broadcom's stocks also all had big third quarters. So, it's possible that Aschenbrenner merely sees these stocks pulling back from their current high valuations in the short term.

A fund to track

Whether Aschenbrenner continues these bets going forward is an open question. It appears that his thinking is somewhat fluid. Even though Aschenbrenner is a technologist, SA also appears to be valuation-sensitive, as evidenced by his flip on Broadcom.

With Aschenbrenner having become an AGI thought leader, and given his recent contrarian bets, Situational Awareness is a new fund all AI investors should track going forward.

Billy Duberstein and/or his clients have positions in Broadcom, Intel, and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Intel, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom and recommends the following options: short November 2025 $21 puts on Intel. The Motley Fool has a disclosure policy.