Why a 24-Year-Old AI Wunderkind Is Betting Big on Bitcoin Miners

If you’re not familiar with the name Leopold Aschenbrenner, you should be.
A 24-year-old wunderkind, Aschenbrenner was hired by OpenAI in 2023 to work on the company’s “superalignment” team, essentially trying to figure out how to keep AI systems safe once they become smarter than the humans building them. After being let go in 2024, he published a 165-page essay called Situational Awareness that went viral in Silicon Valley, Washington and on Wall Street.
His central argument in a nutshell: AI models could become capable of doing the work of AI researchers by around 2027. If that happens, AI begins improving itself, and the timeline to artificial general intelligence—or AGI—compresses dramatically.
Aschenbrenner also created a hedge fund, Situational Awareness LP, specifically to invest in the AGI growth trend. In its 13F filing for the first quarter, the company disclosed it held a respectable $13.7 billion in assets. That’s up from just $254 million at the end of 2024, a head-spinning 54x increase.

I’m happy to report that Aschenbrenner’s fund disclosed a purchase of nearly 3.4 million shares of HIVE Digital Technologies. As many of you know, I serve as executive chairman of HIVE, and on behalf of everyone at the company, I want to express my gratitude in Aschenbrenner and Situational Awareness’s conviction in the HIVE story.
The bigger story, though, is what’s driving Aschenbrenner’s thesis, and how quickly the rest of the world is catching up to it.
The AGI Consensus Is Building Fast
What’s changed since Aschenbrenner published Situational Awareness is that the voices agreeing with his outlook have only multiplied. And these aren’t fringe figures.
Marc Andreessen, co-founder of venture capital firm a16z and co-creator of some of the earliest web browsers, said he believes AGI is already here. On a recent airing of Joe Rogan’s podcast, he claimed that the top AI chatbot platforms (OpenAI, Claude, et al) now give him better answers on any topic than what world-class experts could give him.
Demis Hassabis, CEO of Google DeepMind, claimed at Google’s developer conference last month that humanity is “standing in the foothills of the singularity”—another word for the moment when AI surpasses human cognitive capacity. He now expects AGI to arrive in 2029.
Ambitious forecasts, maybe, but the financial data appears to support this breakneck growth. Microsoft’s AI business alone just surpassed an unbelievable $37 billion run rate, up over 120% year-over-year. Morningstar reports that AI-focused funds attracted over $16 billion in net inflows in 2025, nearly eight times the prior year. Despite broader market turbulence, flows remained strong in the first quarter of 2026.
The Semiconductor Boom Tells the Story
If you want a single indicator of how fast this industry is moving, just look at the chip sector.

The PHLX Semiconductor Index has climbed 82% so far in 2026, its best-ever performance through the first 100 trading days of any year. The previous record was set in 1995. Believe it or not, companies in the index have added roughly $5.7 trillion in market capitalization this year alone.
Last week, memory chipmakers Micron and SK Hynix both crossed the $1 trillion valuation mark. UBS raised its price target on Micron.
Power, Land and Infrastructure
Despite the breakneck momentum, Aschenbrenner’s fund is actually shorting chipmakers.
Instead, he’s going long on companies that own the electricity, data centers and physical infrastructure that AI requires to scale.
His largest holding is the VanEck Semiconductor ETF, but the filing also disclosed significant stakes in Bitcoin miners and infrastructure firms. Beside HIVE, you’ll find IREN, Core Scientific, Riot Platforms, CleanSpark and others.
Why? Because as Aschenbrenner wrote in Situational Awareness:
“The race to AGI won’t just play out in code and behind laptops—it’ll be a race to mobilize America’s industrial might.”
He’s not wrong. Global AI computing capacity is doubling every seven months, according to Epoch AI. Training clusters are on track to cost hundreds of billions of dollars individually by 2028, each requiring power equivalent to a small U.S. state.
Put another way, you can design all the chips you want, but without secured megawatts and physical sites, they have nowhere to run.
It takes roughly three years to build a data center from the ground up. But if you already have the infrastructure from Bitcoin mining, you can cut that to nine months.
That’s the advantage that Bitcoin miners such as HIVE bring to the table. We already control the power contracts, the substations, the cooling system and the land.
New Demand, Old Constraints
I’ve spent my career investing in commodities and natural resources, and I’ve learned that the biggest opportunities tend to emerge when a new source of demand collides with physical constraints. Gold, oil, copper—every great commodity cycle has followed this pattern.
AGI is no different, except the constrained resource this time is electricity and the infrastructure to deliver it.
As Aschenbrenner points out, the timeline to AGI is compressing. The capital flowing into the space is accelerating. And the people who understand the technology best—the builders, the researchers, the fund managers who staked their reputations on it—are placing their bets not on software, but on the physical infrastructure required to make it all real.

—

Originally Posted June 1, 2026 – Why a 24-Year-Old AI Wunderkind Is Betting Big on Bitcoin Miners

All opinions expressed and data provided are subject to change without notice. Some of these opinions may not be appropriate to every investor.

Holdings may change daily. Holdings are reported as of the most recent quarter-end. The following securities mentioned in the article were held by one or more accounts managed by U.S. Global Investors as of (03/31/2026): Micron Technology Inc.

The PHLX Semiconductor Sector Index (SOX) is a widely followed benchmark index that measures the performance of 30 top U.S.-listed companies engaged in the design, distribution, manufacture, and sale of semiconductors.