A 24-year-old fund manager's annual return is 24 times higher! His AI investment portfolio targets the "most scarce resource"

Former OpenAI researcher Leopold Aschenbrenner doubled $225 million to $5.5 billion through his fund. He noted that the bottleneck in AI development is electricity, not chips or models.

Former OpenAI safety researcher, German man Leopold Aschenbrenner, who is only 24 years old, turned his $225 million fund, Situational Awareness, into $5.5 billion in less than a year. While Wall Street money rushed into AI models and chipmakers, he zeroed in on a blind spot that the market overlooked: electricity. By betting precisely on the infrastructure foundations to solve AI’s high power consumption problem, he created astonishing investment returns.

Leaving OpenAI and then diving into the AI investment market

After leaving OpenAI, Aschenbrenner wrote a 165-page report, asserting that artificial general intelligence (AGI) will arrive sooner than everyone thinks, and that the final winners won’t be the companies with the strongest AI models, but the companies that “control electricity.” To that end, he founded the hedge fund “Situational Awareness LP” and poured $875 million into fuel cell company Bloom Energy.

This week, Bloom Energy announced that it signed a 2.8 billion-watt (GW) fuel cell deal with Oracle, sending the stock price soaring 15% after hours, and the paper value of Aschenbrenner’s stake also jumped to nearly $2 billion in an instant.

Portfolio fully revealed: Long infrastructure, short the traditional IT industry

Reports say his investments follow the “electricity first” logic:

• Bloom Energy (BE): Invested $875 million to buy into this fuel cell company. The technology enables data centers to generate power on-site directly, without relying on aging power grids. Benefiting from the 2.8 GW order signed with Oracle, the stock surged, and the paper value of his holdings has already climbed to nearly $2 billion.
• CoreWeave (CRWV): Invested $700 million in this leading AI cloud computing power-supply provider, targeting its scarce infrastructure resources.
• Infosys (INFY): Aggressively shorted this Indian IT outsourcing giant, expecting AI coding agents to completely destroy traditional IT outsourcing businesses.
• Intel (INTC): Used Intel call options to leverage bets, earning multiple times during the stock’s strong 53% rebound.
• Core Scientific (CORZ): Holds a 10% stake. This former Bitcoin mining company is pivoting to convert its existing power facilities into AI data center hosting sites.

The power-hungry monster behind computing: electricity usage doubles year after year

Aschenbrenner said that when looking back at 2022, training GPT-4’s computing cluster consumed about 10 megawatts (MW) of electricity, costing around $500 million. However, AI compute demand is expanding at roughly half a magnitude per year, which means the power demand of the largest training clusters will double every 12 to 18 months.

By 2024, the power consumption of the largest computing clusters has reached 100 MW, equivalent to running 100,000 high-end graphics processing units (GPUs) simultaneously. Now in 2026, leading training clusters require up to 1 GW of continuous power, which is equivalent to the power generation output of a large nuclear reactor.

Whoever controls electricity controls the future of AI

He estimates that by 2028, AI training power consumption will surge to 10 GW—bigger than the total power generation of many U.S. states; by 2030 it will reach 100 GW, consuming as much as 20% of the United States’ current total power generation. This is only for the electricity used to “train” models; if you add the “inference” computing power actually used for public deployment, the amount of electricity required becomes even more unimaginable.

However, over the past decade, total U.S. power generation has only grown marginally by 5%. Reports of massive transformer shortages and data centers that can’t get built are proof that the grid can’t keep up. This is also why he dares to heavily bet on Bloom Energy: the real bottleneck in AI development isn’t chips or software at all, but whether human beings can produce enough electricity.

• This article is reprinted with permission from 《Chain News》
• Original title: 《24-year-old fund manager’s annual return is 24 times! AI portfolio targets the “scarcest resource”》
• Original author: Co2