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

Former OpenAI researcher Leopold Aschenbrenner doubled $225 million into $5.5 billion through his fund. He pointed out that the bottleneck in AI development lies in electricity—not chips or models.

Former OpenAI security researcher, the German man Leopold Aschenbrenner, only 24 years old, turned the $225 million in the fund he managed, Situational Awareness, into $5.5 billion in less than a year. While Wall Street money surged toward AI models and chipmakers, he correctly identified a blind spot the market was ignoring: electricity. By betting precisely on the infrastructure needed to solve AI’s power-hungry problem, he generated astonishing investment returns.

After leaving OpenAI, he pivoted into the AI investment market

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

This week, Bloom Energy announced that it had signed a major fuel cell deal worth 2.8 GW with Oracle, sending the stock soaring 15% after-hours. The paper value of Aschenbrenner’s stake also jumped instantly to nearly $2 billion.

Portfolio fully revealed: Go long on infrastructure, short traditional IT industries

Reports say his investments followed the same logic of “electricity first”:

• Bloom Energy (BE): Invested $875 million to buy into this fuel cell company. The technology allows data centers to generate power on-site directly, without relying on the aging power grid. Benefiting from the 2.8 GW deal signed with Oracle, the stock surged, and the paper value of its holdings has already climbed to nearly $2 billion.
• CoreWeave (CRWV): Invested $700 million in this top AI cloud computing power provider, locking in its scarce infrastructure resources.
• Infosys (INFY): He heavily shorted this Indian IT outsourcing giant. He expects that AI coding agents will completely destroy traditional IT outsourcing business.
• Intel (INTC): Used leveraged trades through Intel call options, earning multiples of returns during the period when the stock price rebounded strongly by 53%.
• Core Scientific (CORZ): Holds 10% equity. This former Bitcoin mining company is now transforming its existing power facilities into AI data center colocation sites.

The power-hungry monster behind computing: Electricity consumption keeps doubling year after year

Aschenbrenner said that looking back at 2022, the computing cluster used to train GPT-4 consumed about 10 million watts (MW) of electricity, costing roughly $500 million. However, AI computing demand is expanding at a speed of about half a quantity grade per year—meaning that the power demand of the largest training clusters will double every 12 to 18 months.

By 2024, the electricity consumption of the largest computing clusters has reached 100 MW, equivalent to 100,000 high-end graphics processing units (GPU) operating at the same time. Now in 2026, the leading training clusters require up to 1 GW of continuous power, equivalent to the electricity generation output of a large nuclear power plant.

Who controls power controls the future of AI

He estimated that by 2028, electricity consumption for AI training will soar to 10 GW, larger than the total generation capacity of an entire number of U.S. states. By 2030, it will reach 100 GW—consuming a full 20% of the United States’ current total electricity generation. This is still only the electricity used for “training” models; if you add the “inference” computing power actually used by the public, the electricity usage becomes even harder to imagine.

However, over the past decade, the United States’ total electricity generation has only grown slightly by 5%. Now reports from around the country about major transformer shortages and data centers that can’t be built are proof that the power grid can’t support the demand. This is also why he dares to heavily bet on Bloom Energy: the real bottleneck in AI development is not chips or software at all, but whether humans can produce enough electricity.

• This article is reprinted with authorization from: 《Chain News》
• Original title: 《24-Year-Old Fund Manager’s Annual Return 24x! AI Portfolio Targets “the Most Scarce Resource”》
• Original author: Co2