The Hidden Winner of the AI Power Grab: A Former OpenAI Researcher Bets $1 Billion on Bitcoin Miners
TechFlow Selected TechFlow Selected
The Hidden Winner of the AI Power Grab: A Former OpenAI Researcher Bets $1 Billion on Bitcoin Miners
He’s betting not on Bitcoin, but on AI-powered electricity.
Navigating Web3 tides with focused insightsAuthor: DLNews
Translation & Editing: TechFlow
TechFlow Intro: At age 24, fired by OpenAI, and managing a $5.5 billion hedge fund—Leopold Aschenbrenner has allocated nearly 20% of his portfolio to Bitcoin mining stocks. But his rationale isn’t bullish on Bitcoin’s price; rather, he sees in miners something AI companies desperately want but can’t access: ready-made, industrial-grade power infrastructure. While new data centers require three to five years to connect to the grid, miners’ grid interconnection rights are becoming more valuable than Bitcoin itself.
Full Article Below:
Leopold Aschenbrenner manages a $5.5 billion hedge fund.
This 24-year-old fund manager has allocated nearly 20% of his portfolio to Bitcoin miners.
Yet his bet is not on Bitcoin—it’s on the infrastructure and grid interconnection rights required by AI.
Bitcoin miners have just received an unexpected $1 billion vote of confidence—from a former OpenAI researcher.
Leopold Aschenbrenner, 24, was dismissed by OpenAI in 2024 over alleged information leaks. Through his $1 billion hedge fund, Situational Awareness LP, he has built a series of positions in the Bitcoin mining sector.
According to the fund’s latest filing with the SEC, Situational Awareness LP currently manages $5.5 billion, approximately $1 billion of which is invested in Bitcoin miners.
Aschenbrenner’s $1 billion bet ranks among the largest institutional investments in Bitcoin miners over the past several months. But analysts say this move signals that the industry’s true asset has never been Bitcoin—it’s electricity.
“A miner’s real value has always resided in its energy infrastructure and grid interconnection,” said Nishant Sharma, founder of mining and hashrate consulting firm Blocksbridge, speaking to DL News. “In today’s market, the valuation of underlying energy infrastructure often exceeds the value of the Bitcoin it might produce.”
As AI firms scramble for power capacity—and prominent mining stocks slump to multi-year lows—Aschenbrenner, formerly part of FTX Future Fund’s charitable team, sees enormous value in Bitcoin companies that already possess gigawatt-scale industrial power supply.
And Aschenbrenner’s timing couldn’t be better.
Following the 2024 Bitcoin halving—which cut block rewards in half—miners’ revenues have remained under pressure. Sparse on-chain activity has further worsened their situation, shrinking transaction fee income.
Miners have therefore pivoted toward AI—riding the wave of this trend—selling Bitcoin and abandoning the business model upon which they were founded.
Bitcoin miners’ shareholders now demand accelerated transitions into AI.
Aschenbrenner did not respond to interview requests.
Aschenbrenner’s portfolio reveals multiple substantial positions in the Bitcoin mining sector.
These include Core Scientific, Iris Energy, Cipher Mining, Riot Platforms, and Hut 8—Bitcoin miners aggressively transitioning into AI, collectively representing roughly $1 billion.
In essence, he is targeting Bitcoin miners that have taken concrete, substantive steps into the AI space.
Core Scientific has signed a 12-year contract with AI cloud provider CoreWeave, projected to generate $10 billion in revenue; Iris Energy aims to achieve over $500 million in annualized AI cloud services revenue by early 2026; and Riot recently shifted focus toward AI and high-performance computing, signing a 10-year data center lease agreement with AMD.
The economic logic behind this pivot is hard to ignore. If sustained, AI hosting will deliver predictable, stable revenue—whereas Bitcoin mining remains dependent on volatile cryptocurrency prices and brutal competition.
“Aschenbrenner’s bet makes perfect sense,” Sharma said.
The Battle for Power
AI faces a major problem: severe power shortages.
It’s reported that training OpenAI’s GPT-4—one of the more popular large language models on the market—consumes over 12 megawatts, equivalent to the electricity usage of roughly 12,000 households.
Future models will likely require even more.
Securing such power is extremely difficult. In the U.S., connecting new data centers to the grid typically takes three to five years—due to layers of environmental reviews, interconnection studies, transmission upgrades, and local permitting.
These timelines are prohibitively long. That’s where Bitcoin miners enter the picture.
“Precisely because traditional data center construction timelines are so protracted, miners’ ready-to-go, power-ready assets become extraordinarily valuable to an industry struggling to keep pace with surging demand,” Sharma said.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
@DLNewsInfo
