Researcher Fired from OpenAI Bets $1 Billion on Bitcoin Mining, AI Computing Power and Electricity Assets Become Core Logic

March 6 News: A young researcher who was previously fired from OpenAI is now becoming a major investor in the Bitcoin mining industry. 24-year-old Leopold Aschenbrenner has heavily invested in Bitcoin mining companies through his hedge fund, Situational Awareness LP, totaling approximately $1 billion. According to the latest filings with the U.S. Securities and Exchange Commission (SEC), the fund’s total size is about $5.5 billion, with nearly one-fifth allocated to Bitcoin mining-related companies.

Public information shows that Aschenbrenner left OpenAI in 2024 amid allegations of information leaks. He had also previously participated in the FTX Future Fund project. Now, he is focusing his funds on several mining companies transitioning into artificial intelligence computing, including Core Scientific, Iris Energy, Cipher Mining, Riot Platforms, and Hut 8. These companies generally possess large-scale power resources and data center infrastructure and are considered key assets in the AI computing industry chain.

Industry insiders point out that the core logic behind this investment is not simply optimistic about Bitcoin prices but rather a bet on “the value of power infrastructure.” Nishant Sharma, founder of the mining and computing consulting firm Blocksbridge, stated that the most important assets for Bitcoin miners have always been energy supply capacity and grid access rights, not the amount of Bitcoin they can mine. Against the backdrop of rapidly growing AI computing demand, mining companies with several gigawatts of power resources are gaining new valuation logic.

Since the Bitcoin halving in 2024, miner revenues have declined significantly, with reduced block rewards and lower on-chain transaction fees impacting profitability. Meanwhile, many mining companies are expanding into AI and high-performance computing businesses, hoping to secure more stable income through computing power hosting. For example, Core Scientific has signed a 12-year partnership with AI cloud service provider CoreWeave, with total revenue expected to reach $10 billion. Iris Energy plans to increase its annualized AI cloud service revenue to over $500 million by 2026.

The AI industry’s demand for electricity is also rapidly increasing. It is estimated that training OpenAI’s GPT-4 model requires over 12 megawatts of power, equivalent to the electricity consumption of about 12,000 households. As next-generation models expand in scale, energy demand is expected to continue rising. However, in the U.S., connecting new data centers to the grid often takes 3 to 5 years, while mining companies already have ready power and site resources.

Therefore, in the context of intensified AI computing competition, the energy and infrastructure of Bitcoin mining farms are being revalued. Aschenbrenner’s investment strategy is based on this trend: as AI computing demand continues to grow, Bitcoin mining companies with power and data center resources may become the most direct beneficiaries.