Bitcoin Mining Operations Redirect Power Resources to Artificial Intelligence Infrastructure
Public Bitcoin miners reduced active mining capacity by approximately 75 exahashes per second during the first half of 2026, with operators reallocating electrical resources to artificial intelligence operations. The displaced mining equipment represented an estimated $1.5 billion in hardware value, while 12 companies recorded over $1.1 billion in related asset markdowns. High-performance computing and AI revenue from mining operators surged 52% quarter-over-quarter during this period.
The shift reflects a fundamental reallocation of computational resources driven by market economics. During 2024 and early 2025, mining operators had invested heavily in expansion, with over $3 billion spent on equipment alone and nearly $5 billion across the full infrastructure cycle. This capacity buildup pushed Bitcoin's network computing power into the zetahash range by 2025. However, within months, the relative profitability calculations shifted, prompting operators to redeploy their most valuable asset—electrical capacity—toward higher-margin AI and high-performance computing services, which grew 52% quarter-over-quarter in the first half of 2026.
The financial impact appears substantial in accounting terms, with IREN and Core Scientific together reporting nearly $900 million in asset write-downs. However, these markdowns represent adjustments to recorded asset values rather than direct cash losses, and they do not capture secondary revenue from equipment sales or redeployment. The Black Pearl facility example illustrates the complexity: its machines generated $57.9 million in operating revenue during 2025 before a $96.1 million markdown upon conversion, demonstrating that asset value reductions do not directly measure profitability.
This transition may reshape infrastructure investment patterns in energy-intensive computing sectors. Mining operators and investors could face heightened uncertainty about long-term asset valuations if profitability shifts rapidly between applications. Electricity providers and grid operators in regions hosting these facilities may experience demand volatility. Conversely, the reallocation could accelerate AI infrastructure buildout and potentially improve operational efficiency by consolidating high-compute operations within existing facilities, though this concentration may influence regional energy markets and supply chains.