OpenAI projects it will spend approximately $856 billion on computing infrastructure and utilize around $278 billion in cash between 2026 and 2030 as the rapid expansion of artificial intelli
OpenAI projects it will spend approximately $856 billion on computing infrastructure and utilize around $278 billion in cash between 2026 and 2030 as the rapid expansion of artificial intelligence accelerates. This growth is fueling major shifts in the data center industry, especially regarding electricity, land, and grid infrastructure—areas in which Bitcoin miners have long held assets and expertise.
The rising demand for artificial intelligence systems has increased the value of existing mining sites. Now, AI projects not only compete with cryptocurrency miners for scarce electricity resources but also drive up the worth of power infrastructure controlled by those miners. This trend highlights a deeper competition over access to physical and energy assets, not just digital ones.
OpenAI launched its Stargate project in January 2025, targeting the deployment of 10 GW of AI infrastructure across the United States by 2029. OpenAI announced in April it had already surpassed this target, adding 3 GW of capacity in just 90 days. Building these facilities requires far more than installing servers. OpenAI emphasized the challenges of securing sufficient energy, suitable land, grid connectivity, permits, construction partners, and skilled workers. In partnership with SoftBank, OpenAI invested $500 million in SB Energy and selected the company for a $1.2 GW data center to be located in Milam County, Texas.
Gartner forecasts that global data center electricity use will reach 565 TWh in 2026, reflecting a 26% rise from 2025. Servers optimized for AI are expected to account for about 31% of this consumption. The International Energy Agency (IEA) predicts usage could double to nearly 945 TWh by 2030, warning that grid limitations could delay up to 20% of new data center projects. VanEck has noted that electricity, more than chips, is now the main limiting factor, while CoinShares identifies land, power, and building shells as critical bottlenecks.
Mini dictionary: IEA (International Energy Agency), an autonomous intergovernmental organization focused on ensuring reliable, affordable, and clean energy for its member countries.
CoinShares, a digital asset investment firm, recently acquired three leased AI facilities in Northern Virginia valued at approximately $3.5 billion, reflecting an average stable infrastructure price of about $27 million per megawatt. In contrast, some Bitcoin miners with energized, but not leased, capacity are valued at $3 million per megawatt. CoinShares linked this difference to difficulty in obtaining new grid connections, citing a US interconnection queue of approximately 2,600 GW. Tighter regulations are also affecting some regions; for example, New York imposed the first statewide ban on constructing new hyperscale data centers in July. As a result, miners with existing grid access see their connections increase in value due to lengthy delays for new entrants.
Facility Type Valuation per MW AI Facilities (leased, Northern Virginia) $27 million Bitcoin Miner (energized, not leased) $3 millionCoinShares reported that firms with contracted AI or high-performance computing (HPC) capacity are trading at an average enterprise value to sales multiple of 12.9, compared to just 3.7 for miners without such contracts. According to S&P Global, many listed mining companies have begun reallocating their resources from Bitcoin mining to AI and HPC workloads.
This disparity has influenced capital decisions at major mining firms. Core Scientific paid $41.9 million to cancel about 15 EH/s of next-generation mining equipment. Keel, previously known as Bitfarms, ceased mining activities on June 29, signaling no mining revenue expected in the third quarter. Cipher Digital is slated to exit mining by the end of 2027, while IREN targets finishing its transition by the end of 2026.
CoinShares estimates that retrofitting mining operations for AI infrastructure comes at a cost of $8 million to $15 million per megawatt, much higher than the $0.7 million to $1 million per megawatt for standard mining. Despite over $100 billion in disclosed AI/HPC contract backlog, actual annualized revenue is only about $1.1 billion, with roughly 550 MW in use compared to more than 4 GW contracted. Operators able to finance retrofits, attract reliable tenants, and bring capacity online swiftly are best positioned for future gains.
Crusoe, a private technology company, initially powered modular Bitcoin-mining data centers using flared natural gas, reducing waste and emissions. The firm sold its mining business to NYDIG in 2025 and shifted to AI infrastructure, reporting over $140 billion in contracted value and 6 GW of gross capacity. Robinhood Ventures Fund I invested $25 million in Crusoe’s $3.9 billion Series F financing round, resulting in a $30.9 billion post-money valuation for the company. This shift illustrates how infrastructure built for cryptocurrency mining is now being repriced as the AI boom places a premium on energy and grid access.
OpenAI’s growing infrastructure demands, combined with new regulatory challenges and soaring energy use, are compelling many Bitcoin miners to redirect capital and resources toward AI and high-performance computing projects, fundamentally reshaping the data center landscape.
The post OpenAI to spend $856 billion on AI infrastructure, Bitcoin miners switch to AI appeared first on COINTURK NEWS.
