AWS, Microsoft Lease Miner Power for AI Data Centers
AWS and Microsoft signed multibillion-dollar leases to convert powered Bitcoin-mining campuses into AI data centers: Cipher’s 15-year, $5.5B for 300 MW; IREN/Microsoft ~$9.7B five-year for 200 MW at Childress.
Amazon Web Services and Microsoft have signed separate multibillion-dollar contracts to repurpose powered Bitcoin-mining campuses into AI data centers. Cipher Mining filed a business update for a roughly $5.5 billion, 15-year lease to provide 300 megawatts of turnkey space and power to AWS, with delivery beginning July 2026. IREN reached an approximately $9.7 billion, five-year GPU-cloud contract with Microsoft to support 200 megawatts of critical IT load at its 750 MW Childress, Texas campus, with NVIDIA GB300 GPUs to be deployed through 2026.
Both agreements rely on infrastructure miners already built: land, grid interconnections, substations and long-term power rights. Those elements supply the physical capacity hyperscalers need and can reduce the time required to bring large AI data-center capacity online.
Analyses from market participants and research firms show a gap between revenue from Bitcoin mining and revenue available from AI hosting. A Fidelity analysis estimated that fleets operating around 20 joules per terahash would need a hash-price level of roughly $60 to $70 per petahash per day to match typical AI-hosting economics. Hashrate Index data from late May put the US-dollar hash price near $35.88 per PH/day. CoinShares reported that public miners’ AI and high-performance-computing contracts had surpassed $70 billion by early 2026, and some listed miners were on pace to derive up to 70% of revenue from AI this year, up from about 30% previously.
AI data-center construction is more capital intensive than Bitcoin-mining facilities. Industry estimates place AI build costs between $8 million and $15 million per megawatt, while Bitcoin-mining infrastructure has generally ranged from $700,000 to $1 million per megawatt. Shifting a powered campus to AI hosting therefore changes operators’ capital requirements, debt structure and execution timelines.
Bitcoin’s protocol includes an automatic difficulty adjustment that lowers the computational work needed to mine blocks when overall network hash rate falls. Recent network indicators show sustained hash-rate drawdowns: 30-day hash-rate momentum was near the 16th percentile and 90-day momentum near the 9th percentile, with reported network difficulty around 136.61 trillion and a 90-day decline of about 5.4%. Modeling of exits shows that a 20% reduction in global hash rate would raise surviving miners’ hash price to roughly $44.85 per PH/day, and a 30% reduction would lift it to about $51.26 per PH/day, figures that remain below the Fidelity AI-hosting crossover range unless Bitcoin price or fees move materially.
Leases that commit power for multiple years prevent immediate conversion back to ASIC Bitcoin mining. The AWS agreement spans 15 years and the Microsoft GPU contract covers five years; one of the contracts contains a delivery-timeline clause that could allow termination if milestones are missed. Operators carrying heavy debt and expecting revenue from AI hosting face execution and timing risk tied to those delivery schedules.
Analyses project a market separation between companies that lease powered campuses to hyperscalers and companies that continue to operate mining fleets at lower-cost, more flexible or stranded-energy sites where AI data centers are less viable. Modeling also shows how network exits affect the Bitcoin price required for mining to reach the $60–$70 per PH/day threshold: under current hash-rate conditions, reaching that hash-price range would imply Bitcoin prices in the roughly $122,000 to $142,000 range; a 20% network exit would lower the implied thresholds to about $98,000 to $114,000, and a 30% exit would lower them to about $86,000 to $100,000.
The AWS and Microsoft contracts provide concrete price reference points for powered, permitted data-center campuses and set multi-year commitments for significant blocks of grid-connected capacity.
Content on BlockPort is provided for informational purposes only and does not constitute financial guidance.
We strive to ensure the accuracy and relevance of the information we share, but we do not guarantee that all content is complete, error-free, or up to date. BlockPort disclaims any liability for losses, mistakes, or actions taken based on the material found on this site.
Always conduct your own research before making financial decisions and consider consulting with a licensed advisor.
For further details, please review our Terms of Use, Privacy Policy, and Disclaimer.








