Bitcoin mining is undergoing a fundamental shift from a cyclical business tied to cryptocurrency prices toward a stable infrastructure play anchored by long-term leases. CleanSpark, a major Bitcoin mining operator, just signed a 20-year lease agreement worth $6.6 billion in base contracted revenue for its Sandersville, Georgia data center campus, with potential total value reaching $11.6 billion when including two optional 5-year extensions. The deal signals a new era where power capacity, land control, and grid access matter more than raw computing hash rate, the measure of mining processing power.

What Is CleanSpark Actually Building?

CleanSpark signed the lease with an unnamed investment-grade global technology company that will deploy production-grade infrastructure for artificial intelligence (AI) and high-performance computing (HPC) workloads. The Sandersville facility is expected to provide 175 megawatts of critical IT load for AI and HPC applications. The first data hall should be ready in the fourth quarter of 2027, with remaining halls ramping up in early 2028. Until that power transfers to the new tenant, CleanSpark plans to continue mining Bitcoin at the site, meaning this is a staged conversion rather than an immediate shutdown.

The deal structure matters significantly. In a triple-net lease arrangement, the tenant covers operating expenses, property taxes, insurance, and maintenance, while CleanSpark retains ownership and control of the physical infrastructure including land, power access, building shell, and mechanical and electrical systems. Management has projected approximately $330 million in average annual net operating income once Sandersville is fully ramped, with near-100 percent NOI contribution margins tied to the lease structure. This transforms CleanSpark's valuation profile from a cyclical Bitcoin miner into a digital infrastructure landlord with predictable cash flow visibility.

Why Are AI Companies Interested in Mining Infrastructure?

The convergence of AI infrastructure demand with physical constraints creates a natural fit between mining operations and AI data centers. Large training and inference workloads require dense power, cooling capacity, and long construction lead times. Mining campuses have already solved much of that problem by securing industrial-scale energy access and navigating permitting challenges.

Mining sites bring several advantages to AI operators. They offer large power positions that may already have utility agreements or development work in place, industrial land where energy-intensive operations are more likely to be permitted, electrical infrastructure that can reduce development timelines, and operational experience managing high-load compute environments. The scarce resource for AI companies is not only servers; it is energized land with a path to utility-scale power. AI companies can purchase GPUs and processors, but they cannot instantly create grid interconnections, permitting history, substations, and local operating relationships.

However, the conversion is not straightforward. Bitcoin mining operations and AI data centers have different operating requirements. ASIC mining fleets, which are specialized computers designed to solve Bitcoin's mathematical puzzles, can tolerate aggressive air cooling, higher intake temperatures, and curtailment windows. AI training clusters require tighter uptime planning, denser rack design, fiber diversity, better fire suppression, and cooling systems that can handle accelerator-heavy racks. One network fabric issue can leave expensive GPUs idle while power still burns, creating a very different operating profile than mining.

How Are Bitcoin Miners Adapting to Post-Halving Economics?

Bitcoin's 2024 halving reduced the block subsidy from 6.25 BTC to 3.125 BTC, cutting mining rewards in half. Miners with high energy costs or older ASIC equipment face significantly tighter profit margins unless Bitcoin price, transaction fees, or efficiency gains offset the reduction. The CleanSpark deal demonstrates an alternative path: converting power assets into contracted infrastructure revenue that does not depend on Bitcoin's price or network difficulty.

This strategy does not apply universally. Some mining sites are too remote, too unreliable, or too poorly connected for AI workloads. A low-cost mining site in a weak fiber location may still be better operated as a traditional mine. However, for operators with multi-hundred-megawatt power footprints in strategic locations, the infrastructure lease model offers more stable returns than mining alone.

CleanSpark's Texas portfolio adds another dimension to this shift. The same tenant signed a letter of intent and exclusivity arrangement covering CleanSpark's Texas operations: 718 acres and up to 885 megawatts of secured and planned power capacity. That is larger than Sandersville by a significant margin. If it converts to a signed lease, it would reinforce the idea that miners with multi-hundred-megawatt power footprints are becoming strategic suppliers to AI infrastructure buyers.

How Will This Reshape the Mining Industry?

• Dual-Use Campus Design: Future mining sites are likely to be planned as dual-use campuses from day one, featuring modular data halls, better grid interconnects, stronger cooling plans, and layouts that can support both ASIC fleets and GPU-heavy workloads. When Bitcoin mining economics are attractive, operators can allocate power to miners. When AI tenants offer stronger risk-adjusted returns, power can move toward colocation or lease models.
• New Valuation Metrics: For years, public miners were compared by hash rate, fleet efficiency, Bitcoin holdings, and power cost. Those metrics still matter, but CleanSpark's deal shows that contracted net operating income, power capacity, tenant quality, and lease duration may become equally important to investors. A miner with 500 megawatts of well-located power and credible data center conversion plans may be valued differently from a miner with the same hash rate but weaker site control.
• Grid Planning Complexity: Sandersville's 175 megawatt load is significant, and the Texas exclusivity figure of 885 megawatts is even more striking. At that scale, these campuses become long-term anchors for regional electricity demand. Expect more scrutiny from utilities, regulators, and local governments. Communities will ask about grid reliability, jobs, water use, tax revenue, and curtailment behavior. Mining operators that want AI tenants will need to speak the language of power planning, not only Bitcoin economics.

The CleanSpark deal provides a working model for life after the latest Bitcoin halving. If block rewards tighten margins, a miner with large power positions can still build a business around long-duration digital infrastructure. That represents a serious shift in how the industry thinks about its core assets and long-term viability.