Crypto mining infrastructure is no longer just about ASIC fleets, hashprice, and cheap power. CleanSpark's 20-year, 6.6 billion dollar AI data center lease at its Sandersville, Georgia campus shows how large Bitcoin miners are turning power rights, land, and grid engineering into long-term infrastructure revenue.

This is not a small pilot. It is a multi-decade lease with an unnamed high investment-grade global technology company, with potential contract value rising to about 11.6 billion dollars if two 5-year extension options are exercised. For miners, AI companies, cloud providers, and Web3 professionals, the signal is clear. The next stage of crypto mining infrastructure will look more like energy-backed data center development than pure Bitcoin production.

What CleanSpark Signed at Sandersville

CleanSpark, a publicly listed Bitcoin miner, has entered a 20-year infrastructure lease for its Sandersville data center campus in Georgia. The tenant is expected to deploy production-grade compute for AI and high-performance computing, while CleanSpark provides the site, power access, and data center shell economics.

The lease is structured as a triple-net arrangement. In plain terms, the tenant is generally responsible for operating costs such as maintenance, taxes, and insurance. That detail is easy to miss, but it changes the economics. CleanSpark is acting more like a landlord and infrastructure owner than a full-stack AI cloud operator.

Reported capacity figures differ. Some reports describe the Sandersville lease as covering 175 megawatts. Others state 250 MW, while analyst commentary refers to roughly 185 MW. That is not unusual in infrastructure reporting. Current load, committed load, and fully planned load often get mixed together in headlines. The practical point is simpler. Sandersville is a large, power-dense campus being repositioned for AI and HPC workloads.

Why the 6.6 Billion Dollar Lease Changes the Mining Model

Bitcoin mining revenue is exposed to hashprice, network difficulty, power cost, block subsidy changes, transaction fees, and Bitcoin's market price. Every miner knows the squeeze after a halving. A long-term AI infrastructure lease behaves differently.

Base contracted revenue: approximately 6.6 billion dollars over 20 years.
Potential value: around 11.6 billion dollars if both 5-year extensions are used.
Expected average annual NOI: about 330 million dollars, according to analyst commentary.
Margin profile: close to 100 percent NOI margins have been discussed because of the triple-net structure.

That last point is the kicker. If CleanSpark can collect infrastructure rent while the tenant carries most operating expenses, the cash flow profile becomes less tied to daily Bitcoin mining conditions. Mining can continue, but it is no longer the only way the site pays for itself.

To be blunt, this is not a miner chasing an AI headline. The valuable asset is the interconnection, power capacity, land control, and the ability to get a high-MW site ready. In real data center work, the queue position and substation plan can be worth more than the building shell. You can buy servers fast. You cannot always buy 200 MW of usable power on a useful timeline.

How CleanSpark Plans the Transition

CleanSpark expects the initial Sandersville data hall to be ready for service in Q4 2027, with additional halls ramping into early 2028. Management has indicated that the remaining halls are contemplated for completion in Q1 2028, pointing to a phased build and commissioning schedule.

The company also plans to keep mining Bitcoin at Sandersville until power is transferred to the AI and HPC data halls. That is a practical bridge. Idle megawatts are expensive. If the mining fleet can run while construction and commissioning continue, CleanSpark can preserve Bitcoin exposure during the transition.

Anyone who has worked around high-density compute knows this is not a plug-and-play swap. ASIC mining containers and GPU data halls may both consume huge power, but their requirements differ. A Bitcoin mining row can tolerate design choices that would make an AI cluster painful: less redundancy, simpler networking, and far less concern about east-west traffic between servers. A GPU training cluster needs serious fiber planning, switch capacity, liquid or advanced air cooling, and tighter uptime expectations. One wrong assumption about rack density can force a redesign of busway, cooling loops, and floor loading.

Texas Exclusivity Points to a Bigger Strategy

The Sandersville lease is only one part of the story. The tenant also secured exclusive rights over CleanSpark's Texas portfolio through a letter of intent and exclusivity arrangement. Reports describe the Texas assets as roughly 718 to 780 acres across multiple campuses, with up to 885 MW of secured and planned power capacity.

The portfolio includes the Sealy campus, around 271 acres with nearly 300 MW of capacity, and the Brazoria campus, roughly 447 acres with existing transmission-level infrastructure supporting 300 MW of initial load and potential expansion up to 600 MW.

No binding follow-on contract has been disclosed for Texas at the same level as Sandersville. Still, exclusivity matters. It suggests the tenant sees CleanSpark's broader footprint as a possible pipeline for future AI or HPC development.

Market Reaction: Investors Repriced the Story

The market noticed quickly. CleanSpark's shares rose more than 20 percent in pre-market trading after the announcement. During the session, the stock reportedly jumped as much as 22 percent before settling lower but still materially positive. One report noted a close up 8.82 percent at 13.45 dollars per share.

That reaction says something about investor expectations. Public miners have often traded as high-beta Bitcoin proxies. Deals like this invite a different valuation lens: contracted infrastructure revenue, creditworthy tenants, project finance, and multi-decade site control.

That does not remove risk. Build delays, permitting, debt costs, grid constraints, and tenant concentration all matter. But it does change the discussion from how many exahashes the miner can deploy this quarter to what long-term cash flow its energy platform can generate.

Why AI Companies Want Mining Sites

AI infrastructure demand is running into a hard constraint: power. Training and inference clusters need enormous electricity, dense cooling, and fast deployment timelines. Bitcoin miners already spent years solving parts of that puzzle.

Miners bring three scarce assets

Power access: large interconnections and utility relationships.
Land: campuses that can support phased expansion.
Energy operations experience: teams used to load management, uptime pressure, and power market volatility.

That does not mean every mining site can become an AI data center. Many cannot. If a site was built around low-cost curtailment with minimal redundancy, poor network access, or weak cooling options, it may stay a mining-only asset. AI tenants will not accept the same design tolerance as commodity mining workloads.

But the best sites, especially those with transmission-level infrastructure and room for purpose-built halls, are now attractive to hyperscalers, AI labs, and specialized HPC providers.

CleanSpark Is Not Abandoning Bitcoin Mining

CleanSpark CEO Matt Schultz has stated in media interviews that AI expansion will not replace cryptocurrency mining. The plan is coexistence. That matters for understanding where crypto mining infrastructure is headed.

Bitcoin mining stays useful because it is flexible load. Miners can ramp down during grid stress or unfavorable power prices. AI workloads are less flexible, especially training jobs with expensive GPUs sitting behind them. A mixed campus could use Bitcoin mining as a power-balancing tool while AI tenants anchor long-term revenue.

That is a strong model when done correctly. It is also more complex. You need better power management, contractual clarity, and engineering discipline. A mining company that only knows how to fill containers with ASICs will struggle. A miner that understands substations, demand response, liquid cooling, and project finance has a real chance.

Liquid Cooling and High-Density Compute Are Converging

CleanSpark's collaboration with Submer, a company focused on high-density and liquid-cooled data center designs, fits the same pattern. AI clusters are pushing rack densities higher. Modern GPU systems move far beyond the cooling profile of older enterprise racks. Bitcoin mining has also experimented with immersion cooling for ASIC efficiency and heat management.

This is where the mining and AI worlds start sharing design lessons. Immersion and liquid cooling can support denser deployments, but they add capex, maintenance skills, and vendor dependence. They are not automatically the right answer for every site. Use them when power density and thermal limits justify the complexity.

What This Means for Blockchain, AI, and Web3 Professionals

If you work in blockchain or AI, the CleanSpark deal is a useful case study. It sits at the intersection of energy, compute, finance, and infrastructure strategy.

For blockchain professionals: Bitcoin mining is becoming part of a wider digital infrastructure business, not a standalone technical niche.
For AI teams: access to power may matter as much as access to chips.
For enterprises: long-term compute planning should include energy sourcing, site risk, and data center density assumptions.
For developers: infrastructure knowledge is becoming a career advantage, even if you mostly write smart contracts or AI applications.

This is where structured learning helps. If you want to connect this topic with a formal path, look at Blockchain Council programs such as Certified Bitcoin Expert™, Certified Blockchain Expert™, Certified Blockchain Developer™, and Certified AI Expert™. The technical overlap is growing. Consensus networks, mining economics, AI infrastructure, energy-aware computing, and enterprise deployment are no longer separate conversations.

The Future of Crypto Mining Infrastructure

CleanSpark's 6.6 billion dollar Sandersville lease points to a future where the strongest miners become energy-backed digital infrastructure providers. Some revenue will come from Bitcoin. Some will come from AI and HPC tenants. Some may come from grid services, demand response, and specialized hosting.

The winners will not be the miners that simply add AI to an investor deck. The winners will be the operators with real power rights, credible build schedules, strong counterparties, and teams that understand both high-voltage infrastructure and high-density compute.

Your next step is practical: study the infrastructure layer, not just the asset price. If you are building a career in blockchain, start with Bitcoin mining economics and energy markets. If you are moving toward AI infrastructure, learn data center design constraints and high-density cooling. A program such as Blockchain Council's Certified Bitcoin Expert™ or Certified AI Expert™ gives you a structured path into the two fields now colliding fastest.