AI wants Bitcoin miners’ power, but the contracts reveal who really gets paid
Bitcoin miners spent years securing cheap electricity, building substations, and negotiating the grid connections needed to power their operations. Now AI companies want those connections so badly they’re willing to make aging mines valuable real estate.
The ASICs that mine Bitcoin are practically useless for training AI models, but the land, substations, and power agreements around them could spare data-center developers years of waiting.
SWI Group took control of Bitcoin miner Genesis Digital Assets earlier this year and announced an €800 million capital increase on Oct. 12 to expand its AI infrastructure business.
Two weeks later, it launched SWI Compute, disclosing roughly $1 billion in NVIDIA equipment orders and its first customers, with some contracts still being finalized.
As part of the fundraising, Genesis is becoming part of a much bigger bet on selling computing power rather than mining BTC.
It’s an irresistible trade on paper: take an industrial property built to consume enormous amounts of electricity and give that electricity to customers willing to pay considerably more for it.
Unfortunately, you can’t turn Bitcoin mining facilities into AI data centers just by plugging different computers into the wall. The conversion requires extensive retrofits, cooling that can handle much denser hardware, sophisticated networking, construction financing, and customers willing to commit to the finished product.
Until the facility is operational, the promised revenue is just that: a promise.
Core Scientific’s agreements with AI cloud provider CoreWeave show where the money goes. CoreWeave helps fund construction but recovers some of that contribution through rent credits, while Core Scientific borrows billions, builds the facilities, and risks penalties for late delivery.
That’s how a mining property can become much more valuable without its owner receiving anything close to that value: electricity draws the customers in, but the contracts determine who gets paid.
Bitcoin mining equipment is losing value
Bitcoin mining is an unforgiving business. Operators run specialized computers known as ASICs, competing for block rewards whose value depends on Bitcoin’s price and the machines’ efficiency relative to the rest of the network.
They require a lot of power to run, regardless of the market, which means companies face the same bills in bull and bear markets. The first wave of public miners were seen as leveraged Bitcoin bets institutions could easily make when access to BTC was hard or impossible.
But the market is a harsh mistress, and these companies now look nothing like leveraged Bitcoin bets. Increased competition and the insatiable hunger for AI have pushed many of them toward becoming data center landlords.
Core Scientific’s second-quarter filing puts a price on the old business losing its appeal.
During the first quarter, the company wrote down $266.5 million of mining equipment and infrastructure as Bitcoin prices, mining revenue per unit of computing power, and secondhand equipment values deteriorated.
Of that charge, $151.6 million related to mining machines and $114.9 million to mining infrastructure. Meanwhile, its newer AI colocation business was already producing revenue.
The company’s mining operation has been losing money at the gross-profit level while AI hosting earned a positive gross margin.
Mining hardware loses value as it ages, but a working grid connection doesn’t. In fact, it can become even more desirable because developers can’t just go out and buy one.
Utilities still need to assess transmission capacity, identify upgrades, and determine who pays for them. Having enough cash for GPUs doesn’t make those steps go away.
Electricity access became the miners’ real prize, but the backlog of proposed projects also makes that access incredibly difficult to value.
In Texas, regulators approved ERCOT’s Batch Zero review process in June, while the state ordered an audit of approximately 474 GW of proposed large-load connections.
Electricity already flowing to a facility differs from capacity approved for future use, and both differ from a place in the connection queue. Grouping them under one headline megawatt figure can make a portfolio appear much closer to generating AI revenue than it is.
SWI’s June announcement about Genesis described land with more than 1.3 GW of energized and approved connections across 15 facilities. By September, its half-year results described approximately 1.2 GW of secured grid connections.
The disclosures don’t reconcile the two figures or provide a site-by-site breakdown separating currently usable capacity from power that requires additional work.
SWI also didn’t simply pay a fixed price for those megawatts. Its initial $500 million investment bought about 38.3% of Genesis’s equity and rights to 77.2% of the value of certain preferred-share liquidation preferences, which govern claims on proceeds rather than ordinary ownership.
SWI subsequently took control, but dividing that first payment by 1.3 GW would tell us little about what it paid for deliverable power.
Even an agreed valuation of the whole company wouldn’t tell us what an operating AI data center costs. Genesis came with a mining business, equipment, land, existing electrical assets, and all kinds of rights to future power.
However, those are just ingredients, and they require a lot of money before they can actually become a profitable AI business.
Turning a Bitcoin mine into an AI data center is an expensive conversion
While Bitcoin mining facilities and data centers might look similar to the untrained eye, they’re two very different businesses.
Bitcoin mines are built to run thousands of independent machines that can be switched off when necessary, while AI data centers need sophisticated cooling, high-speed networking, redundant power systems, and much stricter uptime guarantees to keep densely packed GPU clusters operating reliably. So acquiring an existing facility is just a head start.
Data centers consume electricity beyond what their servers use, particularly for cooling and power distribution. At a power usage effectiveness ratio of 1.25, a site drawing 100 MW from its electrical supply would deliver about 80 MW to computing equipment and use the other 20 MW to support the facility.
Comparing that site’s 100 MW grid connection with a competitor’s 100 MW of usable IT capacity would make two very different properties seem equivalent.
Nvidia’s GB300 NVL72 reference architecture describes liquid-cooled racks with 72 GPUs that draw up to 142 kilowatts each.
Delivering the electricity is only one part of supporting them, since cooling, networking, and building design must also meet standards most Bitcoin mines weren’t built for.
CoinShares estimated in its September mining report that converting mining infrastructure to AI-grade facilities costs roughly $8 million to $15 million per MW, against $700,000 to $1 million per MW for conventional mining infrastructure.
The order of magnitude explains why owning a powered mining site and owning a completed AI facility command such different valuations.
A review of more than $100 billion in AI deals found that only around 550 MW was billing out of more than 4 GW under contract. HIVE’s $350 million AI agreement required an estimated $185 million hardware buildout before most of that revenue could arrive.
Under colocation arrangements, the operator generally provides the buildings, electricity infrastructure, cooling, and other services, while the customer supplies its own computing equipment.
Selling AI computing capacity directly means buying GPUs as well, then finding enough customers to keep them occupied while newer, more efficient hardware makes existing systems less competitive.
Core Scientific earns fees for hosting other companies’ machines, while SWI intends to sell computing itself. Both stand to benefit from former mining infrastructure, but their capital requirements and exposure to hardware costs differ greatly.
Core Scientific’s contracts show how the AI revenue gets divided
We can get a pretty good idea of the math behind these deals from Core Scientific, because some of its converted facilities are already operating and its agreements with CoreWeave spell out where the money goes.
It reported $136.7 million of colocation revenue in the second quarter, with 437 MW billing customers by mid-July.
Its quarterly filing attributes $98.8 million to license fees and $35.1 million to power charges passed through to customers, with the balance coming from other services. The electricity reimbursement also appears as an equivalent expense, so it adds to reported revenue without creating gross profit.
More than a quarter of the colocation revenue therefore represents a utility bill being passed along, rather than money Core Scientific earns for providing the facility.
Inside Core Scientific’s $136.7 million AI colocation quarter
Quarter ended June 30, 2026. Gross profit, non-power colocation costs and electricity reimbursement add to reported revenue. Gross profit is before interest, corporate expenses, taxes and capital spending. Values shown in dollars.
CryptoSlateUnder terms disclosed in April, most of the CoreWeave sites start at a base license fee of $100 per kilowatt per month, subject to 3.5% annual increases, with CoreWeave responsible for utility costs.
CoreWeave also funds or reimburses certain approved construction expenses and, for specified landlord-owned equipment, receives credits against future base fees worth up to $1.5 million per MW. Those credits can cut the relevant monthly rent payment in half until the allowance is exhausted.
Take one megawatt at the initial $100 rate. Its contractual base fee is $100,000 a month, but a 50% construction credit reduces the payment to $50,000.
If the full $1.5 million allowance applies, using $50,000 in credits each month would take 30 months to recover it, before allowing for rent increases and the contract’s other provisions. The customer essentially pays for the data center to be built, then gradually gets that money back through lower payments to the company that owns it.
That kind of customer funding helps Core Scientific avoid raising every construction dollar itself. But the nominal rent in a 12-year agreement doesn’t necessarily arrive in the early years, and accounting revenue isn’t always cash collected. Its filing describes prepayments, construction credits, and straight-line revenue recognition as reasons the figures differ.
The agreements provide for daily late-delivery credits of $20,000 and then $30,000 in specified circumstances, and some delays can give CoreWeave a right to terminate an order. One ground lease depends on the underlying power agreement staying in force, leaving electricity access central to the property’s value even after millions have been spent on construction.
The April disclosures also describe an uncapped completion guarantee from Core Scientific for the specified financed projects if available construction funds prove insufficient.
The financial commitment is already enormous: during the first half of 2026, Core Scientific recorded $954.2 million in capital expenditure, including $180.9 million funded by CoreWeave.
In May, a financing subsidiary also issued $3.3 billion of secured notes at 7.75%, backed by assets and rights associated with specified data-center projects. That interest rate works out to approximately $255.8 million in annual coupon costs on the full original principal before repayments.
Principal installments are structured to begin only after the relevant rent commencement and tenant credits have run their course, but the interest obligations don’t just disappear while those credits are being used.
CoreWeave gets computing facilities without developing all the real estate itself, while recovering certain construction contributions through lower rent. Lenders get interest backed by project assets and expected customer payments.
Core Scientific retains ownership and the potential for a long-term income stream, but also the responsibility for building and financing the facilities. None of that division is obvious from the advertised AI contract value.
All of this has led public markets to begin pricing that financing risk more aggressively. On Oct. 22, Reuters reported that short sellers held approximately $6 billion in unrealized gains across five AI infrastructure companies in October, including companies such as CoreWeave and Core Scientific that were attracting heavy short interest.
Another Reuters report on Oct. 26 described premarket declines of 4.5% for CoreWeave and 1.7% for Core Scientific after Elon Musk warned of lenders’ reluctance to fund specialized AI cloud operators.
While two days of price volatility isn’t the best way to tell whether a business can repay its debt, it does show how quickly enthusiasm for AI contracts can collide with anxiety about the financing behind them.
Debt investors are lending against facilities that must be completed, power that must be delivered, and customers whose payments have to support borrowing. Cipher’s recent $850 million financing shows another way miners can finance the transition while exposing shareholders to new debt obligations or dilution.
SWI’s Genesis properties are much earlier in that process, although the company is assembling the pieces of a broader computing business pretty quickly. SWI reported €631.6 million in first-half profit, largely reflecting the value recognized on its Genesis investment.
On Oct. 26, SWI Compute announced approximately $1 billion in NVIDIA systems ordered and its first customers, while SWI separately announced a 13 MW infrastructure partnership in Spain.
The announcements support a strategy that goes beyond renting space and electricity: SWI wants to own more of the computing operation itself. That gives it a chance to earn revenue from the GPUs as well as the facilities, but requires investment in both layers of the business and creates exposure to equipment utilization and obsolescence.
SWI hasn’t disclosed which Genesis sites will host that equipment, how much of its reported 1.2 GW it can convert economically, or when converted sites might earn AI revenue. Nor has it earmarked a disclosed portion of the €800 million capital increase for Genesis.
Equity can dilute existing shareholders, while debt gives lenders a claim on future cash. Either way, the financing terms affect what shareholders keep.
Bitcoin miners really do have something AI developers desperately want, and existing power connections could be worth far more in a new use than in the mining operations for which they were secured.
The mistake here is treating that potential value as if it automatically belonged to the miner’s shareholders.
Between the grid connection and the final customer payment are construction companies, hardware suppliers, tenants, and creditors, each with its own contract and claim on the economics.
Core Scientific’s filings show what those claims look like once an AI site is operating. SWI has shown how much capital investors are prepared to commit, but not how much of its acquired mining infrastructure can actually earn a return.
Owning the largest power portfolio won’t decide who comes out ahead. The real money will go to companies that can get those megawatts working without giving most of the proceeds to the customers, contractors, and lenders needed to make it happen.
