NextFin

Hyperscale Data Sells Its Last Bitcoin as $74.8 Million AI Funding Gap Emerges

Summarized by NextFin AI
  • Hyperscale Data sold its remaining 1,113 BTC for approximately $74.4 million, while holding only $17 million in unrestricted cash against $231 million of current liabilities.
  • The Michigan AI data center requires another $74.8 million; construction targets 10 MW in December and another 10 MW by March, supporting a potential long-term revenue stream exceeding $1.2 billion.
  • The proposed exchange offer covers up to 192.3 million shares for cash and preferred stock, but its approximately $3.85 million cash component cannot finance the construction gap.
  • Hyperscale’s AI pivot may be structurally valuable, but near-term outcomes depend on binding project financing, timely commissioning, and avoiding excessive dilution or senior preferred claims.

NextFin News - Hyperscale Data has sold its remaining Bitcoin reserve and warned it may need another $74.8 million to complete a Michigan AI data center, but its latest proposed capital measure would not itself provide that amount of project funding. The company intends to offer cash and preferred shares to repurchase up to 192.3 million common shares, leaving investors to weigh a long-term AI revenue opportunity against a near-term liquidity gap.

The company’s regulatory warning said it had $17 million of unrestricted cash against $231 million of current liabilities at Sept. 30 and needed additional funds between November and March. Hyperscale shares closed at $0.157 on Wednesday, up 1.75%, before the warning was filed after the market close.

Hyperscale expects the Michigan project to cost $122.1 million and said it had paid $47.3 million through October. It is targeting 10 megawatts of AI capacity in December and another 10 megawatts by the end of March. The company said that, as of Nov. 8, it had sold all of its remaining Bitcoin reserves at approximately $76,000 per coin, bringing total sales since late July to about 1,113 BTC for approximately $74.4 million.

The company describes a signed 20 MW master-services agreement with a California-based neocloud provider. It projects more than $1.2 billion of revenue if the initial 10-year term and two five-year extensions run to completion. A customer option to raise capacity to 52 MW could lift the company’s projection above $3 billion. Those figures depend on construction, performance, continued customer use and the exercise of the option.

Hyperscale’s proposed exchange offer is a separate and more limited tool. The company said it intended to commence an offer under which shareholders could exchange each 100 common shares for $2 in cash and one preferred share with a stated value of $24 and a 12% annual dividend. The proposal would cover up to 192,307,692 common shares, subject to a minimum tender and other conditions. If fully accepted on those stated terms, the cash component would be about $3.85 million, while the preferred stock’s stated value would not be the same as cash available for construction.

The financing question has therefore become more precise: can Hyperscale convert a conditional AI revenue stream into capital quickly enough, now that Bitcoin no longer provides a liquid reserve?

The AI Revenue Projection and the Cash Requirement Are on Different Clocks

The central mistake is to compare more than $1.2 billion of projected long-term revenue with $74.8 million of near-term project spending. The revenue projection stretches across as many as 20 years. The funding requirement is concentrated between November and March.

Hyperscale’s remaining project requirement is about 1.6 times its unrestricted cash at Sept. 30. Management said it was considering debt, equity, asset sales and other transactions. That ratio does not prove the company must fund the entire amount from cash, but it shows why financing execution now matters more than the headline customer value.

The transmission mechanism is front-loaded expenditure followed by delayed operating receipts. Hyperscale must pay for electrical work, equipment and commissioning before the first 10 MW can be energized. The customer arrangement can begin generating cash only after the facility is available and the customer accepts service.

Bitcoin was the company’s most liquid bridge asset. It could be sold immediately to reduce debt or fund construction. But each sale also removed optionality. After the final Bitcoin sale, Hyperscale no longer has a crypto reserve that can absorb another delay or be pledged against new financing.

The balance sheet compounds the timing risk. Unrestricted cash declined from $36.8 million at June-end to $17 million at September-end, while current liabilities reached $231 million. Current liabilities exceeded current assets by $128 million. Hyperscale said existing funds and financing would not cover its planned spending and obligations over the following 12 months and again disclosed substantial doubt about its ability to continue operating.

The operating base is not yet self-financing. Hyperscale reported a net loss of approximately $62 million for the first nine months of 2026 and first-half operating cash use of $9.9 million. The AI buildout is therefore being funded while the wider business remains cash-consuming.

“Our highest priority is to energize the first 10 megawatts of critical IT capacity at our Michigan data center in December,” Chief Executive Will Horne said in the company’s Nov. 10 release.

The priority is clear. The funding source is not.

The Exchange Offer Is a Capital-Structure Move, Not a $74.8 Million Project Loan

The proposed exchange offer addresses the company’s share count and capital structure, but it does not solve the construction gap on its own. Its cash component is small relative to the project requirement, while the preferred shares carry a stated value and dividend rather than providing immediate cash to the data center.

At the stated terms, the proposed offer would pay $2 for each block of 100 common shares and issue one preferred share with a stated value of $24. The implied common-stock value is $0.26 per share, which the company described as an 82% premium to the Nov. 6 closing price of $0.1429. The company also said at least 100 million shares must be tendered and that the preferred listing requires exchange approval.

Those terms create several uncertainties. If shareholders tender, Hyperscale could repurchase a meaningful amount of common stock while adding preferred obligations. If they do not, the offer may not proceed. If the preferred stock is listed, it could create a new market instrument, but listing and trading are subject to approval and the offer’s other conditions.

The important distinction is between reducing equity-market overhang and raising construction capital. A capital-structure transaction may improve the company’s ability to present a cleaner equity story, but it does not pay the contractor, utility or equipment supplier. The project still requires debt, new equity, asset monetization, customer support or some combination.

The suspended at-the-market program makes the distinction more important. Hyperscale has lost one conventional source of incremental equity issuance, while a deeply discounted financing could transfer more of the future AI economics to new investors. The proposed exchange offer may be intended to address what management calls a gap between market value and balance-sheet strength, but it does not eliminate the going-concern warning.

In other words, the offer may change who owns the company before it changes how the data center gets built.

The AI Pivot Is Structural, but the Liquidity Stress Is Cyclical

Hyperscale’s situation contains two forces that should not be confused. The shift from Bitcoin mining toward AI infrastructure is structural; the immediate cash squeeze is cyclical and can mean-revert if financing and commissioning succeed.

The structural shift comes from the changing value of power. A mining site monetizes electricity through Bitcoin production, exposing revenue to coin prices, network difficulty and hardware efficiency. An AI data center can instead monetize power through multiyear capacity commitments. Existing land, grid access and infrastructure can shorten the path to deployment compared with a greenfield project.

Hyperscale’s own disclosures attach conditions to the Michigan campus’s potential. Expansion depends on financing, infrastructure, engineering, utility agreements and regulatory approvals. Power capacity is not the same thing as operational AI capacity. The missing steps are the expensive ones.

The cyclical element is the financing window. Construction bills arrive before customer receipts, and a small company with large current liabilities has little tolerance for delay. If the first 10 MW is energized in December and the next 10 MW arrives by March, the stress may prove temporary. If the project slips, Hyperscale will have exchanged Bitcoin for an unfinished asset without a dependable replacement cash flow.

The industry logic remains plausible. A long-term AI capacity agreement may be more predictable than Bitcoin-mining revenue, and power-rich sites can be more valuable to AI customers than to miners. But the company-level risk is higher when the operator has limited unrestricted cash, material liabilities and no proven AI operating history.

The shift becomes durable for Hyperscale only if the company reaches operation without surrendering most of the future economics. Otherwise, structural AI demand may benefit better-capitalized operators while leaving the weakest converters behind.

The Second-Order Risk Is Who Pays for the Conversion

The first-order question is where Hyperscale gets $74.8 million. The second-order question is who bears the risk if the project costs more, arrives later or produces less cash than projected.

Debt would protect shareholders from immediate dilution but add fixed claims before the data center operates. Equity would reduce repayment pressure but could transfer future AI economics to new investors if issued at a distressed price. Asset sales could provide speed but shrink the platform supporting future revenue. Customer funding could accelerate construction but give the customer greater leverage over pricing, protections and renewal terms.

The proposed exchange offer adds a different form of capital-structure risk. Preferred stock with a stated value and a 12% annual dividend may not be cash debt, but it creates a senior economic claim relative to common equity. The company would need to show that the restructuring improves its ability to finance and operate the data center rather than simply moving obligations between security classes.

The customer’s bargaining power rises when the operator is short of cash. A customer financing the buildout may demand completion protections, staged payments, performance guarantees or rent concessions. A customer that does not provide capital may still negotiate aggressively because Hyperscale needs the agreement to unlock external financing.

That is the second-order market implication. Investors are not simply valuing AI demand; they are valuing the allocation of construction risk. The same $1.2 billion projection can produce different equity outcomes depending on whether Hyperscale, its lender or its customer pays for delays and cost overruns.

The stock’s Wednesday gain is not evidence that the market resolved the issue. The filing arrived after the close, and the $0.157 price reflects trading before investors could fully digest the warning. The more informative signals will be the financing terms, exchange-offer conditions and construction milestones.

The Counter-Thesis: Capital Restructuring Could Unlock the AI Asset

The strongest argument against a distress interpretation is that Hyperscale is not funding an entirely speculative idea. It has identified a California-based neocloud customer, described 20 MW of initial capacity and set a December energization target. If the customer is financially sound and the company can use the exchange offer to simplify its capital structure, lenders may finance against future contracted cash flows.

The proposed offer also signals management’s belief that the public equity value does not reflect the company’s assets. If shareholders tender and the preferred stock lists successfully, Hyperscale could reduce common shares outstanding and create a security aimed at investors seeking income and asset exposure. A cleaner capital structure might make a later project financing easier to arrange.

Under that counter-case, the Bitcoin sale is a deliberate capital rotation and the exchange offer is a bridge to institutional financing. Hyperscale would emerge with less crypto exposure, an operating AI asset and a financing package tied to customer cash flows rather than volatile coin reserves.

But the offer is proposed, conditional and not a substitute for project funding. The company’s $1.2 billion and more than $3 billion figures are projections that require construction, performance, extensions and option exercise. The exchange offer’s preferred component is not the same as cash, and its small cash component is far below the $74.8 million project requirement.

The falsifying signal is measurable: a binding financing package covering at least $74.8 million, followed by on-time energization of the initial 10 MW without deeply discounted equity or excessive fixed claims. If those milestones arrive, the warning becomes a temporary bridge problem. If Hyperscale completes or extends capital-structure transactions without securing project funding, or misses the March target, the liquidity mismatch will have overwhelmed the structural AI opportunity.

The customer can be real, demand can be durable and the financing can still fail.

Three Time Horizons and Three Possible Outcomes

In the short term, the market will focus on liquidity and transaction mechanics. Investors will watch whether the exchange offer formally launches, whether the minimum tender condition is met, whether the preferred shares can list and whether the company announces funding for the $74.8 million gap. The $0.157 close is a pre-disclosure reference point, not a settled valuation of the warning.

Over the medium term, execution is the test. Hyperscale must move from $47.3 million already spent to a first 10 MW in December and a second 10 MW by March, while meeting $231 million of current liabilities and continuing to fund operations. The critical metrics are energized capacity, customer acceptance, cash collected and project spending against the remaining $74.8 million.

Over the long term, the question is whether a former mining site can become a durable AI platform. That depends on utilization, power reliability, renewal rates, equipment economics and the company’s ability to finance expansion without giving away most of the value it has created.

The base case is a difficult bridge: Hyperscale completes a capital-structure transaction, secures mixed project financing and brings the initial 20 MW online. The upside case requires customer-backed capital, on-time commissioning and exercise of additional capacity options, allowing the company to finance growth against operating infrastructure. The downside case is an unsuccessful offer, a deeply discounted equity raise, a project delay or an asset sale that leaves existing holders with a smaller share of the long-term revenue opportunity.

The beneficiaries of a successful transition would include power-rich infrastructure operators and AI customers seeking near-term capacity. The exposed parties are shareholders facing dilution or preferred claims, lenders financing unfinished assets and miners that sell liquid Bitcoin before replacement cash flows begin.

Hyperscale has not disproved the AI pivot. It has shown that the pivot’s first test is financing, not demand. The company sold its last liquid crypto reserve before the data center became a cash-generating asset; until project capital is committed, the $1.2 billion projection remains a promise with a bill due first.

Explore more exclusive insights at nextfin.ai.

Insights

What caused Hyperscale to sell Bitcoin?

How large is Michigan's AI funding gap?

What does the 20 MW agreement promise?

Why are projected revenues delayed?

How does Bitcoin differ from AI revenue?

What does exchange offer actually fund?

Could preferred shares raise risk?

Which construction milestones matter?

How could customer funding change terms?

What risks face shareholders and lenders?

Can AI demand overcome cash stress?

How does AI compare with Bitcoin mining?

What would prove the AI pivot works?

What could cause the project to fail?

What should investors watch next?

Can site become a durable AI platform?

How does a mining site monetize power?

When will first 10 MW be energized?

Search, ask or analyze anything...