Hut 8’s $7 Billion Cash Mirage: The Restricted Liquidity Behind the AI Hype

Hut 8 reported $7 billion in cash. The number is true. The liquidity is fiction.\n\nThe company’s latest 10-Q draws a bright line between the headline figure and what management can actually spend. Only $233 million sits unrestricted. The other $6.8 billion is locked in subsidiary accounts tied to two AI data center projects, River Bend and Beacon Point. This is not a minor footnote. It is the entire story.\n\nI trace the wallet, not the whisper. When an asset manager sees “$7 billion cash” on a miner’s balance sheet, the first reflex is to treat it as a war chest. The second should be to ask who controls each wallet. Hut 8’s own filing answers: the parent company cannot touch most of it.\n\n## Context: From Bitcoin Miner to AI Landlord\n\nHut 8 began as a bitcoin miner and now wants to be a digital infrastructure provider. It has 17,316 BTC on the consolidated balance sheet. It also has two enormous AI data center developments, funded by subsidiary-level notes totaling $7.5 billion. River Bend carries $3.25 billion. Beacon Point carries $4.25 billion. The notes are obligations of special-purpose subsidiaries, River Bend DC LLC and Beacon Point DC LLC. Hut 8 Corp is not a guarantor. That last sentence is doing more work than any bull-case tweet.\n\nThe financing structure follows a classic project finance pattern. Note proceeds flow into construction reserve accounts and debt service reserve accounts. Interest payments begin in November 2026. Principal repayment for River Bend starts in May 2028; Beacon Point follows in May 2030. Until then, the $6.8 billion of restricted cash is not a source of general liquidity. It is an escrow balance for a construction bet on AI infrastructure.\n\nThe market context matters. Since late 2025, bitcoin miners have been re-priced as AI datacenter plays. Core Scientific signed a long-term hosting deal with CoreWeave. IREN built its own GPU clusters. Bitfarms remained a pure miner. In that landscape, Hut 8 announced two massive project financings and triggered a wave of “AI-transition” excitement. The stock story became: a miner with a giant cash pile and a clear runway into the AI compute boom. The reality, buried in the 10-Q, is that the cash pile is mostly reserved for specific construction projects, and no AI customer has been named.\n\n## Core: The Balance Sheet Is a Narrative Trap\n\nThe core problem is not fraud. It is accounting optics. On a consolidated balance sheet, restricted cash still appears as cash. But the parent’s ability to redeploy that cash for operational survival, bitcoin purchases, or debt service on FalconX loans is effectively zero. The only true buffer is $233 million.\n\nHype is the only asset in a vacuum mint. The gap between the headline and the accessible liquidity creates a classic information asymmetry. Retail investors scan the top line of the balance sheet and see financial strength. Professional investors read the footnotes and see a construction lockbox. This is not a new trick, but in cryptocurrency markets, where “number go up” is a substitute for due diligence, the gap becomes a trading edge for whoever reads the filing.\n\n### Interest Coverage Is an Emergency Signal\n\nLook at the income statement. In the second quarter, interest expense ran $51.2 million. Adjusted EBITDA, excluding mark-to-market digital asset swings, was just $10.4 million. That means the company’s operating earnings cover only about twenty percent of its interest bill. The coverage ratio is 0.2x. This is not a healthy operating business. It is a leveraged construction vehicle.\n\nThe cash flow statement confirms the diagnosis. First-half operating cash flow was negative $32.8 million. The company has no AI revenue yet. It is burning cash while paying interest on a loan backed by bitcoin, and simultaneously trying to fund two construction projects from restricted accounts. This is not a growth story. It is a waiting game.\n\n### The FalconX Loan Is the Hidden Trigger\n\nThe FalconX loan deserves a closer look. Hut 8 owes $200 million at 7%, due April 2027. The collateral is bitcoin. The filing indicates 4,850 BTC are held as collateral, though not necessarily all for FalconX. At a bitcoin price near $100,000, that collateral is worth roughly $485 million against a $200 million loan, implying a loan-to-value around 40-50%. A typical 130% maintenance margin threshold means a liquidation cascade starts if bitcoin falls to approximately $52,000-$65,000. That is a wide safety band, but it is not infinite. And if bitcoin drops, both sides of the balance sheet bleed: digital asset write-downs hit net income, and margin calls hit liquidity.\n\nThe second-quarter results already demonstrate the damage. Hut 8 recorded a $177.1 million net loss, including $138.6 million of digital asset losses. The market prices this stock like an AI growth story, but the income statement still smells like a bitcoin mine. The company’s own incentive structure is tilted: when bitcoin is rising, the FalconX loan becomes cheaper relative to collateral; when bitcoin is falling, the same loan becomes a liquidity vacuum.\n\n### The Subsidiary Sandbox\n\nThe SPV structure deserves more scrutiny. River Bend DC LLC and Beacon Point DC LLC are separate legal entities. The parent is not a guarantor. In theory, that protects Hut 8 Corp if the projects fail. In practice, it also strips value from shareholders. The restricted cash belongs to the subsidiaries, not the public company. If the projects go over budget, the subsidiaries may need equity injections from the parent. If they succeed, cash flows go first to debt service, then to equity returns. The parent’s shareholders are last in line.\n\nThis is the structural fragility that bull-case summaries ignore. When a company raises $7.5 billion in project-level debt, it is not free capital. It is a claim on future cash flows with priority over shareholders. The debt holders are protected by the restricted reserves. The equity holders are exposed to construction risk, execution risk, and the bitcoin price cycle, while receiving none of the immediate liquidity benefits.\n\n## Contrarian: What the Bulls Got Right\n\nNow the contrarian angle. The bulls are not entirely wrong.\n\nHut 8 obtained enormous financing for AI projects before most mining peers closed a single hyperscaler contract. Core Scientific has CoreWeave. IREN has self-built clusters. Hut 8 has commitments of capital. In a capital-constrained AI infrastructure arms race, that is not nothing. The subsidiary SPVs are designed to protect the parent from construction overruns, at least contractually. If River Bend and Beacon Point fail, the noteholders may have recourse only to the project assets, not to Hut 8’s broader bitcoin treasury. That is why the parent could obtain $7.5 billion of project-level debt at 6.13-6.19% without a parent guarantee.\n\nThere is also a plausible hidden assumption behind this financing. Lenders rarely commit billions to a data center project without some evidence of anchor demand. The probability that Hut 8 already has intention letters or preliminary discussions with hyperscalers is high. If a large customer contract is announced, the market will quickly reprice Hut 8 as an AI infrastructure operator and forget the liquidity distinction. The current stock weakness is a valuation recalibration, not necessarily a thesis breaker.\n\nBut the bull case still rests on an unproven bridge. The company has not disclosed customer contracts. It has not published megawatt targets for River Bend or Beacon Point. It has not named the construction contractor. It has not described the technical design for cooling, power delivery, or GPU cluster deployment. The only numbers are the dollar amounts of the notes and the restricted cash balances.\n\nThis is where my audit discipline kicks in. Based on my experience examining protocol treasuries and public company disclosures, I demand a simple forensic test: can the entity that owes the bills access the cash that pays the bills? Here, Hut 8 Corp can access only $233 million. Everything else is trapped in project silos. That is the difference between a balance sheet and a narrative.\n\nWhen the yield is too high, the exit is rigged. In this case, the yield is not a token incentive but a 7% bitcoin-secured loan and 6.1-6.19% project notes. The structure is not rigged against retail buyers in the traditional scamming sense. It is rigged in the sense that the headline cash number does not mean what it appears to mean. The exit, for equity holders, is locked behind a construction schedule and a debt waterfall.\n\n## The Hidden Leaks\n\nThe hidden leak in the structure is time. Interest on the project notes starts accruing in November 2026. The first principal maturity, for River Bend, arrives in May 2028. Construction overruns, regulatory delays, and community opposition are all unquantified. If the projects slip, interest may need to be capitalized further, and if they abort, the parent may have to inject equity capital into the SPVs to protect its reputation or avoid consolidation issues. The $233 million unrestricted buffer is not a rainy-day fund; it is more like a thimble.\n\nThe second leak is the bitcoin price dependency. Hut 8’s real liquidity is a function of BTC price, not the notional cash balance. The 3,090 BTC pledged to purchase miners are an investment in future hashrate, not a source of cash. The 4,850 BTC pledged as collateral create a contingent liability. The 9,376 BTC held in custody may include assets belonging to American Bitcoin or other counterparties, and the filing does not allocate the status buckets between the two entities. Anyone computing net asset value by multiplying total BTC by price is overstating shareholder equity.\n\nThe third leak is governance. The American Bitcoin structure adds complexity. The consolidated group holds 17,316 BTC, with 8,002 attributed to American Bitcoin. But the filing does not explain the precise equity relationship, governance rights, or profit-sharing formula. If American Bitcoin is a separate business partner, then those 8,002 BTC are not fully controlled by Hut 8’s board. The disclosure vacuum is not an evidence of fraud, but it is a reason to discount the net asset value.\n\n## Takeaway: Demand the Footnote, Not the Headline\n\nThe broader systemic lesson is simple. Hut 8’s disclosure is actually more transparent than a typical crypto project. The 10-Q tells us exactly where the cash sits. The problem is that the market does not read footnotes. It reads headlines and marks the stock accordingly.\n\nSo what should an institutional investor demand? Three items. First, a complete breakdown of restricted cash by reserve account, including construction reserve and debt service reserve balances. Second, a clear statement of whether any AI customer contract exists, and if not, the projected timeline for signing one. Third, a sensitivity table showing the FalconX loan’s collateral coverage at various bitcoin prices, including the exact number of BTC pledged specifically to that lender.\n\nWithout those items, the $7 billion cash balance is a mirage. It is not a lie, but it is not a war chest either. It is a set of handcuffs in a safe.\n\nHut 8 has entered a construction phase that will define its next decade. The market is currently pricing the company as an AI infrastructure winner while the filing describes a miner waiting for its first proof of demand. The difference between those two valuations is the difference between $233 million and $7 billion.\n\nI close with a question: when the first interest payment comes due in November 2026, will those restricted dollars have turned into a signed customer, or into a construction site that eats another $233 million? The answer will be written not in press releases, but in the next 10-Q’s footnote.