Galaxy’s $3.5bn data-centre debt topped equity as treasury supplied 87% of Q3 EBITDA

Galaxy Digital entered the third quarter with two businesses moving in opposite economic directions. The group reported $455 million of adjusted EBITDA, yet its Digital Assets segment contributed a loss of $9 million and its Data Centers segment contributed $69 million. Treasury and Corporate supplied $395 million—86.8% of the group total. At the same time, Galaxy disclosed $3.5 billion of senior secured notes for the second phase of the Helios data-centre campus, an amount equal to 116.7% of the group’s reported $3.0 billion of equity.[1]

The combination changes the way the quarter should be read. Galaxy is still branded and valued partly through digital-asset trading, asset management and infrastructure, but the quarter’s adjusted earnings were dominated by treasury activities while the largest disclosed financing commitment was attached to power-intensive computing infrastructure. Neither fact proves that the strategy is unsound. Together they show that the current earnings engine, the capital allocation and the future debt burden sit in different parts of the company.[1]

The debt is not ordinary unsecured corporate borrowing. Galaxy says the notes are senior secured financing for Helios I Phase II, a project with contracted data-centre capacity and future construction. Project security, expected customer payments and ring-fenced assets can reduce direct recourse to the wider group. But comparing the principal amount with group equity remains useful because it measures the scale of one development programme against the capital cushion supporting all of Galaxy’s activities.[1]

The balance sheet expanded around the data-centre build

Galaxy’s total assets rose to $16.4 billion from $10.8 billion in the preceding quarter, an increase of about 51.9%. The most dramatic movement occurred in Data Centers: segment assets rose to $6.8 billion from $2.5 billion, while segment liabilities increased to $5.3 billion from $1.5 billion. Using the rounded figures in the release, assets expanded by about 172% and liabilities by about 253%. The liability increase therefore outpaced the already rapid asset build.[1]

That does not establish a funding gap. Construction projects commonly accumulate debt, contractor obligations and other liabilities before all capacity is placed in service. Galaxy also reported $757 million of data-centre capital expenditure and said 133 megawatts had been delivered. The investigative point is narrower: the group’s balance sheet has become materially more dependent on the execution, customer concentration, power availability and financing structure of Helios than a headline focused only on bitcoin or trading revenue would suggest.[1]

Galaxy’s own allocation disclosure reinforces the shift. It attributed 48% of equity to Data Centers, 33% to Digital Assets and 19% to Treasury and Corporate. Yet the quarter’s adjusted EBITDA composition was almost the reverse: Treasury and Corporate produced most of the result, Data Centers produced a smaller positive contribution and Digital Assets was negative. Capital allocation and current earnings therefore should not be treated as if they describe the same engine.[1]

Why treasury earnings require a separate quality test

Treasury and Corporate adjusted EBITDA included investment gains and other balance-sheet effects that may not recur at the same scale. A strong quarter from treasury can be economically valuable, but it is not interchangeable with contracted data-centre rent, recurring management fees or trading spreads. The filing therefore supports a quality-of-earnings question rather than a claim that the profits are artificial: how much of the $395 million can be expected to repeat without the same market moves and portfolio gains?[1]

The group also reported $8.7 billion of gross revenues and gains alongside $8.0 billion of gross transaction expenses. That presentation reflects the high-volume nature of the digital-asset business and should not be read as an $8.7 billion retained revenue pool. The difference between gross flows and retained economics is another reason to separate business scale from earnings quality. A large gross line can coexist with a much smaller net contribution after transaction costs and hedging.[1]

An alternative interpretation is that Galaxy has deliberately built a diversified model. Treasury gains can support investment while Data Centers moves through a construction-heavy stage, and Digital Assets can recover when trading and asset prices improve. That is plausible and consistent with the company’s strategy. The risk is not diversification itself; it is that investors may use one segment’s current earnings to capitalise another segment’s long-dated cash flows without recognising their different duration, volatility and financing demands.[1]

What the evidence proves—and what remains unresolved

The records prove that one project financing exceeded reported group equity in principal amount, that treasury supplied nearly seven eighths of Q3 adjusted EBITDA, and that data-centre liabilities grew faster than data-centre assets on the rounded quarterly comparison. They do not prove that noteholders have full recourse to all Galaxy entities, that Helios customers will fail to pay, or that the project is overleveraged after considering contracted cash flows, security packages and project-level equity.[1]

The strongest next disclosure would be a project-level bridge showing Phase II committed capital, debt drawdown, sponsor equity, contracted megawatts, customer credit exposure, construction milestones and debt-service coverage under downside cases. Galaxy’s group accounts cannot answer all of those questions because the relevant risks may be ring-fenced inside project entities. Without that bridge, investors can see the scale of the financing but not fully test its resilience.[1]

A second test concerns earnings recurrence. Future quarters can show whether Treasury and Corporate remains the dominant contributor, whether Digital Assets returns to positive adjusted EBITDA and whether Data Centers converts construction spending into contracted operating earnings. A sustained shift toward recurring data-centre cash generation would weaken the concern that current profits and future obligations are mismatched. A reversal in treasury gains without that operating conversion would strengthen it.[1]

Additional documentary tests

The project-finance structure also creates a sequencing question. Debt can be committed before every construction invoice is paid and before every megawatt begins earning revenue. That timing is normal, but it means the headline principal cannot be compared directly with one quarter of data-centre EBITDA. A credible coverage test needs contracted rent, power pass-throughs, construction draw schedules, interest reserves and the dates on which customers become liable to pay. The group release supplies some milestones but not a complete project cash-flow waterfall.[1]

Customer concentration is another missing denominator. A hyperscale campus may have long contracts and high switching costs, yet a small number of tenants can account for most committed capacity. The credit quality of those tenants, termination rights, parent guarantees and the consequences of delayed energisation determine whether secured debt behaves like low-risk infrastructure finance or like a concentrated development exposure. Galaxy identifies the strategic scale of Helios but does not publish enough contract-level data to reproduce that assessment independently.[1]

Power and interconnection are equally material. Delivered buildings do not earn their intended return without reliable electricity, network access and customer equipment. Capital expenditure can therefore be economically stranded for a period even when construction is substantially complete. The filing does not report a project-level sensitivity to power-price changes, curtailment or delayed customer deployment. Those are not allegations of failure; they are the operational variables that connect a secured note to the cash flows expected to service it.[1]

The comparison with equity should also be read dynamically. Group equity can rise through retained earnings or asset appreciation and fall through losses, distributions or valuation changes. Project debt amortises under its own schedule. The 116.7% ratio is consequently a dated scale measure, not a permanent leverage ratio. Its value is to show that one programme is large enough to influence the whole group’s risk narrative even if the legal debt sits in a ring-fenced subsidiary.[1]

The defensible conclusion is therefore not that Galaxy required a rescue or concealed a failing business. It is that the company’s economic identity has become more complex than its digital-asset label: current adjusted earnings were treasury-led, almost half of allocated equity sat in data centres, and a single secured financing exceeded total group equity. Those three facts should be analysed together when assessing valuation, liquidity and execution risk.[1]

The useful unresolved questions for Galaxy are whether the $3.5 billion note package is fully non-recourse to the listed group, what minimum contracted revenue or coverage tests protect the debt, and how management separates recurring treasury income from mark-to-market gains in forward guidance. Clear answers would allow readers to replace the present scale comparison with a more precise project-finance assessment.[1]

Report an error or submit an update