Kip Kool

The $GLXY earnings are out. It was mixed with some good highlights fundamentally but they are far out and pending execution and frankly factors beyond the company's control. And there are bad stuff which tilt me to say that I don't like what I read/heard overall. My conclusion is that the underlying value of Galaxy’s contracted data-center assets improved during the quarter, while my confidence in management’s disclosure discipline and the near-term commercialization of the additional 830MW declined materially. The reported financial results were still weak. Galaxy recorded an $85M net loss, adjusted EBITDA of negative $77M, and adjusted EPS of negative $0.09. Digital Assets and Data Centers together produced $86M of adjusted gross profit and approximately breakeven adjusted EBITDA, while Treasury & Corporate generated negative $78M of adjusted EBITDA. The operating businesses are improving, though the corporate wrapper and balance-sheet exposure continue to absorb much of that progress. Digital Assets showed some resilience. Adjusted gross profit increased 34% quarter over quarter to $66M, Global Markets gross profit rose to $49M despite weaker industry activity, and the average loan book remained around $1.4B. GOFR has already generated nearly $300M of loan originations, and the multi-year BNY engagement includes forward-deployed Galaxy engineers helping design institutional digital-asset infrastructure. These are credible product and distribution developments. Digital Assets still produced negative $11M of adjusted EBITDA, and GalaxyOne remains a small contributor with limited marketing investment, so this segment has yet to become a dependable earnings floor. The contracted Helios business was the strongest part of the quarter. Galaxy completed all 133MW of Phase I critical IT capacity on schedule and on budget. Beginning in Q3, management expects Phase I to produce approximately $80M of quarterly leasing revenue at a project-level adjusted EBITDA margin above 90%. That equates to roughly $320M of annualized revenue and more than $288M of annualized project EBITDA from the first 133MW alone. Galaxy also raised the economic guidance for the full 526MW CoreWeave lease. The company now expects more than $1.2B of average annual leasing revenue and average lease-level adjusted EBITDA margins above 90%. This is a meaningful increase from the previous $1B+ annual revenue language and confirms that the contracted 800MW gross project is more valuable than earlier public disclosures suggested. The accounting still requires care. Leasing revenue includes contractual escalators recognized on a straight-line basis. Operator revenue includes power, utility, and related charges that are offset by corresponding expenses. Project-level adjusted EBITDA excludes overhead. The reported $1.2B+ revenue and 90%+ lease margin therefore cannot be treated as cash directly distributable to GLXY common shareholders. Interest, amortization, maintenance capital, taxes, segment overhead, and future refinancing still sit below that figure. Phase II execution also remains on track. HITT has been on site since April, earthwork is complete, and structural foundation work has started. Seven of the eight Phase II data halls are expected to be online by the end of 2027, with the final hall in early 2028. Phase III is expected to come online progressively during 2028. The $3.5B Phase II financing substantially reduced construction-funding risk. The notes were issued at 85% loan-to-cost, although the 9.875% coupon is expensive and creates a meaningful refinancing requirement in 2031. Management stated that Phase II is fully funded with debt and equity. It also said prior capital planning and expected cash generation have already prefunded the anticipated Phase III equity requirement. Galaxy does not currently expect to need additional equity to complete the original 800MW CoreWeave project, although Phase III will require a separate project-debt financing. This is an important positive for dilution risk. The company’s broader Texas platform also became more tangible. Galaxy now controls more than 5.7GW of potential capacity across Helios, Merlin, Caspian, and Selene. Merlin has an initial 74MW utility path and a longer-term potential of 500MW. Caspian has 700MW of potential capacity and appears to be the most advanced newly acquired site, with executed interconnection agreements and known transmission and distribution funding commitments. Helios III and Selene have entered the Studied Load process, with Galaxy posting $50M and $45M of financial security, respectively. The acquisitions were structured with modest upfront capital and most consideration contingent on future development and leasing milestones. These details support the view that Galaxy is building a real multi-campus development platform rather than accumulating expensive land options. The regulatory backdrop remains unresolved. Governor Abbott’s audit delayed the expected Batch Zero classifications, and management provided no new timetable. Galaxy appears better positioned than many applicants because its projects have identifiable ownership, capital, studies, interconnection work, local agreements, and financial security. That relative advantage does not eliminate the delay or guarantee the final allocation of the unapproved capacity. The central issue in the quarter was the additional 830MW at Helios. In June, Mike Novogratz publicly said the full 1.6GW could be leased by July 4 or, with greater certainty, by the end of summer. That was a highly specific timeline for a material commercial event. It was widely circulated and became part of the market’s near-term investment thesis. The Q2 call described a very different commercial position. Management said the primary gating factor is that the 830MW is scheduled to energize in late 2028. Prospective tenants are currently dedicating much of their limited procurement attention to power available in 2026. Management described the recent discussions as a process of building longer-term relationships until customers are ready to focus on late-2028 capacity. It is also waiting for newer guarantee and lease-wrap structures to develop before committing the asset to certain neoclouds, AI labs, or other emerging counterparties. This is difficult to reconcile with the June statement. The late-2028 energization date was already known months earlier. In the February call, management explicitly said the new 830MW would not be available until late 2028 at the earliest. On the same call, management said major market participants had rapidly shifted from seeking 2026 and 2027 power toward locking up 2028, 2029, and 2030 capacity, and suggested that allowing tenant negotiations to continue for a year would be unacceptable. The sequence is therefore clear. In February, management knew the late-2028 delivery schedule and said customers were actively competing for 2028–2030 capacity. In June, the CEO said the remaining campus would be leased by July 4 or certainly by the end of summer. In August, the company said customers are still focused on 2026 power and that the recent process has largely involved relationship building and evaluating emerging credit-wrap structures. These statements do not form a coherent commercial timeline. There are several possible explanations. A relatively advanced transaction may have failed or been delayed. The CEO may have converted preliminary customer interest into an unjustifiably precise public timeline. There may have been a communication gap between senior management and the team directly responsible for commercial negotiations. None of these explanations is reassuring for a public company whose stock moved materially after the June comments. The call did not directly acknowledge that the earlier guidance was no longer valid. Management did not explain what changed, withdraw the summer-end timeline, or identify a specific commercial process that had broken down. Novogratz answered relatively few questions and did not personally reconcile his prior statement with the current facts. The responsibility for explaining the 830MW process fell largely to Chris Ferraro. This matters beyond one delayed lease. Management credibility is part of the discount rate. Investors must rely on management when evaluating the timing of Helios III and IV, Caspian’s eventual allocation, Merlin’s expansion, future tenant negotiations, the 5.7GW pipeline, and any potential data-center separation or platform transaction. Once a CEO gives a highly confident timeline for a material catalyst and the subsequent disclosure describes a substantially earlier-stage process, future qualitative guidance deserves a lower evidentiary weight. I am therefore removing a summer-2026 lease announcement from my base case. I am also removing a direct investment-grade hyperscaler lease as the default outcome. The 830MW remains a scarce and valuable approved-power asset with interconnection agreements, funded utility upgrades, more than $180M of long-lead equipment orders, and a switching station under construction. Its eventual commercial value remains substantial. Its timing and tenant structure are now much less visible. My base case now includes a broader set of outcomes: a hyperscaler, a neocloud or AI lab supported by an investment-grade wrap, or another emerging tenant structure that gives Galaxy acceptable long-term credit protection. A direct hyperscaler lease remains a favorable outcome, though it should be treated as a bull case until supported by a signed document. I also place less value on the 5.7GW headline than I did before the call. The portfolio is real and increasingly well structured, especially Caspian and the initial Merlin capacity. Much of it remains subject to regulatory allocation, transmission upgrades, customer commitments, financing, and multiyear development schedules. Behind-the-meter generation is being studied, though management said it is expensive, difficult to finance, operationally complex, and outside the company’s present focus. The balance sheet and share count were better than feared. Galaxy ended the quarter with $2.7B of equity and $2.5B of cash and stablecoins. Economic ownership was approximately 391.4M shares, showing no evidence of large-scale ATM dilution through quarter-end. The original 800MW CoreWeave project is now expected to be completed without additional equity funding. These facts reduce the risk that the contracted Helios value is consumed by near-term common-share issuance. My overall view is therefore more cautious after Q2, despite stronger operating assets. The contracted 800MW business is more valuable and more fully funded than previously understood. Phase I is generating cash, Phase II construction and financing are in place, and the company has assembled a credible Texas development portfolio. Those are substantive positives. The near-term 830MW thesis has weakened substantially. The commercial process appears less advanced, the customer decision window is later, and management no longer provides a reliable timeline. The largest negative from the quarter was the inconsistency between the CEO’s prior public statements and the facts subsequently presented on the earnings call. Galaxy can remain materially undervalued while deserving a larger governance and timing discount. Asset value and management credibility are separate underwriting variables. The assets improved this quarter. The credibility attached to management’s unsourced timing statements deteriorated. Until that credibility is repaired, I will give meaningful weight only to executed leases, utility agreements, project-financing documents, filed construction milestones, and actual reported cash flows. I will treat interview-based timelines and broad statements about imminent demand as aspirational. My final assessment is that Q2 strengthened the long-term infrastructure thesis, weakened the near-term stock thesis, and increased the evidence threshold required before assigning value to any uncontracted capacity.