Galaxy Digital: The $85 Million Loss That Doesn't Tell the Real Story

The market is still pricing Galaxy Digital as a crypto bet wearing a data-center costume. Q2's $85 million net loss and the $3.5 billion debt issuance for Helios Phase II confirmed that framing for most of the selling crowd. The stock has fallen roughly 57% from its 52-week high.

But the market is conflating two very different things: unrealized mark-to-market swings on the balance sheet, and a data center that just crossed from construction into revenue generation. Those are not the same risk.

Here's the proof path, and what has to go right for the old story to become stale.

The headline loss is a crypto accounting entry, not an operating decline

Galaxy reported an $85 million net loss and adjusted EBITDA — a rough cash-earnings proxy that strips out interest, taxes, depreciation, and amortization — of negative $77 million in Q2. That looks ugly until you open the segment breakdown.

The Treasury and Corporate segment, which holds Galaxy's own digital asset positions and marks them to market, posted an adjusted gross loss of $42 million and adjusted EBITDA of negative $78 million. That drag is entirely driven by unrealized losses on crypto holdings during a period when digital asset prices fell. It is a balance-sheet accounting entry, not an operating expense. Remove that entire $78 million of negative adjusted EBITDA, and the remaining business is only marginally negative instead of deeply in the red.

The real operating picture is in the two revenue-generating segments. Digital Assets — the trading, lending, and asset management business — lifted adjusted gross profit 34% quarter-over-quarter to $66 million even as trading volumes declined 7%. Management pointed out that the broader industry saw double-digit volume drops. Galaxy outperformed the weakness. That's the kind of detail that suggests fee-based revenue is becoming less sensitive to crypto's boom-bust cycle.

The data center inflection is already here

This is the part the selling crowd missed. The Data Center segment produced $20 million in adjusted gross profit and $11 million in adjusted EBITDA in Q2. Adjusted gross profit grew 560% quarter-over-quarter and the segment swung from negative $0.9 million in adjusted EBITDA to positive $11 million in a single quarter. That is the first full quarter of revenue-generating operations.

The driver is Helios Phase I in West Texas. Galaxy delivered 133 megawatts of critical IT load to CoreWeave under a 15-year lease. Rent commencement in Phase I scales with delivered capacity, which means Q2 was a partial quarter. With the full 133 megawatts online, Phase I is expected to generate approximately $80 million in quarterly leasing revenue starting in Q3. Galaxy projects the project-level adjusted EBITDA margin will exceed 90%.

An $80 million quarterly run rate at 90% plus margins implies roughly $320 million of potential annual adjusted EBITDA from a single phase of a campus that holds 1.6 gigawatts of approved power. The market is anchoring to an $11 billion enterprise value while treating the data center business as a side project.

The $3.5 billion debt is real. So is the capacity pipeline

On July 28, Galaxy issued $3.5 billion in senior secured notes due 2031 to fund Helios Phase II, a 260-megawatt expansion. Data hall deliveries for Phase II are expected in Q2 2027. The company engaged HITT Contracting as general contractor, which management framed as a validation — HITT can be selective, and it chose a 1 GW-plus campus build-out.

The Texas power pipeline now exceeds 5.7 gigawatts across four sites: the Helios campus (1.6 GW approved, with an additional 1.6 GW in the ERCOT interconnection queue), the newly acquired 500-acre Merlin campus in McGregor (initial 74 MW, potential for 500 MW), and two additional sites — Caspian and Selene — with potential capacities of roughly 700 MW and 900 MW respectively.

That pipeline is impressive but it creates the central risk. Management declined to give a timeline for leasing Phase II's 830 megawatts. The CIO acknowledged that prospective customers are focused on 2026/2027 power while Galaxy's next energization is scheduled for 2028. That timing mismatch is real. Peer data center developers have secured deals with investment-grade hyperscalers; Galaxy's anchor tenant remains CoreWeave, which carries a B+ issuer rating from S&P and BB- from Fitch. The 15-year lease is long and the margin is excellent, but the counterparty is not investment grade.

The balance sheet is stretched, not broken

Total debt sits at $7.2 billion. Cash and stablecoin holdings total $2.5 billion. Net debt is $3.1 billion. Total equity is $2.8 billion, giving a debt-to-equity ratio of 167%. Free cash flow over the trailing twelve months is negative $2.0 billion, dominated by $1.4 billion in capital expenditures for the data center build-out.

None of that is attractive on its own. But the structure is project-financed and the cash-flow trajectory is directional. Phase I alone should contribute roughly $80 million per quarter starting next quarter. The question isn't whether Galaxy can service the debt — it's whether Phase II and beyond get leased at terms that justify the capex already committed.

Management said there is no equity need expectation for the CoreWeave build-out of the full 800 gross megawatts. That matters because forced dilution is the mechanism that turns a stretched balance sheet into a broken one.

The market is still pricing the old story

The stock has fallen 19% over the past 20 days and roughly 30% over the trailing year. It trades near $20, well below its 52-week high of $46. The 14% drop on the Q2 print tells you exactly what the market heard: "net loss" and "$3.5 billion in new debt." It didn't hear "data center segment swings to positive EBITDA and is about to step up to $80 million in quarterly leasing revenue at 90% margins."

AInvest's aggregate rating still labels the stock a Buy. The signal is opaque — no contributing analyst names, no disclosed methodology — but at least the consensus isn't overtly bearish. It's a reference point, not proof.

What needs to happen over the next 12 months

The financial bridge is straightforward:

• Q3 2026: Helios Phase I reaches full 133 MW delivery, stepping to the expected ~$80 million quarterly leasing revenue. That alone is a step-function improvement in the data center segment's contribution.
• H2 2026 through 2027: A Phase II lease announcement. Management's language — "advancing" conversations, no commitment — suggests this could come in the second half of 2026 or slip into 2027. The gap between 2027 customer demand and 2028 energization needs to close.
• Digital assets stability: Operating cash flow from the trading and lending business holds at the levels implied by Q2's $66 million adjusted gross profit, reducing reliance on crypto price action.

If Phase I delivers the $80 million quarterly run rate and Phase II signs a credible tenant, the market's $11 billion enterprise value becomes defensible. At that point, the stock stops trading as a crypto-exposure name and starts trading as a leveraged data center developer with a live, high-margin revenue stream.

The tripwire

The setup breaks if CoreWeave's financial condition deteriorates to the point where the 15-year lease becomes a concern, or if Phase II fails to attract a tenant by mid-2027 while debt service on the $3.5 billion notes is mounting. A material delay in ERCOT interconnection for the additional Texas pipeline would also change the math, since Galaxy's valuation premium depends on scale, not just Phase I.

The market is still pricing Galaxy as a crypto story that happens to own some land in West Texas. Phase I says the new story is already underway. The question over the next 12 months is whether the leasing pace catches up to the construction pace.