From Mining to Mining Artificial Intelligence

Written by: Prathik Desai

Compiled by: Block unicorn

Today, we delve into Galaxy Digital's financial report for the second quarter of 2025. This digital asset and data center solutions provider is preparing for a transformation. From its core business—which accounts for 95% of revenue but has a profit margin of less than 1%—to a business model that sounds incredibly good in terms of revenue-to-expense ratio.

Abstract

Galaxy's cryptocurrency trading business achieved only $13 million in profit (profit margin 0.15%) from $8.7 billion in revenue, while paying $18.8 million in salaries each quarter—resulting in negative cash flow from core operations.

Artificial Intelligence Transformation: Signed a 15-year, 526 megawatt contract with CoreWeave, expected to achieve an average annual revenue of over $1 billion in three phases starting in the first half of 2026, with a profit margin of 90%.

In a market with limited supply, controlling the 3.5 GW capacity in Texas, data center demand must quadruple by 2030.

A project financing of 1.4 billion dollars has been secured, validating its commercial viability and eliminating execution risks.

The current model relies on cryptocurrency asset revenue (USD 198 million in the second quarter) to fund operations, as capital-intensive trading yields slim returns.

The stock price rose by 17% and then fell back, as investors do not see incremental revenue before the first half of 2026.

When you look at Galaxy Digital's second quarter data, it's easy to overlook one thing: what happens next. A closer inspection reveals that the company, led by Michael Novogratz, is at a turning point from cyclical cryptocurrency trading to more stable AI infrastructure revenue.

The gold mine of artificial intelligence infrastructure

Galaxy Digital is undergoing one of the largest business transformations in the cryptocurrency industry – shifting from low-margin trading to high-margin artificial intelligence data centers.

Galaxy achieved a net income of 31 million dollars this quarter, with an adjusted EBITDA totaling 211 million dollars after adjusting for non-cash and unrealized expenses.

In its total revenue, the trading business earned only $13 million in profit from sales of $8.7 billion—a profit margin of just 0.15%. Therefore, nearly 95% of its revenue is barely profitable.

In comparison, their new artificial intelligence data center contract promises a 90% profit margin on an average annual revenue of over 1 billion dollars.

While I am optimistic about building artificial intelligence and high-performance computing capabilities, I believe the promised profit margins are overly exaggerated. But don't just take my word for it; you can compare the profit margins reported this quarter by top AI data operators like Equinix and Digital Realty: 46-47%.

However, I believe the direction is correct, purely from the perspective of revenue generation. Currently, most of Galaxy's revenue comes from trading operations that have high costs and low profits. Its majority profit (revenue minus expenses) comes from its asset and corporate divisions.

The asset department includes investments in digital assets and mining activities, equity investments, and realized and unrealized gains and losses on digital assets and equity investments.

Its $2 billion asset pool serves as an investment tool, even as a strategic source of funds when market conditions are favorable.

The department generated $198 million in revenue without accounting for non-cash unrealized gains. Unlike purely cryptocurrency companies, Galaxy's asset business raises funds by strategically selling assets at opportune times.

This is exactly where I believe Galaxy's cryptocurrency asset strategy differs from Michael Saylor's Bitcoin asset strategy. Saylor's "buy, hold but never sell" strategy has achieved an unrealized gain of $14 billion this quarter. However, these are just paper profits, and Saylor's shareholders cannot benefit from them.

The situation with Galaxy is different. It not only buys and holds cryptocurrencies in the asset pool but also engages in strategic sales, bringing in realized profits. This is real money that shareholders can share.

However, I believe that Galaxy's asset division is an unreliable source of income. As long as the cryptocurrency market is in optimal condition, this division will continue to generate profits. But the market does not operate this way, whether in the traditional market or the cryptocurrency market. The market is at best cyclical, which makes these returns heavily dependent on the state of the cryptocurrency market.

This is why Galaxy needs its artificial intelligence transformation to succeed, as the current model is unsustainable.

Market Opportunity

Galaxy has positioned itself at the intersection of two massive trends: the explosive demand for artificial intelligence computing and the long-term shortage of electricity infrastructure in the United States. A McKinsey report indicates that global data center demand is expected to increase from 55 gigawatts in 2023 to 219 gigawatts by 2030, a fourfold increase.

Hyperscale cloud service providers are expected to invest $800 billion in data centers as capital expenditure (CapEx) by 2028, a 70% increase compared to 2025, but this is constrained by power supply.

The advantage of Galaxy lies in its potential capacity of 3.5 GW in the Helios complex in Texas, sufficient to meet the electricity demands of over 700,000 homes in the state. 800 MW has been approved, and an additional 2.7 GW is under study by the Electric Reliability Council of Texas (ERCOT). Galaxy controls some of the largest available electricity capacity in the constrained AI infrastructure market.

Digital rendering of the Helios AI and HPC data center campus of Galaxy located in Texas.

The foundation of Galaxy's transformation is the 15-year commitment made with CoreWeave, which is one of the largest artificial intelligence infrastructure deals in the industry. CoreWeave has committed to providing 526 megawatts of critical IT capacity in three phases.

90% of the projected profit margin is attributed to the asset-light nature of data center operations after the infrastructure is built.

I believe there is a significant risk in CoreWeave's transactions: execution. Just as I was considering the scale of funding that Galaxy needs to raise, planning, and execution capabilities, this company has cleared the first hurdle.

On August 16, Galaxy successfully completed a $1.4 billion project financing for the Helios data center, securing the necessary funding for the first phase of construction. This has increased my confidence in how to eliminate key funding risks and validate the commercial viability of the Helios project.

Cash Flow Equation

Galaxy's current cash flow exposes the instability of its trading business, while highlighting why artificial intelligence infrastructure can provide real financial stability.

The company had $1.18 billion in cash and stablecoins at the end of the second quarter, which sounds like a lot, but the reality is more complicated. Galaxy's trading business operates on a capital-intensive model, and margin loans require large cash reserves. Most of the funds from this $1.18 billion are not freely usable.

The free cash flow generated by Galaxy is negligible. After paying $14.2 million in interest expenses and ongoing operating costs, the core business is barely able to achieve cash flow balance.

This forces Galaxy to rely on the appreciation of the cryptocurrency market, meaning its treasury and mining operations generate funds to support operations amid inherent cyclicality and unpredictability. In contrast, CoreWeave's three-phase contract structure and high-profit business characteristics may immediately generate positive cash flow.

Although the profit margin may not be as high as 90%, even a more conservative profit margin of 40-50% will still be more reliable and stable than the cyclical treasury business.

Unlike trading businesses that require continuous investment in working capital and technological infrastructure, the cash generated by data center operations can be reinvested in expansion or returned to shareholders.

The recent Helios project financing by Galaxy has helped address cash flow issues. By securing dedicated construction funding, Galaxy separates infrastructure development from its operational cash flow needs. This is not achievable in trading business expansion, as the balance sheet capital required for trading business expansion would directly compete with other business needs.

Fee Details

The total expenses for the digital asset department amount to $8.714 billion, with trading fees accounting for the largest share ($8.596 billion). These costs are purely passed on, leaving little room for increases. Galaxy can hardly optimize these costs, as the spread continues to narrow in the commoditized trading business, making these costs unavoidable.

What is even more concerning is that the quarterly compensation expenses include $18.8 million in stock compensation, which must be paid in cash. This means that Galaxy's costs for retaining talent exceed the earnings generated by its core business ($13 million).

The transformation of artificial intelligence infrastructure will change this situation. Once the facilities are operational, the operation of data centers will require minimal variable costs.

To illustrate more intuitively, Galaxy's entire digital asset business generated a $71.4 million adjusted gross profit in the second quarter. At full capacity, the first and second phases of Helios (approximately 400 megawatts) could generate $180 million in quarterly revenue, with operational complexity and costs being only a small fraction of those for trading operations.

Market reaction

Galaxy's stock price rose slightly by 5% within 24 hours after announcing its second-quarter financial report and jumped about 17% within a week, after which investors began to withdraw their funds.

This may be because investors realize that $211 million in earnings includes $180 million from non-cash adjustments and treasury gains, rather than operational improvements.

Investors may not have fully considered Galaxy's complex transition to artificial intelligence infrastructure, as significant data center revenues are not expected until before the first half of 2026.

I remain optimistic about the long-term market sentiment, given the potential for future development.

According to ERCOT research, the addition of 2.7 gigawatts of capacity by Galaxy indicates the company's intention to solidify its position as a long-term infrastructure provider rather than a single-tenant facility operator.

After全面发展, Galaxy's operations in Texas can compete with some of the largest hyperscale data center campuses operated by Amazon, Microsoft, and Google. This scale can provide it with leverage in negotiations with other AI companies, while also improving operational efficiency and thus increasing profit margins.

The company's expertise in the cryptocurrency field provides it with a unique position in the emerging area where artificial intelligence intersects with blockchain technology.

The Road Ahead

Galaxy is undertaking a massive, polarized bet. If the transformation of the artificial intelligence infrastructure is successful, they will transition from a low-margin trading company to a cash-generating machine. If they fail, they will spend billions of dollars in Texas building expensive real estate while their core business gradually shrinks.

The $1.4 billion project financing confirms external confidence, but I'm focusing on two key metrics: can they really deliver 133 megawatts of AI-ready capacity before the first half of 2026? Once they start covering actual operating costs, can this 90% profit margin be maintained?

The current operations provide sufficient cash flow to sustain the business, but strong ongoing performance in the cryptocurrency market is needed to make meaningful growth investments. Opportunities in artificial intelligence infrastructure promise sustained and reliable revenue potential, with success fully dependent on execution over the next 18-24 months.

The recent project financing has eliminated significant execution risks, but Galaxy must now prove that they can successfully transform their crypto mining infrastructure into enterprise-level artificial intelligence computing facilities to attract long-term bets from investors.