I spent three days cross-referencing public annual report framings from crypto financial services companies against the exchange-layer data I work with daily — proof-of-reserve audit dates, CER security scores, verified reserve statuses, daily volume figures. There is a pattern that keeps showing up every time a crypto-adjacent stock moves double digits on what is, by traditional accounting standards, a terrible headline number.
Galaxy stock rallied 11% after its annual report showed core business profitability despite a $241 million net loss. Crypto Twitter celebrated. The finance press ran with the "profitable despite losses" angle without hesitation. And almost nobody stopped to interrogate the machinery underneath that framing: what does "core business profitable" mean when the entity defining "core" gets to draw the perimeter wherever it wants?
I have watched this exact pattern unfold across crypto financial companies repeatedly. A company posts a significant net loss. Somewhere in the filing, a carved-out metric — "adjusted EBITDA," "core operations," "operating business excluding impairments" — shows a positive number. The stock moves. Analysts nod. Retail enters. The entire conversation pivots from "this company lost $241 million" to "the core business works." That reframing is not accidental. It is, structurally, the entire point of how these reports are constructed.
The Core Business Carve-Out
Every time a crypto company posts a loss and simultaneously claims operational profitability, someone upstream has made a decision about where to draw the line between "core" and "everything else."
The practice is not unique to crypto. Traditional finance has been running adjusted-EBITDA gymnastics for decades. But crypto companies operate in an environment where the distinction between core operations and speculative positions is genuinely, structurally blurry. A company that trades crypto, invests in crypto, provides advisory services to crypto projects, and runs asset management products across crypto markets does not have a naturally clean boundary between "the business" and "the market exposure." The line gets drawn retroactively, and it typically gets drawn in the exact place that makes the positive number appear.
I will concede the strongest version of the counterargument here, because it deserves to be stated honestly: some of the separation is legitimate. If Galaxy's core business involves advisory, asset management, and infrastructure services — and the $241 million net loss includes unrealized mark-to-market swings on portfolio holdings — then distinguishing operational revenue from paper losses has genuine analytical value. That is a fair point. I am not pretending the distinction is always meaningless.
What I am dismantling is everything built on top of that concession. The reflex to treat the carved-out number as the real story and the $241 million net loss as atmospheric noise. Because in crypto financial services, the "non-core" items that get excluded from the profitability narrative are frequently the company's largest actual risk exposures. Removing them from the headline does not remove them from the balance sheet. It does not remove them from the counterparty chain. And it does not remove them from the downside distribution that shareholders are actually exposed to. The carve-out is an editorial decision presented as an accounting one. The market, reliably, treats it as the latter.
The Rally-on-Loss Reflex
There is a pattern in how the market prices crypto-adjacent equities that runs precisely opposite to how it prices the exchanges those equities depend on.
When Binance — the largest exchange by daily volume at roughly $18.5 billion — faces a regulatory headline or a negative story, the market treats it as an existential event. Volume shifts. Competitors run ads. Crypto Twitter writes obituaries. Binance carries a CER security score of 9.4, maintains verified reserve status, and its most recent proof-of-reserves audit is dated March 1, 2025. That data is externally verifiable, independently scored, and updated regularly. And yet a single negative headline can move market sentiment against the exchange within hours.
Now compare: a company posts a $241 million net loss, frames it as core profitability, and the stock jumps 11%. The asymmetry is not subtle. Exchanges, which actually provide externally auditable data about their solvency positions, get punished for narratives. Companies that report carved-out metrics from annual filings arriving once a year get rewarded for framing.
Bybit, CER score 9.1, verified reserves, audited March 12, 2025. OKX, CER score 9.3, verified reserves, audited March 1, 2025. Bitget, CER score 8.9, verified reserves, audited February 20, 2025. These exchanges are not publicly traded. They do not have stock tickers that bounce 11% on a well-packaged report. And yet the independently verifiable data they publish about their actual financial positions is, in most measurable ways, more granular, more frequent, and more externally checkable than what a stock investor gets from an annual report designed to produce a narrative.
The rally-on-loss reflex exists because the market has been conditioned by two years of post-FTX recovery psychology. After the collapses, after the cascade, any number that is not catastrophic gets re-coded as bullish. A $241 million loss becomes "only" a $241 million loss. Core profitability becomes "the business works." And an 11% rally becomes the price of narrative momentum rather than fundamental reassessment.
If the exchanges that crypto companies depend on for their trading infrastructure publish more verifiable solvency data than those companies' own annual reports, something in the transparency hierarchy is inverted.
The Exchange Layer Nobody Audits
The part of the Galaxy story that almost no coverage examines is the infrastructure dependency — the exchange layer underneath the entire operation.
Every crypto financial services company, whether it is trading, providing liquidity, managing assets, or offering derivative products, runs its operations through exchange infrastructure. Those exchanges have varying degrees of transparency, and the spread matters far more than most equity analysts appreciate. The data I track daily shows a clear hierarchy: Binance, Bybit, and OKX all maintain verified reserve statuses with recent, dated proof-of-reserve audits. Bitget sits in the same tier. MEXC, with a partial reserve status, its last POR audit dating to December 10, 2024, and a CER security score of 8.5, does not.
When a company reports that its core business is profitable, nobody on the earnings call asks the question that would actually stress-test that claim: which exchanges are you executing through? What is your counterparty exposure to each? When the core operations metric separates trading revenue from mark-to-market portfolio losses, does it also separate out the counterparty risk embedded in where those trades settle?
These are not abstract concerns. Proof-of-reserves audits exist for a reason. Binance's last audit: March 1, 2025. Bybit's: March 12, 2025. OKX's: March 1, 2025. Those timestamps are public. They are verifiable. They represent a specific claim — at this date, this exchange held reserves matching or exceeding liabilities — that can be checked against on-chain data. Where is the equivalent for a crypto company's "core profitability" figure? It does not exist as an on-chain artifact. It exists as a line in a PDF, signed by auditors who may or may not understand the on-chain infrastructure underneath the operation they are attesting to. The collapse of FTX was, at its mechanical core, a counterparty risk event — and every entity that reported profitability up until the moment of failure used the same carve-out framing that excluded the risk factor that actually destroyed them. I am not drawing an equivalence to Galaxy. I am noting that the analytical tools for distinguishing between a healthy operation and a fragile one do not live in annual reports. They live on-chain, in proof-of-reserve data, in real-time reserve monitoring — infrastructure that exchanges provide and that crypto financial services companies, as a class, do not.
The Proxy Exposure Fallacy
Every time a crypto company's stock rallies on an annual report, a particular type of retail investor makes the same analytical error: they treat the stock as a proxy for crypto exposure with the bonus of equity market legitimacy.
The reasoning runs like this: I want Bitcoin exposure but I do not want to deal with exchanges, self-custody, or the perceived operational risk of holding crypto directly. A company like Galaxy is in the crypto business, so its stock should track the sector broadly while giving me the comfort of a regulated equity. This sounds reasonable until you examine what you are actually buying.
A share in a crypto financial services company is not a synthetic long on Bitcoin. It is a levered, opaque position on a bundle of trading books, advisory fee streams, portfolio marks at management's discretion, and counterparty exposures to infrastructure you cannot independently audit — wrapped in annual report packaging that gives you visibility once a year, if that. Compare that to the exchange layer directly. Bybit requires a minimum deposit of $1. OKX and Bitget offer PIX deposits to Brazilian users at zero fees with instant processing. OKX's maker fee is 0.08%. Binance and Bybit charge 0.1% maker and taker. If what you want is crypto exposure, the cost of getting direct exposure has never been lower and the infrastructure for accessing it has never been more verifiable.
Here is a detail that tells you something about the difference in accountability models. Binance carries a Trustpilot rating of 2.3 — the market's largest exchange, punished in user reviews largely for KYC friction and withdrawal process complaints. Bybit sits at 4.5. Bitget at 4.6. OKX at 4.2. These are user-facing platforms rated daily by millions of people executing real trades with real capital. A crypto company's annual report gets "rated" once a year by analysts who may or may not understand the on-chain plumbing underneath. The feedback loops are not comparable in frequency, granularity, or independence.
The fallacy is not that Galaxy is a bad company. The fallacy is that equity exposure to a crypto financial services company is a reasonable substitute for direct market access — when direct market access is cheaper, more liquid, more transparent, and more frequently audited than it has ever been in the history of the asset class.
So What Do You Actually Do
Read the annual report. The actual filing — not the headline, not the analyst summary, not the tweet thread with the chart and the rocket emoji. Find where they drew the "core business" perimeter. Find what falls outside it. Ask yourself whether the excluded items are genuinely non-recurring noise or whether they represent the primary risk factor your capital is exposed to. In crypto financial services, the answer is almost always the latter.
If you hold Galaxy stock because you believe in the advisory and asset management franchise, that is a thesis. A defensible one, potentially. But pressure-test it by asking what percentage of the company's total risk profile is actually captured by the core profitability metric versus what percentage lives in the bucket of items the report asked you to set aside. If the second number dwarfs the first, the thesis has a structural gap that an 11% rally does not fill.
And if you are holding the stock as a proxy for crypto market exposure — reconsider the premise entirely. Open an account on an exchange with verified reserves and a recent proof-of-reserve audit. Bybit, CER security score 9.1. OKX, 9.3. Binance, 9.4. Buy the asset you actually want to own. The maker fee is 0.08% on OKX. It is 0.1% on Binance and Bybit. That is what direct exposure costs. Compare that to the bid-ask spread on a lower-float crypto equity, the management overhead embedded in whatever positions the company is carrying on your behalf, and the analytical opacity of an annual report that presents a $241 million loss as a profitability story.
Whether the "core business profitable" narrative will prove prescient — whether Galaxy's carved-out operations genuinely represent a durable, growing franchise structurally independent of the mark-to-market swings that produced the net loss — is a question that the data available today cannot settle. The market priced in its answer at 11%. I would like to see the on-chain receipts before joining that trade. If you have found a way to independently verify the core-versus-non-core split against observable data rather than taking management's framing at face value, I am genuinely interested. Write.