On September 30, shareholders of Armada Acquisition Corp. II voted to approve a merger with Evernorth Holdings. The press release that followed described a digital asset treasury armed with "approximately $300 million in gross cash," poised to accumulate XRP at scale. Two numbers live inside that sentence, and they contradict each other. The proxy statement โ precise to the decimal, filed under SEC liability โ sums the identifiable new capital available at close to $88.5 million. The gap, roughly $211 million, is not rounding error. It is positioning. When a treasury company quotes its firepower in round numbers and its audited math in decimals, the round number is a marketing instrument and the decimal is the confession.
This is not a story about XRP's price. It is a story about how capital structures are dressed to resemble conviction. The code is silent, but the ledger screams.
Digital Asset Treasuries โ DATs โ are the defining financial mutation of this market cycle. The template is familiar: take a listed shell, load it with a single volatile asset, and sell traditional investors a regulated wrapper around exposure they cannot easily custody themselves. MicroStrategy industrialized the pattern with Bitcoin. A dozen imitators followed, most of them indistinguishable. Evernorth is the first to attempt it with XRP, which makes it a case study rather than a curiosity.
The mechanics are a SPAC merger. Armada Acquisition Corp. II is the shell โ a blank-check vehicle that raised money in an IPO, parked it in trust, and now needs a target to avoid liquidation. Evernorth is the target โ a treasury company whose entire thesis is holding XRP. On completion, the combined entity trades as XRPN on Nasdaq. The structure is Delaware incorporation plus a SPAC roll-up, which means it lands squarely inside the SEC's standard disclosure framework and inherits all the governance quirks of a blank-check deal.
The pitch writes itself: institutional access to XRP without the operational burden of self-custody; a management team fluent in capital markets; the implicit blessing of the Ripple ecosystem through RippleWorks, a related entity contributing roughly 211.3 million XRP as an in-kind subscription. For an investor who wants XRP exposure inside a brokerage account, the story is clean.
The problem is that the story is not the balance sheet. I have spent twelve years reading crypto pitch decks, and the same reflex applies here that applied when I audited the Compound v1 pre-release as a final-year student: ignore the narrative, open the repository, and reconcile every claim against primary documents. Back then, the founders dismissed a critical integer-overflow finding in the interest-rate logic as a "theoretical edge case." The lesson never left me. In this case, the primary document is the proxy, and it tells a story the press release does not.
Let me walk through the capital structure line by line, because that is where the illusion lives.
Evernorth communicates "approximately $300 million" in gross cash. The precise figure is $303 million. That over-rounding โ from $303 million down to "about $300 million" โ is trivial in isolation. What is not trivial is what the $303 million actually contains. When I decomposed the sources, the picture fractured into four pieces, only one of which behaves like money you can spend.
First: a private placement of roughly $224.55 million. This breaks into $214.05 million of prepaid subscriptions and $10.5 million of deferred subscriptions. The prepaid portion was already deployed โ in November 2025, the company used it to buy approximately 84.37 million XRP at an average price of $2.54. That capital is spent. It will not buy a single additional token. A treasury that has already converted its cash into inventory has no purchasing power left; it has a cost basis. This is the single most important distinction in the entire deal, and it is the one the press release blurs most aggressively.
Second: a $30 million convertible note, PIK, paying 4% in kind and maturing in 2031. This is genuinely new money, but it is debt. PIK โ payment-in-kind โ means the interest compounds into principal rather than being paid in cash, a structure that flatters near-term cash flow while quietly inflating the obligation. At 4% compounding, the note is a slow dilution engine, and its 2031 maturity means no near-term principal pressure, which is precisely how leverage hides from casual readers. The debt does not announce itself. It accretes.
Third: trust proceeds of approximately $48 million. This is the SPAC's IPO capital and its investment income โ and it is the single most fragile line item in the entire structure. Trust proceeds are subject to redemption. If Armada's shareholders elect to redeem before close, that $48 million shrinks, potentially to zero. It is money that exists only as long as nobody asks for it back.
Add the identifiable, deployable new capital: $10.5 million deferred subscription plus $30 million convertible plus $48 million trust, minus fees. The gross figure lands near $88.5 million. The net is lower. So the "$300 million" and the "$88.5 million" describe the same company on the same day. The difference is that one number counts money already spent and assets that can be revoked, while the other counts only capital that could actually reach the order book.
Now the position accounting, which is where the second layer of inflation sits. Evernorth touts a post-close holding of roughly 473 million XRP. This number is presented as a single block, but it is a composite of at least three incompatible categories. Approximately 84.37 million XRP were purchased with cash โ realized, irreversible, and struck at $2.54. Roughly 211.3 million XRP were contributed by RippleWorks as an in-kind subscription โ contractual, and revocable if the merger fails. A related-party subscription of 50 million XRP is also contractual and also contingent.
Mixing these into one headline figure is the accounting equivalent of blending revenue, backlog, and pipeline into a single "opportunities" line. In the dark room of DeFi, shadows have names, and here the shadows are "purchased," "committed," and "contributed" โ three words that describe three completely different levels of certainty. An in-kind subscription from a related party is not the same as cash buying pressure. One can be withdrawn; the other cannot. Presenting them as one number is not a lie, technically. It is a blur, and blur is the tool of choice when precision would weaken the pitch.
Then there is the price, and this is where the arithmetic turns cruel. The cash position was built at an average of $2.54. At the time of writing, XRP trades near $1.53. That is a drawdown of roughly 40% on the only portion of the holding that was actually bought with money. On a cost basis of $214 million, that implies an unrealized loss in the neighborhood of $85 million. The company has not disclosed this as a mark, but arithmetic does not require a press release. Every line of code tells a story of greed; every cost basis tells a story of timing, and this timing was late. The purchase happened near a local high, and the market has since repriced the asset the company chose to marry.
The debt compounds the problem. The PIK note accrues, the trust proceeds are redeemable, and the company has no operating cash flow. A treasury vehicle of this kind generates no revenue, runs no protocol, and earns nothing from its asset beyond price appreciation. Its purchasing power is entirely a function of external financing. When the asset it holds is falling and the financing window is narrowing, the structure has no internal engine to fall back on. There is no product, no fee stream, no protocol revenue โ only a balance sheet that breathes in when capital is cheap and suffocates when it is not.
I want to be precise about what this is and is not. It is not, on the current evidence, a Ponzi. A Ponzi requires paying old investors with new investors' money under a false promise of returns. Evernorth is more accurately a leveraged, single-asset bet wrapped in a listing. But the structural similarity that matters is this: the company's capacity to keep buying depends on a rising price and continued access to capital, and those two conditions reinforce each other in both directions. When price falls, financing tightens; when financing tightens, buying stops; when buying stops, the accumulation narrative dies. The feedback loop is not fraud. It is fragility, and fragility is harder to see because it looks like a strategy until the day it does not.
The market math reinforces the point. The ~$88.5 million of gross deployable capital, even if it arrives intact, is small against XRP's daily trading volume, which routinely runs into the hundreds of millions to billions of dollars. The new buying pressure, in the best case, is a rounding error against the tape. So the price impact that the treasury narrative implies is not supported by the size of the wallet. The expectation is being priced, not the flow. And expectations, unlike cash, can be revised downward in a single session.
Now the regulatory layer, because a Nasdaq listing invites scrutiny that a token sale never faces. XRPN the stock is unambiguously a security โ there is no Howey analysis required for a Delaware corporation's common equity. The interesting questions sit elsewhere. First, the disclosure gap between "approximately $300 million" and $303 million is itself trivial, but the framing that surrounds it โ a public narrative of purchasing power that the proxy does not support โ could, if deemed materially misleading, attract SEC attention. I rate that probability low, because round-number summaries are common and the precise figure was disclosed somewhere. Second, the RippleWorks subscription is a related-party transaction with a revocability clause. Related-party terms that favor the contributor over the listed company deserve disclosure scrutiny, and the asymmetry here is stark. Third, and most conspicuously absent: the lock-up schedule. For a SPAC merger, the lock-up governs when insiders can sell, and it is often the single most important variable for post-close price. Its absence from the information available to retail is a hole, not a detail.
The Ripple dependency deserves its own paragraph, because it cuts both ways and the asymmetry is the story. RippleWorks is tied to the Ripple ecosystem. Its contribution gives Evernorth credibility, resources, and a story. But the clause that lets RippleWorks withdraw its in-kind contribution if the merger fails means Ripple holds an option that Evernorth does not. The most important shadow in this room is the one that can walk out the door. Evernorth depends on Ripple for assets and endorsement. Ripple does not depend on Evernorth for anything material. That is a weak bargaining position dressed as a partnership, and it means the listed company carries the downside while its benefactor retains the exit.
Finally, the date anomalies, which I will not paper over. The source material mixes 2025 and 2026 timestamps โ an October 1, 2025 announcement, a September 30 vote, a November 4, 2025 purchase disclosure, and an Armada balance sheet dated June 30, 2026. Either the text is a retrospective written from 2026, or the dates are mislabeled, or it is a scenario exercise. I flag this honestly because it matters for one reason: if the November 2025 purchase at $2.54 and the current $1.53 price are both accurate, the loss is real and recent. If the timeline is scrambled, the loss is still directional โ the cost basis sits above the market. The conclusion survives the uncertainty, which is the only thing an auditor can ask of incomplete data.
Let me steelman the bull case, because a fair audit does not stack the deck.
The first-mover argument is legitimate. Evernorth is the first Nasdaq-listed XRP treasury. If institutional demand for regulated XRP exposure is real โ and the post-ETF Bitcoin experience suggests it is โ then a listed vehicle with Ripple's implicit backing has a genuine head start. Being first in a category that might matter is not nothing. It is the one asset a competitor cannot buy, and category leadership compounds if the category grows.
The access argument is also real. Many traditional allocators cannot hold XRP directly โ custody, compliance, and mandate constraints get in the way. A brokerage-account wrapper solves a real problem for a real audience. The DAT model exists because the demand exists, and dismissing the demand because the vehicle is imperfect is its own kind of error.
And the redemption risk cuts both ways. If Armada's shareholders do not redeem, the trust proceeds hold, and Evernorth enters the market with more capital than skeptics expect. Low redemption is the bull's best card, and it is genuinely unpredictable from the outside.
But here is the flaw in the bull case: none of it addresses the purchasing-power question. First-mover advantage does not change the fact that 84 million XRP were bought at $2.54 and the market is at $1.53. Regulatory access does not change the fact that the "new" capital is closer to $88.5 million than $300 million. The bulls are right about the category and wrong about the arithmetic. A strong position in a growing category is still a weak position if the entry price was 40% above the market and the advertised firepower is triple the reality. Conviction is not a substitute for a cost basis, and narrative is not a substitute for cash.
The verdict is not that Evernorth will fail. It is that the market is being asked to price a narrative โ $300 million, 473 million XRP โ that the underlying documents do not support. The real numbers are roughly $88.5 million of deployable gross capital and a cost basis 40% underwater. The close, around October 7, is the hinge: redemption rates, convertible note settlement, and actual cash on arrival will decide whether the treasury has any powder left. Watch the decimal, not the headline. The round number is written to be quoted; the precise number is written to be audited. And in a bear market, the only question that matters is not how much a company says it can buy โ it is how much it can actually spend when the price it paid has already turned against it. Beneath the surface, the truth is compiled in hex. Read it.