SkyAI's $201M Solana Exit: Decompiling the First Digital Asset Treasury Tender Offer

The Instrument Problem

The word was "tender offer."

Not buyback. Not repurchase. Not burn. Tender offer — a term of art that arrives with a Schedule TO, a commencement date, a withdrawal period, an odd-lot provision, a paying agent, and a termination date. Crypto-native treasuries do not speak this language. They say "we are repurchasing tokens." They say "we are retiring supply." They publish a wallet address and a transaction hash and let the chain do the testifying.

The moment the phrase "tender offer" appears attached to a Solana treasury, the document stops being a crypto announcement and becomes a securities disclosure wearing crypto clothing. Someone paid a law firm. Someone opened a filing account. Someone accepted a reporting obligation that did not previously exist.

That single word is the most reliable datum in the entire story.

Here is what was actually put into the world: SkyAI, a Solana treasury vehicle, sold Solana from its balance sheet and simultaneously announced a $201 million tender offer, framed as a strategic pivot away from its Solana position. Four facts. Zero primary sources. No filing number. No transaction hash. No custodian attestation. No signature block from a general counsel. No date stamp on the announcement itself.

Zero trust is not a policy; it is a geometry. And the geometry of this disclosure has a hole in it exactly where the load-bearing wall should be. When the load-bearing wall is missing, you do not assume the roof holds. You go find the wall.

What follows is not a verdict on SkyAI. It is a map of where the verdict will be found, which variables are decisive, and which of the two mutually exclusive events actually occurred. In a sideways tape, where positioning matters more than direction, the ability to distinguish a signal from a press release is the entire edge.

What a Digital Asset Treasury Actually Is

Start with the category, because the category determines what questions are even legal to ask.

A Digital Asset Treasury — DAT — is a listed or quasi-listed vehicle whose sole product is balance-sheet exposure to a single crypto asset. It does not build. It does not operate infrastructure. It holds, and it monetizes the spread between what the market pays for its shares and what the underlying assets are worth. The 2025 Solana cohort formed fast: Forward Industries, DeFi Development Corp, Upexi, SOL Strategies, Sharps Technology, and a long tail of smaller shells. Most of them announced within months of each other, because most of them were responding to the same arbitrage window.

That window has a name in this business: mNAV. Market-to-net-asset-value. Share price divided by per-share value of the crypto held, net of liabilities.

mNAV above one means the market is paying a premium for your wrapper. That premium is not a bug in the market's pricing engine; it is the product. When mNAV is 2.0, you can issue shares at two dollars, buy one dollar of Solana, and every existing shareholder's per-share NAV rises by fifty percent before a single validator produces a block. Do the arithmetic on a clean book: one hundred shares, one hundred dollars of NAV, a share price of two dollars. Issue one hundred new shares at two dollars. Two hundred dollars of cash arrives, converts into two hundred dollars of Solana. NAV is now three hundred. Share count is two hundred. NAV per share has gone from one dollar to one dollar fifty. That is a fifty percent accretion produced entirely by the wrapper's premium, with zero operating contribution.

This is the flywheel. It is a financial engineering loop, and it is mathematically identical to a currency board operating in reverse.

Now run it backwards. mNAV at 0.7. The wrapper trades at a thirty percent discount to the assets inside it. Issuing shares now destroys value — you would be handing out one dollar of NAV for seventy cents. The flywheel does not slow; it inverts. The only rational response left is to buy the wrapper instead of selling it, because retiring equity at a discount is the mirror image of issuing equity at a premium. Every share you retire costs seventy cents and removes a dollar of NAV.

Every reported fact about SkyAI is consistent with a company executing the inversion. Sell the asset, buy the wrapper. It is the only self-consistent reading of "sold Solana" plus "$201 million tender offer" appearing in the same announcement.

But — and this is the part the commentary class skipped — the inversion is not free money. If the repurchase is funded by selling the underlying Solana, the NAV shrinks as the share count shrinks. The accretion is not the full thirty-cent gap; it is only the discount capture, and the capture scales with how much of the treasury you are willing to liquidate. Retire fifteen percent of the share count at a thirty percent discount and NAV per share rises about five percent. Retire half the share count and NAV per share rises roughly seventy-five percent, and you no longer own a treasury company, you own a shrinking pool of cash with a stock ticker attached.

So the number that matters is not $201 million. The number that matters is $201 million divided by net asset value. That ratio is the entire story, and it is absent from every version of this report I have seen.

The Flywheel, Decompiled

The genealogy here is not speculative. GBTC wrote the manual, and it wrote it in blood.

Between 2021 and 2022, GBTC traded at a premium, arbitrage capital flooded in, the trust kept issuing, the premium turned, issuance stopped, and the discount went as deep as roughly forty-five percent — and stayed there, because the trust had no redemption mechanism. Holders were locked inside a wrapper with no exit. That is the canonical failure mode of a premium-financed holding vehicle: the flywheel's energy source is the premium, and the premium's energy source is the expectation of more premium.

The critical structural difference with a DAT is that the DAT can sell the asset. It has a redemption channel. That channel is real, and it compresses discounts in a way GBTC structurally could not. I will give the bulls that point in full later, because they are right about it and most of the bearish commentary is wrong about it.

But a redemption channel is a capability, not a behavior. Capability becomes behavior only when management decides to use it, and management's decision is not governed by code. It is governed by a board, an incentive package, and a set of lock-ups.

Run the reflexivity math honestly. The DAT model's return composition is roughly six to eight percent annualized from staking rewards on native Solana — assuming the treasury is staked rather than idle — against a share price whose variance is an order of magnitude larger. More than ninety percent of the model's return comes from NAV premium dynamics, not from operations. By any standard decomposition, that is a structure whose revenue base cannot support its valuation, and the sustainability label should read: conditional on continued access to premium issuance.

Strictly speaking, this is not a Ponzi. A Ponzi promises fixed returns to early entrants paid from late entrants' principal. A DAT promises nothing and delivers whatever the wrapper does. But the reflexivity intensity is equivalent. Valuation depends on the expectation of the next premium issuance, and once that expectation breaks, the mechanism self-reinforces downward.

Which brings us to the first genuinely new insight available in this event, and it is not about Solana at all. SkyAI's announcement is the first documented instance of a Solana DAT acknowledging, in public and in a securities-law vocabulary, that its own wrapper was cheaper than its own contents. That is not a balance-sheet operation. That is a confession with a price tag attached.

The Filing That Isn't There

The absence of a primary source is not a minor sourcing complaint. It is the load-bearing finding, and it has a specific shape.

If SkyAI is a reporting company, the tender offer exists in a specific document, and the document's type alone resolves most of the ambiguity. The taxonomy is not decorative; each form tells you a different story about who is buying from whom.

| Filing | What it means | What it would settle | |---|---|---| | Schedule TO-I (issuer tender offer) | The company is buying its own securities | Self-tender; NAV shrinks; remaining holders gain the discount capture | | Schedule TO-T (third-party tender offer) | An outside party is buying the company's securities | Control event; the Solana sale may be a condition or a financing step | | SC 13E-4 | Issuer self-tender, older form path | Confirms buyback intent and the funding source representation | | Schedule 14D-9 | Target's response to a third-party offer | Reveals whether management recommends, rejects, or is neutral | | Form 4 / 13D / 13G | Insider and 5% holder activity | Reveals whether insiders are selling into the event |

I have written this before and I will write it again: the code does not lie, but it often omits. Filings are the closest thing this industry has to code. They omit less. A Schedule TO must state the source of funds, the purpose of the transaction, the identity of the offeror, the consideration per security, and the expiration date. A press release must state none of these things.

So the first question is not "is this bullish or bearish for Solana." The first question is: does a Schedule TO exist, and what is its accession number? If it does not exist, and the "tender offer" language was used loosely, then the entire securities-law reading collapses and we are back to an unverified corporate statement — which is a materially different, and materially worse, evidentiary posture.

There is a strong inferential signal here. A $201 million offer requires a dealer-manager, a depositary, a paying agent, and legal counsel. That execution stack costs low single-digit millions and cannot be assembled retroactively. The existence of the number is therefore weak evidence that real advisors were engaged. But weak evidence is not evidence, and I do not price positions on weak evidence.

When I ran the 2x2x4 audit in 2017, the thing that saved investors was not the elegance of my report. It was that I had a testable artifact — a Python simulation of the flash-loan path, with a reproducible output — that anyone could rerun. The lesson generalized: an accusation without a reproduction is an opinion. An accusation with a reproduction is a finding. Right now this event has an accusation-shaped hole and no reproduction.

Self-Tender or Third-Party: Two Events Wearing One Headline

This is the fork that most coverage has flattened, and flattening it is not a minor analytical sin. The two scenarios point in opposite directions.

Scenario A — issuer self-tender. SkyAI sells Solana, uses the proceeds to buy back its own shares at a discount, and continues as a smaller, more concentrated vehicle. Meaning for Solana: one large holder becomes a seller, and a marginal buyer is removed from the bid. Meaning for SkyAI holders who stay: NAV per share rises, concentration rises, and the company's remaining asset base shrinks. Meaning for the DAT narrative: an admission that the wrapper was mispriced. Direction: bearish for the sector's growth story, mildly bullish for remaining shareholders of this specific vehicle.

Scenario B — third-party tender. An outside party is acquiring SkyAI's securities. The Solana sale is then either a condition precedent, a financing step, or a defensive maneuver. Meaning for Solana: the same directional supply increase, but with a different temporal profile. Meaning for SkyAI holders: a control transaction, which prices at a premium to market and is judged against a fairness opinion. Meaning for the DAT narrative: consolidation rather than capitulation — which is a very different signal. Direction: ambiguous for the sector, potentially bullish for the target's holders, and completely unrelated to the "DAT exit" framing that dominated the commentary.

The observable difference is clean. An issuer tender carries a Schedule TO-I and a Rule 13e-4 filing, and the company is on both sides of the trade. A third-party tender carries a Schedule TO-T plus a Schedule 14D-9 from the target within ten business days, and the target's board must state a position.

If no 14D-9 appears within ten business days of the commencement date, Scenario B is dead and the entire "acquisition" reading was fabricated by whoever wrote the headline. That is a dated, binary, falsifiable test. It is the single most useful thing anyone can do with this story right now.

There is a third, uglier possibility that nobody has raised. A tender offer can be used to concentrate ownership at a discount before a strategic pivot that would otherwise require shareholder approval. Buy the wrapper at seventy cents, take control, and then redirect the corporate purpose — toward AI, given the name — without ever paying a control premium. I have no evidence for this. I flag it because the pattern is common enough in shell-adjacent vehicles that a forensic reader should hold it as a live branch rather than a closed one.

The Unstake Queue Is a Clock

Here is where the technical layer actually lives, and it is not where the headlines put it.

Selling Solana from a treasury is not a token transfer. It is an engineering operation with a latency budget, and the latency is not negotiable.

If the treasury holds native staked SOL — which is the standard configuration for a Solana DAT, because unstaked SOL forgoes six to eight percent and that forgone yield directly damages the mNAV story — then the assets are bound by the epoch mechanism. A deactivation instruction is submitted, it takes effect at an epoch boundary, and the practical elapsed time is on the order of two to three days. During that window the SOL is neither earning nor transferable.

That creates an observable, exploitable information window between the announcement date and the date the supply actually becomes sellable. Anyone who can determine the treasury's stake accounts can watch the deactivation transactions land, timestamp them, and model the settlement date to within a few hours. That is a cleaner signal than any headline, and it is free.

The window is not a Solana flaw. It is a Solana property, and it is a good one — it is the mechanism that prevents bank-run dynamics in the validator set. But properties have consequences, and one consequence is that a large treasury cannot exit silently. The chain timestamps the intention before the market sees the execution. Compiling the truth from fragmented logs is the entire discipline; the deactivation transaction is the first fragment, and it is public.

The alternative configuration changes the problem's shape rather than solving it. If the treasury holds a liquid staking token — JitoSOL, mSOL, or an equivalent — the holder can exit instantly by selling the LST rather than unstaking the SOL. The latency disappears. The cost reappears as a discount.

That trade-off is the most interesting structural fact in this entire event, and it generalizes: you cannot escape the cost of liquidating a large position. You can only choose the currency. You pay in latency, or you pay in discount, and the market continuously prices the exchange rate between the two. A deep LST market means latency is cheap and discount is the tax. A thin LST market means the reverse. For a $201 million position, the LST route likely prices worse than the two-to-three-day queue, because $200 million of single-direction LST flow will push the secondary price below fair NAV and invite arbitrageurs to redeem and unstake themselves — which just moves the queue to someone else's balance sheet.

I spent a week in 2024 mapping EigenLayer's restaking risk surface, and the lesson that transferred cleanly is this: shared security models hide their failure modes in the seams between operator sets. The same is true of shared liquidity. The LST discount is a seam. It is invisible in calm markets and it is where the damage concentrates when a single holder needs to move size.

The Arithmetic of $201 Million

Numbers first, adjectives never.

Solana's market capitalization sits somewhere in the eighty-to-one-hundred-twenty-billion dollar band depending on the week. Daily spot volume across major venues runs roughly three to six billion. Against those denominators, $201 million is about two-tenths of one percent of market cap and somewhere between three and seven percent of a single day's traded volume.

Those are not catastrophic numbers. They are visible numbers. The distinction matters, because the market's reaction function is not linear in size — it is a step function that triggers on the question "is someone large and price-insensitive selling," not on the quantity.

Execution mechanics set the realized cost:

A negotiated block through an OTC desk, pre-arranged with a single counterparty or a small syndicate, prices at roughly one to three percent below mark. The desk warehouses the position and works it out over weeks. The market sees nothing until the desk's own hedging flow shows up in the tape as persistent, directionless selling — which is exactly the pattern that has been misread as "organic distribution" in a dozen prior episodes.

A time-weighted average price program over five to ten days produces a similar band, one to three percent, but with a visible footprint: consistent intraday selling into strength, a flattening of rallies, and a bid that thins whenever price approaches a round number.

A single-day market order produces five to eight percent of slippage and a permanent mark on the chart. Nobody with a $201 million problem and a lawyer on retainer chooses this.

The correct prior is that the liquidation was already substantially complete, or its counterparty already locked, before the announcement was published. The reasoning is not cynicism; it is arithmetic. A company cannot credibly commit $201 million to a tender offer unless the funding is committed, and funding is not committed until the assets are sold or financed. The announcement is therefore more likely a disclosure of a completed operation than a declaration of intent.

That reading has a direct consequence for anyone trying to trade the news: the price impact, if any, already happened before you read about it. What remains is the signal, not the flow.

The Signal Is Not the Size

The market's error is measuring this event in dollars. The correct unit is direction of marginal demand.

A DAT is not a holder. A DAT is a demand channel. Its function in the Solana market is to convert equity capital into spot bids on a schedule. When mNAV is above one, that channel is accretive and it buys continuously, and its buying is indistinguishable from organic institutional demand in every dataset except one: the buying stops the moment the premium stops.

So the event is not "$201 million of Solana sold." The event is "a marginal buyer converted into a marginal seller." Those are different objects. The first is a flow. The second is a change in the slope of the demand curve.

Flows mean-revert. Slope changes do not. A holder who sells can buy back. A channel that inverts does not re-enter until the arithmetic that inverted it is repaired, and the arithmetic is repaired only by either the discount closing or the treasury being exhausted.

My 2020 Curve analysis taught me this in a different register. The veCRV model looked like decentralized governance; the actual mechanism was that a small number of whales could direct reward allocation because voting weight was purchasable and time-weighted in a way that punished small holders. The lesson was not that Curve was broken. The lesson was that complex financial engineering routinely masks a simple power dynamic, and the power dynamic is always visible in the flow of marginal incentives, never in the governance diagram. Applied here: the DAT's governance diagram says shareholders own a treasury. The marginal incentive structure says a small number of managers decide, quarter by quarter, whether the treasury exists.

Custody, Keys, and the Ronin Precedent

A $201 million Solana position does not sit in a wallet controlled by one person. It sits in a custodian account, a multi-signature arrangement, or a staking provider's omnibus. Moving it is an operational event with its own failure surface, and the failure surface is not theoretical.

In 2021 I audited the Ronin sidechain architecture ahead of the Axie Infinity scale-up. What I found was not a clever cryptographic break. It was a validator threshold set low enough that a small number of compromised keys constituted a quorum, combined with key management practices that treated operational convenience as a design constraint. I disclosed confidentially. The disclosure was downplayed. Months later, $625 million left the bridge through exactly the seam I had flagged.

The generalizable finding was not about Ronin. It was that scalability and convenience solutions systematically price security as an externality, and the externality is paid by whoever is holding when the seam opens.

Now apply it. A DAT treasury liquidation requires: authority to instruct the custodian, authority to instruct the staking provider to deactivate, authority to instruct the OTC desk, and authority to move settlement funds into the tender offer's paying agent account. That is four distinct authority paths, each with its own key material, each with its own social-engineering surface, each executed under time pressure because the tender offer has a commencement date.

Time pressure is the single largest contributor to key-management error in this industry. Every major custody failure I have examined involved a deadline.

The observable that resolves this is a custodian attestation — a signed statement identifying the accounts, the balances, and the control arrangement, ideally with on-chain verification of the addresses. Absent that, the public is asked to accept a number from a press release. I do not accept numbers from press releases. After FTX, nobody with a functioning memory should.

Who Votes, and With What Weight

A tender offer is a governance event before it is a market event, and the governance structure determines who actually gets to decide what the vehicle becomes.

Three questions, in order of decisiveness. First: who holds the voting power, and is it concentrated in a sponsor class with supermajority rights? Second: does the tender offer carry a non-participation condition that effectively forces small holders to tender — a "short-form merger" style threshold, or a condition that the offer closes only if a specified percentage is tendered? Third: what do the insiders do during the offer window?

The third question is the one that leaks the truth. Insider participation in a tender offer is disclosed. If insiders tender pro rata, they believe the discount is the whole opportunity and the residual asset is uninteresting. If insiders decline to tender while retail does, they believe the residual asset is worth more than the offer price — which is a bullish signal for the wrapper and a bearish signal for everyone who tendered. If insiders tender disproportionately, the offer is an exit vehicle for the sponsor and the disclosure will show it.

The Form 4 trail is the honest document in any tender offer. It is filed under penalty and it is dated.

This is also where my long-standing position on governance design becomes operationally relevant rather than rhetorical. Optimism's RetroPGF works because the allocation mechanism is legible, the funding is programmatic, and the recipients are public before the money moves. Almost every other grant committee I have examined operates on relationship capital dressed as merit. A DAT board making a $201 million capital allocation decision is structurally the same problem: a small committee deciding the disposition of pooled capital with no continuous, legible mechanism for the pool's owners to ratify the decision in real time. The difference is that the DAT committee's decision is disclosed after the fact in a filing, which is more transparency than most DAOs offer — an uncomfortable comparison that the on-chain governance crowd will not enjoy reading.

The Second-Order Trade: Contagion to the Cohort

The first-order effect is a Solana sale. The second-order effect is a repricing of the entire DAT cohort, and the second-order effect is larger.

Model the cohort's economics. Every DAT in the Solana set depends on the same input: mNAV above one, which depends on the market's willingness to pay a premium for a wrapper, which depends on the belief that other buyers will keep paying a premium for the wrapper. That belief is a shared resource. It is non-excludable and rivalrous — a common-pool asset, in the precise economic sense.

The first credible exit consumes part of that shared resource. Not because $201 million is large relative to Solana, but because it is the first demonstration that the exit is available and that someone chose it. Once the option is exercised, it is priced. Before SkyAI, "the DAT can always sell the asset" was a theoretical comfort that supported premiums. After SkyAI, it is an observed behavior that caps them.

Run the cascade concretely. Premiums compress. A DAT trading at 1.4 drops to 1.1. At 1.1 the accretion from issuing is thin but positive, so issuance continues at reduced pace. At 0.95 issuance stops entirely, and the vehicle's only rational action is to stop buying Solana. The bid that the cohort collectively represented simply disappears — not because anyone panicked, but because the arithmetic that justified the bid no longer holds.

That is the real Solana supply risk: not the $201 million that was sold, but the several billion that will not be bought.

I have watched this exact shape before, in the FTX post-mortem. What made the collapse predictable was not the fraud itself — fraud is always visible in hindsight. It was that the balance sheet's composition was structurally incapable of surviving a withdrawal wave, and the composition was disclosed in fragments that anyone could have assembled. I traced the commingled flows from the exchange to the trading arm using nothing but a block explorer and a spreadsheet. Eight billion dollars of commingling, visible in public data, months before the bankruptcy filing. Nobody looked because nobody wanted the answer.

The DAT cohort's balance sheets are the same kind of artifact. The holdings are, in principle, verifiable on-chain. The mNAV is computable from public prices and disclosed share counts. The entire cascade I just described is arithmetic on public data. The reason nobody is running it is not that the data is hidden. It is that the answer is unwelcome.

Watch three leading indicators, in order of reliability. First, ATM program suspensions announced by other Solana DATs — a suspension is an admission that issuance is no longer accretive. Second, insider Form 4 sales that precede a treasury announcement by one to three months. Third, the discount at which any DAT's shares trade relative to a freshly computed NAV, updated weekly. All three lead the headline by weeks.

The AI Fork

The name contains "AI," and that is not a cosmetic detail. It is the largest unexamined branch in the whole analysis.

If SkyAI is a digital asset treasury vehicle that happens to have an AI-adjacent brand — a common pattern in the 2025 cohort, where the ticker does narrative work the balance sheet cannot — then everything above holds, and the event is a rational capital allocation inside a mispriced wrapper.

If SkyAI is instead an operating AI project that held Solana on its own balance sheet as treasury, then the meaning inverts completely. Selling treasury assets to fund a buyback is no longer discount capture. It is a liquidity event. An operating company that liquidates its reserve to retire equity is telling you something about its cash flow, and what it is telling you is not good. The severity is an order of magnitude higher: it is not a vehicle adjusting its wrapper, it is a business funding itself from its savings.

The two scenarios are distinguished by one question: does the company have revenue, and from what? A DAT has staking yield and nothing else. An AI project has customers or it does not. That question has an answer in any audited financial statement, and it has no answer in a press release.

I raise this not to speculate but because the conflation of "treasury company" and "company with a treasury" is the most common category error in this sector, and every downstream conclusion depends on which one you are looking at. A reader who assumes the wrong one will get the sign of the trade backwards.

What the Bulls Got Right

I have spent several thousand words taking this event apart. Now the part that most of the bearish commentary gets wrong, which is nearly all of it.

First: the redemption channel is real, and it is a genuine structural improvement over GBTC. GBTC holders were trapped because the trust had no mechanism to sell the underlying and retire shares. A DAT can. That capability compresses discounts in a way that was mathematically unavailable to the trust, and any analysis that treats this event as "GBTC 2.0" is importing a failure mode that the structure has actually solved. The discount capture is not a gimmick; it is arbitrage, and arbitrage is how mispricings die.

Second: selling an overvalued asset to buy back undervalued equity is textbook capital allocation. If the wrapper trades at 0.7 and the underlying trades at par, the highest-return use of the underlying is to retire the wrapper. Criticizing this is criticizing arithmetic. A management team that did not do this would be destroying value for the remaining holders, and would deserve the criticism it got for not doing it.

Third: the Solana protocol is untouched. Validator count, client diversity, throughput, DeFi total value locked — none of these variables enter the calculation. A treasury sale changes who holds the asset, not what the asset does. Every claim that this event represents a technical setback for Solana is a category error, and I say that as someone whose entire professional output consists of finding technical setbacks.

Fourth, and least popular: forced selling improves holder composition. A holder base that includes a leveraged, premium-dependent vehicle is fragile in a specific way — it sells at the worst moment because it has no choice. Removing that holder replaces a price-insensitive seller with price-sensitive ones. In a sideways market, that is a structural improvement in the marginal holder's quality, and it usually shows up as reduced realized volatility three to six months later, not as a permanent discount.

Fifth: the base rate for reading one filing as a sector verdict is terrible. One company's treasury action has near-zero predictive power for an asset class. The commentary that treated this as "the DAT era is over" is doing the exact thing it accuses the bulls of doing — extrapolating a narrative from a single data point.

Where the bulls are wrong is narrower and more important. They are right that the mechanism is sound. They are wrong to treat the mechanism's soundness as evidence about management's competence, the filing's existence, or the funding's reality. A correct model executed with unverified inputs produces a confident wrong answer. That is the failure mode here, and it is symmetric: the bears are making the same error with the opposite sign.

The Question Nobody Is Asking

Three falsifiable tests, all cheap, all decisive.

Find the Schedule TO. If it exists, read the source-of-funds section and the purpose-of-transaction section, and the ambiguity evaporates. If it does not exist, the securities-law reading was never warranted and the entire event should be reclassified as an unverified corporate statement. Find the target's Schedule 14D-9 within ten business days of commencement; its absence kills the acquisition scenario permanently. Find the custodian attestation and the on-chain addresses, and reconcile them against the disclosed balance.

None of these require special access. All of them require someone to look, which is the scarce resource in this industry.

What the next quarter should be watched for is not the Solana price. It is whether a second DAT does the same thing. One exit is an event. Two exits is a regime, and a regime is what the mNAV cohort's entire valuation depends on not existing.

The forward-looking question is uncomfortable, and it is the one I would put to the board directly: if the highest-return trade available to this vehicle is to purchase itself, then at what point does the vehicle stop being a treasury and start being a liquidation with a mailing address? And if the answer is "when the discount closes," then the product was never the Solana. The product was the discount. The Solana was just the collateral that made the discount legible.

Security is the absence of assumptions. The assumption here — the one nobody has written down — is that a treasury company will always prefer to hold the asset. This event is the first public evidence that the preference is conditional. Conditional preferences are not commitments. They are positions, and positions get closed.