Three numbers arrived attached to the same event, and they do not reconcile.

The first is 1,665 — the bitcoin Strategy reported acquiring. The second is $143 million, the stated consideration. Divide one by the other and the average execution price is $85,886 per coin. The third is 847,666, the total position, valued in the same disclosure at approximately $70 billion. Divide those and the implied mark is $82,580 per coin.

The gap between $85,886 and $82,580 is 4.0%. Not large. Not zero either. In a disclosure built from arithmetic, a 4% residual between the marginal price and the average mark is a signal — it means one figure is stale, or rounded, or drawn from a different timestamp than the other. A treasury operation of this scale does not publish sloppy numbers by accident. It publishes numbers selected for effect.

Then there is the percentage everyone repeated. Strategy now holds roughly 4% of bitcoin's supply. That is true if your denominator is the 21-million protocol cap. It is 4.26% against the roughly 19.9 million coins actually mined. Against coins that can plausibly change hands, it is north of 5%. Each denominator is defensible. Each produces a different headline. The one that spread is the smallest.

That is the shape of the problem. Not fraud. Not necessarily deception. Selection — a set of individually true statements arranged so the eye lands on the most flattering one.

I spent six weeks in 2020 reverse-engineering the exploit path of a yield aggregator that lost $30 million of user funds. The lesson from that work was not that the code was broken. The code did exactly what it said it would do. The lesson was that the disclosures around the code were structured to make the broken part invisible until it wasn't. Logic does not bleed, but code leaves traces. So do filings. So do press releases.

What follows is a teardown of the 1,665-coin purchase — not as a market event, because $143 million is a rounding error against bitcoin's daily turnover, but as a structural one. The purchase itself is unremarkable. The architecture that makes the purchase possible is not. That architecture rests on three load-bearing assumptions. None of them is audited. All three are currently being tested by a sideways market.

Let me state my prior plainly, because it shapes everything below. I do not think Strategy is a fraud. I think Strategy is a permissionless financial primitive with a centralized operator — a category that has historically not ended well, and that has never been stress-tested at this size. Those are different claims, and I will defend the distinction rather than retreat from it.

Now the arithmetic.

The Instrument and the Machine

Strategy — formerly MicroStrategy, renamed in February 2025 — is the largest corporate holder of bitcoin on earth. That is not a controversial statement. It is a filing.

Its operating business is enterprise analytics software. Its actual business, measured by any variable that moves the share price, is the acquisition and custody of bitcoin financed by instruments sold into the public equity and credit markets. The software division generates revenue in the low hundreds of millions annually. The bitcoin position is valued in the tens of billions. The tail has not just wagged the dog. The tail has eaten the dog.

The mechanism is a reflexive loop, and it is worth stating precisely, because imprecision is where the mythology lives.

Strategy issues equity at a market price. If that price exceeds the net asset value of its bitcoin per share, each new share sold buys more bitcoin per remaining share than it dilutes — so bitcoin-per-share rises. If the price sits below that threshold, the reverse happens. The ratio of market price to net asset value per share is called mNAV. Above one, the company is a machine that converts equity premium into bitcoin accretion. Below one, the machine runs backward.

The funding stack has evolved in three visible phases, and I will return to the sequence because the sequence is the story.

Early acquisitions were financed with convertible notes — coupons between 0% and 1%, conversion premiums set 30% to 55% above the prevailing share price. Convertibles are cheap capital when rates are low and volatility is high, because the buyer is paying for optionality, not for yield. Strategy sold optionality on a stock that behaves like a levered bitcoin wrapper. It worked.

Then came the at-the-market equity program, which allowed the company to drip shares into the tape continuously instead of raising in discrete, priced blocks.

Then, in 2025, the preferred series: Strike, Strife, and Stride — perpetual instruments carrying coupons in the high single digits to 10%, with varying conversion and redemption mechanics.

I will come back to that third phase. For now, note only the direction: the cost of Strategy's marginal capital has travelled from approximately zero to approximately 8–10%, and the market has treated each step as confirmation of the model rather than a symptom of its maturation.

Against that backdrop, the 1,665-coin purchase is small. It is 0.196% of the existing position. It is roughly 0.2% of the position's stated value.

Here is a cleaner way to size it. Bitcoin's post-halving issuance is 3.125 coins per block, 144 blocks per day, or 450 coins per day. Strategy's purchase absorbed 3.7 days of network issuance. Its entire stack of 847,666 coins represents 5.16 years of current emission. Meanwhile bitcoin's daily global spot volume across every venue and pair runs in the tens of billions. $143 million is under half a percent of a single day's turnover.

Gas fees are the price of truth — or, in this market, trading fees are. And $143 million does not move the price of bitcoin. It moves the narrative of bitcoin. Those are different variables, and conflating them is the most common error I see in coverage of this company.

The regime matters here. We are in consolidation. Realized volatility has compressed. New supply is meeting demand that is patient and OTC-mediated rather than urgent and exchange-mediated. In that regime, marginal buys from a single identified buyer matter not for price but for positioning signal — they tell you where a large, informed, long-duration participant is accumulating. That is useful information. It is not a catalyst.

Two structural facts complete the picture, and both are recent.

The first is accounting. Under FASB ASU 2023-08, effective for fiscal years beginning after December 15, 2024, U.S. companies must measure crypto assets at fair value, with changes flowing through net income every period. Before this, impairment-only treatment suppressed volatility in one direction. After it, Strategy's reported earnings are a levered mark-to-market on bitcoin — a 20% quarterly move in spot translating into a multi-billion-dollar swing in reported net income that has nothing to do with operating performance.

The second is custody. Strategy does not hold its 847,666 coins in self-custody. It holds them through a qualified custodian — historically Coinbase Prime — which means the security assumption protecting the largest corporate bitcoin position in existence is not cryptographic. It is contractual. The keys sit inside someone else's operational perimeter, secured by policy, insurance, and third-party controls.

That is a defensible choice for a public company with audit obligations. It is also the precise condition the bitcoin design was built to avoid. The irony is not decorative. It is load-bearing, because it means the largest holder of the asset cannot execute a unilateral, permissionless exit. It can execute a trade — subject to settlement, counterparty, and time.

One note on disclosure mechanics before the numbers. Strategy announces purchases through two channels: a regulatory filing, typically Form 8-K or a periodic report, and a social post from Michael Saylor, the executive chairman, whose framing is consistently promotional. The post is the distribution mechanism. The filing is the disclosure. They describe the same transaction and are not interchangeable. Anyone building a position off the post is trading a marketing artifact. Anyone building one off the filing is trading a fact.

When I autopsied 45 whitepapers in 2017 from projects that raised more than $2 million apiece, my working rule became: read the artifact last and the filing first. This brief carries no source attribution at all. Treat it as a hypothesis, not a datum.

The Arithmetic of Three Numbers

Work the numbers.

$143,000,000 ÷ 1,665 = $85,885.9. Call it $85,886 — the marginal execution price of this tranche, inclusive of fees and slippage as reported.

$70,000,000,000 ÷ 847,666 = $82,579.8. Call it $82,580 — the implied mark on the whole position.

The difference is $3,306 per coin, or 4.0%.

A 4% residual between an execution price and a mark has exactly three plausible explanations, and they are not equally interesting.

One: rounding. The $70 billion is a round number. If the true value were $70.6 billion, the implied mark would be $83,286 and the gap would still be 3.1%. Rounding to the nearest billion cannot account for four points. Discard.

Two: time skew. The $143 million is stamped at execution. The $70 billion is stamped at publication or at a period close. If bitcoin traded at $85,886 when the tranche settled and $82,580 at the mark date, the residual is a calendar artifact — the position simply lost roughly 4% between the two timestamps. Plausible, and boring.

Three: different accounting bases. One figure is a cash cost basis; the other is a fair-value mark. This is the most likely explanation and the least alarming.

I favor three with a time-skew overlay. But notice what that combination produces. The headline 4% supply ratio is computed on a denominator selected to minimize the ratio, while the $70 billion headline is a fair-value mark selected to maximize the headline. Neither statement is false. Both are chosen.

Selection is not unique to Strategy. It is standard procedure for any corporate treasury holding a volatile asset. But a detective's job is to notice when selection is systematic. Across four numbers — 1,665, $143 million, 847,666, $70 billion — the bias runs in the same direction every time.

The Denominator Problem

Supply ratios are the most abused statistic in this industry. I have watched the same protocol claim "X% of supply locked" with four different denominators across four quarters and win a narrative with each one.

For bitcoin the honest answer is that there is no single supply ratio. There is a family of them, and the spread between extremes is wide enough to change the argument.

| Denominator | Value | Strategy's share | |---|---|---| | Protocol cap | 21,000,000 | 4.04% | | Mined supply (est.) | ~19,900,000 | 4.26% | | Mined minus provably lost coins | ~16,000,000 | ~5.30% | | Current annual issuance | 164,250 | 516% |

The final row is the one nobody prints. Against a single year of new supply, Strategy holds 5.16 times the entire annual emission of the network. That is not a percentage. It is a structural position.

The 21-million denominator is the most cited and the least meaningful, because nobody can spend the coins attributed to genesis-era wallets, and several million more are presumed unrecoverable through lost keys, destroyed drives, and abandoned early wallets. Against coins that can actually change hands, the ratio is meaningfully higher than 4%.

I want to be explicit about uncertainty here, because precision laundering is a habit I try not to reward. Chainalysis's 2020 estimate put permanently lost coins near 3.7 million. Later work has revised in both directions. I use a working band of 3.5 to 4.0 million, which produces a band of ratios between roughly 5.2% and 5.4%. That is a band, not a point. Anyone quoting 5.3% to three significant figures is performing confidence, not computing it.

There is a second-order effect that most analysts skip. Supply that is not lost but dormant behaves like lost supply for price purposes. Coins in wallets that have not moved in five years are unavailable at any price short of a regime change. Illiquid-supply metrics — coins held by entities with minimal spending history — have hovered near three quarters of mined supply across the last several cycles. Exclude both lost and long-dormant coins and the effective float that clears price discovery is smaller still, and Strategy's position against it larger.

I will not quote a precise number for that adjusted float, because past a certain point the methodology choices swamp the signal. The direction, however, is unambiguous. Every reasonable adjustment increases Strategy's share of tradeable supply. None decreases it.

The denominator is a variable, not a constant. Whoever selects it selects the conclusion. The 4% figure is the floor of the distribution, not the center of it.

The Flywheel, Derived

Now the mechanism. I want to derive it rather than describe it, because derivation produces a threshold, and thresholds are where positions get made.

Let B be the bitcoin held, N the shares outstanding, P the share price, and S the bitcoin spot price.

Bitcoin per share is B/N. Net asset value per share is (B × S)/N. Therefore:

mNAV = P ÷ [(B × S)/N] = (P × N) / (B × S)

Now suppose Strategy issues ΔN new shares at price P and spends the proceeds buying bitcoin at spot S. New holdings are B + (P × ΔN)/S. New share count is N + ΔN.

Bitcoin per share is accretive if and only if:

(B + PΔN/S) / (N + ΔN) > B/N

Multiply both sides by N(N + ΔN), which is positive:

N(B + PΔN/S) > B(N + ΔN) NB + NPΔN/S > BN + BΔN NPΔN/S > BΔN NP/S > B (P × N) / (B × S) > 1 mNAV > 1

That is the whole model in one inequality. Equity issuance is bitcoin-per-share accretive if and only if mNAV exceeds one. The software business, the convertibles, the preferreds, the marketing cadence — the entire apparatus exists to defend a single line.

In practice the threshold sits slightly above one, because ATM commissions and underwriting run one to two percent. Breakeven lands near 1.02. Below that, issuance is dilutive and the machine runs in reverse.

This is why the purchase size matters less than the premium. 1,665 coins at a favorable mNAV adds more per-share bitcoin than 16,650 coins at an unfavorable one. The governing variable is the ratio, not the tonnage. And the reported 1,665 tells you nothing about the ratio at which the funding was raised.

Here is the part that deserves more attention than it receives. mNAV is an output. Strategy does not control it. The company can influence it — through disclosure cadence, through narrative, through index eligibility, through the composition of each financing — but the premium is set by the marginal buyer of Strategy equity, who is not buying bitcoin. That buyer is purchasing a bitcoin-adjacent instrument with specific convexity, specific index treatment, and specific tax characteristics.

The distinction is the fault line. If mNAV is a function of demand for the instruments, then the flywheel is not a bitcoin supply story at all. It is a capital markets demand story that terminates in bitcoin. And capital markets demand for a levered, single-asset, increasingly coupon-paying instrument can reprice faster than any on-chain metric can register.

I have seen this shape before. In 2021 I spent three months scraping transfer logs from a PFP collection that claimed a $1 billion market cap. The floor held for eleven weeks after the underlying bid had already left. Volume is noise; the wallet cluster is signal. The mNAV equivalent of that cluster analysis is the composition of the buyer base behind each convertible and preferred tranche — and that composition is disclosed far less clearly than the floor price ever was.

The Cost of Capital Inflection

Return to the three phases, because the sequence is the most under-discussed item in this entire story.

Phase one, 2020 through 2021: convertible notes at 0% coupon with conversion premiums of 30% to 50%. The buyer is a volatility fund. Strategy receives cash at a nominal cost of zero, gives up equity optionality above a strike, and pays nothing in the interim. For a company without an investment-grade rating, that is extraordinary capital.

Phase two, 2022 through 2024: at-the-market equity issuance executed against a persistent premium. Shares drip into the tape. Every dollar raised buys bitcoin. The premium is the subsidy.

Phase three, 2025 onward: preferred stock. Strike, Strife, Stride. Coupons in the high single digits to 10%. These are not near-free. They are obligations with a cash cost, in some cases ranking ahead of the common in dividend priority and liquidation.

Read the three phases as a single curve and the direction is unambiguous. Strategy's marginal cost of capital has moved from approximately 0% to approximately 8–10%. That is not a rounding change. It is a regime change, and it happened while the market was calling each step a validation signal.

Why does the cost rise? Because the marginal buyer's required return rises as the risk becomes legible. Early converts priced a call option on a novelty. Preferreds price a yield on a now well-understood, highly volatile, single-asset balance sheet. The first buyer was paid in upside. The second wants to be paid in cash.

The accounting consequence follows directly. Convertible coupons near zero leave the fixed charge negligible and the interest coverage ratio untroubled. Preferred dividends are not negligible. At 8–10% on a preferred stack that has grown into the billions, the annual cash obligation is a real number — payable regardless of where bitcoin trades, and payable before common shareholders see anything.

A company whose asset generates no cash flow, whose obligations increasingly demand cash flow, and whose equity is the bridge between the two, is not running a treasury operation. It is running a duration mismatch. That mismatch is financeable indefinitely — as long as capital markets stay open and the premium stays positive. Both conditions are exogenous to everything Strategy controls.

Gas fees are the price of truth. Here, the coupon is. And the coupon is rising.

The Convertible Arbitrage Layer

There is a layer beneath the capital structure that almost no coverage touches, and it changes how every equity issuance should be read.

Who buys a 0% coupon convertible with a 50% conversion premium? Not a bitcoin bull. A convertible arbitrage desk. The trade is mechanical: buy the bond, short the underlying equity against it, harvest the volatility. The bond supplies convexity. The short neutralizes directional exposure. The spread between implied and realized volatility is the edge.

This means a meaningful share of the demand that appears to validate each Strategy financing is not directional at all. It is hedged. The desk does not want exposure to bitcoin, and it does not want exposure to Strategy. It wants the conversion option's implied volatility to be priced above realized volatility. That is the entire thesis, and it expires with the option.

Two structural consequences follow, and neither requires a conspiracy to be true.

First, the equity shorted against these bonds is a persistent mechanical supply of stock, unrelated to anyone's view of the company. It presses down on the share price. A lower share price compresses mNAV. A compressed mNAV makes the next accretion harder. The same financing that funded the last bitcoin purchase creates a headwind against the next one. This is not a defect introduced by bad actors. It is the arithmetic consequence of using a volatility instrument as a funding vehicle.

Second, the reported raise overstates the durability of the signal. When a company announces a multi-billion-dollar convertible, the market reads institutional demand for Strategy. The more precise reading is institutional demand for the spread between implied and realized volatility on Strategy. Those are different demands, held by different balance sheets, with different holding periods. One is a vote. The other has an expiry.

I am careful here, because this is the kind of claim that gets overextended into nonsense. Convertible arbitrage is legitimate and well understood. Its presence does not make a financing fake. The cash is real, it buys real bitcoin, and the company benefits. What it means is that the durability of the demand is not what the headline implies — and durability is the only variable that matters for a multi-decade accumulation thesis.

In 2020 I mapped an exploit path in which an unaudited oracle feed was the single point of failure for a $30 million protocol. The finding was not that the oracle could be manipulated in the abstract. It was that the protocol's entire economic model was priced off a dependency the team treated as infrastructure. The arbitrage layer occupies that same position in Strategy's architecture. Everyone treats it as plumbing. It is the load-bearing wall.

Where the Trace Ends

Let me do what I actually do.

847,666 coins is not an abstraction to me. It is a set of UTXOs. They arrived at addresses at identifiable times, in identifiable sizes, through identifiable channels. I have traced enough institutional inflows to know the pattern: large OTC desk mediation, settlement in tranches, consolidation into custody clusters, occasionally split across multiple custodians to respect counterparty limits.

What I can see, from a laptop, on a public chain: inflow tranches, their timing, their size distribution, their destination addresses. What I cannot see: anything inside the custodian's perimeter.

That boundary is the finding.

The largest corporate holder of a bearer asset does not hold it bearer-style. The keys sit behind an institutional custody stack — quorum policies, geographically distributed shards, insurance, legal agreements, audit trails. In 2017, analyzing 45 whitepapers from ICOs that had each raised over $2 million, I found mathematical impossibilities in the tokenomics of two major presales, including infinite-supply vulnerabilities. The lesson was that the most dangerous line in any structure is the one where the trace stops and the assurances begin.

For Strategy, the trace stops at the custodian. Everything past that point is attestation. The custodian attests to the balance. The auditor attests to the attestation. The insurer attests to the insurance. Each is a genuine control. Each is also a genuine centralizing dependency, and each introduces a party whose incentives are not identical to the holder's.

The bitcoin network spent a decade being engineered so that no single party can seize, freeze, or counterfeit the asset. Its largest holder has voluntarily reinserted a single party between itself and the asset. For a public company with fiduciary duties, that is rational. It is also the exact shape of vulnerability that smart contract auditors have spent the same decade cataloguing.

The rug is not pulled; it was never tied. Nobody here is alleging a rug. The phrase matters because it names the underlying condition: the trust is not cryptographic. It is contractual, jurisdictional, and operational. Those are three things that can change without a single block being mined.

And beyond custody sits a sharper edge. Regulatory pressure is the one force capable of converting a custodial trust into a custodial constraint. A qualified custodian operates inside a supervisory framework. If that framework changes — on reporting standards, on eligible assets, on segregation, on seizure — the largest holder of bitcoin learns about it at the speed of a compliance memo rather than the speed of the chain. Projects routinely preach decentralization while their foundations and team wallets remain fully traceable. Strategy is the institutional-scale version of that same asymmetry, with the additional twist that its holdings are not merely traceable. They are intermediated.

The Accounting Amplifier

FASB ASU 2023-08 is the most consequential rule change in the history of corporate bitcoin holdings, and it barely made headlines when it landed.

Under the prior treatment, U.S. GAAP required impairment-only accounting for most crypto assets. If bitcoin fell below your cost basis, you wrote it down permanently. If it rose, you did not write it back up until you sold. The result was a one-way earnings filter: losses appeared, gains did not.

Under the new standard, crypto assets are measured at fair value with changes recognized in net income each period. Both directions now flow through the income statement. That sounds like symmetry. Operationally it is leverage.

Take a holder with roughly 847,666 coins. Every $1,000 move in spot is approximately $848 million of mark-to-market. A $10,000 move — which bitcoin produces in a week without anyone calling it a week — is $8.5 billion. That swing hits reported net income. It does not touch cash. It does not touch bitcoin-per-share. It touches the earnings line, and the earnings line is what most of the public market actually reads.

Three second-order effects follow, and they compound.

Reported earnings become a levered proxy for bitcoin with none of the smoothing a fund wrapper would supply. The multiple applied to those earnings becomes unstable, because the denominator is now a moving assumption rather than a business result. And, most importantly for the flywheel, earnings volatility feeds back into the mNAV premium. Equity markets do not pay expanding multiples for unpredictable earnings, even when the unpredictability is mechanical and fully disclosed. If a positive mNAV is the precondition for accretion, and accounting volatility compresses the multiple, then the accounting rule is a direct input into the viability of the model.

I do not think the team was unaware of this. I think they accepted it, because the alternative — being unable to recognize gains until disposition — was worse for the narrative. Acceptance, however, does not neutralize a risk. It schedules it.

A comparable dynamic runs on the preferred side. A cumulative preferred dividend is a fixed charge, and fixed charges are covered by cash flow, not by mark-to-market gains. Strategy's cash flow comes from enterprise software and from financing. If the financing window narrows while bitcoin is quiet, the preferred obligations do not care. They accrue.

This is the part of the structure that a pure bitcoin-appreciation thesis does not model. The obligation side of the balance sheet is denominated in dollars. The asset side is denominated in bitcoin. Those two are volatile relative to each other in exactly the scenario that matters most: a prolonged sideways market in which the premium decays while the coupons keep running.

We are in that market now. That is not a forecast. It is the tape.

The Overhang Flip

Everything above has treated Strategy's holdings as a supply reduction. That framing is correct during accumulation, and it is the entire basis of the bull case. But it is phase-dependent, and phase-dependence is the risk.

Consider the symmetry. A buyer who absorbs 5.16 years of network issuance converts a scarce asset into a concentrated one. Concentration has two faces. During accumulation, concentrated ownership reduces free float and amplifies upside. During distribution, concentrated ownership is a single, identifiable, price-insensitive seller sitting on top of the order book.

847,666 coins is not a quantity the market absorbs quietly. Even sold in tranches through OTC desks, even to pre-arranged buyers, even across quarters, it is a multi-year liquidation that would be visible from the first placement. Every desk would know. Every desk would price it.

That is the overhang flip. The same position that suppresses supply during accumulation becomes the largest known source of potential supply during distribution. There is no state in which that position is neutral. It is always one or the other, and the transition is triggered by conditions the company does not control.

What would trigger it? Not malice. Arithmetic. If mNAV falls below one and stays there, equity issuance stops accreting. If preferred dividends come due against a closed financing window, cash is needed. If a convertible matures into an expensive refinancing market, the company faces the classic triad: repay in cash it does not generate, convert into shares it does not want to issue, or sell the asset. If bitcoin falls far enough for long enough, fair-value losses become the earnings story and the multiple compresses further.

None of these outcomes require a decision to sell. They require a capital market that stops cooperating. The entire architecture rests on that cooperation.

Let me give the honest version of the risk, because overstatement is as useless as denial. Strategy will not dump 847,666 coins into the tape next week. It has years of runway, low-coupon debt with staggered maturities, an operating business with real revenue, and a demonstrated ability to raise capital at a premium. The probability of a disorderly liquidation in any single quarter is low.

But probability multiplied by impact is what matters, and the impact tail is enormous. A single headline about a Strategy sale — not a sale, a headline — would reprice the market's entire model of who the marginal bitcoin buyer is. Five years of price formation have taught participants that a large, patient, price-insensitive buyer exists and absorbs supply. Learning that the buyer has a price after all would be a regime change in the narrative layer, and the narrative layer is where most of bitcoin's current valuation lives.

There is a further asymmetry worth naming. In accumulation, every purchase is announced and celebrated. In distribution, there is strong incentive to be quiet, structured, and gradual — which means the market finds out late, if at all, until the flow becomes undeniable. The information asymmetry runs against the holder of the asset. That condition defines a supply overhang.

I have watched this pattern at retail scale. In 2021 I proved that roughly 60% of a top-tier PFP collection's reported volume was wash trading by a single entity. The floor held for weeks after the bid had left, because the order book looked deep when it was one wallet talking to itself. Concentrated supply looks like strength right up to the instant it looks like weakness, and the transition is not gradual. It is a step function.

What the Bulls Got Right

Now the part where I argue against myself, because a teardown that only cuts is not analysis. It is a mood.

The strongest version of the bull case is not that Saylor is a genius or that bitcoin only appreciates. The strongest version is this: Strategy built a financial primitive that did not previously exist, and it works.

Before 2020, there was no clean way for a public-market investor to obtain levered, index-eligible, retirement-account-accessible exposure to bitcoin without buying futures products or mining equities. Strategy manufactured that exposure out of a capital structure that converted equity premium into asset acquisition. The premium existed because the demand existed. The demand existed because the product was genuinely differentiated. That is not marketing. That is arbitrage between two markets pricing the same underlying differently, and it is a real innovation. Calling it a flywheel undersells it, because flywheels describe momentum while this describes a structural spread.

The bulls are likewise right on a specific claim the skeptics keep botching. The convertibles are not predatory debt. A 0% coupon with a 50% conversion premium is capital transferred from volatility buyers to the company at a cost of approximately zero. Being accused of borrowing for free to acquire a scarce asset is a strange indictment.

They are right about custody, too — more than my own framing above allows. Institutional custody is not amateur hour. It involves quorum key management, distributed shards, insurance, third-party attestation, and audit trails no self-custody arrangement can match. For a public company with fiduciary duties and SEC reporting obligations, self-custody would be the reckless option. The real tradeoff is not decentralization versus security. On the margins that matter operationally — key loss, insider theft, physical coercion — institutional custody is more secure than a hardware wallet in a safe.

And they are right that the overhang argument is frequently overstated. Sellers of this size do not sell on the open market. They negotiate. They find a sovereign wealth fund, a pension, or an ETF sponsor with creation capacity. The bitcoin market has absorbed more than 847,666 coins in single years on the bid side. There is no law forbidding the same on the offer side. Concentration is a risk. It is not by itself a thesis.

So where is the disagreement, if the bulls get that much right?

It reduces to one assumption, and everything else rests on it. The bulls treat the mNAV premium as a property of Strategy. The skeptics treat it as a property of the market. It is the second.

If the premium were a property of Strategy — a structural feature of the instrument, like a coupon — the flywheel would run indefinitely and the only open question would be velocity. If it is a property of the market, then it is a function of conditions: rate environment, risk appetite, index flows, and the supply of investors who want levered bitcoin exposure and cannot obtain it any other way.

That last condition is decaying, and this is where I part company with both camps. Spot bitcoin ETFs now exist. They are cheaper, simpler, more liquid, more tax-transparent for many holders, and available in every major brokerage. The pool of investors who need Strategy specifically, because they have no other route to bitcoin, has been shrinking since January 2024 and continues to shrink.

Strategy's response is to differentiate upward: more leverage, more convexity, preferred yield, index eligibility. That is a real strategy and it has worked. Note what it implies, though. The premium increasingly depends on investors who want something other than bitcoin — yield, volatility, index exposure — and who will leave the moment a better instrument for that something appears. The marginal holder has migrated from wanting bitcoin in a wrapper to wanting a yield instrument that happens to be backed by bitcoin.

Those are different investors with different exit thresholds. The first holds through drawdowns. The second holds until the coupon stops being competitive or the volatility stops paying.

I said at the top that I do not think Strategy is a fraud, and I hold that line. What I think is this: Strategy is a levered, professionally managed, custody-dependent, premium-financed bitcoin accumulation vehicle whose continued function depends on a spread it does not control and is financed increasingly by instruments that demand cash it does not generate. That is not a crime. It is a specification. And specifications come with failure modes.

What to Read Every Quarter

The 1,665-coin purchase is not the information. It is the packaging. The information lives in four numbers published quarterly and almost never read in sequence.

mNAV. Bitcoin-per-share. Fixed-charge coverage on the preferred stack. And the disclosed custody arrangement.

Read those four together, quarter over quarter, and you will know the state of the flywheel before it becomes a headline. mNAV above roughly 1.1 means the machine is still converting premium into accretion. Bitcoin-per-share rising sequentially means the conversion remains net positive after fees. Fixed-charge coverage trending downward means the obligation side is growing faster than the asset side. A change of custodian, or a renegotiated custody agreement, means the trust boundary moved — and trust boundaries are where the trace ends and the assurances begin.

None of that is dramatic. All of it is signal. In a sideways market, where price tells you nothing and positioning tells you everything, scheduled disclosures are the only venue in which a large, patient, identified participant is obliged to tell you the truth on a calendar.

The question I would put to anyone building a thesis here is not whether Strategy buys more bitcoin. It will, for as long as the inequality holds. The question is what the inequality is made of — and whether the market that sets it will still be quoting when the preferred dividends come due.

Imagination is infinite, but liquidity is finite. The coupon does not care which one you believe in.

This analysis is based on public information and should not be construed as investment advice. The underlying news brief carried no source attribution; independent verification against Strategy's SEC filings is recommended. Digital assets carry extreme risk, including total loss of principal.