"Never sold."
Three words. A catechism. Chanted across replies, engraved into quote tweets, repeated until it calcified into scripture.
Then the 10-Q hit.
1,638 BTC. Liquidated. Average sale price: $63,957. Average cost basis on the entire book: $75,419.
The chart didn't cooperate with the narrative. That's an 11.4% discount to average acquisition cost. A realized loss of roughly $18.8 million. On bitcoin. During what the same crowd calls a bull market.
Michael Saylor's defense is structurally elegant: “I never sold bitcoin. Strategy did.”
Technically true. Strategically irrelevant. The entity he founded, chairs, and personally embodies — the one whose public identity is welded to bitcoin maximalism — sold its most sacred asset at a loss. Not to fund research. Not to build products. To service a preferred stock dividend carrying a 12% annual coupon.
Let me translate this into language this market understands. Strategy issued an instrument promising a fixed, dollar-denominated yield. The collateral is bitcoin — an asset that produces zero cash flow. The operating business generates roughly $200 million per year. The preferred dividend obligation is $1.6 billion per year. The gap is covered by the only liquid thing on the balance sheet: BTC.
This is not a conspiracy. It's a cash flow statement. And the cash flow statement just told us something the tweets never did: the one-way treasury is over.
I've spent twelve years inside this market. I've watched yield narratives die in real time. I've seen the gap between story and filing become the only trade that matters. This is one of those moments.
Context
Let me give you the full picture, because context is what separates analysis from panic.
Strategy, formerly MicroStrategy, is a business intelligence software company that became a bitcoin holding vehicle in August 2020. That pivot is now market legend. Michael Saylor converted the company's balance sheet — and his public persona — into a bitcoin accumulation engine. Between 2020 and mid-2025, Strategy accumulated approximately 842,138 BTC. Average cost: $75,419 per coin. Total dollar outlay: roughly $63.5 billion.
The original financing stack was elegant by design. Zero-coupon convertible notes. No interest. Conversion at a premium. The company raised billions without paying a penny in coupon, bought bitcoin, and let the market's rising NAV premium do the rest.
Here's how the loop worked:
Step one: issue convertible notes or equity. Step two: convert proceeds into bitcoin. Step three: watch the market assign a premium to MSTR shares relative to the bitcoin each share represents. Step four: use the premium to justify another issuance. Step five: repeat.
The mechanism ran on a single assumption: the NAV premium would persist. For five years, it did. Sometimes 1.5x. Sometimes 2x. Occasionally 3x. The market paid more for bitcoin inside a corporate wrapper than for bitcoin itself. That spread was the profit engine. Every issuance was accretive because the wrapper traded above its content.
2024 fractured the assumption. The SEC approved spot Bitcoin ETFs in January. Institutional capital suddenly had regulated, efficient, low-fee access to bitcoin. The wrapper premium became contested. Why own MSTR at 1.5x NAV when you can own IBIT at NAV — or below, during discounts? The arbitrage between wrapper and asset began to compress.
Saylor's answer was financial innovation — of a sort. Step forward, STRC. A bitcoin-backed preferred stock carrying a 12% annual dividend, paid quarterly in cash.
The market devoured it. Yield-hungry institutional capital got “bitcoin exposure with income.” The company got fresh dollars. The dollars bought more bitcoin. The loop extended.
But there was a catch embedded in the structure. The dividend is fixed in dollars. Due every quarter. Irrespective of bitcoin's price action. The first quarterly obligation: $400.7 million. Annualized: approximately $1.6 billion. Every quarter, the treasury must find that cash — in dollars, not in sats.
And then, in Q2 2025, something unprecedented happened. Strategy sold bitcoin.
Core
Let's break this down like a trade audit. Numbers first. Narrative later.
The Sale, Deconstructed
Strategy's disclosed holdings: 842,138 BTC as of August 2, 2025. That's down from roughly 843,738 at the end of May. A reduction of roughly 1,600 BTC. The disclosed sale: exactly 1,638 BTC.
Average sale price: $63,957.
Average cost basis on the entire position: $75,419.
The sale executed at an 11.4% discount to average cost. Realized loss: approximately $11,462 per coin. Total realized loss: approximately $18.8 million.
Read that again. The company sold its flagship asset at a loss. To meet dollar-denominated obligations. In a market that many describe as bullish.
Saylor frames this as “capital management.” The footnotes frame it differently. When an asset is sold below its acquisition cost, it's a realized loss. The company just realized $18.8 million in losses. Not by accident. By design.
Gross proceeds: approximately $104.8 million. Stated uses: increase dollar reserves by $250 million (the BTC sale plus other capital sources) and redeem $81 million of STRC preferred shares.
Do the math again. The BTC sale grossed $104.8 million. The STRC redemption consumed $81 million — about 77% of the proceeds. The reserve builds absorb the rest. This is not profit-taking. This is balance sheet triage. The company is using its most liquid asset to retire its most expensive liability.
Here's the part the maxis will skip: 1,638 BTC is 0.19% of the book. Immaterial in size. Material in signal. After five years of one-way accumulation, the company sold. The prior is broken. Every model of Strategy's future flow — for MSTR shares, for the bitcoin market, for the STRC instrument — must now incorporate the possibility of repeated sales.
I've run flow models before. In early 2025, I integrated an open-source AI trading agent into my personal DeFi dashboard. I backtested it against 2020-2024 data. The most important finding wasn't alpha. It was regime detection. Strategies that dominated in one regime produced catastrophic losses in the next. The same logic applies to corporate treasuries. The 2020-2024 regime rewarded accumulation, and Strategy responded accordingly. The 2025 regime is demanding distribution, and Strategy — whatever its public language — is responding. The treasury is adaptive. The ideology is not.
STRC: The Off-Chain Bitcoin Yield Protocol
Strip away the ticker and the prospectus language. STRC is functionally an off-chain bitcoin yield protocol.
The issuer holds bitcoin. The investor receives a fixed yield. The yield does not emerge from bitcoin — bitcoin produces nothing. The yield emerges from the issuer's ability to generate dollars through financing or asset sales.
Let me map the instrument like I'd map a DeFi vault.
Collateral: 842,138 BTC, marked at approximately $64,000 at the time of the filing.
Obligation: $400.7 million per quarter. $1.6 billion per year. Fixed. Non-negotiable.
Yield source: new issuance (equity, debt, or preferred) or bitcoin sales.
This is not a yield product. This is a refinancing obligation with a coupon attached. The “yield” is a claim on future capital inflows, structured to look like an income product.
Now compare STRC to the instrument that preceded it. The zero-coupon convertibles of 2020-2024 paid nothing. The obligation was principal at maturity, and in a rising market the notes converted into equity, dissolving the obligation into dilution. STRC pays cash. Quarterly. At 12%. That's an order of magnitude more expensive than the previous structure.
Why did the company trade cheap optionality for expensive obligation? Because the preferred market demanded yield. Saylor's capital machine needed an instrument that could tap the yield-sensitive institutional pool. STRC was the key. The deal worked at issuance: dollars came in, bitcoin was bought. But the coupon is now a permanent drag. Every quarter, cash leaves the treasury. And when the market's appetite for new issuance wanes, the coupon gets funded the old-fashioned way: by selling the store.
The 1,638 BTC sale is the first evidence that this mechanism has activated.
In 2020, I deployed $5,000 of personal capital into Uniswap v2 liquidity pools and Compound. I learned a lesson that has never failed me: yield does not come from nowhere. If a protocol promises a fixed return on a zero-yield asset, the return is someone else's principal. The question is always: who is the exit liquidity?
For STRC, the answer is becoming visible. The exit liquidity is either future MSTR/STRC buyers, future bitcoin buyers, or the treasury itself — which is now selling bitcoin below cost to make the payments.
The Refinancing Treadmill
Let me model the complete loop, because the machine is more important than the moment.
The machine has three components: an asset (bitcoin), a financing vehicle (equity, converts, preferreds), and a valuation signal (NAV premium).
The loop: issue financing → buy bitcoin → NAV premium justifies next issuance → issue again.
For this loop to continue, the NAV premium must remain positive and ideally stable. The premium is a market verdict on the spread between the wrapper (MSTR shares) and the content (bitcoin). When the wrapper trades above the content, issuance is accretive. When the wrapper trades at or below the content, issuance destroys value.
The STRC coupon changes the economics of the loop. No longer does the company merely profit from the premium at issuance. Now it must also service a 12% coupon. Every point of premium is partially consumed by the cost of carry. The engine is less efficient than it was in 2020-2024.
What breaks the treadmill? Three failure modes, in order of probability.
Failure mode one: premium compression. If MSTR trades at or below NAV, issuance becomes dilutive. The loop stops. The coupon must then be funded from reserves or bitcoin sales. We are seeing the start of this dynamic — not because the premium has collapsed, but because the market is beginning to price the risk of compression.
Failure mode two: refinancing cost. If the market demands more than 12% to hold STRC — due to credit risk repricing, rising rates, or declining bitcoin prices — the cost of refinancing rises. The treadmill burns more capital per cycle.
Failure mode three: bitcoin price decline. Below the average cost basis, sales realize losses. Losses reduce equity. Reduced equity compresses the premium. A compressed premium chokes issuance. Choked issuance forces more sales. This is the geometric sequence that ends badly.
I want to be precise: none of these failure modes is imminent. The company holds $4 billion in reserves. It can service the dividend for multiple quarters without selling another sat. But the direction of travel matters. The company just demonstrated that it will sell bitcoin when it needs dollars. That demonstration is itself information. The market is incorporating it.
In May 2022, I spent 72 consecutive hours dissecting Terra's Anchor protocol — withdrawal queues, mint mechanics, under-collateralization math. The pattern that emerged was simple: a fixed 20% yield on a volatile asset, serviced by new inflows. The moment net inflows turned negative, the yield was exposed as a fiction. The unwind took days.
Strategy is not Terra. The bitcoin is real. The reserves are substantial. The company is not a fraud. It's a leveraged balance sheet with a fixed-dollar dividend and a volatile zero-coupon asset. The structural architecture is different from Terra, and so is the risk profile. But the geometry of a fixed obligation on a variable asset is universal. I've watched that geometry end badly too many times to ignore a quarterly $400.7 million line item.
Cash Flow Reality Check
The asymmetry deserves a hard look.
- STRC quarterly dividend: $400.7 million.
- STRC annualized dividend: ~$1.6 billion.
- Strategy SaaS operating revenue: ~$200 million per year.
- Coverage ratio: 8 cents of revenue for every dollar of dividend.
- BTC sale proceeds (this quarter): ~$104.8 million.
- STRC shares redeemed: $81 million.
- Dollar reserves after increase: ~$4 billion.
The core business covers 8% of the preferred dividend. The bitcoin sale covered roughly one quarter of one quarterly payment. Everything else is funded by capital markets issuance — equity, converts, or preferreds.
Let me restate this in trader language. The company's “breakeven” is not a price. It's a flow. The company requires roughly $1.6 billion per year in external financing just to keep its dividend current. If external financing stalls, the company must choose: sell bitcoin at a discount to basis, or draw down reserves at a run rate that exhausts them in roughly 2.5 years.
That's the real beta of this trade. Not bitcoin volatility. Financing access.
The Book Is Underwater — On Paper
At $64,000 per coin — the approximate spot price during the disclosure window — Strategy's 842,138 BTC position carries an unrealized loss of approximately $9.6 billion against an average cost of $75,419.
Let me say that again. $9.6 billion. Underwater. On paper.
Now — paper is paper. The company doesn't need to realize that loss. It can wait. Bitcoin is a cyclical asset with a secular bias. If the cycle turns upward, the paper loss disappears. The company has time. It has reserves. It has optionality.
But the market prices the paper. The NAV premium is a function of confidence, and confidence is a function of the marked book. A $9.6 billion hole makes the next financing round harder. Underwriters will price the risk. The yield demanded on the next STRC issue will widen. The treadmill gets harder.
At $75,419, the position is breakeven. At $64,000, it's 15% underwater. At $60,000, it's 20% underwater. Every dollar the price spends below cost basis makes future sales more rational — not because it's a smart trade, but because it's the only trade that settles in dollars.
There's a perverse incentive here worth naming. If the company sells bitcoin below its average cost, it realizes a loss. But if the company — or the market — believes the cost basis is too high to be reached again in the near term, selling becomes a form of capital efficiency. The company converts a permanently-underwater asset into dry powder. It's not a capitulation. It's an optimization of the possible.
The Never-Sell Doctrine vs. The Reversible Treasury
Let's talk about the doctrine, because the doctrine is the product.
Michael Saylor built the most recognizable personal brand in bitcoin on a single idea: never sell. “Not your keys, not your coins.” “Only the weak sell.” “Bitcoin is the exit strategy.” The rhetoric was absolute. The opposite of trading. The end of trading.
The doctrine served a function. It signaled permanence. Institutional capital likes permanence. If the treasury never sells, the supply is locked. That belief in permanent lock-up justified the NAV premium. The premium justified the issuance. The issuance funded the purchases.
But a doctrine is not a delivery vehicle. Doctrine doesn't pay dividends. The STRC coupon is a dollar obligation, and obligations are paid with dollars. When the financing window narrows and reserves get thin, the doctrine becomes negotiable.
Which is exactly what happened. The company sold. Not Saylor personally — the company. The distinction matters less than it seems. The corporate veil is a legal formality. The market reads Strategy, Saylor, and the bitcoin stack as a single story. When the story changes — however slightly — the premium reprices. The story did change. The vault is now reversible.
I don't know whether Saylor the individual will ever sell his personal bitcoin. I doubt it. But the entity he controls has demonstrated that it can sell, will sell, and has already sold at a loss. The doctrine survives as a personal statement, but it's dead as a corporate pledge.
Smart money is already pricing this. The question is whether the market fully appreciates what “reversible” means for the supply narrative. Bitcoin bulls point to Strategy's 842,138 BTC as “locked supply.” That assumption is now measured in degrees, not absolutes. The supply is locked only if the price stays above the set of future obligations. That's a conditional lock, and this quarter it failed the condition.
What It Means for Bitcoin's Market Structure
Now let me widen the lens beyond one company.
For five years, Strategy was the largest institutional buyer in the bitcoin market. Its demand was not price-sensitive. It bought at $10,000. It bought at $30,000. It bought at $70,000. It bought when no one else was buying. That relentless, insensitive demand contributed to the structural bid underneath the market.
The reversal — even at 0.19% of the book — changes the market structure's optics. The most dedicated institutional accumulator in the asset class just demonstrated it can flip. Statistically, the move is noise. Narratively, it's a breach in the wall.
Here's what I'm watching on the supply side: if the company needs to raise $1.6 billion per year for STRC dividends, and the financing window constrains, the required BTC sales are not trivial. At $64,000 per coin, $1.6 billion is 25,000 BTC per year. That's roughly 68 BTC per day. Every day. Until the financing window reopens or the price materially improves.
That's not a top-of-market supply flood. But it's a persistent, price-insensitive seller in an asset class that feeds on narrative. And it compounds the reflexive risk: sales pressure → price declines → premium compression → more financing difficulty → more sales.
I ran the numbers through my own models. Through a 2025-2026 horizon, the base case doesn't include forced liquidation. The reserve is too deep. The financing machine is still functional. But the tail scenarios are heavier than the market is pricing. And in options, you buy the tails before they materialize, not after.
I bought the pixel, not the promise, back in the NFT era. I walked away with $12,000 in profit and a $4,000 lesson about execution risk. The lesson applies here: trade what the filings show, not what the narrative promises. The filing shows a company managing a leveraged balance sheet under modest stress. That's not a disaster. It's a repricing.
Hidden Layers — What the Filing Doesn't Say
Let's go forensic for a moment. The filing says what the company wants to say. It doesn't say everything.
First: derivatives. Has Strategy hedged any of its bitcoin exposure? Puts, swaps, collars — anything? The disclosure shows no material positions. But the 10-Q lags reality. Institutional treasuries routinely establish hedging programs before disclosure catches up. The 1,638 BTC sale might be the visible slice of a larger repositioning. Or it might be all there is. My confidence in the “hidden hedge” thesis: low, perhaps 15-20%. But I've audited enough treasury operations to respect the disclosure lag.
Second: custody. Strategy has never disclosed a granular custody architecture. Cold storage? Multi-sig? Third-party custodians? The speed of the 1,638 BTC disposal — roughly 1,600 coins over six weeks — signals operational readiness. The treasury can sell quickly. Under bull market conditions, that's optionality. Under stress, that's velocity. The market should not assume the vault is operationally slow.
Third: the pace of future sales. The current quarter's sale was small. The next quarter's may be larger. Or zero. The company has no stated formula. The only signal is the cash flow gap: $400.7 million per quarter against $50 million per quarter in operating revenue. The gap must be funded somehow. Reserves or sales. The reserve is $4 billion and shrinking with every dividend payment.
The most bullish scenario: bitcoin rallies, the NAV premium widens, issuance reopens, and the dividend is funded by new dollars. The most bearish scenario: bitcoin grinds lower, the premium compresses, issuance stalls, and the company funds its dividend by reversing its own accumulation history.
The trade is not about which scenario happens. The trade is about the market's willingness to price both. The narrative prices one. The filing prices the other. When they diverge, opportunities appear.
Contrarian
Let me play the other side of this trade, because the other side has merit.
The “Saylor sold” narrative is emotionally satisfying but analytically incomplete. Let me lay out the bull case for the seller.
First: selling 1,638 BTC out of 842,138 is a rounding error. The company is still 99.8% committed. It added dollar reserves. It redeemed high-cost preferreds. It improved the balance sheet's resilience. That's prudent capital management, not capitulation.
Second: the STRC redemption is a good trade. Retiring $81 million of 12% paper is a permanent reduction in the dividend obligation. If the company can redeem more STRC at par, it lowers the fixed-dollar burden. That's the opposite of leverage escalation — it's deleveraging.
Third: the reserve build gives the company ammunition. $4 billion in dollars means Strategy can buy the dip if bitcoin corrects violently. A treasury that converts some bitcoin into dry powder at the margin is one that can re-enter the market with force. It's not a one-way door. It's a strategic pivot.
Fourth: the personal conviction of Michael Saylor remains intact. If he truly never sold — and there's no public evidence he did — the individual anchor of the narrative survives. The brand persists. The company's action doesn't erase the founder's conviction. It just complicates the story.
Now the contrarian layer beneath the contrarian. The retail market is reading this as weakness. The institutional market is reading it as a repricing. But the real signal is subtler: the company just told us its priority is the preferred dividend, not the accumulation script. When a bitcoin treasury company issues a yield-bearing liability, it accepts a claim that outranks its own ideology. The rank ordering matters. The dividend gets paid before the stack grows. That's the new hierarchy.
Let me be careful about the P-word. Terms like Ponzi get thrown around carelessly. I've audited actual Ponzi structures. Terra's Anchor was one. The yield was fictional, generated by minting, and distributed to early entrants at the expense of later ones. Strategy is not that. The underlying asset is real. The balance sheet is solvent. The company has actual revenue. The valuation premium is high but not self-evidently fraudulent.
But there is a structural characteristic worth naming. The dividend is a fixed obligation on an asset that produces no cash flow, and the servicing mechanism is new inflows or asset sales. That's not a Ponzi. It's what happens when a zero-yield asset meets a fixed yield product: the yield becomes a claim on future capital. The structure is sustainable as long as future capital arrives. The moment it doesn't, the obligation converts into forced selling.
The market is early in discovering which scenario applies. The 1,638 BTC sale is a single data point. But it's the first of its kind, and first data points set the prior. The variance of the outcome just widened. That's the information.
Retail traders bought the pixel, not the promise. They bought the laser eyes, the “never sell” oath, the idea that Strategy was a permanent sink for bitcoin. The sink is now a tub. It drained slightly this quarter. The market will find out whether the tub drains further or the taps reopen.
Takeaway
I don't trade on promises. I trade on pressure.
Three levels. Watch them.
First: $75,419. The average cost basis. If bitcoin trades below this level for a sustained window, the company's paper loss persists. Every sale below this line is a realized loss. The company will avoid that accounting until it can't. The frequency of future sales is a direct function of how long the price stays underwater.
Second: the STRC market. Watch the yield. If STRC trades above a 12% yield, the market is demanding more compensation for the same balance sheet. That's a canary in the mine. Watch the redemption queue too. Accelerating redemptions mean the company must return principal — likely in dollars. Dollars come from either reserves or bitcoin sales.
Third: the next 10-Q. If holdings drop by another 5,000-10,000 BTC, the reversible treasury is policy, not an event. If holdings stabilize or grow, the 1,638 BTC sale was a bolt from the blue — a one-time adjustment, not a regime change.
The bull case survives. Bitcoin's secular uptrend is intact. The company's stack is enormous. The reserve is deep. This is not a death spiral. It's a structural stress test, and the test results are mixed.
But the model has changed. The one-way door now swings both ways. The narrative is no longer the full story. The filing is part of the story too.
Every candle tells a story of fear. This one tells the story of a treasury that learned the difference between an asset and a cash flow. Liquidity vanishes when the music stops. The music hasn't stopped. But the vault just cleared its throat.
Risk isn't a feeling. It's a quarterly dividend line in a 10-Q. Read accordingly.
