Saylor's Bitcoin Machine Just Started Running Backwards

For four years the loudest argument against Michael Saylor was a question nobody could answer: what happens when the premium goes away. This week the market stopped asking and started pricing it.

• STRC ("Stretch") traded near $74 in late June 2026, roughly 25 to 26 percent below its $100 par value, a record low for the instrument.
• Strategy sold 32 BTC for about $2.5M (average near $77,135) in late May 2026, its first disclosed sale since accumulation began in 2022.
• Corporate mNAV slipped below 1.0 on June 26 to 27, 2026 on a basic reading. Enterprise-level readings sit higher, roughly 1.05x to 1.16x depending on method, so cite this as a range, not a point.
• Dividend coverage compressed from a widely repeated multi-year figure to roughly 14 months (sources range from about 10 to 14).
• Roughly $1.2B in annual preferred dividends against a filing-anchored cash reserve near $900M at May 31 (one later source cites about $1.4B).
• Holdings about 847,363 BTC at roughly $75,650 average cost. With BTC in the high $50,000s, the basis sits around 22 percent underwater.
• Snapshot dated late June 2026. These prints move daily.

The premium was never a side effect. It was the product.

Read the headlines and you get a liquidity scare: a famous Bitcoin buyer sold a little Bitcoin, a preferred stock cracked, a competitor talked his book. That framing misses the machine.

Our read is simpler and less comfortable. Strategy was never a Bitcoin fund. It was a premium conversion machine, and its engine was reflexive. While the common stock and the preferreds traded above the value of the Bitcoin underneath them (mNAV above 1, STRC near or above par), every share Strategy sold bought more Bitcoin than it diluted. Bitcoin-per-share went up. That rising number "justified" the next raise, which validated the premium, which funded the next raise. The loop fed itself.

The detail everyone priced as permanent is the one that just broke: that loop is accretive only above parity. The instant the market value of the common drops below the Bitcoin it represents, the same act of issuing-and-buying starts to lower Bitcoin-per-share instead of raising it. The machine does not slow down. It runs backward. Issuance now destroys value, so the company stops issuing, and the roughly $1.2B annual dividend obligation has to be paid out of a finite reserve or, as it was for the first time since 2022, by selling Bitcoin.

The numbers that mark the inversion

Two prices moved into the same place in the same week. Corporate mNAV slipped below 1.0 on a basic reading (enterprise readings, which fold in the full debt and preferred stack, still sit modestly above parity at roughly 1.05x to 1.16x, so the honest statement is "the whole structure is sitting near parity"). And STRC, the variable-rate perpetual preferred engineered to hover near $100 par via a monthly-adjusted coupon, printed near $74, about a quarter below par, with the dividend pinned at 11.5 percent for a fourth straight month and an effective yield near 15 percent at that price.

Against the stack, the 32-coin sale is almost nothing: 0.004 percent of holdings, a rounding error. The basis tells the real story. The average cost sits near $75,650, BTC trades in the high $50,000s, and the coins that went out the door left below blended cost. The buyer that built the entire narrative on "never sell" sold underwater to make a dividend payment.

STRC is the thermometer, not the fever

Here is why the preferred broke first, and why that matters more than the price of it. STRC is the only instrument in Strategy's capital stack with a hard $100 par anchor the market can score against in real time. The common can drift on sentiment and the convertibles can be hand-waved as "stretched, no near-term wall." STRC cannot. It was explicitly built to sit at par, so the moment the reflexive premium inverts, STRC is simply where the inversion becomes visible and measurable first.

That is the non-obvious point. The discount on a leveraged structure does not sit still the way a discount on a plain asset does. It feeds. The same plumbing that minted Bitcoin-per-share on the way up can drain it on the way down, and the company's own defensive moves (pausing the at-the-market issuance program, touching the stack) read as confirmation of the regime change rather than a cure for it.

There is a serious bull rebuttal to all of this, and it is not weak. There is no lender, no margin call, no maturity wall, no mechanical forced seller. Solvency is years away, not quarters. But that rebuttal answers a question we are not asking. We are not asking whether Strategy goes bankrupt. We are asking what happens to a flywheel everyone priced as self-sustaining once it visibly stalls, and what the company is forced to do at each dividend date to keep it from spinning the wrong way.

The discount on a leveraged structure does not sit still. It feeds. Below the line: the second-order trap, the coupon paradox that makes the only repair lever expensive, the GBTC precedent for discounts that lock in for a year or more, the dividend-date watch-list, and the one number that tells you whether the flywheel re-arms.

The reflexive trap, closing on itself

The risk that matters is not insolvency. It is the loop tightening into a contradiction. To defend STRC's par and protect credibility, Strategy needs to avoid selling Bitcoin. But to fund roughly $1.2B a year without ATM access in a sub-par regime, the only non-dilutive sources are a finite reserve and, when that thins, Bitcoin sales, the single act that breaks the whole story. Each dividend date becomes a public referendum on whether they will touch the stack.

Then there is the coupon paradox, which is the part the bull case quietly steps around. STRC's only built-in lever to pull itself back toward par is to raise the dividend. But raising the coupon does two corrosive things at once. It closes the gap to a roughly 15 percent market yield only partway (an 11.5 to 12.5 percent reset does not reach 15), and it makes the company's own funding more expensive at the exact moment coverage is shrinking. The self-correction mechanism, under stress, makes the underlying problem worse. A June 30 ex-dividend and rate reset is the near-term test of exactly this, with the coupon reportedly expected to step up.

The steelman, stated at full strength

We will not strawman the other side, because the other side is coherent. The preferreds (STRK, STRF, STRC, STRD) are a claim on residual assets, not collateral pledged against Bitcoin. There is no lender to pull a plug and no imminent debt wall (the roughly $6.7B of convertibles is stretched out, with about $15.5B of preferred ahead of the common). This is the structural opposite of the 2022 Celsius and 3AC liquidations, where collateral and margin calls did the killing. Castle Island's Matt Walsh put it cleanly, calling it "a timing and capital markets problem," not "an asset coverage problem today," and that is the honest near-term description.

Reflexivity also cuts both ways, which is the genuinely dangerous part of being bearish here. A single Bitcoin rally back above the roughly $75,650 cost basis re-opens the ATM, re-arms the loop upward, and does it violently, because the same feedback that drained Bitcoin-per-share reverses with equal force. And Brad Garlinghouse, who supplied the week's quotable line by calling the model "financial engineering" and the sub-par STRC "a pretty damning indictment," is a conflicted XRP CEO talking against a competitor while staying explicitly bullish on Bitcoin itself. He is a mechanics critic, not a Bitcoin bear, and "financial engineering" is rhetorically cheap when every treasury company on earth is running some version of it. Read the pro-Strategy coverage adversarially too: the "death spiral" causal chain is an analyst model, not an observed fact, and the bull runway estimates that stretch coverage to decades assume the coupon never climbs and no new capital is ever needed, both of which break precisely in the stress case.

The precedent the bulls do not like to name

The reason the discount deserves respect rather than a shrug is GBTC. Grayscale's Bitcoin trust traded at a fat premium through the euphoria, then flipped to a persistent discount in stress and sat there, near negative 49 percent for more than a year, because there was no redemption mechanism to arbitrage the gap closed. Premiums that look structural in a boom can invert into discounts that look just as structural in a bust, and they can stay there far longer than "it should mean-revert" intuitions allow.

Strategy is not GBTC. The instruments differ and the redemption mechanics differ. But the behavioral template is the warning: a premium is not a property of the asset, it is a function of conditions, and when the conditions break the discount can become self-sustaining with no fast arbitrage to close it. The tail scenario for Strategy is not a cliff. It is a slow grind in which a prolonged stay below cost basis removes both recovery levers at once, balance-sheet growth and multiple re-expansion, and the structure idles for quarters.

The watch-list

This is what we are actually watching, in order of signal value:

• Corporate mNAV versus 1.0. Above parity the flywheel re-arms and the "engineering" looks like genius again. Below it and grinding, every coupon is a live test.
• STRC versus par. As long as it sits below $100, the Bitcoin-accumulation engine is effectively idling regardless of what the common does. Par reclaimed is the all-clear; a wider discount after a coupon bump is the opposite.
• The dividend dates. June 30, then Q3 2026. Each one is a referendum on the "never sell" discipline that already cracked once. Watch whether the reserve is drawn or coins are sold.
• The reset coupon. An 11.5 to 12.5 percent step that fails to pull STRC toward par is the coupon paradox confirming itself in public.

The take

No call here, and no price target. Just the lens.

What changed this week is not a number, it is a belief. The market stopped treating Strategy's preferred flywheel as self-sustaining and marked it to its one hidden assumption: that the premium is permanent. It is not. It was always a function of two conditions, Bitcoin above cost basis and the preferred near par, and both broke in the same seven days.

So the variable to watch is not the Bitcoin price and not Garlinghouse's soundbite. It is whether corporate mNAV reclaims 1.0, and whether STRC returns toward par, before the next dividend dates force another reserve draw or another coin sale. The $2.5M of Bitcoin they sold is trivial. The signal, that a machine built to only buy became a seller, is the entire story. Everything else is just waiting to see which way the loop spins from here.

FAQ

Is Strategy (MicroStrategy) at risk of bankruptcy or a forced Bitcoin liquidation?
Not mechanically. The preferred shares (STRK, STRF, STRC, STRD) are a claim on residual assets, not collateral pledged against Bitcoin, so there is no lender to issue a margin call and no near-term debt maturity wall. Strategy chose to sell 32 BTC, it was not forced. The genuine risk is a slow confidence and refinancing squeeze over time, not a solvency cliff today.

What is STRC and why did it fall below par?
STRC ("Stretch") is Strategy's variable-rate perpetual preferred stock, launched in 2025 at $100 par and engineered to trade near $100 via a monthly-adjusted dividend (held at 11.5 percent). It traded roughly 25 to 26 percent below par in late June 2026 because the market stopped believing the self-correction mechanism and began pricing credit risk rather than just a yield gap. A reset toward 12 to 12.5 percent does not close the gap to a roughly 15 percent effective market yield.

What did Brad Garlinghouse actually say?
The Ripple CEO reportedly called Strategy's preferred funding "financial engineering" that does not drive long-term value, and called STRC trading about 25 percent below par "a pretty damning indictment." Crucially he stayed bullish on Bitcoin itself. He is a mechanics critic, and a competitor, not a Bitcoin bear.

Why does mNAV falling below 1.0 matter more than the dividend burden?
Because it kills the accumulation engine instantly. Above 1.0, issuing equity to buy Bitcoin raises Bitcoin-per-share and is accretive. Below 1.0, the same act lowers it and is dilutive, so the company stops issuing. The roughly $1.2B annual dividend is survivable for years. The premium collapse is what stalls the flywheel that created the premium in the first place. Note that mNAV readings differ by method (a basic reading slipped below 1, enterprise readings sit near 1.05x to 1.16x), so treat it as a range.

What is the single number to watch from here?
Whether corporate mNAV reclaims 1.0, and whether STRC returns toward par, before the next dividend dates (June 30, then Q3 2026) force another reserve draw or coin sale. Above parity the flywheel re-arms upward. Below it and grinding, every coupon is a public test of the "never sell" discipline that already cracked once.

Sources: SEC 8-K filing (32-BTC sale disclosure), CoinDesk (STRC pricing, dividend, mNAV), The Block, CryptoBriefing, news.bitcoin.com, and CNBC (Garlinghouse remarks), cross-checked late June 2026. Figures including STRC and MSTR prices, BTC spot, mNAV, and coverage months are a dated late-June 2026 snapshot and move daily, several are reported as ranges where sources differ. The reflexive "reverse flywheel" mechanism described here is an analytical model, not a proven outcome.

Disclaimer: Goldzweig research is for information only. Nothing here is a buy, sell, or hold recommendation, a price target, or financial advice. Do your own work.