The 3,620-BTC Crack: How a 12% Coupon Broke Strategy's 'Never Sell' Narrative

On July 26, Strategy's most-watched internal metric—per-share satoshis—fell to 203,683. It had been 210,824 less than a month earlier. The drop wasn't triggered by a flash crash or a counterparty default. It was triggered by the mechanics of a preferred stock that pays 12% no matter what Bitcoin does. To fund that obligation, Strategy sold 3,620 BTC in the first seven months of 2026—roughly 1/48th of the 174,895 BTC it bought over the same period. Most coverage called it a rounding error. It isn't. A rounding error doesn't break the founding myth of a company that has spent two years convincing the market it will never sell. A coupon payment does. I have spent the past decade auditing the gap between what financial structures promise and what they can actually deliver. In 2017, I reviewed more than 40 ICO whitepapers and found reentrancy vulnerabilities that killed a €500,000 seed round. In 2020, I watched $2 billion in TVL chase yield that was always going to be someone else's exit liquidity. In 2022, I mapped UST's collapse to the shadow-banking dynamic that eventually dragged down Celsius and Three Arrows Capital. The instruments change. The principle doesn't: when a structure's survival depends on an asset going up forever, the real risk isn't volatility. It's the first obligation that comes due when volatility goes the other way. Strategy has now met that obligation. The market hasn't priced it. Context: What Strategy Actually Is Let's be precise. Strategy is no longer a software company. It's a publicly traded Bitcoin treasury with a preferred-stock funding arm. As of July 26, 2026, the balance sheet held 846,000 BTC, worth approximately $58.45 billion. The public target is to double the number of satoshis per MSTR share every seven years. The entire edifice rests on three financing tools: convertible notes, ATM equity offerings, and STRC, a perpetual preferred stock with a 12% dividend and a $100 par value. STRC is the newest and by far the most important piece. In the first seven months of 2026, Strategy raised $7.53 billion through STRC, pushing its face value from $5.3 billion to $10.5 billion. That is roughly $1.07 billion per month of new preferred capital. The dividend is fixed at 12%; because the shares trade below par at around $89, the effective yield is closer to 13.6%. This is not a small-scale experiment. It is the company's primary marginal funding source. The structure works in theory as follows: Strategy buys Bitcoin using STRC and other financing proceeds. Bitcoin rises, the balance sheet grows, and the company can issue more STRC at increasingly favorable terms. The 12% dividend is supposed to be less than Bitcoin's expected annual appreciation, leaving positive carry. Preferred shareholders receive a fixed coupon and no upside beyond par. Common shareholders receive leveraged Bitcoin exposure through a rising per-share satoshi count. The company gets an almost endless loop of cheap capital into an asset that historically goes up. It's elegant. It's also the same logic as every leveraged yield farm I audited during DeFi Summer, just with a 10-Q instead of a whitepaper. Core: The Preferred That Pays Like a Debt The first thing any auditor notices is that STRC is not equity in the traditional sense. It is a senior, cash-settled claim on a balance sheet dominated by Bitcoin. Preferred shareholders do not participate in Bitcoin appreciation. They get 12% per year, paid in cash, in perpetuity. At $10.5 billion par value, the annual dividend obligation is approximately $1.26 billion. Where does that cash come from? Not from software sales. Strategy's operating cash flow is minimal relative to its balance sheet. The dividend must be funded from one of three places: new securities issuance, cash reserves, or selling Bitcoin. In Q2 2026, with Bitcoin down roughly 40% year-over-year, the company recorded an $8.32 billion digital asset impairment and a net loss of $8.22 billion. The impairment is non-cash, but it highlights that the dominant asset is marked to market. The dividend, by contrast, is cash. It is senior to any future growth. This is the part most cryptocurrency-native readers miss. In a DeFi protocol, a liquidity provider has a claim on a pool of volatile assets and can withdraw in kind. In a corporate structure, a preferred shareholder has a senior, cash-settled claim. When Bitcoin falls, Strategy cannot mark down the preferred and move on. It still owes 12% on the full par value. If it issues more preferred at a discount, it creates new obligations at a less efficient rate. If it buys back the preferred below par, it burns cash. If it does nothing, the STRC price stays below par and future issuance becomes more expensive. The only escape hatch is selling Bitcoin—which is exactly what the company did. Liquidity doesn't negotiate with emotionally attached holders. It observes the senior claim and prices accordingly. Core: The Per-Share Satoshi Metric Is the Tell During Q2, management touted the per-share satoshi count rising from 200,000 to 210,824. Then on July 26, it dropped to 203,683. That was not a rounding error; it was the mechanical consequence of the 3,620 BTC net sale. Here is why this matters. The seven-year doubling metric was designed as a purely accretive number, a way to measure whether each new financing round adds more Bitcoin per share than it adds shares. As long as the satoshi count rises, management can argue that every dollar of preferred capital is being deployed efficiently. The moment it falls, the narrative flips from accumulation to preservation. And the market will notice that the first time the metric moved backward, it was because a fixed-coupon obligation came due. This is the information gain most discussions miss. The 3,620 BTC sale is not significant because of its size relative to an 846,000 BTC treasury. It is significant because it proves the per-share satoshi metric is not a policy commitment. It is a residual variable. It moves in whichever direction the company's cash obligations require. In 2017, I saw projects promise fixed token supply schedules, then quietly mint new tokens to pay exchange listings. The schedule was always real until it wasn't. The satoshi count is the same kind of promise. Useful disclosure, but not a covenant. Core: The Cash Buffer Is a Smoke Alarm, Not a Fire Extinguisher The most cited bullish fact in the Q2 report is the cash restoration. Strategy rebuilt its designated dollar reserve from $871 million to $3.75 billion, extending its coverage period from six months to 2.1 years. Management explicitly acknowledged that it had put too much capital into Bitcoin and allowed cash reserves to shrink. Then it corrected. Let's stress-test that correction. At $10.5 billion par value, the annual STRC dividend is roughly $1.26 billion. A $2.879 billion increase in cash provides slightly more than two years of dividend coverage. But the same cash must also cover the interest on convertible notes, future operating costs, and any buyback activity. And the company is still issuing new STRC, adding more dividend obligations at a rate of over $1 billion per month. The coverage ratio is not improving so much as being reset. The real measure is whether the company can grow Bitcoin holdings fast enough to make the coupon sustainable. In a sideways market, the answer is no. There is also a hidden question: where did the $3.75 billion come from? The most likely sources are new STRC issuance, ATM sales, and convertible notes. Selling 3,620 BTC would produce only a fraction of that amount. So the cash buffer is not a sign of discipline. It is a sign of continued external financing. That's not necessarily bad, but it means the repair is being funded by expanding the very liability that caused the problem. Core: September 8 Is a Stress Test, Not a Promise Strategy has publicly targeted September 8 as the date STRC will return to par. The reference is a 70-trading-day recovery from the $74.57 low to the current $89. But the historical analogy is flawed. During the earlier recovery, Bitcoin was stabilizing after a sharp drawdown. Today, the company is carrying a $975 million buyback authorization against a par-value gap of roughly $1.2 billion. Even if the entire authorization were spent on STRC, it would cover about 81% of the gap between market value and par, assuming no other holders sell into the bid. The auditor blinked; the market didn't. Management can announce a target, but the target doesn't change the pricing of a perpetual preferred that trades below par because the market demands a 13.6% effective yield. To close the gap, either the dividend needs to look more sustainable, the Bitcoin collateral needs to look more stable, or a third party needs to buy the discount. The only lever management directly controls is the buyback, and using cash to buy back STRC at $89 reduces the exact cash reserve needed to pay future coupons. That is self-referential support. It can stabilize the price for a while; it cannot create value. Core: The Shareholder Base Is a Risk Assumption The composition data is as important as the balance sheet. Retail investors hold $7.4 billion of STRC, or 71% of the class. Institutional investors hold $3.1 billion, up from 22% to 29% during Q2. The average retail position is $48,000. The average institutional position is $3.5 million. This is not the typical profile of a preferred stock investor. It looks more like a crypto exchange order book. Retail holders came for the 12% yield and the Saylor story. Institutional holders arrived later, likely because a 13.6% effective yield on a Bitcoin-backed preferred was too high to ignore. Both groups can leave quickly. Retail tends to sell as a herd when the story cracks. Institutional exits are larger and can cascade through a thinly traded preferred class. The increase in institutional ownership is positive in the short run, but it also means the market's exit capacity is not designed for a sell-off; it is designed for accumulation. Core: The Macro Cycle Is the Real Context Strategy's entire model is a leveraged bet on dollar liquidity. When credit is cheap and risk appetite is high, a 12% preferred yield looks like a gift. When the Fed tightens or credit spreads widen, that same yield looks inadequate for a perpetual instrument tied to a volatile asset. The fundamental question is not whether Bitcoin goes up in five years. It is whether the dollar liquidity cycle allows a company with $58.45 billion in Bitcoin and $10.5 billion in preferred obligations to refinance at an acceptable cost. This is where my macro watcher frame kicks in. The current environment is not a crypto-specific bear market. It is a broader repricing of leveraged balance sheets. Strategy's STRC is effectively a high-yield instrument with Bitcoin as the collateral pool. It competes directly with private credit, bank preferreds, and high-yield bond indices. Management has said as much, describing its benchmarks as private credit, bank preferreds, and high-yield bonds, with an eventual ambition to be measured against investment-grade and CMBS markets. That tells you everything: STRC is not a crypto token; it's a fixed-income product with an embedded call option on Bitcoin. And in a fixed-income market, a 12% coupon on a collateral asset with 30% plus annual volatility is not aggressive. It's compensation. The AI-Agent Angle: Who Is Actually Buying STRC? Institutional ownership jumped from 22% to 29% in Q2. That timing is interesting. It did not happen after Bitcoin bottomed; it happened while the company was still marking down assets. What kind of institution buys a Bitcoin-backed preferred during a drawdown? The answer may be systematic. I have spent the past year studying AI-agent behavior in payment and treasury markets, and the STRC inflow pattern looks like a trigger-based strategy: when effective yield crosses 13.5% and the buyback is announced, buy. That's not conviction. That's a momentum algorithm with a compliance layer. This matters because algorithm-driven buying can reverse just as quickly as it appeared, especially if the September 8 deadline passes unmet. Regulatory Compliance Is Not a Howey Test From a legal perspective, STRC is a registered security traded on a U.S. exchange. The Howey test is effectively satisfied, but the regulatory risk is low because the issuer is subject to SEC reporting. The real regulatory issue is investor protection. When 71% of a preferred's holder base is retail and the instrument pays 12%, the SEC is likely to scrutinize whether the marketing and risk disclosure are adequate. There is also the FASB fair value issue: mark-to-market accounting means every major Bitcoin drawdown will create a multi-billion-dollar impairment charge on the income statement. That is not hidden debt, but it makes the company's ongoing viability look worse than a software company with stable revenue. The market does not distinguish between non-cash impairments and operational losses; it just sees red. Governance: Nobody Stress-Tested the Vault From a governance standpoint, Strategy is transparent by public-company standards. It filed a 10-Q, disclosed $8.22 billion in net losses, revealed the 3,620 BTC sale, and announced the STRC buyback. That's more disclosure than any crypto-native treasury protocol has ever provided. But transparency is not the same as independent review. There is no evidence of an external third-party stress test for the combined BTC reserve, STRC dividend, and convertible note obligations. And governance is highly concentrated: Michael Saylor holds a controlling super-voting interest. The CEO, Phong Le, is operationally responsible, but the strategy is still a Saylor thesis expressed on a balance sheet. This creates a structural governance risk. If Saylor's conviction falters, or if he faces legal or health issues, the entire buy-and-hold strategy could lose its anchor. The company's own admission—that it allowed cash reserves to shrink while buying Bitcoin—shows there is no internal circuit breaker. The cash buffer rebuild is a welcome adjustment, but it was only implemented after the market punished the company. That's not risk management. That's learning from a near-death experience. If I were still auditing corporate balance sheets, three items would stand out. One, the 12% dividend coverage is not tested against a bear case. The cash buffer projection assumes no additional buyback expense and no dividend increase. Two, the per-share satoshi metric is unaudited and non-standard. It can be revised at management's discretion. Three, the September 8 target creates a market expectation that is not backed by any commitment. The buyback authorization is discretionary, and the company is not obligated to spend all of it. These three items are not fraud indicators; they are information asymmetries. In a market that worships transparency, those asymmetries are the real edge. Contrarian: The Coupon Is the Collateral, Not Bitcoin The consensus narrative is that Strategy's problem is Bitcoin's price. I disagree. The binding constraint is the fixed dividend on STRC. Bitcoin can go sideways forever; STRC still pays 12%. The more preferred shares the company issues, the more it subordinated its common equity to a senior charge that can only be serviced through Bitcoin appreciation, new issuance, or asset sales. This is the same structural trap as a shadow bank that borrows at 12% and lends to an illiquid asset. It works in a bull market because the asset appreciates faster than the liability. It fails in a sideways market because the liability is compounding while the collateral is static. The decoupling thesis is this: investors should stop watching the Bitcoin price and start watching the STRC discount. The discount is the market's honest assessment of whether the 12% coupon is sustainable. An 11% discount implies there is a measurable probability the company either cannot pay, cannot grow, or will be forced to sell more Bitcoin to maintain the program. If the discount narrows before Bitcoin makes a new high, the market is comfortable with the financial engineering. If the discount widens while Bitcoin is flat, that is a stronger bearish signal than any single sale. There is one more uncomfortable implication. The never-sell ideology was never a covenant. It was a preference held by MSTR common shareholders. STRC holders have a different preference: they want their 12%, and they don't care whether it comes from cash or from selling Bitcoin. When those preferences conflict, the common shareholder narrative loses. The 3,620 BTC sale is the first recorded instance of that preference being enforced. It won't be the last unless Bitcoin resumes an uptrend strong enough to fund the coupon out of new issuance alone. And new issuance at a discount is just another form of selling. The Ecosystem Role: The Marginal Buyer's Dilemma Strategy's role in the Bitcoin ecosystem has been described as a liquidity pump, absorbing traditional capital and converting it into Bitcoin demand. STRC was a key part of that pump. If STRC remains below par, new issuance becomes more expensive, which reduces the capital available for future Bitcoin purchases. The 3,620 BTC sale adds a second effect: it signals to the broader market that Strategy is no longer a pure one-way buyer. It is a conditional two-way participant. Every future Bitcoin rally will now have to carry the weight of a potential dividend-driven sell order. That may also affect imitators. Several public companies have considered Bitcoin-backed preferred stock as a financing mechanism. Strategy's STRC was the template. If the template fails to return to par, the emerging corporate Bitcoin vault narrative suffers. The next company to announce a Bitcoin treasury strategy will face a more skeptical fixed-income market. The cost of capital for this entire category rises when the flagship issuer trades below par. Takeaway By September 8, we will know whether Strategy's cash buffer and buyback can hold STRC at par through a period of Bitcoin consolidation. But the more important date is the Q3 filing. If the per-share satoshi count falls again—if the company's language shifts from accumulate to opportunistically acquire—the strategy has formally moved from accumulation to liability management. The 3,620 BTC sale was a first step, not an exception. The question isn't whether Michael Saylor will sell more Bitcoin. It's whether the 12% coupon leaves him any other choice. A balance sheet is just a collection of promises with different maturities. Strategy promised common shareholders a doubling of satoshis every seven years. It promised preferred shareholders 12% forever. For two years, those promises were complementary. Now they are competing. In a sideways market, the one with seniority always wins. The auditor blinked; the market didn't. If you are long MSTR, you should be asking whether your equity claim has become just a way of paying someone else's coupon.