Strategy's STRC prints roughly $150 million in daily volume. Within ten basis points of the quoted touch, it absorbs about $28 million โ a depth figure that makes every competing preferred security in the market look illiquid by an order of magnitude. On the surface, this is the most liquid instrument in the Bitcoin treasury complex: a floating-rate perpetual preferred trading tight to its $100 reference price, behaving, by reputation, like a cash proxy.
That reputation is the problem.
Remove the issuer's own buyback program from the tape and the picture inverts. Trading capacity falls to roughly 80% of the headline number. The time needed to liquidate a $50 million position stretches out. And the correlation structure exposes what the turnover statistics conceal: STRC's liquidity tracks its own deviation from par โ a coefficient of 0.43 โ far more tightly than it tracks Bitcoin's volatility, at 0.10. The instrument does not trade primarily because the market wants it. It trades because its issuer is standing on the other side. Liquidity is the only truth in a volatile market, and in this case the liquidity has a single author.
To understand what STRC is, separate it from the asset that made its parent famous.
Strategy โ the company formerly known as MicroStrategy โ is the largest corporate holder of Bitcoin. That fact dominates coverage. The more consequential structure is how the position is financed. The balance sheet is not a simple equity-plus-debt stack. It is a layered capital structure: MSTR common at the top of the risk curve, a series of convertible notes beneath it, and a growing family of preferred securities โ STRK, STRF, STRE, and now STRC โ wedged in between. Each tranche targets a different buyer. The common serves those who want Bitcoin beta. The converts serve arbitrage desks. The preferreds serve income and credit funds that want a yield without the drawdown profile of the common.
STRC is the newest tranche and the most engineered. It is a floating-rate perpetual preferred with a $100 reference price. The coupon resets, and the reset is the entire point: by adjusting the payout to prevailing rates, the instrument is designed to trade near par. A fixed-rate perpetual preferred drifts with the rate cycle. A floating-rate one is meant to stay anchored. The design promise is stability โ buy a hundred-dollar instrument, collect a floating coupon, expect the principal to remain intact.
I spent early 2024 mapping institutional flows into this complex after the spot Bitcoin ETF approvals. The result that stayed with me: only about 15% of initial inflows were genuinely new capital. The rest was portfolio rebalancing. That distinction โ new money versus rotated money โ is the same lens that applies here. STRC's depth looks like new liquidity. A meaningful share of it is the issuer's own money, moved from the treasury into the buyback.
STRC's central selling proposition is proximity to par. The floating coupon is supposed to enforce it. The buyback is supposed to backstop it. Strategy authorized a $2 billion repurchase program for the instrument. As of October 4, roughly $1.45 billion had been deployed. That leaves about $547.2 million of remaining authorization โ call it 27% of the original mandate.
For most of the second half of 2025, the mechanism appeared to work. The instrument recovered from the mid-$70s and settled near $99.50, close to its $100 reference. To a credit investor, that looks like a functioning anchor. To an analyst who has watched reflexive structures before, it looks like a price being held in place by a buyer.
The crack in the design: June. In June, STRC traded down to the mid-$70s. Read that number in coupon terms. A move from $100 to the mid-$70s is roughly equivalent to two years of coupon payments โ a deviation that a floating-rate design is specifically built to prevent. The anchor failed. The instrument can and did trade well below par, and the recovery to $99.50 was not the mechanism reasserting itself; it was the buyback doing the work the mechanism was supposed to do on its own.
The floating-rate design does not eliminate price deviation; it only reframes who is responsible for correcting it. The design promise and the buyback program are not the same thing. One is structure. The other is a discretionary, finite, capital-consuming intervention. And the June move may not have been noise. A drop to the mid-$70s in an instrument designed to trade at par carries the signature of a liquidity or credit event, not ordinary volatility. The subsequent recovery via buyback reads less like organic demand returning and more like a repair operation. If that reading is correct, the buyback is not a routine capital allocation tool. It is a contingency instrument that has already been deployed once under stress, and its remaining capacity is now smaller.
Here the Keyrock data is unambiguous, and it is the most important set of numbers in the entire structure. When STRC trades within 1โ3% of par, its liquidity falls by roughly a factor of four. When it deviates by more than 6%, liquidity falls by roughly a factor of eight.
Read that ordering carefully. Liquidity is not merely lower in stress; it collapses precisely when it is most needed. At the moment a holder wants to exit โ when the price is under pressure โ the depth that made the position look safe evaporates. This is textbook procyclical liquidity, the same dynamic that turned the 2022 drawdown into a cascade. In 2022 I modeled correlated exposures between algorithmic stablecoins and lending pools and flagged a 40% potential drawdown in uncollateralized lending venues. The mechanism was different. The pattern is identical: an instrument whose exit liquidity is a function of its own calm.
The best ten percent of trading days do not describe the risk. The worst ten percent do. And on the worst decile of days, STRC's absorbable depth falls to roughly $6.5 million โ a figure that turns a position the market calls liquid into one that is, in practice, trapped. Depth that disappears under stress is not depth. It is a fair-weather option the market writes to itself and forgets it has sold.
This is where the correlation analysis earns its keep, and where the conventional framing breaks. STRC's liquidity correlates with its own price deviation from par at 0.43. It correlates with Bitcoin's volatility at 0.10.
The implication is structural. If STRC were a Bitcoin beta instrument โ a yield-bearing proxy for the underlying โ you would expect its liquidity to track the volatility of the asset it finances. It does not. Its liquidity tracks its own dislocation. The instrument's primary risk factor is reflexivity: the feedback loop between its price, its perceived safety, and the willingness of market makers to quote it.
This matters for anyone who bought STRC as a low-volatility way to hold Bitcoin exposure. STRC is not a Bitcoin instrument with a coupon. It is a credit instrument whose behavior is governed by its own microstructure. The Bitcoin connection is upstream; the risk is local. Risk is not avoided; it is priced and hedged โ and here the market is pricing the wrong risk. It is watching BTC volatility while the actual exposure sits in a self-referential liquidity loop.
Strategy presents itself as a Bitcoin accumulation vehicle. The narrative is singular: convert capital into Bitcoin. But the balance sheet tells a more complicated story, and the week of September 28 to October 4 makes it explicit.
During that week, Strategy spent $176.3 million on STRC buybacks โ $102.6 million in the final three days of September and $73.7 million in early October. Over the same period, it purchased $28.7 million of Bitcoin, or 334 coins. The buyback was roughly six times the Bitcoin purchase. Of the buyback total, $154.1 million came from cash reserves; $13 million funded the Bitcoin buy.
Sit with that ratio. The company whose identity is Bitcoin accumulation spent six dollars supporting its own preferred stock for every one dollar it spent on the asset it exists to accumulate. The buyback and the Bitcoin purchase draw on the same pool of capital. Every dollar defending STRC's price is a dollar not converting into the treasury asset. This is not a rounding error. It is a revealed priority.
The natural reading is that Strategy's operational priority has shifted from offense to defense โ stabilize the financing instrument first, accumulate Bitcoin second. That is a rational sequencing if the preferred is the load-bearing wall of the capital structure. It is also a confession that the instrument cannot stand without being held up.
The buyback program is not a permanent feature. It is a $2 billion authorization with $547.2 million left. During the observed week, buyback activity represented more than 20% of STRC's weekly trading volume, peaking near 28%.
Run the exhaustion arithmetic. At the observed pace of $176.3 million per week, the remaining $547.2 million supports roughly three more weeks of intervention at that intensity. The actual pace will almost certainly slow; the company will ration the authorization. But exhaustion is not a matter of if. It is a matter of when.
This is the pre-mortem question I keep returning to: what happens to STRC's liquidity when the authorization is spent? The headline depth of $28 million within ten basis points assumes a market maker complex willing to quote. If a material share of that willingness is a response to the issuer's own bid, then the depth is not a market property at all โ it is a derivative of a discretionary program with a finite balance. A liquidity profile that depends on a single, depletable buyer is not a liquidity profile. It is a countdown.
Depth is only as durable as the holders behind it. Here the composition is revealing. A meaningful portion of STRC's volume comes from arbitrage firms and high-frequency traders responding to the issuer's activity โ the monthly coupon reset creates windows that these desks harvest. There is institutional participation from preferred-income fund managers, which is genuine. But the marginal, most active layer of the tape is reactive flow, not committed allocation.
The distinction matters for stability. A market with a deep base of long-horizon holders โ the ballast that keeps a bond market functioning through stress โ behaves differently from one dominated by traders who arrived because the issuer was active. The former absorbs shocks. The latter amplifies them. If the issuer steps back, the reactive layer has no reason to stay, and the depth it provided leaves with it. A holder base that arrived because of the issuer will depart because of the issuer. Loyalty to a spread is not loyalty to an instrument.
To be fair to the structure โ and an analyst who only builds the bear case is selling a narrative, not an analysis โ STRC's liquidity advantage in normal conditions is genuine. Keyrock itself acknowledges that STRC remains far more liquid than competing preferred securities. The comparison is not close: STRC absorbs $28 million within ten basis points, while comparable fixed-rate preferreds and Strive's SATA absorb less than $3 million on the same measure. A $50 million position clears in under two days in STRC. The same position in a peer instrument takes roughly five days.
So the lead is real. It is also conditional. Remove the buyback, and the effective trading capacity falls to about 80% of the reported figure, and the exit window lengthens. The advantage narrows. And because the advantage rests on issuer self-purchase rather than network effects or brand stickiness, it is not self-sustaining. A competitive moat built from the issuer's own balance sheet is not a moat. It is a subsidy, and subsidies have budgets.
Strategy is the benchmark issuer in the Bitcoin treasury company complex. Its preferred securities set the reference for how peers โ Strive, Metaplanet, and a lengthening list of imitators โ will finance their own accumulations. That makes STRC's liquidity structure a template, not an isolated case.
If the benchmark issuer's preferred liquidity depends on its own buyback, then the template being exported to the sector is one of manufactured depth. A peer that copies the structure inherits the dependency: it must allocate capital to defend its own preferred to keep its financing cost down. The more issuers adopt the model, the more of the complex's apparent liquidity becomes self-referential. In a downturn, that is not diversification. It is correlation.
I flagged a version of this in 2020, when I modeled Compound's interest-rate algorithms and found a liquidity fragmentation risk that would surface if stablecoin pegs deviated beyond 2%. The lesson then and now is the same: in a system of shared dependencies, the failure of one node is not contained. It propagates along the same wires that carried the confidence. The wires here are the capital markets' willingness to fund Bitcoin treasury companies. If that willingness cools, the reflexive preferreds cool first.
One detail deserves scrutiny on its own. The reporting contains an inconsistency: one figure cites a $1.5 billion preferred repurchase program, another cites the $2 billion authorization. The discrepancy is modest in isolation, but it points to something larger. When an instrument's liquidity depends on a discretionary program, the precise size of that program is not a footnote โ it is the central risk parameter. If the market cannot agree on whether $1.5 billion or $2 billion was authorized, it cannot price the runway. Opacity around a support mechanism is itself a risk factor. You cannot hedge a number you cannot see.
There is a quieter structural issue beneath the liquidity debate: where the coupon comes from. STRC pays a floating dividend. Reverse-engineer the magnitude from the June move โ a drop to the mid-$70s equates to roughly two years of coupon, and a seven-month coupon roughly equates to a typical deviation from par โ and the implied yield sits somewhere in the high single digits to low double digits. That is a substantial obligation across a multi-billion-dollar preferred stack.
Strategy's operating business is software, and its cash flow is trivial relative to the dividend burden on its preferred securities. The dividend is therefore not funded primarily by operations. It is funded by access to capital markets โ by issuing more instruments, by refinancing. This is not a Ponzi scheme in the classical sense; there are real assets underneath, a real Bitcoin treasury and real equity. But the funding structure sits adjacent to a reflexive loop: the coupon is serviced by continued market access, and market access depends on the instrument appearing stable, and the instrument appears stable because the issuer is buying it. A structure is not fraudulent merely because it is reflexive. But it is fragile, and fragility deserves to be priced.
The consensus reading of a buyback is defensive โ capital spent to hold a price, a subsidy with no return. There is a counter-argument that deserves an honest hearing, because it may be right.
If the buyback keeps STRC trading near par, it keeps the instrument's perceived risk low. A low perceived risk supports a lower coupon on future issuance. In that frame, the buyback is not a cost; it is an investment in the company's cost of capital. Every dollar spent defending par buys a lower rate on the next preferred tranche. That is self-referential financing in the most literal sense โ a bootstrapped loop where the issuer's support of the secondary market lowers its primary-market cost.
This is precisely why the structure is fragile rather than fraudulent, and why the distinction matters. A rational cost-of-capital play and a reflexive dependency look identical on a good day. They diverge only under stress โ which is exactly when the market learns which one it was holding.
The blind spot is the denominator. The market reads turnover, the absolute volume that makes STRC look liquid. It does not read the buyback share of that volume โ the fraction of the tape that exists because the issuer chose to be there. One number is a headline. The other is the load-bearing assumption. When those two diverge, the market is pricing the wrong variable, and it will not know until the buyback stops. What looks like broad market participation is, on inspection, a single participant with a mandate.
The variable to monitor is not STRC's daily volume. It is the remaining buyback authorization, and the rate at which it is being spent. At the observed pace, the runway is weeks, not quarters. The question for the next cycle is simple and unanswered: when the authorization is exhausted, does STRC's depth hold on its own โ or was the depth never the market's to begin with? The instrument will answer that question without being asked. It always does, at the worst possible moment.