Digital Credit Is a Transmission Belt. It Is Not a Cap.

A widely followed macro analyst argued that bitcoin-backed preferreds would ensure bitcoin never reaches $1m, by analogy to the paper-gold market of the 1980s. The analogy fails on one fact: these instruments are denominated in dollars, not in bitcoin. What actually governs the structure is a different constraint, and it is arithmetic.

Digital credit is the market for dollar-denominated preferred stock issued by companies that hold bitcoin. Strategy issues several series, including STRC. Strive issues SATA. A buyer hands over dollars, receives a cash coupon, and the issuer spends the proceeds on bitcoin. The instruments pay in dollars and target a price near $100. They do not pay in bitcoin.

On 5 October Luke Gromen, a macro analyst, argued that if digital credit develops it will ensure bitcoin never reaches $1m, and compared it to the London credit-gold market of the mid-1980s. The comparison drew a broad response from bitcoin analysts, most of it arguing the mechanism runs the other way.

The direction of the flow is the whole question, and it is settled by what the instrument is denominated in. Unallocated gold created claims on gold that were themselves payable in gold. Digital credit creates claims payable in dollars and spends the proceeds on bitcoin. That is a different trade, and it carries a different constraint. This piece is about which constraint binds, why June 2026 was the first test of it, and what to measure each month from here.

The claim

Gromen's argument, posted on 5 October 2026, is that "if digital credit is developed it will ensure BTC never goes to $1m", and that this "may even be the point of digital credit". He points readers to the London credit-gold market that began in the mid-1980s, where banks created claims to gold that did not correspond to specific allocated bars.

The mechanism he is reaching for is real, and it did work the way he describes. An unallocated gold account is a claim denominated in gold, settled in gold, and backed by the promise of the dealer rather than a specific bar. Because the claim and the asset were the same unit, the market could create more claims than there was metal, and the marginal price of gold was set by the flow of paper rather than the flow of bars. Supply of claims expanded while supply of metal did not.

The question is whether bitcoin preferreds reproduce that. They do not, and the reason is in the denomination.

The denomination decides the direction

STRC is a dollar-denominated income stream. The holder receives USD distributions, targets a trading price near $100 of par, and holds a residual claim on the issuer as preferred equity. The holder does not receive a claim on a quantity of bitcoin. There is no redemption at par in bitcoin, and no promise of one.

That inverts the credit-gold mechanism rather than repeating it. In the gold case, the buyer of paper wanted gold exposure and received a gold-linked claim, so the claim substituted for the metal. In the digital-credit case, the buyer of paper wants a dollar yield and hands over dollars, and the issuer converts the proceeds into bitcoin on the market. The claim substitutes for a dollar deposit, not for bitcoin. The bitcoin ends up in the same treasury either way.

The best summary of this came from the response that prompted it, by Joe Burnett of III Capital on 5 October 2026: credit gold "expanded the supply of claims on gold", while digital credit "can expand the supply of USD claims while reducing the BTC available to the market". On the direction of the flow, that is correct, and it is the sentence that settles the analogy.

But the response over-reads its own case

"Issuance buys real bitcoin" is true and incomplete. The bitcoin is bought with credit, not with savings. A dollar coupon is an obligation that runs regardless of what bitcoin does, and the collateral behind it is the asset that funds it when times are bad. That makes digital credit a lever on the treasury, and a lever cuts both ways.

June 2026 was the first live test. STRC broke below its $100 par target, with reports of an intraday low near $82.50 and lower in the worst of the move. Strategy sold 32 bitcoin in the week to 31 May 2026, at an average price of $77,135, for approximately $2.5m, and the filing states the proceeds were expected to fund distributions on the preferred stock. It was the company's first disclosed bitcoin sale since December 2022. Holdings were 843,706 bitcoin as of 31 May 2026.

That is the fact that neither framing owns. Below par, new issuance stops and the marginal buyer disappears; the dollar coupon does not. A structure that removes bitcoin from the market when it is rising can put bitcoin back on the market when it is falling, because it sells the reserve to pay the coupon. Gromen's worry is real, but it lives in the coupon, not in synthetic supply. Strategy raised the STRC rate from 11.50% to 12.00% over the period and rebuilt its dollar reserve against coverage.

The constraint is the coupon, and the coupon is a spread

The instrument holds par by paying enough to hold it. A perpetual preferred with a par anchor is a money-market substitute for income buyers, and it competes with short and intermediate Treasury yields. When that spread is too thin, the price falls below par and the issuer loses the ability to fund at par, which is the whole engine.

The numbers are not close. STRC carries a 12.00% coupon against a US 10yr yield of 5.28% on 2 October 2026, per Treasury data. The instrument pays roughly 6.7 points over the 10yr to hold $100. That is the price of the structure. It is not the price of a cap on bitcoin; it is the price of borrowing dollars against a volatile asset.

The cost of the structure is showing in the accounts. Preferred distributions ran $507.7m in the second quarter of 2026 and $758.2m in the first half, against $58.1m in the first half of 2025. The run-rate is roughly $1.7bn a year, and accrued payables on those distributions rose from $27.1m to $155.2m over the half. Mezzanine equity, the preferred sleeve, went from $8.03bn to $15.46bn in 6 months. Every dollar of that is a claim senior to the common.

The engine that is actually closed

The mechanism that both sides of the argument underrate is the equity premium. Bitcoin treasuries accumulate most efficiently when the common stock trades above the value of the bitcoin it holds, because then issuing shares buys more bitcoin per share. Below that line, issuing common to buy bitcoin destroys bitcoin per share, and the engine is closed by arithmetic rather than by policy.

On 6 October 2026 the premium is gone. Strategy traded at $164.55 against a net asset value of $188.51 per diluted share, an mNAV of 0.8729, on 846,000 bitcoin and a diluted share count of 384.23m. At a discount, the common equity is not an accretive funding tool. That is why the flow has moved to buybacks and preferreds rather than common issuance, and why bitcoin purchases slowed to $28.7m in a recent week against a treasury sale redirected toward buybacks.

This is the honest cap on digital credit, and it is not suppression. Throughput is set by how much par-anchored dollar yield the income market will absorb. Each new dollar raised requires a coupon, the coupon stacks behind the convertible and the senior claims, and as the sleeve grows the required spread widens. The structure decelerates on its own arithmetic long before it caps anything.

Where the suppression argument does fit

The gold analogy is not wrong in general. It is misapplied to this instrument and correctly applied to a different one. The mechanism requires the claim and the asset to share a unit, so that claims can multiply without the asset multiplying with them.

That condition is met by cash-settled bitcoin futures, by securities lending against spot bitcoin ETF shares, and by exchange IOU balances that are denominated in bitcoin and redeemable in bitcoin. Each creates bitcoin-denominated claims that can grow without new coins. Those are where paper supply would accumulate if it accumulates anywhere. A dollar preferred with a par anchor is not that instrument, because its unit is the dollar and its buyer is a yield account rather than a bitcoin accumulator.

The falsifier is worth stating plainly. If digital credit were a tool for capping bitcoin, the issuer would not be forced to sell the reserve when the price falls. In June 2026 it was. A cap does not bleed its own balance sheet to hold par.

What to measure each month

The argument is about a rate of change, so it has to be measured rather than settled once. Seven readings decide whether digital credit is amplifying bitcoin demand or grinding against it, and each is available from filings and market marks.

First, the mNAV: whether the common trades at a premium or a discount to net asset value per share, because that sets whether the equity engine is open. Second, the STRC price against $100 of par, because below par the funding engine stalls. Third, the weighted coupon against the 10yr, because that spread is the cost of the whole structure. Fourth, bitcoin per share, which is the only accumulation metric that survives dilution. Fifth, dividend coverage, being the dollar reserve against the quarterly obligation. Sixth, net issuance and whether the company is buying or selling the reserve. Seventh, the shape of the preferred stack, because a deeper stack raises the required spread on the next dollar.

The reading on 6 October 2026 is unambiguous. The premium is closed at 0.8729, the preferreds are funding at par only because the coupon is high, and the reserve has been touched once already. That is a transmission belt between bitcoin and the dollar income market, running in both directions. It is not a cap on bitcoin, and it is not free demand. It is a cost of capital, and this month the cost is rising.