CEBE challenges what the industry measures. Here are the pushbacks we hear most, and where the data lands.
FD BPS divides total BTC by diluted shares. It doesn't subtract debt, preferred, or any claim senior to common. It tells you what the company holds, not what common shareholders own. When premiums compress, the stock lands near CEBE, not BPS. That gap is the claims stack, and it's growing.
It can be. Preferred raises capital without selling BTC. But it adds claims above common equity. Whether it's net positive depends on the spread: BTC growth rate minus the cost of the preferred. If the spread is positive, leverage works for common. If negative, it works against them. CEBE measures the net result.
Fiat-denominated claims do compress as BTC rises. That's real. But BTC-denominated claims (like Capital B's convertibles) are static and don't compress at all. And compression doesn't mean Senior Claims % reaches zero. At $107K BTC, Strategy's Senior Claims % is still 41%. Compression helps. It doesn't eliminate.
The framework applies to any company that holds BTC with senior claims above common equity. Size doesn't change the math. Smaller companies often carry a higher Senior Claims %, more concentrated risk, and less liquid structures. CEBE makes those differences visible.
The path does matter. Every company in this trade is on the same path, accumulating Bitcoin and issuing capital to fund it. The disagreement is the instrument. BTC per share counts every coin on the balance sheet as the common shareholder's, even the coins bought with preferred stock that the senior claim owns. So a company can show a rising BTC-per-share path while the Bitcoin belonging to common stays flat or falls. CEBE measures the same path with the senior claims removed. Read at one moment it is a snapshot. Tracked over time, on the History view of any company page, it is the path of what reached common, and if the snapshot is computed on the wrong unit, every point on the path inherits the error.
Both versions of this reduce to one instruction. Discount a perpetual claim at the rate its own risk commands and you get its market price by definition: STRC's 12% coupon against its 13.81% yield is $86.89, which is where it trades. Discounting instead at Bitcoin's expected return prices this claim off an unrelated asset. The discount to par is real value, and CEBE records it the moment a retirement happens rather than before one. Carry claims below face and the metric moves when sentiment or a forecast moves, so two analysts with different assumptions publish different current CEBE for the same balance sheet. Face is the only convention where they agree, and accrued liquidation preference already moves above face when arrears compound.
That describes gross BPS, and it is the failure CEBE was built to catch. Fund a Bitcoin purchase with preferred and gross BPS rises the day it closes: coins enter the numerator, no shares enter the denominator. CEBE prints zero on the same trade, because the claim added equals the coins bought. Nothing reaches common until Bitcoin outruns the cost of the money. Gross BPS rising while CEBE contracts is the named case, and it is why yield is measured on CEBE.
CEBE is a measurement, and scenarios belong to whoever states the assumptions. It produces one decision rule: issuance adds CEBE per share above 1.0x CEBE mNAV and dilutes below it, tested per deal as sats per new share against standing CEBE. That threshold needs no Bitcoin forecast, which is what lets two people who disagree about Bitcoin agree on whether a specific raise was accretive. Strategy adopted a 1.0x reference point for the same purpose, announced 23 July 2026 and effective the following day.
Over the whole episode it can be, but the loss came from the leverage, not the buyback. The financing leg on its own made money: a claim sold at 100 and retired at 86 is a 14-point gain, less the dividends paid while it stood. The asset leg lost because Bitcoin bought with the proceeds fell, and that loss exists whether or not anything is repurchased. The retirement itself is judged at the margin, today: pay 86 cents to extinguish a dollar of claim and the claim line falls by more than the cash costs, so CEBE rises. What funds it decides the rest: common sold at or above 1.0x CEBE mNAV retiring a below-par claim is accretive on both legs; the same retirement funded by selling Bitcoin at the bottom can go either way. The measurement shows which happened. It does not choose.
The claim's cost never moves: a 12% perpetual costs 12% of face, every year, by contract. What changes is which leg pays it and what that leg costs. Common sold at or above 1.0x CEBE mNAV raises money that costs the residual nothing, so funding the dividend that way spares the treasury without shrinking what the common owns. Sell below 1.0x and the same dividend is paid by diluting the residual, the stated rate plus a dilution cost. So the blended picture can look cheaper or dearer than 12%, but that's the funding leg's arithmetic, not the instrument's. The measurement keeps them separate: the claim costs what the contract says; the funding costs what the issuance price says; each is judged where it happens.