Finish the Math:
ProCap Financial

ProCap is run by a professional communicator, and it shows in the cleanest possible way: the announcements report the part of the equation that looks good, and they are accurate. The press release reports the Bitcoin bought, the face retired, the shares repurchased. The 10-Q reports the cash spent, the collateral moved, the compensation run rate, the shares issued. ProCap's headlines are numerators. The framework finishes the denominators. That is the whole exercise here.

Two quarters of operating history and already four headlines, each true, each leaving the hard half of the math for the footnotes. None of them involve a misstatement. All of them change the answer.

Who They Are

ProCap Financial went public through a SPAC merger on December 5, 2025, led by Anthony Pompliano, who holds 16.43% and takes a $1 salary with equity that vests only above $15 per share. The company calls itself the first publicly traded agentic finance firm, a claim it has since put weight behind: in April 2026 it launched ProCap Insights, an AI research platform built on autonomous agents, and struck a data partnership with the prediction market Kalshi. It launched with 5,000 BTC, $235M in zero-coupon convertible notes, and $195M in cash, and it began buying back its own stock four days after listing. By the end of Q1 2026 it held 5,457 BTC, the stock had fallen 40% to $2.11, and the company traded at roughly 0.44x CEBE mNAV at the Q1 close, a deep discount to the net Bitcoin behind each share. By late June the stock had slid further to around $1.42, near a fresh 52-week low, widening the very discount the rest of this article examines.

That discount is the question the rest of this article answers. The Bitcoin is real. The debt is clean. So why is the market pricing common equity at less than half the net Bitcoin behind it? The answer is not in any single headline. It is in what the headlines leave for the footnotes.

Headline One: We Deleveraged 57%

On February 9, ProCap announced it had repurchased $135.4M of convertible note face for $119.2M in cash, a $16.2M discount, retiring 57% of its debt in six weeks. True. The remaining face is $99.6M. The discount was real and accretive on its own.

Here is the part the announcement did not net. In Q4, those notes were collateralized by cash: $149.9M of ProCap's $194.9M cash position sat restricted as collateral at US Bank. Under CEBE methodology, cash nets against senior claims until it pays them, so the full cash position offset the debt, leaving net claims of just $40.1M. In Q1, ProCap switched the collateral to 3,300 BTC pledged at Anchorage and released the cash, spending it on the discount repurchase, on 457 new coins, and on buybacks. By quarter end, free cash was $26.0M and restricted cash was zero.

The face fell $135.4M. The cash offset fell $168.9M. Net claims rose $33.5M, an 83.5% increase. The Bitcoin pledged as new collateral does not leave the balance sheet, so it stays in the numerator as an asset, but it no longer offsets the claim the way cash did. It is already being counted as Bitcoin the company owns. The protection that used to stand between common equity and the debt is what disappeared.

Fully diluted BPS jumped 20.7% because it assumes the notes convert: retiring face removes conversion shares from the denominator, and with fewer shares against more Bitcoin, the ratio leaps. CEBE assumes the opposite, because at $2.11 against a $13.00 conversion price these notes are not converting. So the face stays a claim and the cash that offset it is gone. The two metrics rose together, but FD BPS rose eight times as fast, and the entire gap is the collateral switch.

None of this is improper. Pledging Bitcoin instead of cash is a reasonable financing decision, and the discount capture was genuine value. But the common equity BTC barely moved: 4,443 coins at the start of the quarter, 4,434 at the end, despite 457 coins added and 2.3 million shares retired. Everything the company gained on the asset side, it returned on the claims side. The deleveraging headline describes a real event. It just is not the event that determined what common equity owns.

The claim is still fiat-denominated, which means it compresses as Bitcoin rises. ProCap's claims share is fiat_compressive, with no static floor and no FX layer. Held at the Q1 capital structure:

CEBE here is shown at each row's Bitcoin price, the one place in this piece not normalized to $72K.

The higher claims share is real, but it is the temporary cost. Raise the Bitcoin price and it compresses, because the fiat claim shrinks against a larger reserve. The break-even is the permanent cost. Before the collateral switch, Bitcoin could have fallen to $8,028 before common equity was zeroed. After it, the floor is $13,494, sixty-eight percent higher, and unlike that share, it does not move with price. It stays there until the notes retire in December 2028. The claims share is what common equity pays while it holds; break-even is how far Bitcoin can fall before common equity owns nothing. A rising price erodes the first and leaves the second untouched. The deleveraging headline raised that floor by two thirds, and it is the one figure from this quarter that survives every move in the Bitcoin price between now and maturity.

Headline Two: We Bought the Dip, Twice

This is the headline where the math finishes in the company's favor, and it deserves to be stated as plainly as the first.

ProCap ran two accretive programs at once in Q1. It acquired 457 BTC for $36.0M while the price was depressed, and it repurchased 2.67 million shares for roughly $8.0M at an average near $3.00, below NAV per share. Buying Bitcoin below your cost basis and buying your own stock below its net Bitcoin backing are both accretive to the holders who remain. The stack grew 9.1% and the basic share count fell 2.7%.

That is the reason CEBE rose at all. Against a claims position that worsened by $33.5M, the accumulation and the buybacks together pushed common equity Bitcoin almost exactly flat and lifted CEBE 2.6% at $72K, 3.2% at the live price. Without those two programs, the collateral switch would have driven CEBE down. The buyback discipline is the offsetting force.

There is a structural logic to running buybacks when the claims share is high. Claims-share compression is front-loaded: the move from $45K to $72K takes 11 points off the share, the move from $72K to $100K only 5. Repurchasing shares cheaply while it is elevated concentrates the eventual compression into fewer shares. If the program continues and Bitcoin recovers, the holders who stayed capture a steeper per-share gain. The honest read on ProCap's first full quarter is that management made one decision that raised the claim and two decisions that fought it, and the fight was close to a draw.

The Footnote: The Wrapper Was in the Stub

At inception, ProCap looked like it might carry almost no wrapper fee at all. Zero-coupon notes mean no cash interest. No preferred means no dividends. A $1 CEO salary and a board paid entirely in performance equity mean almost no cash compensation. The early estimate was on the order of 0.17%.

The Q1 10-Q told a different story. ProCap carried 8.22 million unvested RSUs at year end, but the partial-period 10-K booked only $442K of stock-based compensation across the inception stub, which understated the run rate badly. The full quarter showed $3.54M in SBC, or $14.16M annualized. That annualization is a Q1 run-rate, not a verified yearly figure, and it most likely understates the forward number, because the incoming CTO's $1M RSU grant lands in Q2 and is not in the Q1 base. Read the 4% as a floor the Q2 10-Q will confirm or raise, the same single-period caution this piece applies to the company's 0.17%. Measured against the reserve's market value of roughly $372M at the quarter-end Bitcoin price, or about $393M normalized to $72K, that is a wrapper near 4% either way.

The cash items are as light as advertised. The number is near 4% because of equity compensation that only became visible at full-quarter resolution, more than twenty times the 0.17% the cash items alone implied. That reserve figure is market value, not cost: ProCap's average acquisition price sits above $104K per coin, putting its cost basis north of $500M, well above the current market value of the stack. The $5M one-time signing bonus for the incoming CTO is a separate event, not an annual cost, but it consumed close to a fifth of the $26M cash position. And the CFO's compensation is still unknown, because the annual proxy that would disclose it is delinquent. The 0.17% first impression was not a misstatement. It was a stub period reporting a fraction of a year.

Headline Three: A Shareholder-Friendly Acquisition

On February 9, the same day as the deleveraging, ProCap announced it would acquire CFO Silvia, an AI agent lab, in an all-stock deal to become the first publicly traded agentic finance firm. Management framed it as shareholder-friendly: the Silvia sellers receive nothing unless the stock more than triples to $9. The merger closed April 6.

The framing describes the earnout accurately. It does not describe the closing. 8,416,951 shares were issued at close, with full voting rights immediately, in exchange for no Bitcoin and no reduction in claims. That is roughly 10% dilution of common equity on day one, with up to 9 million more contingent on the $9 trigger. Institutional Shareholder Services recommended voting against the merger. ATG Capital, a shareholder, opposed it publicly and made the CEBE point directly: the closing shares dilute every holder's proportional claim on the Bitcoin treasury from the moment they issue, regardless of whether the earnout ever fires. The vote passed 68% to 32%.

The transaction was led by Pompliano as the company's controlling shareholder and chief executive, approving an acquisition whose earnout milestones track the same price thresholds that govern his own equity compensation. The incoming CTO received a $700K base, a $300K bonus target, a $1M annual RSU grant, and the $5M signing bonus paid from the thin cash position. Whether the Silvia business generates revenue that offsets the wrapper is the Q2 10-Q's question to answer, and the company spent the quarter building toward it. ProCap Insights launched in April as an agentic research product, drawing on Silvia's free consumer app that tracks more than $30 billion in user assets. A podcast followed, then data partnerships with Kalshi and, in June, Ornn. The cadence shows activity. The next filing shows whether it shows revenue. What is not open is the dilution at close: 8.4 million shares of Bitcoin-backed equity, issued for an AI platform, diluting the per-share claim the day the deal closed.

This headline, like the others, was true. The acquisition is real and may create value. The math the announcement left unfinished is the 10% that left common equity's hands before any milestone is met.

Headline Four: We Sold Bitcoin to Buy Stock

On June 1, ProCap announced it had sold roughly 52 Bitcoin to repurchase two million shares at approximately a 50% discount to NAV. For a Bitcoin treasury company, selling Bitcoin is the one move the entire category is built to never make, and it is the first time ProCap has done it. The announcement framed the sale as accretive. This time the finished math agrees.

As of May 29, ProCap held about 5,405 BTC against 88,730,996 shares, with NAV near $3.47 per share and the stock at $2.15, a 38% discount that the buyback executed against at a wider 50% point near $1.74 per share. The 52 coins sold are roughly 1.0% of the stack. The two million shares retired are roughly 2.2% of the count. Retiring a larger fraction of the equity than the fraction of Bitcoin sold leaves every remaining holder with more Bitcoin per share than before. Pompliano put it plainly: converting a small amount of Bitcoin into repurchased shares increased the Bitcoin owned by all remaining shareholders. That sentence is the CEBE argument, stated by the company itself.

The trade is accretive only because the stock trades far below NAV. Selling an asset at one hundred cents on the dollar to retire a claim on that same asset priced at fifty is accretive by exactly the size of the discount. At NAV the move would be neutral. Above NAV it would be dilutive to Bitcoin per share. The engine that makes this work is the discount, and the discount exists, as the rest of this article documents, in part because of the governance and disclosure questions the market is weighing.

The accretive buyback is powered by the same gap the buyback is trying to close.

Walk the circle once. Documented capital-structure and disclosure decisions are part of what the market is weighing: the collateral switch that removed the cash buffer, the 8.4 million shares at close, the proxy that has not been filed. The discount makes every repurchase accretive per share, including the ones funded by selling Bitcoin. The accretion raises NAV per share, and when the price does not follow, the gap widens, which makes the next repurchase more accretive than the last. This is a flywheel running in reverse. The forward version converts a market premium into Bitcoin accumulation. This one converts a market discount into per-share concentration, and the fuel it burns is the absolute stack. Each turn pays the holders who remain without addressing what the market is weighing. The wheel spins as long as the discount stays open, and it stops in only one of two ways: the gap closes, or the Bitcoin runs out.

ProCap notes it has nearly twenty years of runway at current expense levels with no revenue and no Bitcoin appreciation, so the wheel is turning by choice, not necessity. For a holder who measures what they own per share, each turn is accretive and rational. For the premise that a Bitcoin treasury company exists to accumulate an ever-larger pile of Bitcoin, it is a quiet reversal. Both readings hold at once, which is the recurring shape of this company.

What the Headlines Got Right: The Note Is Clean

A forensic read owes the company its clean findings as loudly as its complicated ones, and structurally, ProCap's convertible notes are about as benign as a fiat-denominated senior claim gets.

The conversion price is fixed at $13.00, confirmed in 10-Q Note 7, stated as 76.9 shares per $1,000. There is no formula, no VWAP discount, no most-favored-nation reset, no PIK accretion, and no amendment history. The coupon is zero, so there is no cash interest at all. The collateral is structurally protective: 3,300 BTC pledged against 2,929 required, with the company retaining sole discretion, no rehypothecation, and no forced-liquidation trigger. This is the opposite of a toxic convertible. A skeptic looking for a hidden dilution mechanism in the debt will not find one.

The Claims Grade rubric puts a letter on that. ProCap's notes are its only senior claim, so the instrument grade is also the company rollup, and as of late June with the stock near $1.42 it reads B · Moderate. The four dimensions are Permanence 1, Moneyness 3, Downside 1, Cost 0. Three of the four sit at their most benign. The notes mature rather than persist, so Permanence is a 1. They carry no cash cost, so Cost is a 0. They expand only the way any fiat claim does as Bitcoin falls, with no trigger, no compounding, and a maturity well outside the stress window, so Downside is a 1. The only weight in the grade is Moneyness, scored a 3 because the stock trades at roughly 11% of the $13 conversion price. That score is not a flaw in the instrument. It is the stock. At a conversion-clearing price the same note grades A · Minimal. The deep discount that defines ProCap's equity is the only thing standing between its debt and a clean A, which is the reverse flywheel showing up one more time: the gap that stays open is what pushes even the cleanest piece of the structure from near-equity into a treasury liability.

That matters for the contingent picture too. ProCap carries 46.9 million contingent shares against an 82.1 million base, a 57% overhang that looks alarming as a headline. But only the RSUs dilute at the current price, at roughly 500,000 shares a year. The warrants strike at $11.50, the sponsor earnout at $10.21, the Silvia earnout at $9, the converts at $13. Everything except the RSUs requires a recovery of 4.5x to 6.5x to activate. The overhang is real but almost entirely price-contingent, and the dilution curve is non-linear: a rally the equity would celebrate brings progressively heavier dilution with it. The one non-contingent item to flag is the 10 million shares issued at close to Inflection Points, an entity under common control with Pompliano, for a four-year consulting agreement, already in the outstanding count.

The $150K That Never Came

ProCap's capital structure was built for a market that expected $150K Bitcoin. The notes were struck at $13.00, with company-forced conversion available near $16.90. At $150K the stock would plausibly have cleared those levels, the notes would have converted to equity, the claim would have vanished, and ProCap would have exited the leverage phase with a clean balance sheet and the full stack intact.

At a $2 stock, the $13 conversion is more than 6x away, and the notes will mature as debt in December 2028 unless the equity recovers first. Run the Q1 balance sheet at $150K and the structure improves but does not pivot: net claims of $73.6M become 491 BTC, the claims share falls to 9.0%, CEBE rises to 6,052 sats. There is no conversion event, no maturity wall that resolves at a price threshold, no instrument that flips from claim to equity. The remaining $100M simply compresses against a larger reserve, the way any fiat claim does.

This is where ProCap separates from the companies whose $150K story is a structural pivot. For Nakamoto, $150K is the difference between a collateral cascade and survival. For Capital B, conversion triggers retire claims at price thresholds. ProCap has none of that. Its $150K is purely arithmetic compression, which is also why management did not wait for it.

The decision archaeology is the most revealing thing about this company. Rather than bet the balance sheet on $150K arriving, management retired $135M of face at 88 cents on the dollar while it had the cash, locking in a smaller permanent debt regardless of price. The cost of not waiting was the collateral switch that removed the cash buffer and raised the claims share at every price. So the trade ProCap made was a large cash buffer in exchange for a smaller absolute claim that compresses away faster. At $150K that trade looks wise, because the claim shrinks toward irrelevance and the buffer would not have mattered. At $45K it looks expensive, because the buffer is exactly what common equity wants when the claims share is climbing and break-even is rising. The company chose the path that pays off if Bitcoin rises and costs more if it falls, which is a coherent bet for a Bitcoin treasury company to make, as long as it is named as a bet and not sold as a certainty.

The Verdict

Normalized CEBE rose 2.6% from inception through Q1 2026, from 5,268 to 5,404 sats at $72K, and 3.2% at the live price. By every fully diluted measure the quarter was excellent. The distance between those two readings is the entire reason this series exists.

ProCap is the rare case where the math is clean and the open questions sit on the governance and disclosure side. The convertible notes are pristine: fixed conversion, zero coupon, protective collateral, no toxic mechanics anywhere in the debt. The Claims Grade rubric reads them B · Moderate, an A but for the underwater stock. The claims share compresses honestly as the Bitcoin price rises. And yet common equity owns almost exactly what it owned at inception, the wrapper fee is near 4% rather than the 0.17% it first appeared, the share count grew 10% for an AI acquisition that added no Bitcoin, the proxy is delinquent, the FY2025 10-K flagged a material weakness in internal controls, and the controlling shareholder ran the acquisition. None of those facts is a verdict on its own. Together they are what the market is weighing. The 0.44x CEBE mNAV at the Q1 close was not pricing the Bitcoin or the debt. It was pricing the distance between the announcements and the footnotes. And the company has built an engine that runs on that distance, paying shareholders out of the gap instead of closing it.

Every ProCap headline this quarter was true. The deleveraging happened. The dip got bought, twice. The acquisition is shareholder-friendly above $9. The Bitcoin sale did raise per-share exposure. The framework's only job was to finish the math each one started, and the finished answer was reliably different from the headline: smaller for the deleveraging and the acquisition, larger for the buyback and the Bitcoin sale, and in every case dependent on a variable the announcement did not foreground. That is not an accusation. It is the difference between what a company chooses to announce and what common equity actually owns, which is the one number this series was built to measure.

What to Watch

Several of these are brakes on the wheel described above. The reverse flywheel spins only while the discount stays open, so the items that could close it are the ones that matter most.

Convert retirement versus buybacks. The remaining $99.6M matures December 2028. Retiring it early drives the claims share toward zero, the CEBE-optimal outcome. But the June 1 Bitcoin sale shows management's current preference: it sold coins to fund buybacks rather than to retire debt. That is a defensible choice while the discount is this wide, since buybacks at half of NAV are more accretive per dollar than debt paydown, but it leaves the higher post-switch claims share in place. Watch whether any future Bitcoin sales or cash get directed at the notes instead.

Pace of Bitcoin-funded buybacks. The June 1 sale of 52 coins may be the first of a recurring program rather than a one-off. Each repeat raises per-share exposure and lowers the absolute stack. The rate at which the stack shrinks, and whether management caps it, is now a live capital-allocation question that did not exist a quarter ago.

DEF 14A filing. The annual proxy is delinquent past the April 30 deadline. When it lands it discloses the CFO's compensation, the last open input in the wrapper fee, and the expanded equity plan authorized at the March vote. A clean, on-time filing is one of the brakes.

Controls remediation. The FY2025 10-K flagged a material weakness in internal controls. Evidence that it is being remediated, in the Q2 10-Q or a dedicated 8-K, is the kind of governance repair that narrows the discount the flywheel runs on. Continued silence keeps the wheel spinning.

Q2 10-Q, approximately August 2026. The first filing with Silvia consolidated. It will show the true post-merger share count, the CTO's RSU terms, any Silvia revenue, and integration costs. Silvia revenue is the variable that matters most, because real income would be the first genuine wrapper-fee offset in the structure.

Open-market buyback pace. The program retired 2.67 million shares in Q1 through open-market repurchases, separate from the June Bitcoin-funded buyback. By May 29 the post-merger count stood at 88,730,996. Whether ordinary buybacks continue depends on available cash and the discount to NAV, and the program remains the main force offsetting the higher claims position.

Sponsor earnout window. 8.33 million shares vest if the stock reaches $10.21 for 20 of 30 trading days, in a window closing around December 2027. Deep out of the money today, but a recovery that triggers it adds roughly 10% dilution at exactly the moment the equity is healing.

$15 compensation threshold. The CEO's and board's equity vests only above $15. At current prices the cash wrapper stays near zero. Crossing $15 unlocks that equity and rewrites the wrapper story a second time.