Finish the Math:
Capital B

Capital B is the only Bitcoin Treasury Company in the CEBE Tracker whose drag does not move with price. The rest have curves. Capital B has a staircase. Drag steps down when management makes it step down. Price will not do the work. The reason is one architectural decision the rest of the industry did not make. They borrowed Bitcoin in Bitcoin.

Most Bitcoin Treasury Companies borrow in fiat. Dollars, yen, euros, pounds. Fiat debt compresses in BTC terms when Bitcoin rises and expands when Bitcoin falls. Drag moves continuously with price because the fiat numerator stays fixed while the BTC denominator changes. Capital B's debt is denominated in the asset on the balance sheet. Numerator and denominator are in the same unit. The 1,060 BTC claim is 1,060 BTC at every price. Net drag varies modestly with cash offset (currently 32.2% at $72K, would be 34.2% at $150K), but the variation comes from cash sitting in fiat, not from the claim expanding. The claim does not expand under any market condition.

That removes one expansion vector from the structure. Every other capital structure in the tracker has at least one path for drag to grow during a drawdown, and Capital B does not. The market priced this feature across 553 consecutive days. Capital B has never closed below 1.0x mNAV in its entire treasury history. The worst single day was 8.04% above NAV.

The team did not stop at the patient structure. They built a second mechanism that runs through institutional capital markets. The first cycle came on September 16, 2025 as a EUR 58.1M Accelerated Bookbuild, Europe's first ABB for any Bitcoin Treasury Company. The second came on May 11, 2026 as a EUR 15.2M private placement attached to a 92-million-warrant self-liquidating cascade. The warrant strikes climb from EUR 0.86 to EUR 1.46. The trigger thresholds climb from EUR 1.118 to EUR 1.898. The company controls the accelerated exercise window. Unexercised warrants inside that window go void. The cascade is built to convert capital market momentum into Bitcoin accumulation, with no maturity date forcing any decision regardless of price.

Two coupled engines. One retires BTC-denominated debt through price-triggered conversion. One raises new equity at premium valuations through accelerated bookbuilds and warrant cascades. Their outputs feed each other through stock price. Neither carries a calendar wall. The whole capital structure runs on price signals and human execution, not clocks.

I want to walk through the architecture, the people who built it, the three transactions that proved it works between September 2025 and May 2026, the market's read across 553 days, and the design intent visible in the sequence.

The Watermill

The CEBE framework categorizes capital structures into three drag types. Fiat-denominated debt (most BTCTCs) compresses with BTC price because the fiat face stays fixed. FX-denominated debt (a small subset) has two compression vectors, BTC appreciation against USD plus FX dynamics. BTC-denominated debt is one company. Capital B.

The mechanical implication separates clearly. Capital B's gross BTC claim was 1,060 BTC at $50K BTC. It is 1,060 BTC at $72K BTC. It will be 1,060 BTC at $200K BTC, until a conversion event eliminates it. The claim curve is a horizontal line.

This is why the mechanism behind the company is a watermill.

Picture an overshot waterwheel. Water enters at the top from a sluice. A gate controls how the water hits the wheel. The wheel itself is built with buckets along the rim. Water fills a bucket. The bucket gets heavy. Eventually it tips, empties, and continues around. As long as water flows and the gate is set, the wheel turns.

Capital B's OCA system maps cleanly onto this image.

The sluice is macro liquidity. Institutional allocations, sovereign treasuries, ETF flows, retail accumulation. The channel that delivers capital toward Bitcoin. Outside the company's control. Can run full or thin. Between October 2025 and March 2026 it ran thin.

The water is liquidity itself. Liquidity raises BTC, raises NAV, raises the stock floor, raises Capital B's share price toward the conversion trigger on each bond.

The gate is capital management. Conversion prices, trigger conditions, repricings, sweeteners. The company's direct lever. March 17 the gates were widened on TOBAM tranches A-03, A-04, and A-05 (conversion prices halved). March 30 the gates were widened again on Blockstream B-01 and B-02 (trigger gates removed). May 4 the gate was widened on Adam Back's B-04. Each widening means a bucket tips at a lower fill line than it did before.

The buckets are OCA tranches. Eight active buckets are built into the wheel. The B-01 remainder held by Blockstream Capital Partners. The B-01 remainder held by TOBAM. The B-02 tranche held by Blockstream, the largest single bucket at 57% of remaining face. A-03, A-04, and A-05 held by TOBAM. B-03 held by Moonlight Capital. B-04 held by Adam Back personally. Each bucket holds a fixed BTC claim. The aggregate is 1,060 BTC.

Gravity is time. Capital B's conversions require duration, not just price level. Force conversion needs 20-day VWAP above 130% of the conversion price. The force conversion authority itself does not unlock until year three of each tranche. April 2028 for B-01. June 2028 for B-02. July 2028 for B-04. Time pulls every day a bucket sits full.

The tip is a conversion event. The bucket empties. The BTC claim is eliminated. Drag steps down by subtraction from the claims numerator.

Liquidity fills the buckets. Time tips them. The wheel turns at the speed of Bitcoin.

This system was operational by mid-2025. The team that built it and the strategic coalition behind it shaped how it would actually run.

The Coalition

Capital B adopted the Bitcoin standard on November 5, 2024 as the first European corporate to do so at this scale and the first listed company on Euronext Growth Paris to make Bitcoin its treasury asset. Saylor had demonstrated the playbook in the United States. Capital B answered from Europe with a different engine pursuing the same goal through different mechanics.

The team is small.

Alexandre Laizet is Deputy CEO and Director of Bitcoin Strategy. He named the structure publicly in March 2026 as Bitcoin-denominated convertible debt. The annual report filed April 30 carries the same language. Laizet confirmed the BTC-denominated character of the obligation in correspondence on April 3, 2026. The staircase and watermill framings used in this piece are analytical, derived from the mechanics of the structure rather than attributed to him.

Jean-Philippe Casadepax-Soulet is CEO. His compensation is not disclosed in the FY2025 annual report. That gap is worth noting as a governance item.

Jean-Francois Descaves is Chairman and ESG Director (EUR 150,000 employment contract verified). Ludovic Chechin-Laurans rounds out the board.

The strategic investor coalition is where the engineering signal sits.

Adam Back holds B-04 personally. That single-instrument concentration is the conviction signal in the coalition. The CEO of Blockstream and only person cited in the Bitcoin whitepaper put his own euros against a specific OCA tranche rather than a passive position in basic equity. The B-04 strike was repriced from EUR 5.174 to EUR 2.59 on May 4, 2026. He subscribed to EUR 1.1M of BSA 2026-02 warrants the same day and to the May 11 institutional private placement a week later. His maximum potential position across all instruments runs above 54 million shares. The conviction is the structural read, not the resume.

TOBAM holds across the OCA stack: the B-01 remainder, A-03, A-04, and A-05 tranches, plus 18.3M of the BSA 2026-01 warrants and an allocation in the May 11 placement. That distribution is an alignment signal. It is also a concentration risk worth naming. A single counterparty across multiple tranches means a single counterparty's decisions affect multiple steps on the staircase. Yves Choueifaty (TOBAM's founder) and Laizet appear together on podcasts and at conferences with notable frequency. TOBAM is not just a holder. It is a co-presenter of the thesis with reputational capital tied to instruments management designed.

Blockstream Capital Partners holds the single bucket that transforms the steppe. B-02 at EUR 55.3M is 57% of remaining OCA face, gate-free since March 30. Its conversion drops gross drag from around 30% to around 10.5% in one event. The largest structural step on the wheel sits in one institutional holder's hands.

Moonlight Capital holds B-03 (EUR 4.6M). UTXO Management carries OCA exposure totaling 7.99M shares on a fully diluted basis. Both sit on the watermill in positions analogous to Blockstream's, smaller in scale.

In September 2025 the coalition expanded through institutional capital markets. Cantor Fitzgerald Europe served as Global Coordinator and TP ICAP Europe SA served as Joint Bookrunner for a EUR 58.1M private placement. Capital B's own press release framed it explicitly as Europe's first Accelerated Bookbuild for a Bitcoin Treasury Company. 35 institutions across 10 countries subscribed, including Blockchain.com as a notable strategic participant. TOBAM added to its existing position. The pricing was EUR 1.55 per share at a 9.1% discount to 5-day VWAP, which translated to approximately 2.3x mNAV at execution. Capital B raised institutional capital at a premium that quarter during a period when the broader BTCTC sector was compressing toward parity. Swissquote Bank Europe SA handled the proceeds, custody routed through Swiss company Taurus, the same partner used elsewhere across the company's holdings.

On May 11, 2026 the coalition extended again, this time through a US-aligned channel. Maxim Group LLC of New York served as lead placement agent for the EUR 15.2M ABSA institutional private placement. Marex S.A. served as European co-manager. Maxim Group does not work Euronext Growth micro-caps unless it sees a path to real US institutional capital formation.

Two institutional channels operated in seven months. Cantor Fitzgerald Europe and TP ICAP delivered global capital at 2.3x mNAV in September 2025. By May 2026, Maxim Group had extended the engine to US institutional formation at 1.3x mNAV. The team did not just open a new sluice on the watermill. They started a second engine in 2025, and they ran it twice.

March 30: The Watermill Turned

The staircase is the watermill's projected path down. Each step is a conversion event.

B-02 is the transformative step. One conversion drops gross drag from around 30% to around 10.5%. The first step was taken on March 30, 2026.

Through summer 2025 the watermill was spinning fast. Liquidity was strong, the OCA stack was scaling, buckets were filling toward triggers. Then from October 2025 to March 2026 the sluice ran thin. Bitcoin fell from $126K to around $60K, the watermill slowed, and buckets stopped tipping. The first bucket on the wheel did not move until BTC began its recovery.

That morning, Blockstream Capital Partners converted the original Fulgur position in B-01. EUR 19.9M of bonds became 41.8M new shares. Basic shares rose from 229,727,727 to 271,572,205 in a single transaction.

Working the claim reduction at $72K BTC, the bond face was originally calibrated to a BTC reference at issuance rather than to today's price. The proportional walk from the Q4 2025 anchor of 1,279 BTC claim to the post-conversion 1,060 BTC claim is 219 BTC. That is the BTC quantity eliminated by the March 30 conversion. The face was EUR 19.9M and the BTC obligation behind it was set when the bonds were priced. Converting at any later BTC price retires the originally calibrated BTC quantity.

CEBE per share landed approximately neutral at $72K normalization. The 219 BTC of claim reduction was offset by the 41.8M of new shares spreading common equity BTC across a larger denominator. Drag stepped from approximately 44% to approximately 35%. Common equity ownership rose from approximately 56% to approximately 65%. The aggregate BTC backing common equity grew by 219 BTC. The wheel turned.

This is the Finish the Math moment.

FD BPS shows nothing meaningful on this event because it had already loaded the conversion shares into the denominator from the day the bonds were issued. Sats per share at the FD level moves only when actual BTC changes hands. The 219 BTC claim reduction is structural, not transactional, so FD BPS reads no change.

CEBE per share also stays steady because the claim reduction and share dilution roughly cancel at this price anchor. What moves is the ownership composition. The 9 percentage point step in common equity ownership is the structural read. Drag stepped by the same amount in the opposite direction. The reader who looks only at sats per share sees nothing. The reader who looks at the composition behind those sats sees the wheel turn.

At $72K BTC and 1.17 EURUSD, FD BPS sits at 742 sats and CEBE sits at 723 sats. They are 2.6% apart at this price. The convergence looks like agreement. It is arithmetic coincidence. FD penalizes the denominator by inflating it with potential conversion shares. CEBE penalizes the numerator by subtracting the BTC claim. Two different penalties producing similar outputs at one ratio. The two metrics diverge under stress.

March 30 was the proof. Same event, same day. The composition shift was visible in drag and common equity ownership. The per-share metrics on both sides showed nothing. The framework that tracks claim composition caught the structural improvement. The framework that tracks per-share counts did not.

The watermill works when liquidity is flowing.

The Flywheel

Two engines with opposite mechanics. The watermill retires claims through conversion, the flywheel adds Bitcoin through institutional capital. Capital B built both, and ran the flywheel before the drawdown started.

Cycle 1: September 2025. On September 16, 2025, Capital B announced a EUR 58.1M institutional private placement at EUR 1.55 per share through an Accelerated Bookbuild. Cantor Fitzgerald Europe served as Global Coordinator. TP ICAP Europe SA served as Joint Bookrunner. 37,508,088 new shares were issued at a 9.1% discount to 5-day VWAP. 35 institutions across 10 countries subscribed, including Blockchain.com as a new strategic participant alongside existing strategic investor TOBAM. Net proceeds funded the acquisition of 530 BTC at an average cost of EUR 99,272 per coin. By September 22, total group holdings reached 2,800 BTC and BTC Yield was 1,651.2% YTD.

The transaction was Europe's first Accelerated Bookbuild for any Bitcoin Treasury Company. It was executed at approximately 2.3x mNAV during a period when the broader BTCTC sector was compressing toward parity. Capital B priced institutional capital at a premium when peer companies could not.

That was the flywheel's first cycle.

From October 2025 through March 2026 Bitcoin fell from $126K to around $60K. Capital B paused both engines and held position for the duration of the drawdown.

Cycle 2: May 2026. On May 11, 2026, Capital B announced a EUR 15.2M institutional private placement at EUR 0.66 per share, a 1.51% premium to the prior close. Maxim Group LLC served as lead placement agent. Marex S.A. served as European co-manager. The mNAV at execution was approximately 1.3x. The structure closed at a different shape from September.

The raise. 23,038,844 new shares. Net proceeds of approximately EUR 14.4M deploy to BTC at market prices, expected at approximately 182 BTC of new exposure for the common stack based on management's near-term execution plan. The structure adds no new debt and no new senior claim. Drag falls because the BTC numerator grows while the static claim stays flat.

The cascade. Every new share carried four warrants spread across three tranches.

The tranches unlock sequentially. Each one requires the stock to trade at or above 130% of its strike for the required period before the company can force accelerated exercise. Warrants unexercised inside the accelerated window go void. If all warrants exercise, the structure delivers EUR 99.1M in additional capital, EUR 114.3M total when combined with the initial raise.

The cumulative track record. The two cycles raised EUR 58.1M in September 2025 at 2.3x mNAV and EUR 15.2M in May 2026 at 1.3x mNAV. That is EUR 73.3M of institutional capital deployed into approximately 712 BTC of additions to the common stack. An additional EUR 99.1M of warrant capital is contingent on the price path. That is 35 institutions in 10 countries via the European channel plus US institutional placement via the New York channel. The institutional base of Capital B is not concentrated. It is global and broadening.

The accretion test. 182 BTC of new exposure divided by 23,038,844 new shares yields 790 sats per new share for the May 11 raise. Existing CEBE at $72K is 723 sats. The new shares deliver more Bitcoin per share than the existing shares carried. Accretive at the CEBE level by 67 sats per new share. FD BPS reads the same event as dilution because the 92 million new warrants attach to the FD denominator while the 182 BTC of expected deployment is small against the larger FD total. Same event, opposite reads. Same pattern as March 30.

The flywheel. This is structurally different from the watermill. The watermill's buckets carry BTC claims and tip via conversion to subtract the claim. The flywheel's raises carry no claims. Each cycle is an equity issuance at a premium to the prior close, optionally attached to a warrant cascade that fires when the stock crosses a 130% VWAP threshold of strike. When the warrants fire, capital comes in, BTC goes onto the pile, and drag steps down through numerator growth instead of claim subtraction.

The two mechanisms are not parallel. They are coupled.

Every flywheel exercise brings cash that buys BTC. New BTC lifts NAV. Higher NAV lifts the stock. Higher stock brings the next OCA conversion bucket closer to its trigger. The watermill's next bucket tips. The conversion eliminates a claim. CEBE rises. Stock rises further. The next flywheel tranche approaches its trigger. Output of one engine is input to the other.

The shared signal is stock price. Everything in this capital structure is price-triggered. Nothing carries a maturity date forcing a decision regardless of where the stock sits. The watermill's force conversion authority is price-conditional (130% VWAP) and time-gated (year 3 of each tranche). The flywheel's warrant exercise is price-conditional (130% VWAP of strike) and company-controlled (the accelerated window). No calendar walls anywhere in the structure.

Three transactions in eight months took the structure from theoretical to demonstrated. The September 16, 2025 Accelerated Bookbuild established the flywheel at 2.3x mNAV. Six months later, the March 30 Blockstream conversion proved the watermill at the start of recovery. Six weeks after that, the May 11 ABSA raise ran the flywheel a second time with a cascade attached. The team did not build these engines in response to the drawdown. They built the flywheel before it started, weathered the drawdown without firing either engine, and ran both engines coming out of it.

553 Days Above Par

The CEBE Tracker covers nine Bitcoin Treasury Companies. Today three trade at premium to net asset value. Six trade at parity or discount. Of the three at premium, two have traded at a discount earlier in their treasury cycles. Capital B is the only one that has never closed below 1.0x mNAV across its entire treasury history.

553 consecutive days. November 5, 2024 through May 11, 2026. Zero closes below 1.0x on management's reported FD mNAV methodology. The absolute floor was 1.0804x on October 15, 2025, two weeks before Bitcoin began its drawdown from the November high. The peak was 20.21x on November 21, 2024, in the first weeks after Bitcoin standard adoption.

The trajectory is a maturation curve. Early premium reflected speculation against a small balance sheet. Mid-2025 premium normalized as the OCA stack scaled and BTC holdings grew. Late 2025 premium compressed alongside the broader market drawdown. The current band sits around 1.30x to 1.46x depending on the methodology used. The CEBE Tracker shows 1.46x at $72K normalization, slightly above Capital B's reported FD 1.28x because the CEBE methodology counts more items as senior claims in the NAV calculation.

The drawdown window is the structural test. From October 2025 to March 2026, Bitcoin fell from $126,000 to around $60,000. 182 days. A peak-to-trough drawdown of roughly 52% in the underlying asset. Capital B's mNAV range across the same window was 1.08x at the floor to 1.91x at the peak. The worst single day was 8% above NAV. The market did not give it back.

This is what the math predicts. Fiat-denominated debt structures have at least one expansion vector for drag in a drawdown. As BTC falls, the fixed fiat value of the claim grows in BTC terms. Drag expands at the same moment that sentiment is weakest. The premium compresses to absorb the larger claim. The structure is price-vulnerable.

Capital B's structure has no such vector. The BTC claim is 1,060 BTC at $50K BTC and 1,060 BTC at $150K BTC. Drag does not expand on the way down. The premium does not compress to absorb growing claims, because the claims do not grow.

The market priced this structural feature before the framework named it. 553 days of unbroken premium is the market saying it understands what borrowing Bitcoin in Bitcoin does to risk. The patient structure is the safe structure. The market read the design.

The $150K That Didn't Come

This section appears in every deep dive in the series. It usually explains how the capital structure behaves at higher BTC prices. For fiat-leveraged companies, the $150K story is automatic drag compression, and Capital B's structure inverts that pattern.

At $150K, the 1,060 BTC claim is still 1,060 BTC. Cash offset shrinks in BTC terms from 113 BTC (at $72K) to 54 BTC (at $150K), so net drag actually rises from 32.2% to 34.2%. The claim does not move. The drag moves slightly, the wrong direction, because the cash offset is fiat-denominated and shrinks in BTC terms as BTC rises. The staircase has not moved. Price did not do the work.

The decision archaeology is here. In 2024 and 2025, Capital B could have issued fiat-denominated convertible debt against its Bitcoin stack and let price compress the drag automatically as BTC rose. That is the standard architecture and Saylor proved it works. The team chose BTC-denominated debt instead. The cost was static drag through every market condition, including the drawdown that compressed peer structures below NAV. The benefit was structural insulation from that same drawdown. The market priced the second part of the trade across 553 days. The first part requires the watermill to be operated. The team has been operating it.

Between October 2025 and March 2026, Bitcoin fell from $126K to $60K. The upstream slowed to a trickle. The stock dropped below every conversion trigger. In a fiat-leveraged structure, that drawdown would have expanded drag as fixed dollar claims grew in BTC terms. Capital B's drag did not expand because it could not. 1,060 BTC was 1,060 BTC at $126K and 1,060 BTC at $60K. Static claims absorbed the drawdown without amplifying it.

Management used the low-price window to open the gates by halving conversion prices, removing trigger gates, and attaching sweeteners. The gate was widened while the water was low. When gravity pulls the stock floor back up, and it will if the CAGR thesis holds, the buckets are shaped to tip at lower levels than they required six months ago.

The company's investor presentation discloses an 8-year plan targeting 1% of total Bitcoin supply, or 170,000 to 260,000 BTC by 2033. The current stack is approximately 3,125 BTC pro forma the May 11 deployment. That gap requires the watermill to cycle continuously across OCA generations, and now the flywheel to recycle equity capital through repeated cascade triggers. The arithmetic of the plan implies it. New tranches will replace converted ones. Drag may not trend to zero. It may oscillate as the wheel keeps turning.

The wrapper fee is verified at EUR 2.9M net, around 1.2% of the BTC reserve value. Zero-coupon debt means zero interest payments. No preferred dividends. Those two structural zeros keep friction in the mill low. CEO Casadepax-Soulet's compensation is not disclosed in the annual report, which is a gap worth noting.

The Verdict

Capital B's scorecard position is 723 sats per share (normalized CEBE at $72K, 275,910,417 basic shares per April 27, 2026 PR, pre-May 11 ABSA close). FD BPS at the same anchor is 742 sats. The 19-sat gap between them is misleading. FD BPS inflates the denominator with potential conversion shares. CEBE subtracts the BTC claim from the numerator. Two different penalties producing similar outputs at one ratio. The two metrics diverge under stress.

All CEBE figures normalized to $72K BTC. The FX rate used is 1.17 EURUSD throughout. Basic share count is used per framework methodology. Q3 2025 figures are derived from HY 2025 Financial Report Note 3.15 and reconciled against the September 30 share register. Q4 2025 figures come from the FY2025 annual report filed April 30, 2026, with 1,279 BTC claim and 226,955,116 basic shares both VERIFIED. Q1 2026 figures come from the March 30, 2026 conversion PR and the April 27, 2026 PR confirming the post-conversion share register. Current figures are pre-May 11 ABSA close, capturing the announcement state.

The improvement from Q3 2025 to Current did not come from BTC price. The price anchor is fixed at $72K throughout. The improvement came from the March 30 conversion eliminating 219 BTC of senior claims and from cash buildup increasing the offset against the static claim. CEBE per share at the snapshot moved modestly (710 to 723 sats across the window). The structural improvement is more visible in net drag (45.8% to 32.2%, a 13.6 percentage point reduction) and in common equity ownership (54.2% to 67.8% over the same window).

FD BPS read the same period entirely differently. From Q3 2025 to Current, FD BPS fell from 1,310 sats to 742 sats at $72K, a 43% drop. Over the same window where CEBE captured the structure improving, FD BPS captured the structure looking dramatically worse, because the convention loads every potential conversion share into the denominator from the day the bonds are issued. FD BPS has been counting the dilution from the start. It has nothing left to recognize when the claim actually clears.

At $72K BTC, 182 BTC of expected deployment across 23,038,844 new shares delivers 790 sats per new share, above the existing CEBE of 723. CEBE recognizes the new shares as accretive by 67 sats per new share. FD BPS reads the same transaction as dilutive because the 92 million new warrants attach to the FD denominator while the 182 BTC of expected deployment is small against the larger FD total. The flywheel does not amplify the watermill's mechanic. It runs in parallel and feeds back through stock price. Finish the math.

What to Watch