On a single board resolution, Metaplanet deleted 131.3 million potential shares and roughly $220 million of accrued insider upside. It did not sell a coin. It did not buy one either.
That asymmetry is the story. The Tokyo-listed bitcoin treasury company is cutting its Series 10 stock acquisition rights pool from about 319.5 million shares to 188.2 million — a 41% reduction — and the one number that moved in shareholders' favor is bitcoin per fully diluted share, up roughly 8.8%. No sats were acquired. No treasury operation occurred. The metric improved because the denominator shrank.
Now set that against price. Metaplanet equity is down more than 43% this year. Bitcoin is off approximately 15%. Strategy is off about 20%. Across two sessions in the week before the announcement, the stock shed roughly 17% after CEO Simon Gerovich's first public response failed to satisfy the people who own the residual claim.
In a bear market, survival is the only performance metric that compounds. This is what a company looks like when it is triaging its own capital structure — not because governance worked, but because the marginal buyer disappeared.
Series 10 rights are a convertible claim dressed as compensation. Each right entitles the holder to buy shares at 10 yen, a strike so far below spot that the instrument carries no optionality worth modeling, only torque. The conversion ratio had crept from 410 shares per right to 696, and the reason is structural: the pool was never sized as a fixed grant. It was sized as a percentage of fully diluted capital. Every raise that funded a bitcoin purchase enlarged the insiders' claim and diluted everyone else in the same transaction. The pool inflated from roughly 46 million shares to about 319 million, an expansion exceeding 600% that ran in parallel with the share count it was feeding on.
In September 2025, Metaplanet completed an international share offering. Management has now identified that transaction as the moment capital raises stopped being strongly accretive. The board reset the conversion ratio to 410, the level immediately before that offering. The remaining warrants become exercisable in thirds in 2029, 2030, and 2031. Shares received on exercise stay locked until August 2031. A plan to migrate 20% of the warrants into a new employee incentive pool was scrapped outright; those rights form part of the 41% cancellation.
The concession is real but incomplete. Gerovich, who recused himself as a Series 10 holder, retains 64 million shares obtained through an August 28 exercise under the old, more generous terms, plus the right to acquire another 49.1 million. The announcement did not address his economic interest in MMXX Ventures. Matthew Sigel, VanEck's head of digital assets research, put the forgone amount at roughly 79 million shares worth about $123 million and called the package a meaningful realignment of management and shareholder interests. Meaningful is carrying a lot of weight in that sentence.
The company says a replacement compensation program will be designed with an outside consultant.
Run the structure, not the narrative. What shareholders received is not a cash return, not a buyback, not a bitcoin purchase. It is a reduction in forward supply. That distinction matters because forward supply is the last thing a treasury vehicle can still control once its model inverts.
Start with the 8.8%. Bitcoin per fully diluted share is a ratio — treasury holdings divided by share count including every exercisable claim. Reduce the claim by 131.3 million shares and the ratio improves without a single on-chain transaction. This is denominator engineering, and it cuts both ways: the same mechanism that inflated insider claims on the way up deflates them on the way down. A treasury company's per-share bitcoin metric is not a measure of accumulation. It is a measure of who is permitted to dilute whom.
Then price the surviving warrants honestly. Roughly 188.2 million shares at a 10 yen strike, vesting in thirds across 2029, 2030, and 2031, locked until August 2031, is a deep in-the-money call option with a five-to-six-year tail. Cancelling 41% of it removes headline dilution, but the residual remains a claim on future float sitting beneath an equity that has already lost 43%. Deferral is not deletion. The board shifted the supply curve to the right on the calendar; it did not remove it from the chart.
I have audited this exact failure mode before. In late 2017 I worked through 45 ICO whitepapers for a university finance seminar, rebuilding each token distribution schedule against a conventional equity cap table. Roughly 80% carried inflationary schedules that made the founders' claim grow mechanically with every milestone. The tell was always identical: a pool defined as a percentage of something that grows, rather than a fixed quantity. Metaplanet's Series 10 was sized that way from inception, from 46 million shares to 319 million. Structure precedes value; chaos destroys both. Nobody designed a ratchet. The ratchet was the default setting.
The 2020 lesson rhymes. I built a Python scraper across Uniswap V2 pools to map $200 million in TVL and correlate yield sources, and what it exposed was that incentive-driven liquidity is rented, not owned, repricing the instant the denominator changes. Warrants behave the same way. They are liquidity inside the float, struck at 10 yen, and their supply is what the marginal seller works against.
The institutional flow lesson from January 2024 applies directly. I spent four weeks after the spot bitcoin ETF approvals reconciling BlackRock and Fidelity net flow data against historical commodity ETF curves, and the model pointed to six months of consolidation driven by allocator profit-taking, not by price action. Flow told the story then. It tells this one now: the pool that got cut was never really a compensation program. It was structured supply — and structured supply is what breaks first when institutional demand thins.
The consensus read is that shareholder pressure worked. Weeks of public anger, a recused CEO, a consultant-led redesign, a 41% haircut. Governance functioning as designed.
The structural read is colder. In the absence of alpha, volatility is just noise, and the noise here is governance. The actual driver is market-to-NAV compression. When a digital asset treasury trades beneath the value of its coins, issuing equity to buy more bitcoin stops being accretive and starts destroying bitcoin per share. September 2025 is now officially marked as that inflection. The board did not cut the warrants because it was persuaded. It cut them because the pool had become the largest single source of forward supply overhanging a stock already down 43%, and removing it was the cheapest available way to manufacture an 8.8% improvement in the only metric the market still watches.
The most dangerous debt is the kind no one sees. Metaplanet's Series 10 never appeared as a liability. It appeared as compensation, as alignment, as a call option granted to people the company trusted. It was leverage on the cap table, denominated in dilution, compounding silently for as long as the equity rose. The MMXX Ventures question left unanswered in the announcement sits in precisely the same blind spot.
The timing is not generous either. Warrants deferred to 2029 and 2031, locked until August 2031, arrive after this cycle resolves in one direction or the other. That is a structure built to survive a bear market, not to reward one. Liquidity is merely trust, tokenized and flowing — and trust in this vehicle is being repriced in real time by people who can read a fully diluted share count.
Watch the replacement program's construction, not its size. If the next incentive pool is a fixed share grant with a fixed strike and a fixed term, the September 2025 reset becomes a genuine structural break. If it is defined once more as a percentage of fully diluted capital, this was a deferral with better optics.
The other number to track is the disclosure Metaplanet chose not to make: what, precisely, remains of management's economic interest through MMXX Ventures. Bitcoin per share improved 8.8% this week. It can be engineered once. Engineering it twice is called dilution.
