There is a clause inside Metaplanet's equity structure that does something no compensation committee would ever approve in a room with the lights on. It resets. Automatically. Every time the company issues new shares, the option pool re-inflates to track twenty percent of everything the company is allowed to issue โ not twenty percent of what it issued last year, not a fixed number of shares, but a live percentage anchored to a moving ceiling. No board vote. No shareholder vote. No performance gate. The only condition attached to the grant is that the recipient remains employed.
On September 18, VanEck took that clause, along with the rest of Metaplanet's compensation architecture, and scored it against the nine other public companies that treat bitcoin as their primary treasury reserve. Nine of those ten came back Good or Pass. One came back Bad. Metaplanet โ 3350.T on the Tokyo Stock Exchange, MTPLF for the American retail investors who bought it as the Asian Strategy โ is the only company in the top-ten digital asset treasury cohort to carry that word.
The stock is up roughly 1,300 percent since the pivot. Bitcoin per share is up forty-fold. The company holds 43,000 BTC, second only to Strategy among corporate holders. By every metric the market actually quotes on a chart, Metaplanet is a triumph. And that is exactly why the rating matters. Because the only thing the chart does not price is the mechanism that produced the number โ and in a treasury vehicle, the mechanism is the asset.
The ledger remembers every trembling hand. It also remembers every hand that wrote a formula specifically so it would never have to be seen doing the taking.
What Metaplanet is now, and what it was when the clause was written
To understand why a compensation scorecard turned into a governance event, you have to know what Metaplanet is now versus what it was when the clause was authored.
Metaplanet did not begin as a bitcoin company. It began as Red Planet Japan, a hotel operator, and by early 2023 it was a distressed one. The February 2023 rescue restructuring is where the option pool language VanEck is now flagging was first set. Read that sentence twice. The autoreset mechanism was not designed for a bitcoin treasury strategy. It was designed for a company that needed to retain employees while it fought for survival, at a moment when the equity was worth so little that almost any incentive design was defensible and almost no shareholder was paying attention.
Then in 2024 the company pivoted. It renamed itself Metaplanet, brought on Dylan LeClair, adopted the Strategy playbook, and began converting its balance sheet into bitcoin through a combination of equity, debt, and preferred instruments. The pivot worked. The stock re-rated. The company raised. The pool followed.
By the time anyone outside Japan looked closely, the structure read like this. Fully diluted share count had traveled from 153.9 million to roughly 1.35 billion โ a two-year expansion of about 776 percent. The option pool had traveled from 46 million to 319.5 million shares, roughly 595 percent growth. The pool as a share of fully diluted stock stood at 14.7 percent, against a peer median of 4.0 percent. Named executive exposure sat at 8.2 percent, against a peer median of 0.8 percent. Ten times the cohort. That is not a rounding difference. It is a different species of company.
Two partial rollbacks followed, on August 18 and September 11. The pool was cut 41 percent, down to 188.2 million shares, which knocks the ratio to something near 13.9 percent. That still leaves 105.4 million potential new shares hanging over the register, roughly seven percent of the company. And it still leaves Metaplanet far outside the range of any comparable treasury vehicle. VanEck's summary of the post-fix position was blunt: better, and still worse than any peer.
The company's public defense has been the performance record โ the 1,300 percent move, the forty-fold BTC-per-share growth, two major acquisitions. Here I want to flag something from my own history, because I have stood on both sides of this exact rhetorical move. In 2017 I was trading ICOs off token distribution curves, and the story every weak project told was the chart. The chart was always true. It was also always backward-looking. I learned the hard way that a return figure describes the conditions that produced it, not the durability of the structure underneath it. When Terra collapsed, I spent three months tracing Anchor and UST flows, and the most useful thing I extracted was not the failure point. It was that the failure point had been legible in the incentive design eighteen months before price acknowledged it.
Governance is the same kind of data. It just does not print a candle.
Decomposing the mechanism
Bad rating is not an explanation. So let me decompose the machine.
The formula. Under the disputed terms, the option pool benchmarks to twenty percent of the company's issuable shares. Two words carry the entire problem. Issuable is not issued. A cap tied to the issued count is a fixed claim: you dilute once, the pool grows once, the number stops. A cap tied to the issuable count is a ratchet: the board authorizes more shares, the pool expands in lockstep, and the claim reproduces itself. Metaplanet's pool is the second kind. During a treasury expansion cycle, the company issues constantly by design โ that is how a treasury vehicle grows โ which means the pool expands constantly by design. The dilution is not a side effect of the strategy. It is wired into it.
LeClair's own framing of the cleanup is that it cut the option pool by 41 percent. Accurate, and also the most misleading true statement in the coverage. A 41 percent reduction in a number that is auto-indexed to a rising base is not a structural fix. It is a one-time haircut on a mechanism that will re-inflate on the next raise. If the board issues shares next quarter โ and it will, because that is the business model โ the pool begins climbing again. You cannot cut your way out of a ratchet. You can only cut the ratchet.
The execution layer. VanEck's language here is the part to read slowly: the increase is executed by formula rather than decided by a committee. That distinction is the whole ballgame. A committee can be lobbied, but a committee can also say no, and a committee's decisions are minutes shareholders can read. A formula says yes forever, and its minutes are an arithmetic identity. The board did not approve each expansion. The board approved the rule that makes each expansion automatic and unremarkable. That is not governance. That is the outsourcing of governance to a spreadsheet, and the spreadsheet owes no fiduciary duty to anyone.
My audit background is worth something here. In 2021 I wrote Python to crawl metadata links across more than a thousand NFTs and found a fifteen percent broken-image rate that founders had not disclosed, because the disclosure burden had been shifted to a marketplace with no incentive to look. The pattern in that incident is the pattern here. Nobody lied. A rule was written so that nobody had to look. Silence is the only honest metadata โ and the silence around an auto-executing dilution clause is more informative than any disclosure document the company has filed.
The performance gate that is not there. There is no performance threshold on the pool. None. The only requirement is continued employment. That means the mechanism is fully decoupled from the metric that supposedly defines the company: bitcoin per share. This is the structural absurdity at the center of the affair. A treasury company's entire pitch is that you are buying a wrapper whose per-share bitcoin content grows over time. Management's incentive should be welded to that number. Metaplanet's is welded instead to issuance volume โ to the size of the raise, not the accretion it produces. And here is the knife: at scale, dilution mechanically reduces bitcoin per share. The pool is not a reward for growing the metric management claims to optimize. It is a tax on it. In a rising market the tax hides behind the price. In a flat market it becomes the entire story.
The math of the transfer. VanEck's estimate is that roughly eighty percent of the bitcoin accretion reaches shareholders, with management dilution absorbing about a fifth. That is the cleanest way to state the leak: for every five units of value the treasury strategy generates, one is diverted before it reaches the register. Not fraud. Friction with a name attached to it. And 8.2 percent named-executive exposure means the name is not abstract.
Between share count, pool growth, and insider grants, the two-year picture is this. 153.9 million shares became roughly 1.35 billion. 46 million option shares became 319.5 million. 82.8 million shares were issued to insiders. And the pool the clause created along the way has been estimated in the neighborhood of 273 million shares. That last figure sits awkwardly against the post-cut 188.2 million pool, and the discrepancy matters to anyone doing forensics. Either the 273 million number includes both issued and potential claims, or the counting basis differs between documents. I would not resolve that discrepancy by assuming the smaller number is the real one. I have spent too many hours reconciling Anchor and UST flows to trust an unaudited reconciliation only one side has seen.
Put those figures against the peer group and the isolation is total. A 14.7 percent pool against a 4.0 percent median. An 8.2 percent executive slice against a 0.8 percent median. A mechanism no other company in the cohort uses. And one plain, unflattering fact: shareholders never voted on the growth of the pool or on the terms themselves. Not once. The thing that moved six hundred percent in two years has never appeared on a ballot.
The rollbacks and what their timing implies. August 18 and September 11 sit eight weeks apart, both after the pool had already become the defining governance question in the treasury-company conversation, and both executed as reductions rather than restructures. If the board believed the clause was sound, it would have defended it. If it believed it was indefensible, it would have replaced it. It did neither. It shaved the number and left the ratchet. That is the signature of a decision sized to the pressure rather than to the problem.
I hold a low-confidence read here, and forensics means labeling your uncertainty rather than hiding it. The sequencing of the two cuts tracks the cadence of institutional engagement almost exactly. My base case is that at least one significant holder or prospective holder asked a question in a room, and the answer was a haircut. If that is right, it will eventually surface in the register.
The part nobody is writing
Here is the angle almost nobody is publishing, because it requires explaining why the obvious defense is actually the prosecution.
The consensus read is that Metaplanet got caught being greedy. I think that is downstream. The upstream fact is that the clause was never designed for this company, and the pivot locked it in place. The February 2023 rescue was authored for a near-bankrupt hotel operator, where aggressive retention equity was a reasonable, even humane, instrument. Nobody financing a collapsing hotel in Tokyo in early 2023 expected the paper to become the balance sheet of the second-largest corporate bitcoin holder on earth. Then the strategy changed, the equity re-rated, the issuable base exploded, and a survival clause became a wealth-transfer clause by pure path dependency. No villain required. Just a formula that no longer matched the firm it was attached to, and nobody with the authority or the incentive to cut it.
That framing changes the remedy. If this were greed, you would demand accountability and move on. If it is a structural mismatch, the only fix is structural: delete the autoreset, replace it with a fixed pool, attach it to bitcoin per share, and put the pool before shareholders. Which is โ not coincidentally โ almost exactly the four changes VanEck enumerated.
Which brings me to the second contrarian point. Everyone is treating the absence of those four changes as inertia. I think the absence is the finding. VanEck specified the remedy, publicly, with a date stamp, and as of this writing none of the four has been implemented. That is not a slow board. That is a board that has priced the cost of compliance against the cost of reputational damage and concluded the damage is cheaper. And the reason it can reach that conclusion is that the stock has been going up. There is no forcing function in a bull market. Infinite leverage, finite patience โ the pool can afford to wait for the price to stop covering it.
Third point, and this is the one I would watch if I had a single screen. The auto-execution design is a liability shield. A committee that repeatedly approves an expanding pool generates a paper trail a plaintiff or a regulator can read. A formula that approves it automatically generates a rules-based explanation: the board did not do it, the terms did. I cannot prove intent, and I will not pretend to. But I have watched enough structures to know that when the same outcome could be produced by a person or by a rule, and the rule is chosen, the rule is chosen partly because it cannot be blamed. Logic chains break where greed connects โ but they break cleanest where the connection is automated.
And the last blind spot. Everyone is reading this as a Metaplanet story. It is a treasury-company story. Metaplanet is the most visible bitcoin treasury vehicle in Asia; it is the listing Asian allocators use as their reference implementation. If its governance is the cohort's worst, that does not stay contained. Strategy spent two years establishing that a treasury company can be institutionally respectable โ Good rating, governance benchmark, the model everyone copies. Metaplanet spent two years establishing that you can copy the strategy and skip the governance. The next Asian treasury company to pitch an allocator will be asked about this. The cost is not Metaplanet's alone.
And the deeper lesson is the one the sector keeps refusing to learn. A treasury vehicle is not a protocol. It is a capital structure wearing a ticker. It has no consensus mechanism, no code audit, no slashing. It has a board, a pool, and a register. The market has spent two years pricing these things like tokens and analyzing them like tokens. That works until the price stops doing the arguing.
What to watch
First, the pool terms. Not the number โ the terms. If the autoreset language survives the next raise, the 41 percent cut was theater, and every subsequent capital action re-inflates the claim. If the pool gets re-anchored to a fixed share count or to bitcoin per share, that is the first real governance signal the company has produced.
Second, index and ETF eligibility. Governance ratings are not legal constraints, but they are screening inputs. If allocators begin treating the Bad rating as a filter, the effect is passive outflow that no bitcoin rally offsets โ the kind of cost that shows up two quarters late and gets attributed to something else entirely.
Third, the register itself. Two rollbacks eight weeks apart, both sized to the pressure rather than to the problem, is a pattern that usually has a counterparty. If a holder of size surfaces in the filings, August and September will retroactively read as the first act rather than the fix.
Metaplanet holds 43,000 BTC and just recorded a 1,300 percent year. Nothing in this story matters while the price is doing the work. Everything in it matters the first quarter the price is not. Speed wins the trade, clarity wins the war โ and this war is over who owns the bitcoin per share. The clause has been answering that question for two years without ever appearing on a ballot.