The anomaly was in the voting ratio before it reached the wallet.
Ninety point six three percent of Satsuma Technology Plc shareholders voted for the capital return. Ninety point five nine percent voted for the delisting. Two resolutions, separated by four hundredths of a percentage point, both carried — and both carried against the recommendation of the majority of their own board.
Read that again. The shareholders of a UK-listed Bitcoin treasury company used a ninety percent vote to override the directors and force the sale of every coin in the treasury. Not a partial hedge. Not a debt-driven trim. The entire stack — 668.48 BTC as of June 30 — is pointed toward the exits. The board has been authorized to prepare the closure of trading activities and the sale of the Bitcoin. The indicative target: a sale on or around Aug. 3.
The vote converted what was a conditional proposal in the middle of July into an approved motion. But approval is not clarity. As of July 30, the official London Stock Exchange issuer page still carried the July 20 result as its latest RNS. Execution date? Undisclosed. Venue? Undisclosed. Amount, price, net proceeds? Undisclosed.
I have sat through enough UK corporate actions to know what that silence means: the money moves before the narrative does. And in this case, the money is 668.48 coins bought at an average of £84,026 each, marked at $58,353 in late June, and now scheduled for sale at whatever the August tape gives. This is not a market story. It is a wiring diagram.
The broader pattern comes first, because Satsuma is not an isolated incident. It is a category.
The listed Bitcoin treasury company was one of the most confidently marketed structures of the last cycle. The formula: raise equity at a premium to net asset value, stack the proceeds into Bitcoin, and tell a public-market audience that this is the reg-compliant way to own the hardest asset on earth. The pitch depended on two unstated assumptions. First, that Bitcoin's trajectory would be kind enough to justify the acquisition price. Second, that the wrapper's own architecture — listing fees, governance, investor relations, legal counsel — could be financed invisibly in the background.
Bull markets forgive the second assumption. Bear markets audit it.
The audit year is 2026. A US Bitcoin treasury company sold its entire position in early July after the combined pressure of debt covenants and Nasdaq listing requirements closed in. In late June, Capital B and BTC AB were publicly testing how much dilution their shareholders would tolerate to buy more Bitcoin. In May, the prevailing social-media doctrine — 'never sell' — cracked as dividend obligations and buyback costs turned the sacred reserve into a liquidity cushion.
Satsuma's vote is the cleanest expression of this cascade, because there is no debt on the balance sheet. No lender is pulling a covenant. No exchange is threatening delisting for non-compliance. The company reported no material liabilities as of June 30. The only trigger was the relationship between the value of the coins and the price of the wrapper.
That relationship, by June 30, was broken. The market capitalization of Satsuma stood at roughly 0.80x the market value of its Bitcoin holdings. In plain terms, the market said the company was worth twenty percent less than the coins it held in escrow. Every day the board kept chanting 'never sell,' the discount to mNAV widened, and every day the discount widened, the probability of a forced corporate action increased. The shareholders who voted the resolutions were not buying Bitcoin exposure at a discount. They were buying a convergence event.
Satsuma's own financial snapshot explains the bitterness of the transaction. The company held 668.48 BTC at the end of June, valued at £29.44 million using a reference price of $58,353 per coin. The average acquisition cost across the treasury was £84,026 per Bitcoin. That left an unrealized loss of £39,984 per coin — approximately £26.7 million of red ink against the purchase program that the treasury-company pitch had once described as prudent and disciplined.
There were no disposals during June, according to the report. That is worth remembering. The company had not begun to unload when the shareholder vote made its future. The pre-vote silence was not a sign of confidence; it was the calm before the scheme's machinery was unlocked.
Now let me read the record the way a forensic document should be read, because the vote percentages are the least informative part of this file.
The resolution carried with 90.63% of votes cast in favor of the capital return and 90.59% in favor of the delisting. The phrase 'votes cast' is doing heavy lifting. This is a UK corporate action, and UK corporate actions count turnout, not the register. A ninety percent mandate from those who bothered to vote is not a ninety percent mandate from those entitled to vote. The announcement does not disclose how many qualifying shares were present in person or by proxy. It does not disclose abstentions. It does not tell us whether a small concentration of event-driven capital moved the literal result.
Based on my audit experience with wind-ups of this kind, the controlling vote almost certainly did not belong to retail true-believers in Bitcoin doctrine. It belonged to someone who bought the paper knowing the discount would eventually force the board's hand. The people who pushed this resolution are not the people who bought Satsuma at an £84,000-equivalent cost per coin. They are the people who bought Satsuma at 0.80x NAV and stood to harvest the difference on the way out.
The two percentages themselves contain a quiet hint. 90.63 on the return. 90.59 on the delisting. Four hundredths of a point apart. The divergence means at least one meaningful holder voted for the return and against the delisting, or the reverse. Ticket-splitting in a near-identical pair of resolutions. It is the kind of detail that reads like administrative noise but tells you this was not an emotional mob. It was a position being settled with different instructions attached to each ballot.
Now the numbers that matter.
Average acquisition cost: £84,026 per Bitcoin. June 30 mark: $58,353 per Bitcoin, which the company translated into a £29.44 million valuation for its 668.48-coin treasury. Implied mark: approximately £44,040 per coin. Unrealized loss at June 30: £39,984 per coin.
But the unrealized loss is not the loss shareholders will realize. The liquidation drags costs with it. The scheme deducts approximately £2 million of retained working capital and approximately £2.7 million of estimated transaction and termination costs. That is £4.7 million of dead weight against a £29.44 million treasury — 16 percent of the asset base. Spread across 668.48 coins, the cost burden is £7,031 per coin.
Now do the ledger the way the market will actually experience it. If the sale executes at the June 30 mark of £44,040 per coin, the gross distribution pool is about £29.44 million. After the £4.7 million deduction, the distributable pool drops to roughly £24.74 million. That is £37,009 per coin. Hold that against the £84,026 average cost, and the true per-coin loss is £47,017 — not the £39,984 the headline writers are using. The aggregate realized loss is roughly £31.4 million.
The gap between the headline loss and the realized loss is the entire thesis of this incident in miniature: the wrapper's termination costs are the hidden tax on 'never sell.' Volume spikes lie; liquidity flows tell the truth. And in corporate actions, the flows include the invoices.
Of course, the sale price is the central missing input. If the August tape is kinder and the coins clear at £50,000, the net recovery rises to £42,969 per coin. If it is meaner and they clear at £35,000, the recovery falls to £27,969 per coin. The entire outcome range is a function of one undisclosed variable: the execution price. The June 30 mark is a photograph from six weeks before the liquidation. The realized number is a live broadcast.
The indicative timetable fixes the record date and the sale date on the same day. At 6 p.m. UK time on Aug. 3, the register closes: every ordinary share held at that moment becomes entitled to one B share. Warrant holders who want to participate in the capital return must exercise their instruments before that cutoff. And 'on or around' the same day, the board intends to sell the Bitcoin.
This is a bold compression. The fixing of entitlement and the pricing of the asset in the same twelve-hour window means shareholders are locking their participation in a transaction whose proceeds are not yet known. It is the functional equivalent of signing a sales contract with the price left blank. The company clearly wants a clean narrative: by Aug. 3, we know who gets the money and we have made the money. But the market is left to infer that the two events actually line up. The public record admits slippage — 'on or around' is doing deliberate legal work. If the sale slips, the record date holds, the entitlement pool freezes, and the distribution math floats on an undisclosed execution schedule.
The legal sequence after that is standard but slow. A directions hearing on Aug. 13. A confirmation hearing on Sept. 8. The return effective on Sept. 11. Cancellation of the listing at 8 a.m. on Sept. 14. Payments on or before Sept. 28. That is eight weeks between record and payment, with the sale proceeds sitting in a sterling account earning nothing while the High Court blesses the machinery.
The requirement for High Court confirmation is worth understanding. A capital return structured through B shares is a scheme that requires the court's endorsement, primarily to protect creditors. Satsuma has no debt and no material liabilities, so the creditor-officer angle is minimal. The confirmation hearing is likely a formality. But it is a formality with calendar risk. Court hearings are listed, adjourned, and rescheduled. Every adjournment pushes the distribution later. The market context is a bear market with capital already rotating out of treasury vehicles; delayed cash is opportunity cost for the arbitrageurs who financed the vote.
The warrant holders are the quiet actors in this scheme. The return calculation includes 'any warrant exercise proceeds.' That phrase turns the final per-share amount into a function of warrant-holder behavior — and every warrant holder faces the decision alone.
Imagine the position. You hold warrants in a treasury company that has just voted to liquidate. Your warrants can be exercised before Aug. 3 to create ordinary shares, which will then convert into B shares and receive the cash return. Exercising adds cash to the pool — the warrant strike price — but also adds shares to the denominator. If the cash received per share is above the strike plus the pro-rata cost drag, exercise is rational. If not, it is a donation to the other shareholders.
The problem is that no single warrant holder knows whether the aggregate exercise will clear the threshold. Early exercises influence later math. Late exercises can be irrational. This is a coordination game played against a clock, and in bear-market treasury structures, the historical pattern is that most warrants end as dead weight. They expire unexercised, the pool recovers only the base cash, and the arbitrageurs take the spread.
No one in the coverage is talking about the warrants. They are the second-order variable in the distribution, and they are entirely discretionary. Watch what warrant holders do in the first 48 hours after this scheme goes live; their behavior will signal the real expected recovery before the board says a word.
From my surveillance seat, the most telling detail is not the vote. It is the absence of movement.
The June 30 report says no disposals in June. The July 20 announcement authorizes preparations. As of July 30, the public record still shows the July 20 result as the latest development. The company has told the market what it intends to do, but it has not told the market the execution method, the venue, or the price. This is unusual for a UK plc winding down a concentrated asset position. Most boards front-run their own liquidation narrative with at least a transactional outline, because a sudden sell-down into a thin book is the classic way to realize the worst possible price.
We know what the coins are worth at spot. We do not know how they will be sold. A block trade through an OTC desk can absorb a $39 million countervalue in a matter of hours, typically at a discount of 50 to 100 basis points to prevailing spot, with the exact haircut depending on the desk's inventory and the direction of the flow. A programmatic sell into public order books across multiple sessions is slower and more transparent, and it gives the market time to front-run. In August — historically a low-liquidity month across crypto venues — a 668-coin position is not trivial. It is roughly one percent of daily global BTC spot volume on a good day. On a bad day, it is five percent of what a single venue will print. The slippage on a badly executed liquidation is not a rounding error; it is another tranche of the shareholders' already-shrunken recovery.
The company has not disclosed whether the execution will be a single block, a series of auctions, or a dealer-crossed transaction. It has not named the counterparty. It has not said whether the sale will be subject to minimum-price protections or fallback provisions. All of that matters more than the vote. The vote established the authority; the execution establishes the price. And the price, not the mandate, is what the shareholders will receive.
I have watched this play before. In July 2020, I caught the Curve Finance treasury drain by tracking the outbound transaction before the team confirmed it. The discipline is the same: address first, statement second. The Satsuma wallet cluster will speak before the RNS does.
Now the part of this story that every headline has gotten wrong.
The coverage wants this to be a narrative about shareholder democracy triumphing over a stubborn, diamond-handed board. The majority of Satsuma's board recommended against the vote. The shareholders overrode them at ninety percent. The cover version: the board believed in Bitcoin and lost; the shareholders wanted cash and won.
That is true at the level of resolutions. It is false at the level of economics.
The real story is that the board's recommendation was the only coherent option available to a structure that had already failed. Let me walk the logic. The board's 'no' recommendation was not a thesis about the future of Bitcoin. It was a refusal to accept the wrapper's terminal condition. A company holding 668 BTC with no revenue, no debt, and a twenty percent NAV discount is not a going concern. It is a mark-to-market corpse waiting for an auditor to notice. The board's choice was not 'hold forever' versus 'surrender.' It was 'signal confidence' versus 'admit the wrapper's overhead is a permanent value leak.'
The chart on the trading terminal shows BTC's price. The chart does not show the cost-drag line. It never does. But every treasury company has one — an invisible, compounding drain of listing fees, audit fees, legal fees, board insurance, and investor-relations overhead that must be financed out of the balance sheet every single quarter. The chart of the coin looks like price discovery. The chart of the wrapper looks like decay. At 0.80x mNAV, the decay line had crossed the value line, and no amount of Bitcoin conviction could restore it without a bull-market rescue.
The shareholders' vote was not an act of rebellion. It was an act of mark-to-market discipline. The people who bought the paper at a discount understood that the liquidation costs — £2 million retained, £2.7 million in transaction costs — are the price of converting the wrapper back into its underlying asset. They did the arithmetic: buy at eighty percent of NAV, absorb the sixteen percent cost drag, collect the residual. The gross spread is thin — roughly four percent of asset value — but for an event-driven fund, a four-percent gross over a two-month holding period annualizes into the high twenties, and the trade is practically riskless once the votes are counted.
That is the insight the 'crushing loss' framing obscures. Yes, the long-date holder who bought at the top of the acquisition program loses. That holder loses £47,000 per coin after liquidation costs, not the £39,984 the headlines use. But the capital that closed the discount does not lose. It harvests.
The board's real failure was not in voting the wrong way. It was in failing to offer an alternative that could close the discount without a fire sale. A buyback at 0.80x NAV would have been the disciplined move: repurchase paper below intrinsic value, retire shares, increase per-share BTC exposure. The board did not propose it. Perhaps the company lacked the cash; a treasury-company balance sheet is mostly coins, and selling coins to buy back shares defeats the purpose. Perhaps the board underestimated the concentration of event-driven holders. Either way, the outcome was deterministic. Once the discount exceeded the cost of the exit, the exit became inevitable. 'Never sell' is a marketing slide. It was never a governance commitment.
The 'shareholder rebellion' story also misreads the vote's size. Ninety percent mandates are not retail phenomena. Retail investors do not coordinate ninety-percent votes against a board's recommendation in a forty-day window. They do not split tickets by four hundredths of a point. A ninety-percent vote at these levels is the signature of a concentrated register cultivated by an activist who bought the discount, filed the proposal, and rallied the other early-bird sellers into the box. The paperwork says 'shareholders of Satsuma Technology Plc approved.' The provenance says 'a convergence trade structured as a governance event.'
And this is the deeper warning for every other treasury company reading the tape. The mechanism that killed Satsuma is not a law. It is an exploitable spread. When an asset wrapper trades at a discount wider than its own liquidation costs, the market gains a built-in arbitrage: buy the paper, vote the wrapper out of existence, collect the difference. The exploit is not a vulnerability in code. It is the open architecture of the plc itself. Speed is safety when the exploit is already live — and the exploit here is the corporate concept of the Bitcoin treasury fund with marketing overhead and a live listing.
The next twelve hours matter more than the next twelve press releases.
If the plan holds, the 668.48 coins move on or around Aug. 3. The sale will be visible on-chain before it appears in an RNS. The wallet cluster Satsuma has disclosed through its custody or self-custody reporting will light up when the first transfer hits the mempool. I will be watching the size of the first outbound transaction, the recipient address, and the timing relative to the close of the order book on the venue that ultimately executes. The market will learn the execution price from the block data before the company confirms it. That is the race.
The precedent is the larger story. Satsuma has now demonstrated, with the formal machinery of a UK plc, that 'never sell' is one shareholder vote away from 'sell everything.' Every Bitcoin treasury company trading at a discount to NAV now carries this sword over its own head. The US company that sold in July did so under debt and listing pressure; Satsuma did it with a clean balance sheet and a ballot box. The second is a stronger precedent than the first. It announces that the wrapper itself is the exit trigger.
I will be tracking two things in the coming weeks: the warrant exercise rate, and the discount register on the remaining treasury peers. If arbitrageurs start rotating out of ex-Satsuma positions into the next 0.75x mNAV wrapper, this will not be an event. It will be a cycle.
Bitcoin does not need a corporate wrapper to survive. The question is whether the corporate wrapper can survive another shareholder meeting. The chart does not show the cost-drag line, but the shareholders who voted this motion read it anyway. First the coins move. Then the regalia falls away. Then we watch which treasury company is next to learn that a vote is a different instrument from a promise. We don't need an RNS to know when the first coin leaves. The mempool speaks first. We only have to be watching.