At Least 8.93% of Bitcoin Sits Behind Somebody Else's Keys
One fund, one state and ten companies reveal a minimum of 1.793 million BTC held through institutions rather than as bearer property. The true figure is higher; the rest is unknown.
Bitcoin made property strangely simple again: either you can produce the key, or you cannot. Finance, naturally, has spent sixteen years putting receipts back on top.
Satoshi Gazette's new custody ledger can now prove how much of that receipt-layer Bitcoin is visible. At block 964,000, one disclosed spot fund, one disclosed state and ten public companies accounted for 1,792,978.35158655 BTC. Against 20,075,003.125 BTC issued under the protocol's subsidy schedule, that is 8.9314%.
Roughly one bitcoin in every eleven already sits behind somebody else's keys in just those three disclosed groups. The holder may own a share, a political promise, or exposure to a balance sheet. What the holder does not own is the key.
That number is a floor, not an estimate of the whole market. It deliberately leaves out everything SG cannot establish from direct disclosure. No exchange-wallet clustering. No guesses about unlabelled addresses. No assumed custody for coins that have not moved. The method trades completeness for a number that can be audited.
Live Data Desk · observed Aug 31, 2026Bitcoin Held as a Claim9.0683%Current evidenced floorAt least ₿ 1,820,737.35158655 is directly evidenced as claims on companies, funds or states.9.0683% evidenced claims90.9317% unclassifiedOpen the live custody ledger and methodology →What the floor counts
| Disclosed category | BTC counted | What the ordinary holder owns | Who controls the keys |
|---|---|---|---|
| Ten public-company treasuries | 1,013,582 | Shares in companies that own bitcoin | The company or its custodian |
| BlackRock's IBIT | 771,640.9773 | A beneficial interest in a trust | The trust's appointed custodian |
| El Salvador's ONBTC wallet set | 7,755.37428655 | No direct property claim; these are state assets | The state or parties acting for it |
| Disclosed floor | 1,792,978.35158655 | Different claims, none equivalent to holding a private key | Somebody else |
The largest slice comes from ten companies in SG's public Bitcoin Treasury Ledger. A shareholder owns stock in a company. The company owns bitcoin. Those are related economic interests, but they are not the same property. A shareholder cannot point to a transaction output, produce a key or redeem a share for a particular coin. Exit means finding another buyer for the stock.
The second slice is one fund: the iShares Bitcoin Trust ETF. BlackRock's official holdings file reported 771,640.97730 BTC on August 25. Its SEC filing is equally plain about the structure: each share is a fractional beneficial interest in the trust's net assets; the trust's bitcoin is held by a custodian; and redemption orders belong to authorized participants operating in large baskets. An ordinary shareholder can sell a share. The shareholder cannot withdraw the underlying bitcoin to a personal wallet.
The smallest disclosed slice is El Salvador's government-labelled ONBTC wallet set. Its official endpoint returned 20 addresses whose funded outputs minus spent outputs total 775,537,428,655 satoshis. Those coins are state property. A citizen does not gain a redeemable claim on them by living in the country, paying taxes or holding the currency the state issues. Whether the state keeps or spends them is a political decision.
Three structures, three legal wrappers, one common fact: the person looking at the exposure is not the person who can sign the transaction.
The missing 91.0686% is not a self-custody victory lap
The ledger leaves 18,282,024.77341345 BTC unclassified. That does not mean 91.0686% of issued bitcoin is safely held by individuals. It means SG does not know the split and refuses to pretend.
The unclassified remainder includes exchange balances, custodians that publish nothing, direct individual holdings, lost coins, dormant coins and coins whose controller cannot be established from public evidence. Calling all of it self-custody would be the mirror image of bad institutional research: a comforting story dressed as a statistic.
The floor is almost certainly below the true intermediary-held share because it counts only one spot fund, one government disclosure and ten companies. It excludes other disclosed products not yet compiled into the method, as well as every intermediary that offers no auditable quantity. But a floor across three categories cannot establish a trend. It cannot tell us whether direct ownership is rising or falling, whether custody is becoming more concentrated, or how many people consciously chose each arrangement. Those questions need repeated observations and broader coverage.
The strongest case for the receipt
There is a serious counterargument, and it deserves more than a ritual sentence near the bottom.
Many people choose regulated intermediaries knowingly. A fund can fit inside a retirement account, reduce operational mistakes, simplify accounting or let a fiduciary obtain price exposure under rules that forbid direct custody. A company can hold bitcoin more securely than an unprepared shareholder could secure a seed phrase. A state can hold reserves on behalf of a public balance sheet without claiming that every citizen owns the coins.
For some people, delegated custody is a rational security choice. Losing a private key is not liberation. Neither is turning self-custody into a purity test that humiliates newcomers or ignores disability, inheritance, tax and institutional constraints. The common person's right to Bitcoin includes the right to choose assistance.
But informed choice requires honest nouns. A fund share is a security. A company share is a claim on a corporate enterprise. State bitcoin is state property. All may deliver economic exposure. None gives the holder bearer control of the underlying bitcoin. Convenience can be valuable without being renamed ownership.
That distinction becomes most important precisely when everything stops being convenient. A broker may restrict an account. A fund can change terms within its legal framework. A company can sell treasury assets. A state can reverse policy. A custodian can fail, be ordered to act or become unavailable. These are not predictions that every intermediary will misbehave. They are descriptions of where veto power sits.
Adoption needs a key test
Institutional demand can deepen markets and make Bitcoin easier to buy. It can also persuade a generation that seeing a number on a regulated screen is the finished form of adoption. Wall Street has always been excellent at selling the feeling of ownership while keeping the object itself somewhere behind the desk. Orange branding does not repeal that business model.
A people-first adoption test asks harder questions:
- Can the holder move the bitcoin without a broker, board or state approving the transaction?
- Can the holder verify the asset and its rules independently?
- Which custodian, solvency assumption, surveillance layer, fee and legal jurisdiction sits between the person and the network?
- What happens to exit when the intermediary is closed, unwilling or insolvent?
The answer need not always be self-custody today. It should always be legible.
The 8.9314% floor is useful because it measures something institutional enthusiasm usually blurs: the difference between Bitcoin as bearer property and Bitcoin as a claim. It is not a score of good holders and bad holders. It is a map of dependency built from what can actually be proved.
And it is only the beginning. The true figure is higher. The ledger will grow as more categories gain direct evidence, and its limits will remain visible as it does. Bitcoin was not invented to make financial receipts scarcer. Its unusual promise is that an ordinary person can hold the thing itself. A serious adoption story should at least count how often that promise has been delegated away.