About 11.1 million bitcoin, roughly 55% of the 20.1 million coins mined so far, sit in addresses that cannot be tied to any known owner, according to Galaxy Research’s ownership data, as reported by CoinDesk on Monday 26 October. Galaxy, the research arm of Galaxy Digital, can link the other 45% or so, close to 9.0 million BTC, to exchanges, exchange-traded funds, governments and the early blockchain and earns a reward, which makes rewriting history very expensive.">mining pattern usually credited to Satoshi Nakamoto. The rest is “unattributed”. That word is the whole story: it means Galaxy cannot put a name on those addresses, not that the coins are lost, and not that they are about to be sold.
With bitcoin at about $80,300 on Coindaro’s market ticker at 11:42 UTC on 26 October, the unattributed pile would be worth around $891 billion (our calculation). Most of what is said about who holds bitcoin rests on labels applied by a few analytics firms and public trackers. Their totals are far apart, as the comparison further down shows.
Unattributed bitcoin: what Galaxy’s October snapshot says
The figures come from what Journal du Coin called Galaxy’s October snapshot of bitcoin ownership, which Galaxy head of research Alex Thorn shared on X. We could not open Galaxy’s own page or chart, and Thorn’s post could not be loaded either, so every Galaxy number in this article is relayed by the outlets that carried it: CoinDesk (Omkar Godbole, 03:18 EDT, 07:18 UTC), crypto.news (08:07 UTC) and Journal du Coin (midday), all on 26 October.
The date of the snapshot itself is unclear. The Chinese outlet PANews ran an article on Thorn’s claim dated 14 October, which suggests the post is about two weeks older than the wave of coverage. Galaxy’s cutoff date for the data has not been published in any of the reports we could read.
What the coverage gives is short. Unattributed: about 11.1 million BTC, or 55% of supply. Attributed: about 45%, or roughly 9.0 million BTC. The largest single attributed entry is labelled “Satoshi Nakamoto (Patoshi Pattern)” with 1,123,658 BTC, about 5.6% of supply. The other attributed groups named are exchanges, ETFs led by BlackRock’s fund, and the US government. None of the articles gives the amount Galaxy puts in each of those groups.
“More than half of all BTC supply is sitting in addresses with no known attribution,” Thorn wrote, as quoted by CoinDesk. He added that “still a ton of bitcoin (maybe even most) is held by individuals in self-custody”. CoinDesk itself was more careful, noting that the 55% does not necessarily belong to individuals holding their own coins for the long term. Xapo Bank answered Thorn that retail holders have been there for more than a decade.
A measurement limit, not a mystery
Attribution works from the outside in. Analysts start with wallets whose owners are already known, for example an exchange’s published cold-storage addresses, an ETF custodian’s disclosed holdings or coins seized by a government. They then follow the public transaction history and apply clustering rules: addresses that are spent together in one transaction are likely controlled by the same party, and change outputs usually go back to the sender. Each new address that joins a known cluster gets the cluster’s label.
Whatever never touches a labelled cluster stays unlabelled. A person who bought bitcoin on an exchange in 2015 and withdrew it to a hardware wallet may show up as one hop from an exchange, then nothing. A fund that keeps coins with a custodian Galaxy has not mapped will look the same as a pensioner in Lyon with a paper backup. So will an over-the-counter desk, a family office or a company that never disclosed its holdings.
Journal du Coin gives its own examples of who could sit in the unattributed bucket: a small saver, a long-standing whale, a company whose wallets nobody has identified, institutions, and coins that can no longer be reached at all. It adds two warnings. One actor can spread its bitcoin over many addresses, so counting addresses is not counting investors. And a custodian may hold coins for thousands of clients, so naming the custodian says nothing about the people who own the coins. The proportions between these groups are not known, and Galaxy has not claimed to know them.
For a refresher on why holding the keys is what ownership means on Bitcoin, see our explainer on private and public keys and how they work.
What “known” covers: exchanges, ETFs, companies, governments
Galaxy did not publish its category amounts, but the public tracker BitcoinTreasuries.net does publish its own. Its page, read around 12:00 UTC on 26 October, adds up holdings it can source from filings, fund disclosures and official statements. The live figures carry no per-line date.
| Holder group (BitcoinTreasuries.net) | BTC | Share of ~20.1M mined |
|---|---|---|
| Public companies | 1,289,903 | 6.4% |
| Private companies | 284,652 | 1.4% |
| Government entities | 645,609 | 3.2% |
| ETFs and exchanges | 1,615,524 | 8.0% |
| DeFi and other | 374,559 | 1.9% |
| Total of the five groups | 4,210,247 | 20.9% |
Shares are our calculation against 20.1 million coins; BitcoinTreasuries shows 6.142% for public companies because it uses its own supply figure. The government line covers every state the site tracks, the United States among them, and we found no dated figure for US holdings alone.
Some names inside those lines moved this month. Strategy bought 267 BTC on 19 October, according to BitcoinTreasuries’ news feed. crypto.news, citing recent ETF holdings data, put BlackRock’s iShares Bitcoin Trust (IBIT) at about 814,263 BTC, and wrote that the investment products tracked by the on-chain analytics account Lookonchain held 1,284,312 BTC on 22 October. We have not checked either figure with iShares or Lookonchain, and neither is Galaxy’s. Tesla and SpaceX together hold 30,221 BTC and sold none in the third quarter, as we reported in our piece on the SpaceX and Tesla bitcoin holdings.
Why the public trackers see 4.2 million and Galaxy about 9 million
The gap is large: roughly 9.0 million BTC attributed by Galaxy against 4.2 million summed by BitcoinTreasuries, a difference near 4.8 million BTC. About 1.12 million of it probably comes from the Patoshi coins, which do not fit any of BitcoinTreasuries’ five group names (companies, governments, ETFs and exchanges, DeFi and other). That still leaves around 3.7 million BTC that Galaxy labels and the public tracker does not.
The two lists measure different things. BitcoinTreasuries counts what holders disclose. Galaxy labels addresses on-chain, so in principle it can tag an exchange’s cold wallets or a custodian’s address clusters even when nobody publishes a balance. Coindaro’s inference, not a sourced finding, is that exchange cold wallets and custodians make up much of the remaining 3.7 million. None of the sources we read reconciles the two totals, and Galaxy has not released the breakdown that would confirm or refute that reading.
The order of magnitude matters for readers. An investor who reads that “companies and ETFs own a fifth of bitcoin” and another who reads that “45% is known” are looking at two methods, not two contradictory facts.
The Patoshi coins: 1.12 million BTC on a pattern, not a proof
The biggest attributed entry is also the least certain. The Patoshi pattern is a statistical fingerprint in the way some of the earliest blocks were mined, described by researcher Sergio Demian Lerner in a blog post of 17 April 2013. The coins mined with that fingerprint have mostly never moved, and they are usually credited to Satoshi Nakamoto, as Galaxy’s own label does. Journal du Coin stresses that this attribution remains an analytical hypothesis rather than a demonstrated identity.
Estimates differ. Galaxy’s label carries 1,123,658 BTC. Lerner’s own 2013 estimate, made by eye from a chart, was around 1 million BTC, and he wrote that he could not be fully certain every block in the pattern belonged to Satoshi. Nobody has proved who controls those keys, or whether anyone still does. Our guide to Satoshi Nakamoto and bitcoin’s early history covers what is actually documented.
Unknown is not lost
A coin is lost when nobody can sign for it any more: the private key was forgotten, the backup was destroyed, the owner died without passing it on. A coin is unattributed when an outside analyst cannot tell who owns it. The two sets overlap, but they are not the same set, and the second is far bigger.
Long inactivity does not prove a loss either, as Journal du Coin points out. Plenty of owners leave coins untouched for a decade by choice. Galaxy itself gave no estimate of lost coins in the coverage we read.
Lost-coin estimates are old and far apart
The numbers usually quoted for lost bitcoin come from other firms, and none of them from Galaxy. BitGo, the custody company, has compiled several of them in a blog post on lost coins, which opens with an overall range of 2.3 million to 4 million BTC. The oldest figure in its table is a Chainalysis study from November 2017, which put losses at 2.8 million to 3.8 million BTC. The newest come from Ledger and Unchained.
| Estimate | Lost BTC | Date |
|---|---|---|
| Chainalysis | 2.8M to 3.8M | November 2017 |
| River Financial | 3.0M to 4.0M | September 2023 |
| Ledger | 2.3M to 3.7M | November 2025 |
| Unchained | 3.0M to 3.8M | April 2026 |
Figures as listed by BitGo; we did not open each firm’s original study. The same post cites a CoinDesk count of as much as 5.6 million BTC that has not moved in over a decade, which measures dormancy, not loss.
Put side by side with Galaxy’s 11.1 million, even the highest of these lost-coin figures, 4 million, would leave about 7.1 million unattributed BTC that someone may still be able to spend. That arithmetic only holds if the estimates are right, and no one knows how many coins are truly gone.
Old coins keep waking up
The best evidence that dormant does not mean dead is on-chain. Galaxy tracked six wallets that moved 553.59 BTC, worth about $40.15 million at the time, in August 2026 after sitting idle for roughly 12 to 15 years, with coins dating from 2011, 2012 and 2014, according to crypto.news, which added that researchers could not establish who controlled most of those addresses or why the coins moved. On 18 October we reported on a 2013 stash of 471 BTC swept to 15 new addresses. Each such move takes coins out of the “maybe lost” column and, often, into a new set of addresses that nobody has labelled yet.
Custody is shifting at the same time, according to crypto.news. Citing its own August report, the outlet wrote that Citigroup plans to offer bitcoin custody before the end of 2026, and it noted that BNY Mellon and Coinbase Custody already serve ETF issuers. The same article reports that Jameson Lopp, chief security officer of Casa, published a letter on 22 October warning holders not to give up self-custody after a run of hardware wallet compromises. He recommended multisignature wallets with keys kept on separate devices and in separate places, and planning for an owner who can no longer manage the coins.
What the 55% figure changes for readers
For anyone following the supply-shock argument, the snapshot is a caution. A common claim is that ETFs and treasury companies are absorbing a shrinking float. Galaxy’s data says the float that analysts can see and name is smaller than the one they cannot. If a large share of the 11.1 million BTC is spendable, the supply that could reach the market is larger than the “known holders” tables imply. If a large share is lost, it is smaller. The data does not say which.
There is also a plain arithmetic limit. Mempool.space reported a chain tip at block 973,231 around 12:00 UTC on 26 October; adding up the block subsidy through that height gives 20,103,850 BTC issued (our calculation), which matches Galaxy’s 20.1 million. About 896,150 BTC remain to be mined before the 21 million cap. Miners now receive 3.125 BTC per block, and that subsidy halves again at block 1,050,000.
For the price backdrop on Monday, our Asia morning report on bitcoin dipping under $80,000 covers the moves ahead of Wednesday’s Fed decision. Bitcoin was down 2.9% over seven days at 11:42 UTC, according to the same ticker. This article is not investment advice.
What is still unknown
Several points cannot be checked from the public record yet. Galaxy has not published, in any source we could reach, the date its snapshot was taken, the amount it assigns to exchanges, ETFs and the US government, or the method notes behind its clustering. All coverage traces back to one post by Thorn, so there is no independent second count of the 55% figure.
The next test is simple. If Galaxy releases a full report with category totals, the 3.7 million BTC gap with BitcoinTreasuries can be compared line by line. Until then, the 11.1 million figure is best read as the share of bitcoin that one well-resourced analytics team cannot put a name on, as of a date it has not stated.