I watched the silence break the noise of 2021, and I have been listening for a different kind of silence ever since. On October 14, Alex Thorn, the research director at Galaxy Digital, released a figure that most of the market scrolled past without pausing: more than half of all Bitcoin in existence sits in addresses whose ownership no one can name. Not anonymous by design. Unknown by limitation. The chain — our ledger of perfect memory — was quietly admitting that it does not actually know who holds the majority of its own asset.
I sat with that number for a long time. In the winter of 2021 I had buried myself in the CryptoPunks and Bored Ape communities, interviewing forty artists and collectors while the tickers screamed, and I learned then that the loudest claims are usually the least examined. This was the inverse. This was a quiet number, a load-bearing number, resting at the foundation of the one asset the entire industry describes as transparent.
To understand why this matters, you have to hold two stories in your head at once, because they are in open tension and have been for years.
The first is the institutional story. Spot Bitcoin ETFs arrived in early 2024, corporate treasuries began stacking coins, and the dominant narrative of 2024 and 2025 became: the institutions have taken over. I tracked that shift closely with a small research team, mapping how language migrated across the accounts of traditional finance influencers. The narrative shifted from "store of value" to "institutional yield play," a change we logged across two hundred handles and turned into a framework I called the Institutional Narrative Bridge. That story is real. Real money flowed in. Real custody infrastructure was built around it.
The second story is older and quieter: Bitcoin as the asset of the self-sovereign individual, where the private key is the entire point, where ownership is a fact you hold rather than a permission you are granted. This story has no marketing budget and no quarterly filings. It simply sits there, in cold storage, in hardware wallets, in seed phrases stamped onto steel plates and buried in gardens.
Bitcoin has cycled through these two stories for a decade and a half. The halving narrative, the store-of-value narrative, the payments narrative that died and was quietly buried — each rose, peaked, and settled into the sediment of the next. The institutional narrative is simply the most recent layer, and like every layer before it, it is already being tested by the data underneath.
Alex Thorn's observation is interesting precisely because it stands at the intersection of both stories and refuses to fully serve either. If more than half of Bitcoin sits in unknown-ownership addresses, then the "institutions have taken over" narrative is, at minimum, overstated. And the mechanism behind that number — the thing that makes ownership unknowable — is not a flaw in Bitcoin. It is the feature that made Bitcoin worth caring about in the first place.
Before I go further, I want to be careful about the epistemics, because the source material here is thin. This is a single expert's observation, reported as a fast-fact, without a published methodology, without a defined sample, without independent cross-verification. I will separate what was explicitly stated from what I am inferring from professional experience, and I will flag where I am reaching. That discipline matters more than the headline. A single respected voice is not the same thing as a verified dataset, and the industry's habit of treating the former as the latter is one of its most persistent failures.
Here is the machinery, because the machinery is the actual story.
When an analyst says an address has "unknown ownership," they are not reporting a cryptographic fact. They are reporting the outer edge of a heuristic. On-chain attribution works by clustering: software groups addresses together and guesses they belong to the same entity based on behavioral patterns. The oldest and most relied-upon rule is the common input ownership heuristic — the assumption that if several addresses are signed together in a single transaction, they share a controller. It is a reasonable guess. It is not a proof. There is no cryptographic link between an address and a human being. There never has been.
That single gap explains the entire statistic. The "more than half unknown" figure is not a measurement of hidden whales or dark pools. It is a measurement of how far clustering heuristics can reach before they fall apart. And they fall apart predictably, in exactly the places that matter. CoinJoin transactions deliberately merge inputs from unrelated parties, breaking the common-input assumption by design. Mixers obscure provenance. Exchange batch withdrawals push thousands of users' coins through a single outbound transaction, blurring the line between the exchange and its customers. Every privacy tool and every operational convenience erodes the map a little further.
So when I read that half of Bitcoin is "unknown," my first instinct is not to imagine a shadowy cabal. My first instinct is to ask: unknown to whom, and by what method? The number is a horizon, not a fact.
I have spent enough hours inside these dashboards to know their texture. The label "unknown" appears in a clean gray font, as if it were a settled category, when in reality it is a shrug rendered in software. Independent providers — Chainalysis, Glassnode, Arkham — all wrestle with the same ceiling, and their numbers diverge because their heuristics diverge. When three reputable firms give three different answers to the same question, the honest conclusion is not that one of them is right. It is that the question exceeds the instrument.
Now let me push into what is actually likely sitting inside that unknown half, because lumping it all together destroys the meaning. Based on my years working through on-chain data alongside independent providers, I would expect three very different populations. First, coins that are permanently lost — private keys gone, hardware destroyed, owners dead without leaving recovery instructions. Historical estimates for this graveyard run between roughly 1.5 million and 3.7 million BTC, and that supply will never move again. Second, dormant addresses that are not lost but sleeping — the most famous being Satoshi's roughly 1.1 million BTC, untouched since the earliest days of the network. Third, active self-custody addresses that simply have not yet been clustered by any analytics firm, because the holder is a careful individual who does not reuse addresses and does not transact in patterns that invite grouping.
These three categories have nothing in common except that our instruments cannot name them. One is a permanent hole in the supply. One is a philosophical monument. One is the beating heart of the self-sovereign thesis, invisible only because its owners are competent. Calling all three "unknown ownership" is technically accurate and analytically useless.
The dormant category deserves its own warning. If a meaningful share of the "unknown" supply is early-era coins — the so-called Satoshi-era hoards that have not moved in over a decade — then the unknown half is not purely a story of dispersed retail conviction. It is also a latent overhang. A dormant address is not a lost address; it is a loaded one. The probability that any single one moves is tiny, but the aggregate is a permanent theoretical supply shock that the market has learned to live with and, occasionally, to fear irrationally.
This is where the narrative work begins, and it is worth mapping the regulatory endpoint backward, because the endpoint is already visible on the horizon. A high self-custody share is not just a technical fact. It is a policy problem waiting to be weaponized. The Financial Action Task Force's Travel Rule and the European Union's MiCA framework both lean toward imposing reporting obligations on self-custodied wallets — the logic being that a transfer into an unhosted wallet is a transfer into a blind spot. A statistic showing that the majority of the world's most valuable crypto asset sits in exactly those blind spots is a gift to regulators who want to tighten the rules. I can already imagine the memo: over half of Bitcoin is beyond our visibility, therefore we must require disclosure.
There is a bitter irony here that connects to something I have argued for years. Most project-level KYC is theater — buying a handful of wallet holdings bypasses the entire charade — and the compliance costs are passed straight to the honest user, while the sophisticated mover slips through the heuristic's blind spots. The same is true at the protocol level. The individual who holds her own keys, pays her taxes, and transacts plainly is the one who becomes legible, and therefore regulatable. The whale who understands clustering and deliberately avoids it stays "unknown." Legibility is not a function of guilt. It is a function of naivety, and that is a deeply uncomfortable thing for a system that claims to be fair.
Now the token economics, because Bitcoin is genuinely different here and it deserves precision rather than slogans. Bitcoin has no team allocation, no vesting cliff, no unlock schedule, no protocol revenue. Its value capture is a monetary premium — a belief premium — not a discounted cash flow. When a token has a treasury and an unlock calendar, its holders are structurally waiting for later buyers to absorb the supply, which is why I have long argued that most governance tokens are non-dividend equity at best and a slow Ponzi at worst. Bitcoin is not that. The self-custody-dominant structure is, from a pure supply standpoint, a genuinely positive signal: the coins are dispersed, and there is no single centralized entity sitting on a cliff of locked tokens waiting to dump onto the market.
But — and this is the part the enthusiasts skip — self-custody does not mean no selling pressure. It means unpredictable selling pressure. A foundation with a vesting schedule is a known quantity; you can model the cliff, you can anticipate the unlock. A million individual key-holders are not a known quantity. They sell into euphoria and they panic into crashes, and their aggregate behavior is harder to forecast than any single institution's. The absence of a central seller is a gift to the decentralization thesis and a headache to anyone trying to model volatility. The ETF didn't fail. It simply didn't do what the narrative promised, which was to concentrate control into a handful of identifiable custodians who could be reasoned with, regulated, and predicted.
That distinction — dispersible versus predictable — is the real fault line this data exposes. And it is why the ecosystem around self-custody matters more than the headline suggests. Hardware wallets, multisig services, and self-custody tooling sit downstream of exactly this structural fact. If the majority of Bitcoin genuinely lives outside custodial reach, then the infrastructure that makes self-custody usable is not a niche curiosity. It is load-bearing plumbing. Meanwhile, the custodial businesses — exchanges, ETF providers, custody banks — face a mild narrative headwind, because their "asset reservoir" argument weakens if the reservoir was never as full as advertised. None of this moves price today. All of it shapes which infrastructure earns the right to exist tomorrow.
And here is where the current market mood matters. We are in a sideways grind, the kind of tape where narratives do the work that price refuses to do. In a trending market, no one reads methodology notes; they read candles. In a consolidating market, everyone becomes a researcher, because positioning is the only game available. That is precisely when a figure like "half of Bitcoin is unknown" travels — not because it changes anything, but because it gives a directionless market a frame to think with. The danger is that a frame becomes a belief, and a belief becomes a position, long before anyone checks the arithmetic.
Here is where I want to break from the consensus reading, because I think most people are drawing the wrong lesson from this number.
The comfortable interpretation is: "Good news — Bitcoin is still decentralized, retail still rules, the institutions haven't eaten it." I understand the appeal. It flatters the founding myth. But it mistakes a measurement artifact for a discovery. The real finding is not that individuals hold most of Bitcoin. The real finding is that we do not actually know who holds most of Bitcoin, and we probably never will — and that the asset's defining property is precisely the reason we cannot know.
Think about what it would take to close this gap. You would need every self-custodial holder to voluntarily identify themselves, which is the opposite of what self-custody means. Or you would need the protocol to embed identity, which would require changing the very thing that makes Bitcoin Bitcoin. Bitcoin's governance is so conservative that any "constitutional" change of this kind is effectively impossible. So the blind spot is not a transitional state we will grow out of. It is permanent, structural, and self-reinforcing. History doesn't reward the asset that knows everything. It rewards the asset that cannot be forced to tell.
There is also a source-bias worth naming, gently. Galaxy Digital runs asset management, mining, and ETF-adjacent businesses. Its research director is credible and widely respected, and I have no reason to doubt the intent behind the figure. But a firm with a stake in Bitcoin's long-term legitimacy has a structural incentive to emphasize the decentralization story — because decentralization is the argument that keeps Bitcoin out of the securities bucket and keeps its premium intact. I am not accusing anyone of spinning. I am noting that the same number reads very differently depending on who is holding the pen, and that a single expert's observation, however respected, is not the same thing as a peer-reviewed measurement. The honest position is that this is a hypothesis wearing the costume of a fact.
So what do we do with a number that tells us the ledger does not know its own contents? I think we stop treating on-chain data as a window and start treating it as a lens — one with a focal length, a field of view, and a blind spot you have to learn to see around. The next phase of this industry will not be won by whoever holds the most data. It will be won by whoever is honest about what the data cannot reach.
If half of Bitcoin is a silence, then the question I am left holding is this: in a system built so that no one can be forced to speak, how do we ever assemble a true picture of who owns it — and should we even want to?
I keep coming back to the human layer, because that is where the numbers eventually land. Behind every "unknown" address is a person who made a choice: to hold their own keys, to accept responsibility for their own security, to opt out of the custodial comfort on which most of the financial system is built. Some of them are dead, and their coins are monuments. Some are careful, and their coins are a quiet act of faith. A few are lost forever, and that is a kind of grief the ledger will never record.
The uncomfortable ethical question is not whether self-custody is good. It is whether a world that demands visibility from the honest and rewards invisibility to the clever is a world we should keep building. Bitcoin, for all its flaws, is one of the few systems that lets an ordinary person say: this is mine, and I do not need anyone's permission to prove it. That is worth protecting, even when protecting it means admitting — publicly, uncomfortably — that we cannot see who holds the majority of it. I watched the silence break the noise of 2021. I am still listening to the quieter silence underneath. It has more to teach us than the noise ever did.