Somewhere around three million coins are almost certainly gone — dead drives, forgotten passwords, keys thrown out with the junk mail. Here's how that number is estimated, why nobody can pin it down, and what it actually changes.
All 21 million coins, each square 10,000 BTC. The dark band at the bottom is the part nobody is ever getting back.
Here is the uncomfortable thing about a bearer asset: the ledger cannot tell the difference between a coin held with iron discipline and a coin whose owner died without telling anyone where the backup was. Both look the same — an unspent output sitting still. That single ambiguity is why "how much bitcoin is lost" has no exact answer, and why every number you have seen quoted is an estimate wearing a confident face.
The credible range is 2.3 to 3.7 million BTC, or somewhere between 11% and 18% of the entire supply that will ever exist. Analysts get there by looking at coins that have never moved since the era when bitcoin traded for cents and nobody backed anything up — plus the roughly 1.1 million coins mined by Satoshi Nakamoto that have sat untouched since 2009. Against the
19.91 million coins issued so far, that means the real circulating float is closer to 17 million than 21.
The losses are not evenly distributed across bitcoin's history, and that shape is the most interesting part of the data. Almost all of it happened before 2013. In 2010 a coin was worth less than a dollar; people mined thousands on a laptop, reinstalled the operating system, and thought nothing of it. James Howells threw away a drive with 8,000 BTC and has spent a decade trying to dig up a landfill. Stefan Thomas locked himself out of a hard drive holding 7,002 BTC with two password guesses remaining. These are not edge cases — they are the era.
After 2014, the loss rate collapses. Not because people became more careful by nature, but because the tooling changed:
seed phrases replaced raw key files,
hardware wallets made backups a mandatory step instead of an afterthought, and the price got high enough that people started treating storage seriously. The modern loss story is less "I forgot" and more "I died without a plan" — an inheritance problem, not a technical one.
Does any of this make the remaining coins more valuable? Mechanically yes, and practically almost not at all. Lost coins permanently shrink the effective float, so the same demand chases fewer coins. But the market has been trading around a float that excludes them for more than a decade. Those coins are not a supply shock waiting to happen; they are already in the baseline. The genuinely useful conclusion is the boring one: the scarcity people cite when arguing for bitcoin is quietly tighter than the headline 21 million, and it gets tighter every time someone loses a key.
Which points at the part you actually control. Every failure mode in the historical record reduces to one of three things: no backup, an untested backup, or a backup nobody else could find. That is a solvable list — and solving it is the difference between owning bitcoin and merely having held it for a while.
BITCOIN ALMANACK ANALYSIS
Where the 21 million actually sits
Midpoint estimates, July 2026
~3.0M likely lost forever
14.3% of total supply
~4.3M dormant 5+ years
held, or lost — unknowable
~12.6M circulating float
exchanges, ETFs, active wallets
~1.1M not yet mined
through roughly 2140
Ranges, not precision: lost-coin estimates span 2.3M–3.7M depending on methodology. Sources: Chainalysis and Glassnode dormancy studies, bitcointreasuries.net, blockchain.com · Chart: Bitcoin Almanack
WHY IT MATTERS
Every other asset has a recovery path — a bank, a registrar, a court. Bitcoin has arithmetic. That is precisely what makes it censorship-resistant and precisely what makes a lost key final. The three million missing coins are the running total of everyone who learned that the hard way, and the strongest argument for spending an afternoon on your backup plan.