Encyclopedia Supply, markets and the chain · Entry 419

How many bitcoins are lost forever? What the estimates can and cannot tell you

Degrees of Satoshi editorial projectAI-assisted research · separate automated verificationCorrections

Theme: Supply, markets and the chain
Sources: 4 cited records
Reading time: About 4 minutes
Automated verification: 2 October 2026
Substantive update: 2 October 2026

In this article

At a glance

Key facts

Key facts for How many bitcoins are lost forever? What the estimates can and cannot tell you
Fact	Detail	Source
Global total	No exact total is established by the cited evidence; inactivity cannot reveal whether a private key survives	[1][2]
Genesis output	The original 50 BTC subsidy is absent from the spendable-output database under current rules	[3]
Unspendable scripts	An output beginning with OP_RETURN is a different category from a normal address with no known private key	[4]
Howells case	The claimant alleged access to 8,000 BTC was on a discarded drive; the claim against Newport City Council was dismissed on 9 January 2025	[1]
Patoshi observation	Lerner reported 99.9% of his selected blocks unspent in April 2019; the classification and attribution are not a proof of lost keys	[2]

Separate unspendable outputs from missing keys

An output is a quantity of bitcoin locked by spending conditions. Some conditions cannot be satisfied under the rules: an OP_RETURN output is an example. The original genesis subsidy is another special case, because the software never inserted it into the spendable-output database.

A normal output protected by a key is different. If every usable copy of the required key is lost, nobody can currently authorize that spend. The public chain does not record how many key copies were created, who retained them or whether a recovery is possible.

Silence is observable; loss is an inference

Consider two outputs created on the same day for the same amount. One owner destroys the only backup; the other keeps a working backup and chooses not to spend. Until a spend occurs, an observer can see the same absence of activity in both cases. An age threshold cannot distinguish them.

Lerner’s 2019 Patoshi analysis is a useful illustration of scope. He reported 99.9 percent of the blocks in his selected early-miner set as unspent. That percentage concerns his classification at that date. It does not prove that the keys were lost, that every block belongs to Satoshi, or that the same percentage holds today.

What the Howells judgment does establish

The January 2025 judgment records James Howells’s claim that a discarded hard drive held access to 8,000 BTC and his effort to recover it from a Newport landfill. The court granted judgment for the council and dismissed the claim.

The allegation about the drive’s contents is not an independent blockchain verification that the claimant controlled that exact amount. A court deciding property rights, time limits and the viability of a claim is answering different questions from a forensic audit of a private key.

Loss estimates are not the issuance schedule

The supply schedule limits how much subsidy blocks may claim. Spendable supply then depends on unspendable outputs and keys that remain usable. A published estimate must explain which categories it includes and why it treats inactivity as evidence of loss.

Calling an address a burn address usually means it was deliberately constructed without a known private key. That is a computational assumption about finding a key, not the same validation rule as OP_RETURN or the genesis output. Keep the reason for claimed unspendability attached to the number.

Direct answers

Questions people ask

Can an explorer prove that coins are lost?

It can show that an output is unspent and display its locking script. It cannot reveal private backups. A provably unspendable script and an old output with an unknown key are different evidence categories.

Does the Howells case establish an exact lost balance?

The judgment records the claimant’s allegation about the drive and resolves his claim against the council. Treat the stated 8,000 BTC as part of that allegation, not a separate verification of key control.

Inspect the evidence

The answer and key facts have stable claim links. These records retain the scope and qualification when reused.

The blockchain cannot distinguish a lost private key from a key whose owner chooses not to spend. Some outputs are unspendable under the protocol, including the original genesis subsidy; other coins are only believed lost. Court records and historical analysis document particular claims and assumptions, not an exact global lost-coin total. A deliberately constructed burn address with no known key is not the same proof of unspendability as an impossible script.