Bitcoin is a chain of decisions, not a single date · August 15, 2010

184 Billion BTC That Should Never Have Existed

A transaction in block 74638 created more than 184 billion BTC. The faulty chain was stopped, the software was repaired, and the impossible coins disappeared from Bitcoin's active history.

A strange block

On August 15, 2010, Jeff Garzik noticed something extraordinary in block 74638. Two outputs of one transaction contained roughly 92.2 billion BTC each. This was not a display error: an early version of the software had accepted the transaction and included it in the chain.

The creator of the transaction has never been reliably identified. We only know that an unknown participant found a weakness in validation and used it.

How the impossible became valid

Bitcoin amounts were represented as whole numbers of satoshis. Each output was checked separately, but adding two extremely large values overflowed the numeric type. The sum crossed its limit and appeared to the program as a different value.

The supply rule existed. The failure was in the way the software checked whether that rule had been respected.

The network did not repair itself

A person first noticed the unusual data. Other participants checked the block independently. Developers found the cause, prepared a fix, and warned users not to trust transactions after block 74638. Node operators had to install the new version, while miners had to build on the repaired rules.

Decentralization does not mean the absence of decisions. It means a decision cannot merely be declared: network participants must adopt and enforce it.

A repair within hours

The first public report appeared roughly an hour and a half after the event. A patch followed within hours, and Bitcoin 0.3.10 was released. It rejected overflowing sums and outputs above the permitted limit.

For a short time, two chains existed. The old one accepted block 74638; the repaired one rejected it. Once the repaired chain accumulated more proof of work, nodes converged on it. The impossible transaction was no longer part of active Bitcoin history.

It resembled a rollback, but no administrator deleted it

Nobody opened a central database and erased a row. Participants installed software with a stricter rule and began building another chain. The outcome depended on their compatible actions.

Bitcoin was a small network in 2010, Satoshi was still actively involved, and there were far fewer users and miners. Such a repair was much easier then than it would be across a mature global network.

The limit does not exist apart from validation

The faulty software briefly accepted an impossible amount, but those BTC did not survive in the active chain. The supply limit is not a statement on paper. It is a rule that nodes check in every block.

What block 74638 left behind

This incident does not prove Bitcoin is invulnerable. It demonstrates the value of open code, independent verification, and public analysis of failure. The block vanished from the main chain but remained in Bitcoin history as a reminder: software rules are resilient only when participants can detect a violation and act together to defend them.

Reliability is not the absence of errors. It is the ability to see an error, understand it, and prevent it from becoming the new rule.

Historical and educational material. Not financial advice.