How to destroy bitcoin (and what happened to 107 burned coins)

If you want to destroy some fiat money, you can set fire to it. Easy to create, easy to destroy.

If you want to destroy gold, it’s much harder. You need a nuclear reactor and a lot of energy, which is why pretty much all the gold that’s ever been mined still exists. Side note, you can dissolve gold in aqua regia - I guess that’s the retro version of a coinjoin.

But what about bitcoin? How hard is it to destroy?

It’s actually quite easy - and there are several ways you can do it. In May of 2026, someone destroyed 107 bitcoin, worth about $8.3 million at the time. I thought it would be fun to dig into how they did it, how else they could’ve done it, and to speculate wildly about their motivations - including one detail that most of the coverage missed. Let’s go.

1. What actually happened
2. You can’t delete a bitcoin
3. Twelve years, then a burn
4. Why would anyone do this?
5. If you’re going to burn, burn properly

1. What actually happened

On 25th May 2026, block 950962 confirmed five near-identical transactions. Each spent a handful of inputs from one legacy address and produced exactly one output. Each paid a fee of 1.81 sats/vb - higher than the 1 sat/vb the rest of the block was paying, and under $2 apiece - and each set its locktime to block 950958, strongly suggesting that all five transactions were built by the same wallet in one sitting.

The destination - 1111111111111111111114oLvT2 - is a well-known burn address. It’s been collecting coins for over a decade and now holds a staggering 809 BTC across ~260,000 transactions. None of the coins sent here will ever move again.

I’ll save the why question for later. First, what does it even mean to destroy a bitcoin?

2. You can’t delete a bitcoin

Every bitcoin transaction that’s ever taken place is permanently inscribed on the network’s immutable ledger. This ledger only ever grows. So if you want to destroy some bitcoins, what you’re really asking to do is make those specific coins permanently unspendable.

You could do this by generating a wallet, deriving a fresh address, sending bitcoin to this address, then destroying the seed and any backups. No-one has the keys to that address, so the funds will be permanently locked there. This is, in effect, what’s happened to countless coins over the years.

But - no-one else can prove that the coins are unspendable. From an outsider’s point of view, they see coins locked to a normal address and staying there. Much like Satoshi’s coins, no-one can guarantee that they won’t move again. Can we instead send coins to an address which provably has no private key?

This is a legacy P2PKH address - Pay To Public Key Hash. For a full technical primer, head to Learn Me A Bitcoin. For now, you’re probably wondering why it starts with such a long run of ones. Let’s build one and find out.

Making an address like this takes two steps.

1. Take a public key and put it through a one-way hash function. This produces a public key hash; a shorter, random-looking string, with no way to run it backwards to recover the public key.

2. Format that public key hash into an address, which is what you share to receive bitcoin.

Our 1111…oLvT2 address is what the second step produces when the public key hash you feed it is 0000000000000000000000000000000000000000 - twenty bytes of nothing (or, to be more precise: a 00 version byte, twenty zero bytes, and a checksum).

You’ve probably figured this out already, but the 1111…oLvT2 address was not generated from a pre-existing public key. Someone skipped a step, started with an impossible all-zero public key hash and encoded that into a P2PKH address. There is no known public key for it.

To spend from 1111…oLvT2 you’d have to produce a public key that hashes to twenty zero bytes - and, crucially, the private key used to generate it. So the search is: pick a private key, derive its public key, hash it, and see how close you get to all zeroes.

Every attempt is a fresh, genuine key pair, and every one of them misses. For what it’s worth, I have coded this component to stop upon finding a forty-zero match and highlight your achievement in a suitably celebratory shade of green. Just bear in mind that the combined work of every miner in Bitcoin’s history equates to around 10²⁹ hashes. So don’t hang around.

A word of warning, in case it needs saying: never send coins to a burn address unless you intend never to see them again.

3. Twelve years, then a burn

A lot of the coverage of the May burn focused on the age of the coins. All five addresses were first funded on 10th April 2014, within twenty-five blocks of one another, when bitcoin was around $364. This means someone held onto returns of over 20,000% before deciding to burn what was left.

Shortly after the burn, Galaxy Research published a visualisation highlighting the second-smallest address - 16g5hMo…99bVQt - as a notable exception to the pattern. It was marked as holding fresh, 2026-era coins - which is true of the coins themselves, but overlooks an interesting point.

This address received 20 bitcoins on the 10th April 2014, worth about $7,300 at the time. They stayed there for twelve years until 18th March 2026, when the entire balance was sent in a single transaction to a P2SH address. Three weeks later, almost exactly the same amount came back, in three payments of 7, 7 and 6 BTC. Seven weeks after that, it was burned.

That P2SH destination, 3FZity…MxsQd, never rests. Every deposit is swept within an hour or two by a ~50-input consolidation transaction gathering dozens of 2-of-3 addresses built to the same template into one big output. This looks to me like a per-customer deposit address at a multisig custodian. Interestingly, our 1PkWqW1…BeiRT6V address - the smallest of the burned five - had been paying into this same deposit address since May 2022, some sixty payments of between 0.15 and 0.62 BTC each, draining a total of 19.6 BTC out of a 2014 stack.

I’m speculating here, but perhaps someone was gradually liquidating the smallest address in their stack through a custodian, decided to liquidate a big chunk more, changed their mind and returned the funds to their own wallet, before finally burning the lot.

4. Why would anyone do this?

Burning on purpose is rarer than losing by accident, but it happens. Some of the most common explanations are:

1. Tax loss harvesting. In some jurisdictions, provably destroying an asset lets you claim a capital loss you can offset elsewhere. But the loss you can claim is what you paid, not what it’s worth - and these coins were bought at around $364 each. Our burner would have set fire to $8.3 million to write off about $39,000. Nobody does that on purpose.
2. Proof of burn. Doesn’t appear to be applicable here, but some projects have used burning as a costly signal. Counterparty launched its token in 2014 by sending bitcoin to a dedicated burn address - there’s 2,131 BTC sitting at 1CounterpartyXXXXXXXXXXXXXXXUWLpVr. Side note, I’ll write another article sometime on how these kinds of vanity addresses are created.
3. Tainted coins. If the stack was the proceeds of something illegal, and every route to spending it runs through someone who’ll ask where it came from, then destroying it does at least end the problem of owning something you can’t safely use.
4. AI mishap. Plenty of people now trust their entire digital lives to agents, and an agentic blunder could plausibly make coins unspendable. But a well-chosen burn address and five matching transactions sharing a locktime doesn’t look like a blunder to me.
5. Coercion. Someone under threat of violence is forced by an attacker to destroy their wealth rather than hand it over, out of spite. I hope it’s not this one.
6. Making a point. Burning coins as protest, as art, or simply to be remembered in the ledger’s lore. There’s also the possibility that a very wealthy individual without heirs decided to permanently burn their coins (thereby publicly reducing the total bitcoin supply), rather than just destroying their keys.

The on-chain data is a signal here. The three-week round trip (covered in section 3 above) tells us somebody was thinking about this well before May, and five transactions built in one sitting with matching fees suggests a deliberate action.

Remember that Bitcoin’s supply is capped at 21 million, but that’s the nominal figure. Millions are already gone thanks to forgotten keys, hard drives in landfills, etc. A burn is just the deliberate version of this, and is - in some sense - a gift to every other holder.

5. If you’re going to burn, burn properly

There are multiple ways to make an output permanently unspendable, with some being less burdensome to the network than others.

• An eater address is the crude method, and it’s what May’s burn used. 1111…oLvT2 looks like a normal address because it is one - it just commits to a hash somebody picked rather than one derived from a key. You can be confident nothing will ever move from it, but your node can’t be so sure. All it sees is an ordinary P2PKH output waiting for a signature, so those 809 BTC sit in the UTXO set indefinitely - which is why Adam Back called May’s burn an “accidental quantum bounty”.
• OP_RETURN is a somewhat tidier method. Rather than locking coins to a hash, you give the output a script beginning with the OP_RETURN opcode (0x6a), which is designed to fail the moment it runs. There’s no key to search for and no vanishingly small chances to argue about. Your node marks the output provably unspendable, and never adds it to the UTXO set at all.

To burn coins with OP_RETURN, you have to put them in that output. Default relay policy expects those outputs to carry zero value, a precaution which exists precisely to stop people destroying coins by accident. With bitcoins in the output, the transaction becomes non-standard: valid by consensus, and a miner could still include it in a block, but regular nodes won’t relay it. So this method is really only open to miners, or to anyone who can hand a transaction straight to one.

Since nothing after the opcode is ever executed, you get to attach a small data payload. In my example below, that’s four bytes of ASCII spelling “GONE”. Exactly what belongs in that payload has become a heated topic among Bitcoiners, but that’s a blog post for another day.

May’s burn used the first method. The coins are equally gone either way, but only one of the two leaves every node on the network storing outputs that will never move again.

There are other ways to permanently shrink the supply, too. For one, a miner can simply claim less than the subsidy and fees they’re eligible for. Nobody else can claim the difference later - it’s gone for good, no output required.

Wrapping up

I hope you’ve enjoyed this journey down a very specific branch of the Bitcoin rabbit hole (even if I am nearly 3 months late to the original news of the burn). This post felt so niche - even for a website like this - that I’ve stuck it on my blog instead of over in Learn, on the basis that it’s got little practical utility! But if people enjoy this kind of article, I’d certainly like to write more of them.

Anyway - spare a thought for the 107 bitcoin that’ll never be used as money again, trapped in the UTXO set of every full node around the world.

And if by some miracle you do manage to unlock them using my tool above, do us a favour - sling some of the sats my way ;)

Any donation is gratefully received and will help keep these posts coming. Thanks for reading!