The 51% Attack: Could Anyone Take Over Bitcoin?

Category: Mining and the Network · Published 2026-07-01

Security reviewed and sourced on 2026-07-03

What people mean by "taking over Bitcoin"

The phrase 51% attack sounds like a master key — as if whoever controlled a majority of the network's mining power could seize Bitcoin and do whatever they liked. The reality is narrower, and far more interesting. Bitcoin holds two separate kinds of power. There is the power to order transactions — to decide which valid transactions go into the next block, and in what sequence. And there is the power to decide what is valid in the first place — the rules that a signature must match, that no one may spend coins they do not own, that only twenty-one million coins will ever exist.

Miners hold the first kind of power. They do not hold the second. A 51% attack is an attack on ordering, not on the rules. Once you see that line, nearly every scary headline about Bitcoin being "hacked" or "taken over" sorts into one of two buckets: things a majority miner can genuinely do, and things it physically cannot.

What a majority attacker could actually do

Suppose someone assembled more hashrate than the rest of the network combined. Because they can produce blocks faster than everyone else on average, they can build a longer chain in private and then publish it, overruling the chain the honest network was working on.¹ With that ability they could:

• Reorganize recent blocks — erase the most recent few blocks and replace them with their own version.
• Double-spend their own coins — pay someone, let the transaction confirm, take the goods, then rewrite history so the payment never happened and the coins return to them.
• Censor transactions — refuse to include particular transactions in the blocks they mine, delaying them for as long as they hold the majority.

That is the whole arsenal, and notice what it has in common: it touches only recent, shallow activity. The attacker can rewrite only their own payments to their benefit — but other people's transactions caught in the replaced blocks can be dropped, delayed, reordered, or censored until they are mined again. This is precisely why merchants wait for several confirmations on large payments. Each block buried on top of a transaction makes reversing it exponentially more expensive, until rewriting it would mean re-mining a mountain of work.

What it could not do — the hard limits

Here is where the myth collapses. A 51% attacker cannot do the things people fear most, because those things are forbidden by rules that every honest node enforces independently. No amount of hashrate buys an exception. Specifically, an attacker cannot:

• Steal your coins. Moving bitcoin requires a valid digital signature from the owner's private key. Mining power is not a key.² A block full of unsigned theft is rejected by every node as invalid.
• Spend coins that are not theirs. Same reason. They can reorder transactions, but they cannot forge them.
• Change the 21 million cap or print extra coins. The supply schedule is a rule, not a vote. A block paying the attacker too much is invalid and ignored.
• Rewrite old, deeply buried history. Reversing a transaction from years ago would mean re-doing every block since while outpacing the whole network — a cost so astronomically large and ever-growing that it is treated as practically unachievable.¹

A useful way to hold this in mind: miners propose, but nodes dispose. The thousands of independent computers running the Bitcoin software each check every block against the rules and throw out anything that breaks them — whoever mined it. The attacker can shuffle the deck, but cannot invent new cards.

Why the economics work against an attacker

Intuition says a bigger, more valuable network makes a juicier target. The economics say the opposite. To out-mine the honest network you must match and exceed its total hashrate — and that total currently represents a vast amount of specialized hardware and electricity (though hashrate can fall as well as rise). You would need to acquire or build comparable machines, which are not sold in such quantities and take time to manufacture, then power and run them, all to win a fight that pays very little.

Because the prize is so limited — reversing a handful of your own recent transactions, not looting the network — the spend rarely justifies the cost. Worse, a visible attack would undermine confidence in the very thing the attacker paid a fortune to attack, eroding the value of any coins gained and the hardware bought. For a rational, profit-seeking adversary, mining honestly is almost always the better business — though the incentive argument says less about attackers who do not care about profit. The system is defended less by a wall than by an incentive: it is cheaper to play along than to break in.

Why big chains resist it — and small chains don't

There is no widely accepted public evidence of a sustained majority-hash attack reorganizing Bitcoin's main chain for a successful double-spend. That is a narrow, specific claim: it is not the same as saying Bitcoin has never had worrying mining concentration — it has, at times — and a temporary majority-hashrate estimate is not the same thing as a confirmed attack.³ Absence of public evidence is not proof that an attack is impossible.

Smaller proof-of-work coins are a different story. When a network's total hashrate is modest, an attacker may not need to build anything — they can rent enough hashing power for a few hours to overwhelm it, and several smaller chains have been double-spent in exactly this way.⁴ A small chain does not share Bitcoin's security profile. 51% resistance is a function of how much honest work stands behind a chain — not something automatic. This is also why broad participation in mining, including home miners, supports decentralisation.

A clearer mental model

Picture Bitcoin's ledger as a stone tablet that many scribes copy at once. A 51% attacker is a scribe who can carve faster than all the others combined. With that speed they can argue about the last line or two — whose transaction was recorded, in what order — and occasionally win that argument by force. What they cannot do is carve a line that breaks the rules every other scribe checks: no forging signatures, no inventing coins, no exceeding the supply, no erasing the deep, settled pages.

So could anyone take over Bitcoin? Not in the way the phrase implies. The worst a majority attacker can manage is to double-spend their own recent payments and disrupt — delay, reorder, or censor — other recent transactions, at great cost, for a small and self-defeating reward — while the rules that actually protect your coins stay untouchable. The honest answer turns out to be the reassuring one.

Key takeaways

• A 51% attack is an attack on transaction ordering, not on the consensus rules.
• It can reorg recent blocks, double-spend the attacker's own coins, and censor transactions.
• It cannot forge signatures, steal arbitrary coins, mint beyond the cap, or change the rules by hashrate alone.
• No widely accepted public evidence shows a sustained majority-hash reorg/double-spend of Bitcoin's main chain — which is not a proof of impossibility.
• Smaller PoW chains have been double-spent by rented hashrate; they don't share Bitcoin's security profile.

Sources

1. Bitcoin: A Peer-to-Peer Electronic Cash System (Nakamoto, 2008) — bitcoin.org
2. Bitcoin developer documentation — Bitcoin Project
3. Weaknesses (majority-attacker capabilities enumeration) — Bitcoin Wiki (secondary/optional reading)
4. Ethereum Classic hit by third 51% attack in a month (Aug 2020) — CoinDesk

Related educational reading

• How Proof of Work Secures Bitcoin
• What Is a Bitcoin Full Node?
• Soft Fork vs Hard Fork, Explained
• How Bitcoin Transactions and UTXOs Work