What Is a 51% Attack on a Blockchain? (And How It Is Prevented)
Last updated: 3 July 2026 · By Spencer Li, CFTe
A 51% attack happens when a single miner or group gains control of more than 50% of a blockchain’s mining power (its “hash rate”), which lets them mine faster than everyone else and rewrite recent transaction history. With that majority, an attacker can stop new transactions from confirming and double-spend their own coins (spend the same coin twice). What they cannot do is steal coins out of your wallet or reverse a transaction that is already deep in the chain. So the damage is real, but it is mostly temporary and limited to recent, unconfirmed activity.
The catch is scale. On a huge network like Bitcoin or Ethereum, buying more than half the mining power is so expensive that it is, in practice, not worth it. The real-world victims have always been smaller coins (Grin, Vertcoin, Bitcoin Gold, Ethereum Classic), where hash rate is cheap enough to rent. The two main defences are simple: keep mining power spread out so no one passes 50%, or switch the network to Proof of Stake, where an attacker has to risk their own money to misbehave.
Here is how the attack works, what it can and cannot do, the real cases, and how networks defend against it.
What is a 51% attack?
A 51% attack occurs when one party in a blockchain network controls more than 50% of its computational power. That majority lets them mine new blocks faster than the rest of the network combined.
Why does that matter? Because the chain follows the longest valid version. If you can build blocks faster than everyone else, you can produce a longer chain and force the network to accept your version of events. You can alter recent data or stop transactions from confirming without getting consensus from anyone else.
Blockchains are secure by design, but they are not perfect. There is no way to guarantee every participant is honest. On Bitcoin, miners decide which transactions go into the next block, so a miner could choose to leave out transactions that do not suit them.
For most large cryptocurrencies, including Bitcoin and Ethereum, this stays theoretical. The computing power needed to out-mine an established network is impossibly large. You mostly see selfish miner behaviour in smaller skirmishes like “penny wars,” where small players spam low-value transactions to push up the per-kilobyte fee they earn. To handle this, most blockchains build in hard fork protocols that can change the rules if needed.
What can a 51% attacker actually do?
This is where most explanations overstate the danger, so let me be precise. A majority attacker can do a specific list of things, and there is an equally specific list they cannot do.
| Can do | Cannot do |
|---|---|
| Double-spend their own coins (spend the same coin twice) | Steal coins from your wallet directly |
| Block or delay new transactions from confirming | Reverse a transaction already buried deep in the chain |
| Reorganise recent blocks and run an alternate history | Create coins out of thin air or change old, settled balances |
| Redirect their own transactions (broadcast one, mine another) | Break the network’s cryptography |
So an attacker with majority hash rate can form their own consensus, run a different version of recent history, and double-spend. They can also block transactions they do not want confirmed. All of this hurts end users, because someone with majority power can bend the chain’s near-term rules in ways nobody signed up for.
This is one reason some people prefer off-chain solutions like Segregated Witness (SegWit) or the Lightning Network for everyday Bitcoin payments. Those move activity off the main chain, so they do not depend on heavy on-chain hashing to settle.
How a 51% attack disrupts the network
Most large networks, including Bitcoin and Ethereum’s original design, run on Proof of Work (PoW). Under PoW, the network aims to add a new block roughly every 10 minutes. To add one, a miner has to solve a hard mathematical puzzle, which is near-impossible without serious computing power.
In a 51% attack, the attacker uses their majority to interfere with unconfirmed blocks and transactions. That is the disruption: not stealing settled funds, but jamming and rewriting the recent edge of the chain. Users can lose access to digital assets in the short term, which understandably shakes trust in the network’s reliability.
Is the damage permanent?
Do note that the damage is usually temporary. A 51% attacker can disrupt the chain and invalidate recent transactions, but they generally cannot rewrite settled history.
Here is the part that calms people down. If you already sent Bitcoin to another person or a merchant and that transaction is buried in the chain, an attacker cannot reverse it. They can stop future transactions from confirming, and they can double-spend their own coins, but those are different things from reaching into a settled payment and undoing it. Once the attack ends, normal access to funds returns.
So even if it looks like money has vanished during an attack, it does not stay gone forever.
51% attack vs 34% attack: what is the difference?
People sometimes confuse the two, so here is the clean line between them.
| 34% attack | 51% attack | |
|---|---|---|
| Share of mining power | Less than half, but a large minority | More than half (majority) |
| What it can do | Tamper with the ledger to a limited degree | Run an alternate chain, double-spend, block transactions |
| Severity | Limited | Far more severe |
| Why it works | Enough power to interfere | Enough power to outpace the entire main network |
A 34% attacker can meddle with the ledger but cannot dominate it. A 51% attacker, with more than half the power available, can create competing versions of the chain and out-mine the main network. That is the difference between vandalism and control.
Real 51% attacks: the case file
The theory matters less than the track record. Every real attack has hit a smaller coin, because smaller coins have cheaper hash rate. Here are the four from the original post.
| Coin | What happened | Network response |
|---|---|---|
| Grin | An unknown miner took 57% of Grin’s hash power. The attacker’s intent was never clear. | Grin shut off payouts, told miners to pause, then re-established the network with extra safeguards. |
| Vertcoin | Attacked several times. In one, attackers replaced genuine blocks with their own. | Switched to a more robust PoW system and cut out powerful mining chips to keep mining community-based. |
| Bitcoin Gold (BTG) | Uses Equihash and GPU mining instead of ASICs. An unknown miner took over 51% of hash rate in 2018; another attack in 2020 caused two chain reorganisations in two days, with a large sum double-spent. | The community pushed for a more secure algorithm. There was suspicion that hidden ASIC devices were involved. |
| Ethereum Classic (ETC) | Attacked three times in a single month in 2020. ETC’s decentralised PoW makes 51% attacks hard to mitigate. | Prices were not badly hit, but user trust in the network fell. |
Notice the common thread: privacy coins, forks, and minority networks. None of these were Bitcoin or Ethereum proper. The size of the network is itself the security.
Can a 51% attack be prevented?
Blockchain is decentralised, but it is not immune to attack. There are two practical defences.
Keep mining power spread out. The simplest protection is to make sure no single miner or pool ever holds more than 50% of total hash rate. If no one can pass the line, no one can run the attack. This is partly social and partly built into how a healthy network distributes its mining.
Use Proof of Stake (PoS) instead of Proof of Work. Under PoS, validators (not miners) produce blocks, and they have to lock up their own cryptocurrency as a stake. To attack the network, they would have to risk forfeiting that entire stake. The bigger the stake at risk, the less rational it is to misbehave. This is a large part of why Ethereum moved to Proof of Stake.
Hence, the strongest network is not the one with the cleverest cryptography. It is the one where attacking it costs more than it could ever pay, either because the hash rate is too expensive to corner, or because the attacker’s own money is on the line.
Where the human edge comes in
A scanner can tell you a coin’s hash rate is concentrated, or that its market cap is small enough to rent an attack against. That data is free now. What it will not do is supply the judgment to size your exposure to a thin, attackable network smaller than you would a blue-chip one, or to treat “cheap to attack” as a real risk rather than a footnote. The chain’s security is the easy part to look up. Pricing that security into your own position is the judgment, and judgment is the first of the Five Edges no tool can trade for you.
FAQ
What is a 51% attack in simple terms?
It is when one party controls more than half of a blockchain’s mining power and uses that majority to mine faster than everyone else, letting them rewrite recent transactions and double-spend their own coins. They cannot steal funds from your wallet or reverse settled payments.
Can Bitcoin suffer a 51% attack?
In theory, yes. In practice, no one has done it, because buying more than half of Bitcoin’s mining power would cost an impossibly large amount. The networks that have actually been attacked are much smaller coins like Grin, Vertcoin, Bitcoin Gold, and Ethereum Classic.
Is my money safe during a 51% attack?
Mostly. An attacker can block new transactions and double-spend their own coins, but they cannot reverse a payment that is already deep in the chain or steal coins out of your wallet. Once the attack ends, normal access returns.
What is the difference between a 51% attack and a 34% attack?
A 34% attacker has a large minority of mining power and can tamper with the ledger to a limited degree. A 51% attacker has the majority, which is far more severe: they can run an alternate chain, block transactions, and double-spend.
How do you prevent a 51% attack?
Keep mining power spread out so no single party passes 50%, or switch the network to Proof of Stake, where attackers must risk forfeiting their own staked cryptocurrency to misbehave.
Now that you know what a 51% attack can and cannot do, can you tell which small tokens are most exposed to one? Let me know in the comments.
And if you want the full picture of how blockchains and crypto fit together, read the pillar: The Ultimate Guide to Blockchain and Cryptocurrencies.
Want a calmer way to trade crypto and any other market? Grab the free 15-Minute Swing Trading Starter Kit. It is the exact routine I use to scan once a day and trade any market in 15 minutes.
About the author. Spencer Li is the founder of Synapse Trading and a Certified Financial Technician (CFTe) with 15 years of trading across stocks, forex, crypto, commodities, and bonds. His trade log is public, 404 trades, losses left in. He teaches low-risk swing trading in 15 minutes a day, one system for any market.
Education, not financial advice. Synapse Trading is not licensed by MAS to advise on investment products. Trading carries risk of loss; past performance is not indicative of future results.
Related
The Ultimate Guide to Blockchain and Cryptocurrencies (pillar) · What is Proof of Stake vs Proof of Work · What is a blockchain fork · How to value a cryptocurrency
