How 51% Attacks Work in Proof of Work Blockchains

How 51% Attacks Work in Proof of Work Blockchains

51% Attack Cost Estimator

Calculate Attack Viability

Example: Bitcoin Gold has ~1.5 TH/s (1,500 GH/s)
Example: $100,000

Hash Rate Conversion Guide

1 EH/s = 1,000 TH/s = 1,000,000 GH/s = 1,000,000,000 MH/s

Bitcoin's hash rate: ~400 EH/s
Bitcoin Gold: ~1.5 TH/s (as of 2020)

Imagine you’re at a public ledger where everyone writes down transactions. Everyone agrees that the longest version of this ledger is the truth. Now imagine one person secretly writes a different version - faster, longer, and hidden from everyone else. When they finally show it, the whole network switches to their version. Suddenly, money you sent is gone. That’s a 51% attack.

What Exactly Is a 51% Attack?

A 51% attack happens when a single miner or group controls more than half of a blockchain’s total computing power - called hash rate. This isn’t about hacking code or breaking encryption. It’s about brute force. If you have more power than everyone else combined, you can outmine the rest of the network.

This only works on blockchains using Proof of Work (PoW), the original consensus system Bitcoin uses. In PoW, miners compete to solve complex math puzzles. The first to solve it gets to add the next block and earns a reward. The more computing power you have, the more often you win.

But here’s the catch: the network trusts the longest chain. That’s the rule. So if you secretly build a longer chain - even if it’s fake - the network will accept it as real. That’s how you reverse transactions, double-spend coins, or block others from mining.

How It Actually Happens

Let’s say you want to double-spend 100 Bitcoin Gold coins. Here’s the step-by-step:

  1. You send 100 BTG to an exchange to buy something valuable - say, Ethereum.
  2. The exchange waits for 12 confirmations (blocks) before releasing your purchase. That’s normal.
  3. While the exchange waits, you start mining your own secret chain in private. You don’t include your original transaction in this chain.
  4. You keep mining faster than the public network. You’re using rented hash power from a site like NiceHash, spending $2,000 for 4 hours.
  5. After 30 minutes, your secret chain is now 35 blocks long. The public chain is only 30 blocks long.
  6. You broadcast your longer chain to the network. Every node sees it and says, “This is longer - it must be the real one.”
  7. Your original transaction disappears. The exchange thinks you never sent the 100 BTG. You keep your Ethereum. They’re out $100,000.

This is called a chain reorganization - or “reorg.” The network rewrote history. And it happened because you had more power than everyone else.

Why Small Blockchains Are Easy Targets

Bitcoin’s network has over 400 exahashes per second (EH/s). That’s 400 million terahashes. To control 51% of that, you’d need hardware worth billions. It’s impossible.

But look at Bitcoin Gold (BTG). Its hash rate? About 1.5 terahashes per second. That’s 266,000 times smaller than Bitcoin’s. Renting enough power to attack it costs under $2,000 for a few hours. In January 2020, attackers did exactly that - stole $70,000 in double-spent coins.

Same thing happened with Verge (XVG) in 2018. Attackers reversed 215,000 coins - worth $1.7 million at the time - by building a 300-block-long secret chain. The network had no way to stop it.

These aren’t rare. Between 2019 and 2023, MIT’s Digital Currency Initiative recorded over 40 major reorgs on small PoW chains. Many were under $100 million in market cap. The pattern is clear: low hash rate = easy target.

Cartoon attackers renting hash power from a marketplace to attack a small blockchain.

Hash Rate Rental Markets Made It Too Easy

Before 2018, launching a 51% attack meant buying thousands of ASIC miners. Expensive. Hard to hide. Now? You go to NiceHash. Click “Rent Hash Power.” Pay in Bitcoin. In 10 minutes, you have enough power to crush a small blockchain.

Chainalysis estimates that $533,000 worth of hash power is rented monthly for malicious purposes. That’s not just one attack - that’s dozens. Attackers don’t need to be tech geniuses. They just need cash and patience.

And exchanges? Many still only require 12-24 confirmations for altcoins. That’s not enough. After the Verge attack, some exchanges raised their threshold to 60-100 confirmations. But many still don’t.

Proof of Work vs. Proof of Stake

Ethereum switched from Proof of Work to Proof of Stake in September 2022. That wasn’t just a tech upgrade - it was a survival move.

In PoS, you don’t need computing power. You need coins. To attack a PoS chain, you’d need to own 51% of all staked ETH. That’s over $10 billion. You’d be buying up the entire market. And if you tried to double-spend, the network would slash your stake - burning your own money.

That’s the key difference. PoW attacks cost money to rent. PoS attacks cost money to buy - and you lose it if you fail.

Today, over 63% of enterprise blockchain projects use PoS. Only 12% use PoW. Why? Because PoW’s biggest strength - decentralization through mining - is also its biggest weakness. Mining pools already centralize power. Now, with rental markets, anyone can buy that power.

What’s Being Done About It?

Some projects are fighting back. Vertcoin added “checkpointing” - trusted nodes freeze the chain every few hours. Even if someone builds a longer chain, the network ignores it past the last checkpoint.

Bitcoin Gold tried switching to a GPU-friendly algorithm called Autolykos in 2022. It worked for months. Then attackers used rented GPU farms and attacked again.

Exchanges are slowly catching up. Binance users now manually wait for 60+ confirmations on small PoW coins. Coinbase doesn’t warn users - but users are learning the hard way.

MIT’s monitoring system now checks 150 PoW chains in under an hour. That’s a big improvement. But detection isn’t prevention. By the time you spot the attack, the damage is done.

Cartoon comparison of vulnerable small blockchain vs secure Bitcoin with massive mining power.

Is Bitcoin Safe?

Yes. Bitcoin has never had a successful 51% attack. Not even close.

Its hash rate is so massive, and its value so high, that attacking it would cost more than you could steal. Even if you spent $10 billion on mining gear, you’d lose it all if the community rejected your chain. The economic incentive doesn’t add up.

Bitcoin’s security doesn’t come from being unbreakable. It comes from being too expensive to break. That’s why it’s still the gold standard.

What Should You Do?

If you’re a user:

  • Never trust a small PoW coin with fewer than 60 confirmations.
  • Wait 100+ confirmations for anything over $1,000.
  • Check the coin’s hash rate. If it’s below 10 TH/s, assume it’s vulnerable.

If you’re an exchange or business:

  • Require at least 60 confirmations for any altcoin under $500 million market cap.
  • Monitor hash rate changes. A sudden 20% drop could mean an attack is coming.
  • Don’t rely on “standard” confirmation numbers. They’re outdated.

If you’re building a blockchain:

  • Avoid Proof of Work unless you have massive, distributed mining.
  • Consider hybrid models - PoW with checkpoints or staking requirements.
  • Or just go PoS. It’s cheaper, faster, and immune to this exact attack.

Final Reality Check

A 51% attack isn’t science fiction. It’s a business decision. If the cost to rent hash power is less than the value you can steal - someone will do it. And they’re doing it right now.

Proof of Work was brilliant in 2009. But today, it’s a relic for the biggest networks only. For everything else, it’s a ticking time bomb.

The future isn’t about more mining power. It’s about making attacks too expensive to even try.

Can a 51% attack steal Bitcoin?

Technically yes, but practically no. Bitcoin’s hash rate is over 400 exahashes per second. Renting enough power to attack it would cost billions - far more than you could steal. Even if you succeeded, the community would fork the chain, invalidate your blocks, and you’d lose your investment. The economics don’t work.

How long does a 51% attack take?

It depends on the target. For a small coin like Bitcoin Gold, an attacker can launch a successful attack in under an hour. For Bitcoin, it would take years - if it’s even possible. Most attacks last between 30 minutes and 6 hours, long enough to reverse transactions and cash out before the network reacts.

Can you detect a 51% attack before it happens?

You can spot warning signs. A sudden drop in network hash rate, multiple orphaned blocks, or unusual spikes in mining activity on rental platforms like NiceHash can signal an attack is brewing. MIT’s monitoring system detects these in under an hour. But by the time you detect it, the attacker may already be halfway through.

Why don’t miners just stop the attack?

Miners don’t control the chain - the protocol does. The network automatically follows the longest chain, no matter who built it. Even honest miners will switch to the attacker’s chain if it’s longer. There’s no manual override. The system is designed to be trustless, which is also its weakness.

Are there any blockchains that are immune to 51% attacks?

Proof of Stake blockchains like Ethereum, Solana, and Cardano are immune to traditional 51% attacks. To control them, you’d need to own 51% of all staked coins - which would cost billions and destroy the value of your own holdings. Hybrid models with checkpointing (like Vertcoin) also add layers of protection against chain reorganizations.