Proof of stake vs proof of work comes down to what a network makes participants put at risk. Proof of work (PoW) makes miners spend electricity and hardware racing to solve a hash puzzle, and the chain with the most accumulated work wins. Proof of stake (PoS) makes validators lock up the network's own coin as collateral that can be destroyed if they cheat. Bitcoin still runs on proof of work, with a new block roughly every 10 minutes, per the Bitcoin FAQ. Ethereum switched from proof of work to proof of stake on September 15, 2022, per Ethereum's Merge page.
​

This page compares the two mechanisms on five points: how blocks get produced, what an attack costs, how finality works, how much energy each uses, and who can participate. Background reading: what a blockchain is and what a layer 1 blockchain does.
​

Proof of work secures a blockchain with computation: miners burn energy guessing values until one finds a valid block, and honest chains outpace attackers because they hold more computing power. Proof of stake secures it with capital: validators deposit coins, get chosen to propose and vote on blocks, and lose part or all of that deposit for provable misbehavior.
​

Both are consensus mechanisms, the rules that let thousands of independent computers agree on one transaction history without a central operator. The Bitcoin whitepaper describes PoW as hashing transactions "into an ongoing chain of hash-based proof-of-work," where the longest chain is proof that it "came from the largest pool of CPU power," per the Bitcoin whitepaper. Ethereum's documentation defines PoS as a way to prove that validators "have put something of value into the network that can be destroyed if they act dishonestly," per Ethereum's proof-of-stake docs.
​

The practical split: PoW turns electricity into security, and PoS turns locked-up coins into security. Everything else in this comparison (energy, finality, hardware, attack cost) follows from that one design choice.

Proof of work secures a chain by making block production expensive and verification cheap. Miners repeatedly hash block data with a changing value called a nonce until the result falls below a difficulty target. Any node can check the winning answer instantly, and rewriting history means redoing all that work faster than the honest network.
​

On Bitcoin, the network raises or lowers difficulty so the average time to find a block stays at 10 minutes as miners join or leave, per the Bitcoin FAQ. The whitepaper frames this as a moving average "targeting an average number of blocks per hour," per the Bitcoin whitepaper. Block rewards also issue new coins: Bitcoin issuance halves over time until it stops at 21 million bitcoins, per the Bitcoin FAQ.
​

Ethereum ran its own PoW algorithm, Ethash, until 2022. Miners pushed a dataset that required the full chain through a mathematical function to find a mixHash under the difficulty target, and after the Constantinople upgrade a winning miner earned 2 newly minted ETH plus part of the fees, with ommer blocks paid 1.75 ETH, per Ethereum's proof-of-work docs.
​

Mining stopped being a home-computer activity long ago. In Bitcoin's early days anyone could find a block with a CPU, but rising difficulty means specialized hardware is now the only cost-effective option, per the Bitcoin FAQ.
​

The core security argument is a race. An attacker who wants to replace recent blocks has to out-produce every honest miner combined, which is why the whitepaper puts its guarantee this way:
​

"As long as a majority of CPU power is controlled by nodes that are not cooperating to attack the network, they'll generate the longest chain and outpace attackers." (Bitcoin whitepaper)
​

Ethereum's proof of stake replaces miners with validators who lock 32 ETH in a deposit contract. Time is cut into 12-second slots, a randomly selected validator proposes each block, and committees of other validators vote on it. Honest work earns ETH rewards; dishonest votes or double proposals get part of the stake destroyed.
​

To run a validator, a user deposits 32 ETH and operates three pieces of software: an execution client, a consensus client and a validator client, per Ethereum's proof-of-stake docs. New validators wait in an activation queue. Once active, they re-execute the transactions in each proposed block and send a signed vote called an attestation.
​

The schedule is fixed rather than set by mining difficulty. Slots last 12 seconds, an epoch is 32 slots, one validator is randomly picked to propose in every slot, and committees split the validator set so every active validator attests once per epoch, per the same Ethereum's proof-of-stake docs. When validators see competing heads of the chain, the LMD-GHOST fork-choice rule picks the fork with the greatest weight of attestations.
​

Penalties do the work that electricity does in PoW. Proposing two blocks in one slot or submitting contradictory attestations triggers slashing: an immediate penalty of 1/4096 of the validator's effective balance (up to 0.5 ETH), a correlation penalty on day 18 that scales with how many validators were slashed at the same time, and ejection on day 36, per Ethereum's proof-of-stake docs. A lone slashed validator can lose less than 0.1% of its stake, while a mass slashing event can destroy 100%.
​

Both resist attack by making it expensive, but the cost lands differently. A PoW attacker must rent or buy more than half the network's hash power and pay for the energy. A PoS attacker must acquire more than half the staked coin, and the protocol can then destroy that stake. Ethereum's documentation argues PoS is the more economically secure design.
​

Under PoW, a malicious miner needs over 51% of mining power to keep producing valid but malicious blocks, and the energy spent might outweigh the gains, per Ethereum's proof-of-work docs. Under PoS, the 51% threat still exists, since an attacker with 51% of staked ETH could make their fork canonical. The difference is the response: honest validators can keep building on the minority chain and the community can remove the attacker and destroy their staked ETH, per Ethereum's proof-of-stake docs.
​

Reverting a block that Ethereum has already finalized is harder still. An attacker would have to commit to losing at least one-third of all staked ETH, and if a one-third minority stalls finality, an inactivity leak starts after four epochs and drains stake from validators voting against the majority, per Ethereum's proof-of-stake docs.
​

The counterpoint comes from the same documentation. Ethereum's own pros-and-cons table lists PoS as "younger and less battle-tested" and more complex to implement than PoW, per Ethereum's proof-of-stake docs. The companion page credits PoW with keeping Bitcoin and Ethereum secure and decentralized for many years, per Ethereum's proof-of-work docs.
​

Proof of work offers probabilistic finality: a transaction becomes harder to reverse as more blocks stack on top, but never strictly impossible. Ethereum's proof of stake offers explicit finality: once two-thirds of staked ETH votes for a pair of checkpoints, the earlier one is finalized and reversing it would burn a large share of the stake.
​

Under PoW, two miners can find valid blocks at the same moment, creating a temporary fork that resolves when one branch grows longer. Transactions on the losing branch can be reversed, so finality "refers to the time you should wait before considering a transaction irreversible," per Ethereum's proof-of-work docs. Bitcoin block discovery is itself probabilistic, with no guaranteed minimum or maximum delay around the 10-minute average, per the Bitcoin FAQ.
​

On PoS Ethereum, the first block of each epoch is a checkpoint. When a pair of checkpoints attracts votes representing at least two-thirds of total staked ETH, the newer one becomes justified and the older one finalized, a process run by Casper the Friendly Finality Gadget, per Ethereum's proof-of-stake docs. For anyone moving funds, that difference sets how long to wait before treating a deposit as settled. A longer treatment of the topic sits in settlement finality across crypto networks.
​

Proof of work uses orders of magnitude more electricity, because security comes directly from energy spent hashing. Proof of stake needs only enough power to run ordinary servers. Ethereum's switch from one model to the other is the clearest before-and-after measurement for one network, and it shows electricity use falling to a tiny fraction of the old level.
​

The Crypto Carbon Ratings Institute (CCRI) estimates The Merge cut Ethereum's annualized electricity consumption by more than 99.988%, per Ethereum's energy report. The current PoS Ethereum estimate is about 2,601 MWh (0.0026 TWh) per year across the global network, corresponding to about 870 tonnes of CO2e, from a bottom-up study by CCRI cited on Ethereum's energy report. The same page reports that Ethereum's carbon footprint fell by about 99.992%, from 11,016,000 to 870 tonnes CO2e. Shortly before the switch, Ethereum miners were consuming about 70 TWh per year according to Digiconomist on July 18, 2022, per Ethereum's proof-of-work docs.
​

For Bitcoin, ethereum.org's comparison table lists 149 TWh per year, sourced to the Cambridge Bitcoin Electricity Consumption Index and accessed in July 2023, per Ethereum's energy report. The live figure moves daily: the Cambridge index publishes a theoretical lower bound, an estimate and an upper bound, updated every 24 hours with a 7-day moving average. ethereum.org also notes that energy use does not map directly to environmental footprint, because the mix of energy sources differs by network.
​

One common shortcut is misleading. Per-transaction energy figures imply fewer transactions would use less energy, but the energy to propose and validate a block is independent of how many transactions it contains, per Ethereum's energy report.
​

Proof of work is open to anyone with hardware and electricity, and needs no coins to start. Proof of stake needs the coin itself: 32 ETH to run an independent Ethereum validator, or a staking pool for smaller amounts. In return, PoS validator hardware is modest enough for a normal laptop, while competitive mining needs specialized machines.
​

ethereum.org lists PoW's neutrality as a strength: a miner can start from 0 ETH and earn a positive balance from block rewards. Its listed weaknesses are high energy use, a large up-front investment in specialized equipment, and the risk that mining pools come to dominate, per Ethereum's proof-of-work docs.
​

For PoS, the same documentation says a validator node can run on a normal laptop, staking pools let users participate without 32 ETH, and economies of scale do not apply in the way they do for mining. The trade-off is operational: validators run three pieces of software and must stay online, or they miss rewards, per Ethereum's proof-of-stake docs. Validators also earn priority-fee tips from users, which ties into the ordering game described in what MEV is.
​

Neither mechanism is free of trade-offs. Proof of work's weak points are energy consumption, hardware costs and pressure toward large mining pools. Proof of stake's weak points are a shorter track record, more protocol complexity, and a requirement to own the network's coin before taking part in consensus.

For most users the mechanism matters through three practical effects: how long to wait before a payment is final, what staking options exist, and the network's energy profile. Bitcoin holders live with PoW's probabilistic settlement and 10-minute block average. Ethereum users get 12-second slots, checkpoint finality and the option to stake ETH.
​

Choose a PoW chain if the priority is the longest operating record and a security model that needs no coin ownership to join. Choose a PoS chain if lower energy use, explicit finality and a staking yield matter more. Developers deploying smart contracts on Ethereum or its layer 2 networks inherit PoS settlement from Ethereum mainnet. Related comparisons in this series: hot wallet vs cold wallet and web2 vs web3.
​

No. Bitcoin uses proof of work, with miners using specialized hardware and network difficulty adjusting to keep the average block time at 10 minutes, per the Bitcoin FAQ.
​

Ethereum switched from proof of work to proof of stake in September 2022, in an upgrade called The Merge, per Ethereum's Merge page.
​

A solo validator deposits 32 ETH into the deposit contract. Staking pools let users participate with less, per Ethereum's proof-of-stake docs.
​

Yes. Proposing two blocks in one slot or signing contradictory attestations gets a validator slashed, with losses from under 0.1% of stake for an isolated case up to 100% in a mass slashing event, per Ethereum's proof-of-stake docs.
​

Ethereum's documentation argues the opposite: an attacker needs 51% of staked ETH and risks having it destroyed, and reverting a finalized block costs at least one-third of all staked ETH, per Ethereum's proof-of-stake docs.
​

Mechanism details, penalties and pros and cons come from ethereum.org's developer documentation. Bitcoin figures come from bitcoin.org and the Bitcoin whitepaper. Energy figures come from ethereum.org's energy page, which cites CCRI for PoS Ethereum and the Cambridge index for Bitcoin; the Cambridge live dashboard updates daily, so the 149 TWh figure is dated to its July 2023 access. All pages were read on October 13, 2026.