Staking
Oct 9, 2026

Proof of Stake vs Proof of Work: Key Differences Explained

Proof of stake vs proof of work compared: how each secures a blockchain, energy use, 51% attacks vs slashing, finality, and how Sui, MultiversX and Radix do it.

Proof of Stake vs Proof of Work: Key Differences Explained

Proof of work (PoW) secures a blockchain by making block producers spend computing power and electricity; proof of stake (PoS) secures it by making them lock up the network's own token, which can be destroyed if they cheat. Both solve the same problem: getting thousands of computers that don't trust each other to agree on one transaction history. The difference is what an attacker has to put at risk. Under PoW it is hardware and energy; under PoS it is capital held inside the system.

Bitcoin is the best-known proof-of-work chain. Ethereum switched from PoW to PoS in September 2022, and newer networks such as Sui, MultiversX and Radix were built on proof of stake from the start. This guide compares the two on mechanics, energy use, security, decentralisation and finality, with examples from each.

What a consensus mechanism does

A blockchain has no central server deciding which transactions count. Every node keeps its own copy of the ledger, so the network needs a rule for three things: who gets to add the next block, how everyone checks it, and what happens if two versions of history compete.

The hard part is stopping a "Sybil" attacker who spins up thousands of fake identities to outvote honest nodes. Both PoW and PoS answer that by tying influence to something scarce and costly. One vote per unit of computing power, or one vote per unit of staked token, rather than one vote per computer.

How proof of work works

In proof of work, block producers called miners race to solve a puzzle. They repeatedly hash a candidate block with a changing number (a nonce) until the result falls below a target set by the network's difficulty. Finding a valid hash takes enormous trial and error; checking one takes a fraction of a second. The Bitcoin whitepaper, published in 2008, set out this design using SHA-256 hashing.

The winning miner adds the block and earns newly issued coins plus fees. If two miners find blocks at once, the chain temporarily forks, and nodes follow the chain with the most accumulated work. That is why PoW finality is probabilistic: each extra block on top makes a transaction harder to reverse, but no block is ever formally final. As ethereum.org puts it, under PoW "the more blocks were mined on top of a specific block," the more confident you could be it would not be reverted.

The scale of this race is huge. Bitcoin's network hashrate was about 1,000 exahashes per second, or roughly a billion trillion hashes every second, on 9 October 2026, according to mempool.space.

How proof of stake works

In proof of stake, block producers called validators lock up tokens as collateral. The protocol picks validators to propose and vote on blocks, usually weighted by stake, and pays them rewards for doing it correctly. If they break the rules, for example by signing two conflicting blocks, part of their stake can be slashed (destroyed). We cover penalties in detail in what is slashing in crypto.

Ethereum is a useful reference. According to ethereum.org, a solo validator deposits 32 ETH and runs three pieces of software. Time is split into 12-second slots and 32-slot epochs. One validator proposes a block each slot, committees attest to it, and a block becomes final once validators holding two-thirds of the stake vote for it at checkpoints.

Most PoS chains let ordinary holders take part without running a node. In delegated proof of stake, you assign your tokens to a validator, share its rewards and keep ownership of your coins. See what is crypto staking for the basics.

Proof of stake vs proof of work at a glance

Proof of workProof of stake
Who produces blocksMinersValidators
Scarce resourceComputing power and electricityStaked tokens
SelectionWhoever solves the hash puzzle firstChosen by protocol, weighted by stake
Penalty for cheatingWasted energy and lost rewardsSlashing of staked tokens
FinalityProbabilistic (wait for confirmations)Usually explicit, once two-thirds of stake agrees
Energy useVery highLow
HardwareSpecialised mining machinesOrdinary servers
ExamplesBitcoinEthereum, Sui, MultiversX, Radix

Energy use

This is the most visible difference. PoW is designed to be expensive: security comes from the cost of the electricity miners burn. PoS replaces that cost with locked capital, so validators can run on normal hardware.

Ethereum gives a clean before-and-after comparison because it ran both. According to ethereum.org, The Merge on 15 September 2022 cut Ethereum's energy consumption by about 99.95%. Shortly before the switch, Ethereum miners were using roughly 70 TWh a year, "about the same as the Czech Republic," while post-Merge Ethereum is estimated at around 0.0026 TWh a year, per its energy consumption page.

Bitcoin's PoW defenders argue the energy is the point: it makes rewriting history physically costly, and miners increasingly use stranded or surplus power. Whatever your view, the gap in energy use between the two designs is not in dispute.

Security and decentralisation

51% attacks vs slashing

Both systems can be attacked by whoever controls most of the scarce resource. The question is what that costs and what happens afterwards.

Under proof of work, an attacker with more than 50% of the hashrate can outpace honest miners, reorder recent blocks and double-spend. They need to buy or rent the hardware and pay for the power, but if the attack succeeds, the hardware still works afterwards and can be used again.

Under proof of stake, the attacker needs a large share of the staked token, which they would usually have to buy on the open market, pushing up the price as they go. Ethereum's documentation explains that to revert a finalised block, an attacker "would commit to losing at least one-third of the total supply of staked ETH," because the conflicting votes are provable and slashable. An attacker with a third of the stake could stall finality, but an "inactivity leak" gradually drains the stake of validators who refuse to vote with the majority until the chain can finalise again.

PoS has its own weak points. Ethereum's attack and defence page describes "long-range" attacks that use old validator keys to build an alternative history, which is why PoS chains rely on finality checkpoints. And because stake can be concentrated in a few large staking providers, PoS security depends on how widely that stake is spread.

Decentralisation trade-offs

Neither model is automatically more decentralised. Each concentrates power in a different way.

  • PoW rewards cheap electricity and economies of scale in specialised hardware. In practice, block production flows through a handful of mining pools: over the month to 9 October 2026, the three largest pools found about 61% of Bitcoin blocks, per mempool.space. Individual miners can switch pools, though.
  • PoS rewards whoever holds the most tokens, and stake can cluster around exchanges and liquid staking providers. It also has a low barrier to entry for small holders, since anyone can delegate or use a liquid staking token without buying hardware.

The useful questions for any chain are concrete ones. How many independent validators or pools would need to collude to halt or rewrite it? Can an ordinary user afford to run a node to verify the chain themselves? How quickly do the rules change, and who decides?

Examples: Bitcoin, Ethereum, Sui, MultiversX and Radix

Bitcoin (PoW). The original design: SHA-256 mining, probabilistic finality and no staking.

Ethereum (PoS since 2022). Moved from PoW to PoS at The Merge. Validators stake 32 ETH, finality comes from two-thirds checkpoint votes, and misbehaviour is slashed, from under 0.1% of stake for an isolated offence up to 100% in a mass slashing event, per ethereum.org.

Sui (delegated PoS). According to Sui's documentation, Sui uses delegated proof of stake to choose validators and set their voting power, and a DAG-based consensus protocol called Mysticeti, which lets several validators propose blocks in parallel. A quorum needs more than two-thirds of voting power, and the validator set and stakes stay fixed for each epoch of about 24 hours. Validators can be penalised through slashing of stake and rewards for misbehaviour.

MultiversX (Secure Proof of Stake). MultiversX's consensus docs describe Secure Proof of Stake (SPoS): validators stake at least 2,500 EGLD, are reshuffled across shards every 24-hour epoch using verifiable randomness, and sign blocks with aggregated BLS signatures. A block needs signatures from at least two-thirds of the shard's validators and is final as soon as that proof is broadcast.

Radix (delegated PoS). On Radix, XRD holders delegate stake to validators, and the top 100 validators by stake form the active set, according to the Radix docs. The current Babylon network uses HotStuff BFT with deterministic finality and epochs of about five minutes. Radix's sharded Cerberus consensus is planned for its future Xi'an upgrade rather than live today.

Where JewelSwap fits

Proof of stake is what makes liquid staking possible. On the three PoS networks JewelSwap supports, you can mint a liquid staking token that stays usable in DeFi while the underlying coins are delegated to validators: JWLSUI on Sui, JWLEGLD on MultiversX and JWLXRD on Radix. Each uses a dual-token model, in which you stake the base token to receive an S-token (SJWLSUI, SJWLEGLD, SJWLXRD) that rises in value against it as validator rewards arrive. According to JewelSwap's docs, redeeming back to the native coin takes an unbonding period of about 10 days, during which you hold a transferable claim NFT. The usual staking risks still apply, including validator performance and smart contract risk.

Frequently asked questions

What is the main difference between proof of stake and proof of work?

Proof of work selects block producers by having miners compete with computing power and electricity. Proof of stake selects validators according to tokens they lock up as collateral, which can be slashed if they cheat. Both aim to make attacking the network more expensive than following the rules.

Is proof of stake more secure than proof of work?

Neither is strictly more secure; they protect against different things. Proof of work makes rewriting history cost real-world energy. Proof of stake can destroy an attacker's stake and offers explicit finality, but depends on stake not being too concentrated. Each network's design and distribution matter more than the label.

How much energy did Ethereum save by moving to proof of stake?

According to ethereum.org, The Merge on 15 September 2022 reduced Ethereum's energy consumption by about 99.95%, from roughly 70 TWh a year under proof of work to an estimated 0.0026 TWh a year under proof of stake.

Is Bitcoin proof of work or proof of stake?

Bitcoin uses proof of work. Miners compete to find SHA-256 hashes below a difficulty target, and Bitcoin has no staking at the protocol level. Its hashrate was about 1,000 exahashes per second on 9 October 2026, according to mempool.space.

Are Sui, MultiversX and Radix proof of stake?

Yes. Sui uses delegated proof of stake with the Mysticeti consensus protocol, MultiversX uses Secure Proof of Stake with a 2,500 EGLD validator minimum, and Radix uses delegated proof of stake with HotStuff BFT on its current Babylon network.

What is a 51% attack?

A 51% attack is when one party controls a majority of a network's mining power or stake and uses it to reorder recent blocks, censor transactions or double-spend. On proof-of-stake chains such as Ethereum, attempts to revert finalised blocks can be detected and the attacker's stake slashed.

Keep reading

This article is educational and isn't financial advice. Ethereum figures are from ethereum.org, Bitcoin hashrate from mempool.space as of 9 October 2026, and Sui, MultiversX and Radix consensus details from each network's official documentation as checked on the same date. JewelSwap issues the JWLSUI, JWLEGLD and JWLXRD liquid staking tokens mentioned in this article.

About the author.

Co-Founder at JewelSwap & Chief Strategy Officer at iDenfy. Viktor brings his successful track record of superb development & project management.