Proof of stake

Proof of stake (PoS) is the consensus mechanism Ethereum and most new blockchains use to agree on the order of transactions, replacing the proof-of-work model Bitcoin still uses.

The basic idea: validators lock up cryptocurrency as a stake (currently 32 ETH per Ethereum validator). The protocol selects validators pseudo-randomly weighted by stake to propose and attest to new blocks. Validators earn fees and protocol rewards for honest participation. Validators who attempt to cheat (double-signing blocks, going offline for extended periods) lose part or all of their stake — “slashing.”

The security argument: an attacker who wants to corrupt the chain must accumulate enough stake to dominate validator selection (roughly 1/3 for liveness attacks, 2/3 for safety attacks). Acquiring that much stake is enormously expensive on a high-cap chain, and the moment the attack starts, the value of the stake itself collapses — making the attacker punish themselves.

Compared to proof-of-work, PoS:

• Uses dramatically less energy. Ethereum’s energy use dropped ~99.95% after switching from PoW to PoS in September 2022.
• Has lower hardware barriers. A validator is a normal computer; PoW needs ASIC farms.
• Concentrates differently. PoW concentrates in cheap-electricity regions. PoS concentrates in whales who hold large stake (or staking pools aggregating it).
• Allows in-protocol penalties. Slashing is a feature only PoS can really implement; PoW just lets cheaters fail to mine valid blocks.

Major PoS chains as of 2026: Ethereum (post-merge), Cardano, Solana, Polkadot, Cosmos, Tezos, Avalanche, plus essentially every new L1 launched after 2022. Bitcoin remains as the canonical PoW holdout.

Liquid staking and the centralisation worry: Ethereum’s 32-ETH threshold puts solo validation out of reach for most holders, so “liquid staking” protocols (Lido, Rocket Pool, EigenLayer) emerged to pool smaller deposits into validators while issuing tradeable receipts (stETH, rETH). The convenience is real — users earn staking yield without locking up 32 ETH or running a node — but the result is concentration: as of 2025, Lido alone controlled ~30% of all staked ETH. This is the single most-debated risk in modern PoS — if any single staking provider crosses the 33% threshold needed to halt block finality, the chain loses its security guarantee against that operator. Ethereum’s long-term roadmap (single slot finality, distributed validator technology) aims to lower the barrier and reverse the concentration.

The “nothing-at-stake” objection — once a theoretical problem, now solved: early PoS proposals were criticised for not penalising validators who signed contradictory blocks on competing forks — without “skin in the game”, validators might rationally sign every fork to maximise reward chances. Modern PoS designs solve this through slashing: signing two blocks at the same height is a slashable offence that destroys most or all of the validator’s stake. The penalty makes the “sign everything” strategy unprofitable in expectation, restoring the convergence properties PoW gets for free from energy costs. See how cryptocurrency mining actually works for the PoW comparison. Reference: Ethereum Foundation — The Merge.