This post is part of a series on how everyday investors can understand the word "consensus" in the context of blockchain networks.

Comparing blockchains to the internet is one of the best ways to make them easier to understand. The internet is a distributed set of protocols for transferring information. No one company owns it, and it doesn't run on a single computer. Blockchains are similar. They are decentralized protocols for owning, verifying and moving assets without relying on any single institution. In both systems, shared rules allow computers around the world to coordinate with one another.

The computers that make this coordination possible on a blockchain are called nodes. A blockchain node is a computer running software that connects to a blockchain network and follows its rules. That software downloads network data, checks whether transactions and blocks are valid, shares information with other participants, and helps keep the ledger synchronized across the world. In simple terms, a node is one of the machines helping the blockchain stay alive and agree on what is true.

This is where nodes connect directly to consensus. Consensus is the process by which a blockchain network reaches agreement on the state of the ledger. Nodes are the participants carrying out that process. They receive new transactions, verify them against the rules, relay valid data to one another, and update their local view of the chain. Without nodes, there is no network to agree, no shared ledger to maintain, and no way for consensus to happen in practice.

This still leaves the question, how do thousands of independent nodes decide who gets to propose the next block? Also, how does the network confirm that proposal is legitimate? That method is called a consensus mechanism, and one of the most widely used today is proof of stake.

Proof of stake is consensus with a security deposit

Every consensus mechanism answers the same question: how do we coordinate thousands of independent nodes so they agree on one history? In proof of stake, the answer is capital. Validators lock up funds that can be taken away if they break the rules. That locked capital is the security deposit that earns them the right to participate in consensus. That's the core mental model. You don't need to prove you did expensive work. You need to prove you have something valuable to lose.

Who are the "validators," really?

In proof of stake, the nodes doing consensus work are called validators. A validator is a node operator running validator software plus a required stake locked in the protocol. On Ethereum, running your own validator requires depositing 32 ETH, though you can participate with less via pooled options. Validators have two main jobs. They propose blocks when the protocol selects them, and they vote to confirm that blocks they see are valid. It's not that only one validator checks the work. Everyone checks. The system just assigns who gets to propose.

How proof of stake actually runs (Ethereum as the example)

Ethereum is the simplest mainstream example of proof of stake. In Ethereum's proof of stake, time is split into slots and epochs. A slot is 12 seconds, and an epoch is 32 slots. In each slot, one validator is randomly selected to propose a block. In the same slot, a committee of validators is chosen to vote on the block's validity. That's the rhythm. One proposer says, "Here is the next block." A committee responds, "Yes, that block follows the rules" (or doesn't). And those votes don't just help pick the head of the chain in the moment. They also build toward something proof of work doesn't give you explicitly.

Slashing is the what keeps validators honest

Proof of stake works because it makes honesty the best business decision. Validators get rewarded for doing their jobs, proposing new blocks and checking new blocks. If they go offline, they miss rewards and can incur smaller penalties. But if they do things that signal malicious intent, like signing conflicting blocks or conflicting votes, they can be slashed.

Ethereum outlines three core slashing categories: proposing two different blocks for the same slot, "double voting" (confirming conflicting candidates at the same time) and "surround voting" (a vote pattern that contradicts prior checkpoint votes). Slashing destroys part of the validator's stake and ejects them from the validator set, with penalties that can scale up if many validators are slashed together. So the system doesn't just hope validators behave. It makes misbehavior expensive in a way that's enforced automatically by the protocol.

Benefits of proof of stake

Proof of stake is energy efficient by design. Validators don't need specialized hardware or massive electricity budgets. They can run on modest machines with minimal power draw. When Ethereum switched from proof of work to proof of stake ("The Merge"), its own roadmap page estimated the transition reduced energy consumption by roughly 99.95%. Proof of stake can also punish bad actors directly. Because a validator's influence comes from locked capital rather than external hardware, the protocol can destroy that capital if the validator misbehaves. Ethereum's proof of stake FAQ makes this point clearly: dishonest validators can be slashed and ejected, costing them substantial ETH. The penalty isn't just lost revenue. It's lost principal.

That direct accountability feeds into one of proof of stake's strongest properties: finality with economic teeth. Once a checkpoint is finalized through supermajority voting, reversing that history would require an attacker to trigger slashable offenses that destroy at least one-third of all staked ETH. Investors often talk about "security" like it's abstract. Proof of stake makes it concrete. Rewriting finalized history is technically difficult and a business mistake. The game theory of blockchains comes into play.

Drawbacks of proof of stake

Proof of stake isn't "free security." It's a different set of tradeoffs.

Proof of stake protocols are complex. Ethereum's own comparison page notes that proof of stake introduces additional attack vectors and failure modes, including the fact that having a pre-selected proposer can create denial-of-service targets. Complex systems can be robust, but complexity always raises the bar for implementation and coordination.

There is also a natural centralizing pressure in proof of stake. Rewards scale with stake. That's not necessarily unfair (everyone gets roughly the same percentage), but it does mean large pools can become influential. Ethereum's FAQ specifically calls out the growing influence of liquid staking derivatives (LSDs) as a current decentralization concern. This is one of the most important "real-world" proof of stake conversations today: not whether the protocol allows anyone to validate (it does), but whether economic and product realities pull stake into a few dominant venues.

Finally, proof of stake makes transaction ordering more complex. The moment you have a predictable block production schedule and a known proposer, the incentives around ordering and inclusion become problematic. There is ongoing protocol work to separate builders from proposers and improve censorship resistance, noting upcoming changes that introduce constraints around transaction inclusion (who gets to decide what gets added to the chain). Proof of stake affects the whole blockchain network design.

Conclusion

Proof of stake is consensus built around a simple idea: put up a security deposit, do the job, get rewarded, cheat and lose the deposit. Cryptography makes verification objective. Distributed systems keep thousands of nodes synchronized. Game theory makes honesty the rational strategy, not because people are good, but because the protocol makes dishonesty unprofitable.

The internet gave us open protocols for information. Proof of stake is part of what makes blockchains feel like open protocols for value, systems that can run globally, continuously and publicly, without needing a single institution to keep the ledger honest.