Consensus Mechanisms (PoW/PoS & More)

A consensus mechanism answers a single question: who may write the next entry, and why should the others accept it? The three common answers do not differ in what they achieve, but in the currency they are paid in — electricity, capital or contract.

What the mechanism protects against

Digital money has a basic problem that cash does not: a file can be copied. Without an entity keeping the balance, the same amount could be spent twice. In a blockchain or any other distributed ledger, the order of events therefore has to become binding — and binding in a way that cannot be cheaply rewritten afterwards.

Every mechanism solves this through a majority assumption. The BIS states the condition for open networks plainly: the system works as long as honest nodes control a majority of computing power. What changes between mechanisms is what that majority is measured in — and how expensive it is to buy.

Proof of work: security paid for in electricity

Under proof of work, participants compete to be first to solve a computationally expensive task. Whoever solves it may write the next block. An attack on the past would have to redo all of that computational work — so the ledger's security is tied directly to a physical expenditure.

That is the strength of the mechanism and equally its invoice: security arises through energy consumption, not as a side effect but as a design principle. Anyone who wants to know how high that consumption actually is will find a continuously updated estimate for the Bitcoin network at the Cambridge Centre for Alternative Finance — useful when somebody in-house needs a defensible figure rather than a newspaper headline.

Proof of stake: security posted as capital

Under proof of stake, nothing is computed; something is posted. Whoever wants to propose and attest blocks locks up their own assets in the network. Behave against the rules, and part of that is taken away. Security no longer depends on electricity costs but on the capital that is lost through misbehaviour.

What that changes on the energy bill can be read from a switch that actually happened. The Ethereum network moved from proof of work to proof of stake in September 2022. The Ethereum Foundation puts electricity consumption before the switch at around 21 terawatt hours per year and afterwards at roughly 0.0026 — a reduction of more than 99.9 per cent. The reasoning is the mechanism itself: it secures the network with posted assets instead of computational work.

The price sits elsewhere. Those who can post more carry more weight, and the question of how power is distributed moves from the electricity bill to the balance sheet.

Permissioned mechanisms: security resting on contracts

In a permissioned network the circle of writers is known. Then neither computational competition nor posted assets are needed: a defined majority of named participants votes, and an entry stands once that majority carries it. Such mechanisms are fast, frugal, and give a hard statement about the moment of finality — it is reached when the majority has signed, not probably after a few blocks.

In exchange, security rests neither on physics nor on capital, but on contracts and selection: who is in the circle, who may extend it, what happens after misbehaviour. That is not a disadvantage, but a different kind of risk — legal and organisational rather than computational.

What a company notices

A supplier making around sixty cross-border payments a month does not face this as a question of principle, but when setting up the payment route — and notices it in three places. First in time: how long until a payment may count as final, and does that fit the release of goods? Second in cost: in open networks the network fee moves with demand, which is more awkward for calculating small amounts than the absolute level. Third in the ability to answer: if the bank, the auditor or your own sustainability reporting asks about the energy profile of the network in use, the answer is a property of the consensus mechanism — and should not be researched only at that point.

The real risk is less technical. Choosing a network binds you: addresses, approvals, posting logic and counterparties all align to it, and switching later is not a configuration step but a project. Which is why the question about the mechanism belongs at the start of a selection process rather than in detailed technical planning — even though it rarely comes up by itself in conversations with providers.

What follows from this

No mechanism is inherently superior. Each shifts the same question — who prevails in case of doubt — onto a different foundation: energy, capital or a contract. Which foundation fits is decided by the use case, and by which kind of assurance you want to trust in a dispute.

How a single payment runs through such a network is covered in How Does a Blockchain Work?; whether a shared ledger is the right tool at all, and whether it should be open or permissioned, is the subject of Blockchain & DLT – Fundamentals.

For your own selection, then, the interesting point is not which mechanism a network uses. It is which assurance about finality, cost and energy profile can be derived from it — and whether the provider will repeat that assurance in writing.