What Is a Bitcoin Node, and What Is One For?

What a node is

Every rule in Bitcoin is written as code. How many new bitcoin a block is allowed to pay out, which signatures count as valid, whether a coin has already been spent: each of these is a condition the software can check. A node is a program that holds that rulebook and runs every incoming block and transaction past it.

If a node spots a rule being broken, it throws the block out, no matter who sent it or how much mining power stands behind it. That independent check is the whole point. Relaying data and answering other nodes are useful extras, not the core job.

What a node does on the network

The Bitcoin whitepaper sets out how the network runs in a handful of steps: new transactions go out to every node, each node gathers them up, checked blocks are passed around, and a node only accepts a block if every transaction in it is valid and unspent.

On top of that, the bitcoin.org guide to running a full node notes that most nodes also serve lightweight wallets: they take those wallets’ transactions and let them know when a payment touches their balance. If too few nodes offered this, light wallets would be pushed onto centralised services instead.

So the network is a large set of independent checkers that owe each other nothing. There is no vote on whether a block is good. Each full node judges every block on its own, and among the chains that pass its checks it follows the one with the most mining work behind it. A block that breaks the rules is rejected by a full node regardless of how many others accept it.

Full node versus light wallet

Many everyday wallets are light wallets, even if nothing tells you so. The Bitcoin whitepaper calls this simplified payment verification. A wallet like that does not download the whole blockchain, only the block headers plus a short proof that a given transaction sits inside a particular block. A full node, by contrast, validates the entire chain itself. Run in pruned mode, it does not keep all the old block data on disk afterwards, and it is still a full node.

• A light wallet cannot judge a transaction’s validity for itself. All it sees is that a node put the transaction in a block, and that more blocks have piled on top.
• That holds up as long as most nodes are honest. If an attacker out-muscles the network, simplified verification is easier to fool than the full kind.
• A full node does not carry that dependency. It checks every rule itself and notices at once if someone hands it an invalid block.

The Bitcoin whitepaper says plainly that a business taking payments often will probably still want to run its own node, for security that does not depend on anyone else and for faster checks.

Node does not always mean the same thing

The word node turns up in very different settings. What actually gets checked, and what you are trusting, varies a lot between them.

The distinction matters because most claims about “the number of Bitcoin nodes” mean only the full nodes that verify independently. They are the ones checking, day in and day out, that the rules are being kept.

What running your own node costs and gives you

You do not need your own node to use Bitcoin safely. It is an extra step for people who want to check the rules themselves, not just hold their own keys. It sits on the same idea as self-custody: relying on third parties as little as possible.

The bitcoin.org guide to running a full node lists these rough figures for a Bitcoin Core full node, as read on 17 September 2026: an ordinary computer; at least 7 gigabytes of free disk space with pruning turned on, or several hundred gigabytes without pruning for the full chain; a one-off download of around 740 gigabytes on the first run; 2 gigabytes of memory; a broadband line with at least 400 kilobits of upload; and about six hours a day of uptime. For traffic it calls 200 gigabytes of upload or more a month common, plus around 20 gigabytes of download.