Why run a node

Open your wallet and it shows a balance. Where did that number come from? Not from the bitcoin network in the abstract — networks do not answer questions. It came from a specific computer, owned by somebody, that your wallet asked. Unless you run a node, that somebody is a stranger, and their answer is what you are looking at.

This does not mean you are being lied to. It means the option of being lied to exists, and that you have no way to notice.

1What your wallet does without one

A wallet with no node of its own has to get its information somewhere. In practice that is a server run by the wallet's developer, a public Electrum server, or a block explorer's API. Your wallet tells it which addresses to watch, and the server reports back balances and history.

Some wallets do better and verify block headers, checking that the chain they are being shown carries real proof of work. That rules out an outright fabricated history but not much else — a header is a summary, and confirming that work was done is not the same as confirming the transactions underneath were valid.

Two distinct problems come out of this, and they are worth separating because people usually only think about the first.

2Problem one: you are trusting their answer

Nothing here lets a server steal from you. Your keys are yours, and a signature it did not produce is one it cannot forge. What it can do is tell you things that are not true.

• It can show a payment as confirmed when no such transaction exists, which matters if you hand over goods on that basis.
• It can hide transactions from you, so your displayed balance is simply wrong.
• It can feed you a stale view of the chain, or fail to relay a transaction you asked it to broadcast.
• If the consensus rules ever changed contentiously, it decides which chain's version of reality you are shown — and you would have no independent way to check.

None of this is common. All of it is possible, and the entire architecture of bitcoin exists to remove exactly this class of dependency. Keeping it at the last step is an odd place to stop.

3Problem two: you are telling them everything

This one is not hypothetical, not rare, and happens continuously.

To watch your coins, the server must know what to watch. Many wallets simply hand over the account's extended public key — the xpub — which lets the server derive every address you will ever use in that account. Others query address by address, which reveals the same information a little more slowly.

Either way the operator can see your complete balance, your full transaction history, and your future receipts as they arrive, correlated with the IP address you connected from. They did not need to break anything. You told them, and there is no way to take it back.

A wallet with no node of its own is a privacy leak with a nice interface. The keys stay safe; everything else is on display.

4What changes when the node is yours

A full node downloads every block and checks all of it for itself — signatures, whether each input exists and is unspent, the subsidy, the script rules, the proof of work. It builds its own picture of who owns what, from the raw data, trusting nothing.

• You verify instead of asking. Your balance is a conclusion your own machine reached, not a number a stranger sent you.
• You enforce the rules yourself. A block breaking the supply schedule or any other rule is rejected by your node regardless of how much work is behind it. This is the mechanism, not a metaphor.
• You stop leaking. Your wallet queries your own machine, so no third party learns your addresses, balances, or IP.
• You broadcast privately. Your transactions enter the network from your own node rather than through a service that knows they are yours.
• You are not asking permission. Nobody can rate-limit you, log you, or decline to serve you.

A small single-board computer sitting on a shelf beside a router, one LED lit, an unremarkable domestic corner with a power cable and an ethernet lead. Deliberately boring.

Image to come

5What it actually costs

Less than people expect, but not nothing, and the honest figures are worth knowing before you start.

• Disk. The full chain is around 750 gigabytes, growing by roughly 50 to 100 gigabytes a year while blocks stay full. Bitcoin Core's own guidance is to have at least 1TB free; a 2TB SSD means not thinking about it again. Check the current figure before buying — it only moves one way.
• Initial sync. Verifying the whole history from scratch takes anywhere from several hours to a couple of days depending on hardware and connection. It happens once.
• Bandwidth. Modest for your own use; higher if you let other nodes download blocks from you, which is the neighbourly default and can be limited.
• Hardware. A Raspberry Pi-class machine is enough. Any desktop from the last decade is more than enough.
• Attention. Occasional updates. It otherwise sits there and does its job.

If the disk requirement is the obstacle, a pruned node solves it. It verifies every block exactly as a full node does, then discards the old ones once it has finished checking them, keeping the total under about ten gigabytes. You get complete validation and complete rule enforcement; what you give up is the ability to serve history to other nodes, and rescanning an old wallet becomes awkward. For most people it is the right trade.

6Getting one running

There are three routes, in rising order of convenience and falling order of control.

• Bitcoin Core on a computer you already own. Free, official, and the reference implementation everything else is measured against. Install it, let it sync, point your wallet at it.
• A node distribution — Umbrel, Start9, myNode, RaspiBlitz and similar — which wrap Core in a friendly interface and bundle the extra services wallets want. Easier, at the cost of trusting the packaging.
• A prebuilt node appliance, bought ready to plug in. Fastest, most expensive, and worth checking what the vendor can see.

One detail catches people out: most wallets do not speak to Bitcoin Core directly for address lookups. They expect an index layer — Electrs, Fulcrum, or similar — sitting alongside it. The node distributions include this already; a manual Core install usually needs it added. Sparrow and Electrum both connect happily once it is there.

7Does everyone need one?

No, and pretending otherwise puts people off unnecessarily.

If you hold a modest amount, use a mobile wallet, and are comfortable with the privacy trade, you are making a reasonable decision. The security of your keys does not depend on running a node, and a hardware wallet without one is still enormously better than an exchange.

But the case strengthens quickly as the stakes rise:

• Holding an amount that would hurt to lose. At some point the balance you are shown deserves to be one you verified.
• Receiving payments regularly, especially in business, where being shown a false confirmation has direct consequences.
• Caring about privacy at all — this is the single largest improvement available to most people, ahead of anything more exotic.
• Running multisig or a serious setup. If you are careful enough to hold keys in three places, asking a stranger what they contain is an odd gap.
• Wanting the network to keep working like this. Nodes are what enforce the rules. A network where few people run them is one where fewer people are checking.

The short version

Your keys prove you own bitcoin. Your node tells you what you own, and whether the rules were followed. Without one, the second half is outsourced to somebody you have never met, who also learns your entire financial history in exchange for the service.

It is a cheap old computer, a large disk, and an afternoon of syncing — and it converts the most repeated slogan in bitcoin into something you are actually doing.