Imagine a town that keeps its property records not in a single town hall vault but on photocopies held by every resident. To forge a deed, you would have to break into every house simultaneously. That is, more or less, how a blockchain works — and the "photocopies" are called nodes.
In Plain English
A node is any computer that runs a blockchain's software and stores a copy of its transaction history. There is no head office, no central server, and no permission required to join. Right now, roughly 21,000–25,000 Bitcoin nodes are spread across dozens of countries, each holding an identical copy of every transaction ever recorded. They talk to each other directly, check each other's work, and collectively decide what counts as valid. No single one of them is in charge.
What a Node Actually Is
Nodes come in a few varieties, and understanding the difference matters.
A full node downloads, validates, and stores every block and transaction ever recorded on the chain. It checks every rule independently — no shortcuts, no trust placed in anyone else. When a new block arrives, the full node verifies it from scratch before passing it along.
A pruned node does the same validation but discards older transaction data once it has been verified, keeping only recent blocks. This saves disk space while still protecting the network.
A light node, sometimes called an SPV (Simplified Payment Verification) node, stores only block headers — a compact summary of each block — and relies on full nodes for the actual transaction data. Mobile wallets commonly use this approach to avoid downloading hundreds of gigabytes.
Anyone can run a full node. There is no application form, no fee, and no minimum stake. You download the software, connect to the internet, and you are part of the network.
How Nodes Talk to Each Other
Blockchain networks use a peer-to-peer (P2P) architecture: nodes connect directly to each other, with no central server routing messages. The result is a mesh that has no single throat to choke.
When a new node starts up for the first time, it queries hardcoded DNS seeds — special domain names that return IP addresses of active full nodes. From those first contacts, the new node builds its own persistent peer database and keeps expanding it over time.
Once connected, nodes exchange short "inventory" messages announcing new blocks and transactions, then request the full data from peers that have it. A new transaction spreads node-to-node across the entire network in seconds via a process called a gossip protocol — the same way rumours spread through a crowd. If a peer goes offline or starts misbehaving, the network simply routes around it.
The Job Every Node Does: Enforce the Rules
Mining gets most of the attention in blockchain coverage, but miners only produce blocks — full nodes are the ones that decide whether those blocks are accepted.
Every full node independently checks every transaction and block against the network's consensus rules before relaying it. These rules include: no coins created beyond the scheduled block reward, no double-spend (spending the same output twice), and a valid cryptographic signature on every spend. If a block violates any rule, the node rejects it. The block stops spreading and vanishes from the network.
Crucially, these rules are encoded in open-source software that anyone can read. No company decides what is valid. There is no appeals process — a bad block is simply dead.
Miners propose; nodes dispose. The economic incentive for miners is to produce blocks that the nodes will accept, because rejected blocks earn no reward.
Why Thousands of Copies Make the Network Resilient
Because every full node holds an independent, identical copy of the ledger, there is no single server to shut down, hack, or subpoena. An attacker wanting to rewrite history would need to disable all reachable nodes simultaneously — a practically impossible task.
As of early 2026, Bitnodes reported between roughly 21,000 and 25,000 reachable Bitcoin full nodes. Approximately 59–60% of those are accessible over Tor (the anonymous network, via .onion addresses), meaning that even aggressive state-level interference with the public internet would leave thousands of nodes operating.
Nodes also exist in different legal jurisdictions. Even if one country bans running a node, nodes elsewhere continue operating unaffected. The ledger's immutability is similarly protected: rewriting even one historical block would require redoing all subsequent proof-of-work AND convincing the majority of nodes to accept the rewritten chain — both at the same time.
What Goes Wrong Without a Distributed Node Network
A blockchain where only a handful of entities run nodes is functionally centralised. Those operators can collude to censor transactions or alter the rules without meaningful resistance.
History offers a concrete example. In 2017, a coalition of large miners backed a proposal called SegWit2x that would have changed Bitcoin's rules in ways most users opposed. The proposal failed not because of the miners but because the user-operated node network refused to enforce those new rules. Miners quickly found that blocks following the new rules were rejected by the nodes users were actually running.
Contrast that with a centralised database — a bank, a payment processor, a government registry. Efficient, yes. But a single regulatory action, company failure, or security breach can freeze or erase balances. The trade-off of decentralisation is speed and complexity: thousands of nodes reaching consensus is slower than a single database write. That trade-off is worth making only when trust-minimisation genuinely matters — which is precisely the use case blockchains are designed for.
Running Your Own Node
Running a node is the strongest possible relationship you can have with a blockchain. Instead of trusting a third party's copy of the ledger, you verify your own transactions independently.
Bitcoin Core, the reference implementation, requires roughly 650+ GB of disk space for the full chain as of 2026. A pruned installation reduces that to a few gigabytes. Hardware requirements are modest — a Raspberry Pi or any always-on home computer handles the load comfortably.
Running a node also improves privacy. Wallet apps that connect to a third-party server reveal which addresses you are watching. Connect that same wallet to your own node, and the visibility disappears: you query your own copy of the ledger.
Just as securing your private key is the foundation of on-chain self-custody, running your own node is the foundation of self-verification. And just as a hardware wallet removes the need to trust a software intermediary with your keys, a personal node removes the need to trust an intermediary with your view of the chain.
Key Takeaways
- A full node independently validates every transaction and block — it does not trust miners or other nodes, it checks for itself.
- The P2P gossip protocol and DNS-seed bootstrapping mean the network is self-organising and has no single point of failure.
- As of early 2026, roughly 21,000–25,000 reachable Bitcoin nodes spread across dozens of countries and ~60% accessible over Tor make a coordinated shutdown essentially impossible.
- Miners produce blocks; nodes enforce whether those blocks are accepted — the real power rests with whoever is running validation software.
- Running your own node is the deepest form of participating in a blockchain network: you verify rather than trust.
