A Bitcoin node is a computer that runs Bitcoin software, connects to other computers doing the same, and independently verifies every transaction and block against the network's rules. Together, these nodes form the Bitcoin network itself. There is no head office and no master server. Just tens of thousands of machines around the world, each holding the network to the same standard.
Every balance you have ever checked in a wallet app was reported to you by somebody's node. If it wasn't yours, you were trusting theirs. That single fact explains why nodes matter, why people run them at home on hardware as small as a Raspberry Pi, and why "don't trust, verify" became Bitcoin's unofficial motto.
This guide covers what a Bitcoin node is, how nodes work, the different types, how they differ from miners, and exactly what it takes to run one yourself in 2026.
Manage your Bitcoin securely with the self-custody Bitcoin.com Wallet app.
Key Takeaways
- A Bitcoin node is a computer running Bitcoin software that validates transactions and blocks and relays them across the peer-to-peer network.
- Full nodes independently enforce Bitcoin's consensus rules, including the 21 million coin supply cap. Miners propose new blocks; nodes decide whether those blocks are valid.
- Running a node earns you no bitcoin. The payoff is verification, privacy, and independence from third parties.
- As of September 2026, more than 20,000 nodes are publicly reachable, and the true total, including nodes behind firewalls and Tor, is estimated to be several times higher.
- You can run a full node on an old laptop, a Raspberry Pi, or a plug-and-play device. The blockchain is roughly 700 GB as of September 2026, but a pruned node can validate everything with as little as a few gigabytes of storage.
What Is a Bitcoin Node?
A Bitcoin node is any computer that runs Bitcoin's protocol software and participates in the network by storing, validating, and sharing blockchain data. If Bitcoin is a shared ledger, a node is one of the many independent bookkeepers who each keep their own copy and refuse to write down anything that breaks the rules.
Compare that with a bank. A bank keeps one ledger, on its own servers, and you have to trust that it is accurate. Bitcoin flips the model: thousands of nodes each hold the ledger, check every entry themselves, and reject anything invalid. No single machine is in charge, which is what makes the network censorship resistant and gives it no single point of failure.
The concept dates back to the beginning. Satoshi Nakamoto's 2008 white paper describes a network of nodes that timestamp transactions and enforce rules by consensus, and the first node came online when the network launched in January 2009. In the white paper, Satoshi used "node" loosely to describe miners. Today the term usually refers to a validating node, and when people say they "run a node," they almost always mean a full node: one that checks every transaction and block for itself rather than taking anyone's word for it.
How Does a Bitcoin Node Work?
A node's job sounds complicated but boils down to a repeating cycle: listen, check, and pass along. Here is what happens under the hood.
1. Connect to Peers
On startup, a node discovers other nodes and opens connections to them, typically over port 8333. This mesh of peer-to-peer connections, with no central coordinator, is the Bitcoin network. Information spreads through it like gossip: each node tells its neighbors, who tell theirs.
2. Validate Incoming Transactions
When someone sends bitcoin, their wallet broadcasts the transaction to nearby nodes. Each node checks it against Bitcoin's consensus rules before doing anything else. Among other things, it verifies that:
- The digital signatures are valid, proving the sender controls the coins
- The coins being spent actually exist and have not already been spent (no double-spending)
- The transaction follows all format and script rules of the protocol
3. Hold Valid Transactions in the Mempool
Transactions that pass inspection wait in the node's memory pool, or mempool, which is the queue of unconfirmed transactions. Miners draw from this pool when building new blocks, and wallets use mempool data to estimate appropriate fees.
4. Validate New Blocks
When a miner finds a new block, nodes put it through an even stricter inspection. Every transaction inside is re-checked, the proof-of-work is verified, and the block reward is confirmed to be no larger than the protocol allows. This last check is how nodes collectively enforce Bitcoin's 21 million coin limit. A block that breaks any rule is simply rejected, no matter how much computing power produced it.
5. Relay and Repeat
Valid transactions and blocks are passed along to peers, and the cycle continues. Because every node performs the same checks independently and reaches the same conclusions, the network stays in agreement about a single valid blockchain without anyone being in charge.
A brand new node has to catch up first. It downloads and verifies the entire blockchain from the very first block in 2009 to the present, a process called the initial block download (IBD). On a modern computer with an SSD and decent internet, this takes somewhere between six and 24 hours. On a Raspberry Pi, expect several days.
Types of Bitcoin Nodes
Not all nodes do the same job. The main distinction is how much data they store and how much they verify for themselves.
| Node type | What it stores | Typical storage (September 2026) | Verifies independently? | Best for |
|---|---|---|---|---|
| Archival full node | The entire blockchain since 2009 | ~700 GB and growing | Yes, everything | Serving data to the network, block explorers, developers |
| Pruned full node | Recent blocks plus the current state of coin ownership (the UTXO set) | As little as ~10 GB in practice | Yes, everything | Users with limited storage who still want full verification |
| Light node (SPV) | Block headers only | Under 1 GB | No, relies on full nodes | Mobile wallets, low-power devices |
| Mining node | Varies (usually full blockchain data) | ~700 GB+ | Yes | Building and proposing new blocks |
| Lightning node | Bitcoin blockchain plus payment channel data | ~700 GB+ (runs alongside a full node) | Yes | Fast, cheap payments on Bitcoin's Lightning Network |
Two of these deserve a closer look.
Pruned nodes are the underrated entry point - A pruned node downloads and verifies the whole chain during initial sync, then discards old block data it no longer needs, keeping storage to a user-defined cap. Crucially, it enforces every consensus rule exactly like an archival node. If a 700 GB download is what has been putting you off, pruning removes that objection almost entirely. The trade-off is that a pruned node cannot serve historical blocks to other nodes that are syncing.
Light nodes trade independence for convenience - Using a method called simplified payment verification (SPV), they download only block headers and ask full nodes to confirm that transactions were included in blocks. Most mobile wallets work this way. It is fast and lightweight, but it reintroduces trust: you are relying on someone else's full node to tell you the truth.
Bitcoin Nodes vs. Miners: What's the Difference?
The terms get mixed up constantly, but nodes and miners play different roles. The cleanest way to remember it: miners propose, nodes dispose.
| Bitcoin node | Bitcoin miner | |
|---|---|---|
| Main job | Verify transactions and blocks against the rules | Assemble transactions into new blocks and compete to add them |
| How it works | Runs validation software, checks everything it receives | Expends energy on proof-of-work computations |
| Earns rewards? | No | Yes, block subsidy plus transaction fees |
| Hardware | An ordinary computer | Specialized ASIC machines |
| Power over the rules | Enforces them; rejects invalid blocks | None; even a valid proof-of-work block is worthless if nodes reject it |
Miners spend enormous amounts of energy to win the right to add the next block. But mining power buys no authority over Bitcoin's rules. If a miner produced a block that paid itself too large a reward or included an invalid transaction, every full node on the network would reject it automatically, and the miner would have burned all that electricity for nothing. Nodes are the referees; miners are players trying to make a valid move.
This division of power is a core part of Bitcoin's security model. It means the rules are enforced by tens of thousands of cheap, widely distributed computers rather than by whoever owns the most expensive hardware.
Why Run a Bitcoin Node? (And Do You Earn Money?)
Let's answer the most searched question first: no, running a regular Bitcoin node does not earn you bitcoin. There is no reward, no yield, and no payout for validating. Miners get paid for producing blocks; node operators get paid in something less tangible but arguably more valuable. Here is what you actually get:
- Verification without trust: Your node checks your incoming payments and your balance against the real blockchain. You never have to take an exchange's or wallet provider's word for it, and you can personally verify that the 21 million supply cap is intact.
- Privacy: When your wallet uses someone else's node, that node's operator can potentially link your addresses, balances, and transactions to your IP address. Your own node keeps that information at home, and routing it over Tor hides your node's location too.
- Independence: No third party can feed your wallet false information, censor your transaction broadcasts, or go offline and take your access with it.
- A stronger network: Every independent node makes Bitcoin more decentralized and more resistant to attacks, bad data, and rule changes pushed without broad agreement.
- Better infrastructure for your own tools: A personal node can serve as the backend for your wallet, a Lightning node, or self-hosted payment tools.
One honest caveat: if you run a Lightning node on top of your Bitcoin node and route payments through well-funded channels, you can earn small routing fees. For most hobbyists these amount to pocket change, not income. The reason to run a node is sovereignty, not profit.
How to Run a Bitcoin Node
Running a node in 2026 is easier than its reputation suggests. There are three well-trodden paths, depending on how hands-on you want to be.
| Path | Hardware cost | Difficulty | Best for |
|---|---|---|---|
| Existing computer | Free (uses what you have) | Easy to moderate | Trying it out with zero investment |
| Raspberry Pi or mini PC | ~$150-300 including SSD | Moderate | A dedicated, always-on node that sips 5-15 watts of power |
| Plug-and-play device (Umbrel Home, Start9 server) | ~$400-900 | Easiest | Non-technical users who want one-click apps and Lightning |
Whichever path you choose, the baseline requirements are the same:
- Storage: A 1 TB SSD is the practical minimum for an archival node, and 2 TB gives you years of headroom. The blockchain is roughly 700 GB as of September 2026 and grows by around 50 GB per year. An SSD matters far more than a fast CPU; syncing on a mechanical hard drive is painfully slow. A pruned node needs only a fraction of this.
- Memory: 2 GB of RAM is the floor; 4-8 GB makes syncing much faster.
- Internet: A stable, ideally unmetered connection. A node can move hundreds of gigabytes per month if it serves data to peers, though upload limits can be configured.
The Setup, Step by Step
- Download the software: Most node operators run Bitcoin Core, the protocol's reference implementation, available for Linux, Windows, and macOS from bitcoincore.org. As of September 2026, the latest release is version 31.1.
- Verify your download: The project publishes cryptographic checksums and signatures for every release. Checking them takes five minutes and protects you from tampered software, which is a real and documented attack vector against Bitcoin users.
- Install and start the initial block download: Your node will fetch and verify the entire chain, block by block, from January 2009 onward. Budget six to 24 hours on a modern machine with an SSD, or several days on a Raspberry Pi. This is a one-time cost.
- Configure to taste: In the configuration file you can enable pruning to cap storage, set bandwidth limits, and route traffic through Tor for privacy. The official documentation at bitcoin.org walks through each option, including how to allow inbound connections so your node contributes data back to the network.
- Connect your wallet: The final step that makes it all worthwhile: point your own wallet at your own node, so every balance check and every broadcast goes through infrastructure you control.
If steps 1 through 4 sound like too much, the plug-and-play route with Umbrel or Start9 reduces the whole process to connecting a device to your router and clicking install. You give up a little configurability and pay a premium for the hardware, but the node underneath is just as real.
What It Costs to Keep Running
Almost nothing. A node is not a mining rig. A Raspberry Pi setup draws about as much power as an LED bulb, working out to roughly $5-15 per year in electricity. The real ongoing costs are bandwidth and the occasional software update.
How Many Bitcoin Nodes Are There?
Trick question: nobody knows exactly, and the commonly quoted numbers only count part of the network.
Trackers such as Bitnodes crawl the network for reachable nodes, meaning nodes that accept incoming connections on a public address. Throughout 2025 and into 2026, that figure has generally sat above 20,000, at times exceeding 23,000.
But reachable nodes are the tip of the iceberg. The majority of nodes run behind home routers, firewalls, or Tor, where crawlers cannot see them. These unreachable (non-listening) nodes validate everything for their owners just the same; they simply don't advertise themselves. Estimates of the hidden population commonly exceed 100,000, with some researchers putting the total network several times higher than the reachable count. As of September 2026, "at least tens of thousands, likely well over a hundred thousand" is the honest answer.
The distinction matters for how you read decentralization statistics. A dip in the reachable count does not necessarily mean people stopped running nodes; it can simply mean more of them went private.
Bitcoin Core, Bitcoin Knots, and Why Node Software Choice Matters
For most of Bitcoin's history, node software was a non-topic: nearly everyone ran Bitcoin Core. That changed in 2025, and it is now one of the most interesting things happening at the node layer.
Bitcoin Core remains the reference implementation, maintained by a large open-source contributor base. But an alternative implementation called Bitcoin Knots grew from a rounding error to a meaningful share of reachable nodes during 2025 and 2026. The disagreement driving that growth is about relay policy: which transactions a node is willing to pass along. Knots ships with stricter default filters against transactions that embed non-financial data (images, text, and other arbitrary content) in the blockchain, while Core's maintainers moved toward relaying whatever miners will realistically include anyway, arguing that filtering at the node level is ineffective and distorts fee signals. By late 2025, public trackers showed Knots running on roughly a fifth of reachable nodes, and its share has remained a closely watched statistic since.
Two things are worth understanding about this debate, whichever side you find persuasive:
- Relay policy is not consensus - Both Core and Knots nodes enforce identical consensus rules and follow the same blockchain. This is a disagreement about what to gossip, not about what is valid.
- Implementation diversity cuts both ways - Multiple implementations reduce the risk that a single software bug or a single development team's decisions define the network. On the other hand, coordination across implementations is harder. Either way, the fact that node operators can vote with their software choice, and visibly did, is Bitcoin's governance model working exactly as designed: the rules are whatever the nodes collectively run.
For a newcomer, the practical takeaway is simple. Start with Bitcoin Core, and know that alternatives exist and that your choice of software is itself a small vote on the network's direction.
Trade-Offs and What to Watch
Running a node is genuinely low-cost, but it is not zero-cost, and it is worth being clear-eyed about that. The initial sync takes patience. The blockchain's growth of roughly 50 GB per year means storage planning is a real, if minor, chore for archival nodes. Bandwidth usage can be substantial if you serve data to peers. And a node quietly demands the same discipline as any home server: occasional updates, occasional troubleshooting.
Against that, the barrier to entry keeps falling. Storage prices continue to drop faster than the chain grows, pruning removes the storage objection entirely, and plug-and-play devices have turned a weekend project into an evening one.
Three things worth tracking from here:
- The reachable node count and its software mix, visible live on Bitnodes and similar trackers, as a rough pulse of network health and the ongoing Core vs. Knots story
- Chain growth versus hardware costs, the quiet race that determines how accessible full validation stays for ordinary users
- Node-in-a-box adoption, since every step that makes self-hosting easier tends to move users from trusted third parties to self-verification
Conclusion
A Bitcoin node is a computer that independently verifies every transaction and block against Bitcoin's rules, and the tens of thousands of them running worldwide are what make the network decentralized, censorship resistant, and trustworthy without requiring trust. Nodes, not miners, are the ultimate enforcers of Bitcoin's rules, including the 21 million coin cap. In 2026, running one is within reach of anyone with a spare computer or a couple hundred dollars, and the reward is the ability to check the money for yourself. That idea, verification over trust, is the whole point of Bitcoin, and a node is how you practice it.
