A blockchain is a shared digital ledger that records transactions across many computers at once, so no single party controls it and no past entry can be quietly changed. Think of it as a record book that thousands of people hold identical copies of, where anyone trying to fudge their copy is exposed by everyone else's.

Table of Contents

What a blockchain actually is

The US National Institute of Standards and Technology defines a blockchain as a tamper-evident, tamper-resistant digital ledger kept across a distributed network. At its simplest, then, it is a database with two unusual properties. It is distributed, meaning many independent computers each keep a full copy, and it is append-only, meaning you can add new records but cannot rewrite old ones without everyone noticing. Those two traits are what make it useful for recording ownership and value without a central authority such as a bank or a government registry.

Traditional databases live on servers owned by one organization. That organization can edit rows, reverse entries, or restrict who sees what. A public blockchain removes that single point of control. The rules for adding data are enforced by software running on every participant's machine, and the history is visible to anyone who wants to check it. This is why blockchains are often described as "trustless." Trust does not disappear, but you no longer have to trust a specific middleman, because you can rely on the network's rules and math instead.

It is worth being precise about what that does and does not buy you. A blockchain is slow, expensive, and redundant by design: every node repeats the same work so that no node has to be trusted. That is a terrible trade for most software. If one organization already controls the data and nobody disputes its right to do so, an ordinary database is faster, cheaper, and easier to fix. The technology only earns its cost when several parties need one shared record, have reason to distrust each other, and cannot agree on a neutral referee to hold it.

That test explains why the first durable use was money. Two strangers on opposite sides of the world have no shared bank, no shared court, and no reason to trust one another, yet both need to agree on who owns what. Before 2009 the only answer was an intermediary. A blockchain replaced the intermediary with a rulebook that every participant can check for themselves.

To understand where blockchain fits in the bigger picture, it helps to see it as the machinery underneath cryptocurrency as a whole and specific assets like Bitcoin. The coins are the visible product, and the blockchain is the machinery underneath.

Transactions are grouped into batches called blocks. Each block contains a list of recent transactions, a timestamp, and, crucially, a cryptographic fingerprint (a "hash") of the block that came before it. That backward reference is what turns a pile of blocks into a chain.

A hash is a short string of characters produced by running data through a one-way mathematical function. Change even a single character of the input and the output hash changes completely. Because every block embeds the previous block's hash, altering an old transaction would change that block's hash, which would break the reference in the next block, and the next, all the way to the present. To rewrite history, an attacker would have to redo every block after the one they tampered with, faster than the rest of the network is producing new ones. On a large network, that is prohibitively expensive, which is exactly the point.

Inside a block, the transactions themselves are also hashed together in a structure called a Merkle tree: transactions are hashed in pairs, then those hashes are hashed in pairs, and so on, until a single hash remains. That final value, the Merkle root, sits in the block header and acts as a fingerprint for every transaction in the block at once. It is a genuinely useful trick. It lets a lightweight device confirm that one specific transaction is included in a block without downloading the block, which is how a phone wallet can verify a payment without storing hundreds of gigabytes.

This chaining is why people call blockchain records "immutable." The data is not physically impossible to change, but changing it without detection is economically and computationally impractical. The distinction matters: immutability here is an economic property, not a law of physics. It holds because rewriting history costs more than anyone stands to gain from it, and it weakens on small networks where that cost is low.

How a transaction travels

The abstract description gets much clearer when you follow a single payment from start to finish. Say you send someone a small amount of crypto from your phone.

  • You sign it. Your wallet builds a message saying, in effect, "move this amount from my address to theirs," and signs it with your private key. The signature proves the instruction came from the owner of the funds without ever revealing the key itself. Nothing has left your device yet, and nothing has been spent.
  • It is broadcast. Your wallet sends the signed transaction to a nearby node, which passes it to the nodes it knows, which pass it on again. Within a few seconds it has spread across the network. This waiting area of valid-but-unconfirmed transactions is usually called the mempool.
  • It is checked. Every node that receives it verifies the signature, confirms the sending address actually holds the funds, and checks the transaction against the network rules. An invalid transaction is dropped on the spot and never reaches a block. No central authority makes this call; each node decides independently, and they reach the same answer because they run the same rules.
  • It is included in a block. Whoever is building the next block, a miner under proof of work or a validator under proof of stake, selects transactions from the mempool, typically favoring those offering the highest fee, and packages them into a candidate block.
  • The block is accepted. The new block is broadcast, other nodes verify it, and those that accept it build the next block on top. Your transaction is now part of the shared history.
  • It gets buried. Each subsequent block makes reversal exponentially more expensive. This is what a wallet means by "confirmations": not that the payment became more valid, but that undoing it became less affordable.

The whole sequence usually takes seconds to minutes depending on the network. Note what is absent: no clearing house, no business hours, no settlement window, and no one who can reverse the payment afterward on your behalf.

Nodes and consensus

The computers that store and validate the blockchain are called nodes. When someone submits a transaction, it is broadcast to the network, checked against the rules, and eventually bundled into a block. But if thousands of independent nodes are all maintaining the ledger, how do they agree on which block comes next? That agreement process is called consensus.

Not all nodes do the same job. A full node downloads every block and independently verifies every transaction against the rules, which is what lets it refuse anything invalid no matter who sent it. An archive node goes further and keeps the complete historical state, which is mostly useful to explorers and analytics tools. A light client stores only block headers and asks other nodes for the specific data it needs, verifying inclusion with those Merkle proofs, which is what most phone wallets do. Miners and validators are full nodes that additionally compete for the right to propose blocks.

The number of independent full nodes is one of the more honest measures of how decentralized a network really is, because each one is a party who can reject an invalid rule change. A chain with thousands of nodes across many countries and operators is far harder to capture than one whose nodes are mostly rented from the same cloud provider.

The two best-known consensus mechanisms are:

Proof of Work (PoW). Nodes called miners compete to solve a hard mathematical puzzle. The first to solve it earns the right to add the next block and receives a reward. Solving the puzzle takes real electricity and hardware, which is what makes cheating expensive. Bitcoin uses this model, and that puzzle-solving race is what people mean by mining.

Proof of Stake (PoS). Instead of burning electricity, validators lock up (or "stake") the network's own currency as collateral. The protocol selects validators to propose and confirm blocks, and anyone caught behaving dishonestly can lose their stake. Ethereum switched to this model in 2022. The two approaches are weighed against each other in our guide to how the consensus models compare.

Both approaches solve the same problem: getting strangers who do not trust each other to agree on a single shared history without a referee. They also share a failure mode. If one party ever controls a majority of the network's mining power or staked value, they can, within limits, reorder or block recent transactions. This is the well-known 51% attack. It cannot mint coins from nothing or spend funds the attacker does not have, because those actions break rules every full node checks independently. What it can do is reverse the attacker's own recent payments, which is enough to defraud an exchange. Large networks defend against this with sheer cost; small ones have genuinely been attacked this way.

This is also why finality differs between designs. Under proof of work, a block is never mathematically final, only progressively more expensive to reverse, which is why exchanges wait for a number of confirmations before crediting a deposit. Many proof-of-stake networks add explicit finality instead: after a certain point, reversing a block would require destroying an enormous amount of staked collateral, so it is treated as settled. The practical difference for a user is simply how long you wait before a payment is safe to rely on.

Public vs. private blockchains

Not every blockchain is open to the world.

Public blockchains (Bitcoin, Ethereum, Solana) are permissionless. Anyone can run a node, read the ledger, and submit transactions. They are transparent and censorship-resistant, though slower and more resource-intensive because everyone must agree.

Private (or permissioned) blockchains are controlled by a company or consortium that decides who may participate. A supply-chain network shared among a few manufacturers is a common example. These are faster and more private, but they give up much of the decentralization that makes public chains distinctive.

Consortium blockchains sit between the two. Rather than one company holding the keys, a group of them, say a handful of banks settling payments between each other, jointly operate the network and share governance. No single member can rewrite history alone, but the general public cannot join either. In practice this is where most serious enterprise deployments land, because it matches how those industries already work: a known set of participants who trust each other only partly.

The honest summary is that permission and decentralization trade against each other. The more you restrict who can participate, the faster and more private the network gets, and the more it comes to resemble a shared database with extra steps. That is not automatically wrong. It is only misleading when a permissioned system borrows the credibility of a public one.

Most of the crypto assets an everyday investor encounters live on public blockchains. Private chains are mainly an enterprise tool, and the difference between the two is worth understanding properly if you ever read a corporate blockchain announcement; we cover it in public vs. private blockchains.

What blockchains are used for

The original use case was digital money, but the technology now supports a lot more:

  • Cryptocurrencies, recording who owns which coins, as with Bitcoin and other cryptocurrencies.
  • Smart contracts, self-executing agreements that run automatically when conditions are met, powering decentralized finance and much of the Ethereum ecosystem.
  • Tokens, digital representations of assets, memberships, or rights, which come in several distinct types.
  • Supply chain and provenance, tracking goods from origin to shelf.
  • Digital identity and record-keeping, tamper-evident logs for credentials or land titles.

The common thread is any situation where multiple parties need a shared, trustworthy record but do not fully trust one another to maintain it.

The inverse is just as useful to know. Blockchains are a poor fit for storing large files, for anything requiring personal data to be deleted on request, for problems where one trusted party already exists and is not disputed, and for any process that needs mistakes reversed. Supply-chain tracking is the instructive case: the ledger can prove a record was not altered after someone wrote it, but it cannot prove the person scanning the crate told the truth. Where a claim enters the system from the physical world, the blockchain secures the record, not the reality behind it.

Risks and limitations

Blockchain is powerful but not magic, and it comes with real trade-offs:

  • Scalability. Requiring every node to process every transaction limits speed. Solutions like layer-2 networks help, though the base layer is still a bottleneck.
  • Energy use. Proof-of-work chains consume significant electricity, which raises environmental questions.
  • Immutability cuts both ways. If you send funds to the wrong address or fall for a scam, there is usually no "undo."
  • "Garbage in" still applies. A blockchain guarantees a record was not altered after it was written, but it cannot guarantee the original information was true.
  • User error and complexity. Managing keys and wallets safely is a genuine hurdle, and a mistake can be irreversible.
  • Regulatory uncertainty. The record may be permanent, but its legal treatment is not settled, and it varies by country and keeps moving.

Several of these pull against each other in a way that is hard to escape. Developers often describe it as the scalability trilemma: a network can push hard on decentralization, security, or scale, but improving one usually costs another. Want more throughput at the base layer? Raise the hardware requirements for running a node, and fewer people run one, which concentrates the network. Want maximum decentralization? Keep nodes cheap to run, and accept a hard ceiling on transactions per second. Most credible scaling work does not dissolve this tension so much as move it somewhere more tolerable, which is the actual purpose of layer-2 networks.

Common misconceptions

"Blockchain and Bitcoin are the same thing." Bitcoin is one blockchain, and the first one that worked at scale. Blockchain is the general structure it introduced. Thousands of networks now use that structure for different purposes, and most look nothing like Bitcoin.

"Transactions are anonymous." On public chains they are the opposite: pseudonymous and permanently visible. Every payment an address has ever made sits in an open ledger anyone can read forever. Nothing links an address to your name until something does, and once one transaction is tied to you, the trail runs in both directions.

"Data on a blockchain must be true." It only proves nobody altered the record after it was written. Whether the claim was accurate when someone entered it is entirely outside the system.

"Nothing on a blockchain can ever change." The rules themselves can change if enough participants agree, and networks have deliberately reversed or rewritten history after major incidents. What resists change is any one party changing it unilaterally.

"It is decentralized because it is a blockchain." Decentralization is a spectrum, not a property you get for free with the label. A chain run by five nodes belonging to one company is a blockchain, and it is centralized.

How to get started

You do not need to run a node or understand cryptography to benefit from blockchain, since most people interact with it simply by holding crypto assets. A sensible on-ramp looks like this: learn what the assets actually are, understand where they live, and then take a small, deliberate first step. Curiosity about the technology is often what makes the rest click.