What is blockchain and how does it work?

A blockchain is a global, transparent, distributed digital record of transactions that is stored across thousands of computers around the world, rather than on one company’s server. When someone submits a new transaction, the network’s decentralized computers, called nodes, check it and must collectively agree it’s valid before it’s added to the record, also known as the ledger. The process of collective agreement is called consensus. Once a transaction is recorded, traditional blockchain transactions can’t be changed or deleted without redoing the work of every record that came after it, which is why blockchain data is considered immutable, or tamper-proof. Because no single company or government controls the record, blockchain is often called “trustless” infrastructure: participants rely on math and network verification rather than any one institution.

Disclaimer: This guide is for educational purposes only. It is not financial advice, not a solicitation, and not for UK audiences. Blockchain-based digital assets are risky and not suitable for all users.

How a blockchain network differs from a traditional centralized digital database

A conventional, centralized digital database (the kind powering your bank’s app, your social media feed, or a government records system) lives on servers controlled by one organization. An administrator can add, modify, or delete records at will. You’re trusting that organization to maintain the data accurately and protect it from tampering.

A blockchain flips this model. Instead of one entity controlling the record, identical copies of the entire transaction history live across thousands of independent computers, called nodes. No single node can rewrite past records on its own. Changes require network-wide agreement, and any attempt to alter historical data would immediately conflict with the copies held by every other node.

This architecture is called a distributed ledger. “Distributed” means the ledger exists simultaneously on many machines in many locations, not in one place. “Ledger” reflects its core purpose: recording transactions in sequence, like an accounting book, except this one is shared, public, and append-only.

Historically, the trade-off was speed. Traditional databases process thousands of transactions per second because they run from centralized infrastructure, and early blockchain networks were significantly slower. That gap is closing. Layer 2 networks, parallel execution, and newer chain architectures now handle thousands of transactions per second with sub-second finality. Today, the primary distinction is architectural: blockchain networks prioritize decentralization, transparency, and tamper resistance by design, which is why they now underpin everything from global payments to decentralized finance to digital identity.

Blocks: how data gets organized

The name says it all: information is grouped into units called blocks, and each block links to the one before it, forming a chronological chain. Each block typically contains 3 things.

First, a batch of transaction data, meaning records of value transferred between addresses on the network. On Ethereum, for example, a single block can contain hundreds of individual transactions, each documenting a transfer of ETH or an interaction with a smart contract (a self-executing program stored on the blockchain).

Second, a timestamp recording when the block was created.

Third, and this is the part that makes the whole system secure, a cryptographic hash. A hash is a fixed-length string of characters produced by running data through a mathematical function. Even a tiny change to the input produces a completely different output. Each block contains the hash of the previous block, which creates a chain of mathematical dependencies stretching back to the very first block. If someone tried to alter a transaction in block 1,000, the hash of that block would change, breaking the link to block 1,001 and every block after it. Every node on the network would see the tampering instantly.

This chained hashing structure is what gives blockchain its immutability. Retroactive data manipulation becomes computationally impractical because an attacker would need to recalculate every subsequent hash faster than the rest of the network produces new ones.

Nodes: the computers running the blockchain network

A node is any computer that runs the blockchain’s software and maintains a copy of the ledger. Nodes validate new transactions against the network’s rules, relay valid transactions to other nodes, and store the complete transaction history.

More nodes means more resilience. If some go offline or try to broadcast invalid data, the remaining honest nodes keep operating normally. As of early 2026, Ethereum operates ~14,000 execution nodes across more than 70 countries and nearly 900,000 active validators, according to Etherscan and beaconcha.in. No government or company can shut the network down by targeting a single data center.

Here’s what makes this different from most technology you use daily: anyone with the right hardware and software can run a node on most public blockchains. No one needs to approve your application. This open participation is central to decentralization. Nobody decides who gets to join or who gets to see the data.

Consensus: how the network agrees on what’s true

When a new batch of transactions is proposed as the next block, the network needs a way to agree that those transactions are valid and that the block belongs on the chain. That mechanism is called a consensus algorithm, and it solves a genuinely hard problem: how do thousands of independent computers, some of which might be unreliable or even malicious, reach the same conclusion about what’s legitimate?

Two consensus mechanisms dominate public blockchains today.

Proof of Work (PoW) requires participants called miners to spend computational energy solving a complex mathematical puzzle. The first miner to find a valid solution earns the right to add the next block and collects a reward in the network’s native cryptocurrency. The puzzle’s difficulty adjusts automatically to keep block production steady. Bitcoin uses Proof of Work, producing a new block every ~10 minutes. The energy expenditure makes attacking the network prohibitively expensive, because you’d need to control more computational power than every other miner combined.

Proof of Stake (PoS) works differently. Instead of competing with raw computing power, participants called validators lock up (or “stake”) a quantity of the network’s native cryptocurrency as collateral. The protocol selects validators to propose and confirm new blocks, and validators who act dishonestly, for example by trying to approve fraudulent transactions, risk losing their staked collateral through a penalty called slashing. Ethereum has operated on PoS since its September 2022 Merge, which reduced the network’s energy consumption by ~99.95% from its previous PoW consensus model. More broadly, PoS is now the dominant consensus mechanism across the industry.

Both PoW and PoS reach the same outcome: network agreement without a central coordinator. They differ in resource requirements, security assumptions, and environmental footprint.

Immutability: why you can trust a blockchain  ledger

Once a transaction is confirmed and added to the blockchain, the record of that transaction can’t be altered or reversed without the consent of the network. This property, immutability, comes from the combination of cryptographic hashing, distributed storage, and consensus working together.

An important distinction: the transaction record itself is immutable, but smart contracts can be designed with upgradeable logic or mutable state that authorized parties can modify through new transactions. The history showing that a change was made remains permanent and auditable, even when the contract’s current state has been updated. Immutability applies to the ledger’s record of what happened, not necessarily to the behavior of every program running on it.

Think about what it would actually take to change a historical transaction. An attacker would need to recalculate the cryptographic hash of the altered block, then recalculate the hash of every block after it, and then convince a majority of the network’s validators to accept the rewritten chain, all before the honest network produces new blocks and moves further ahead. On a network like Bitcoin or Ethereum, that’s economically and computationally infeasible.

Immutability isn’t absolute in a theoretical sense. It’s a practical guarantee. A blockchain with very few nodes or minimal staked value could be more vulnerable. The security of immutability scales with the size, distribution, and economic commitment of the network’s participants.

This property is what makes blockchain valuable for applications where historical record integrity matters: financial transactions, property registries, supply chain documentation, and digital identity systems. Once data is confirmed on a blockchain, it becomes a shared source of truth that no single party can retroactively edit.

Decentralization: why no single entity is in charge

Decentralization describes how control and decision-making are spread across a network’s participants rather than concentrated in one place.

In a fully decentralized blockchain, no company, government, or individual controls which transactions get processed, which blocks get added, or what rules the network follows. Protocol changes typically require broad community agreement: developers propose changes, node operators choose whether to adopt them, and validators enforce the agreed-upon rules.

The practical benefits are significant. Censorship resistance means no single authority can block a valid transaction. Fault tolerance means the network keeps running even if a large percentage of nodes go offline. Permissionless access means anyone can send transactions, deploy smart contracts, or use applications built on the network without needing a gatekeeper’s approval.

This is where blockchain stops being abstract and starts becoming something you interact with directly. For example, when you sign a transaction with a self-custodial wallet like MetaMask, it’s broadcast straight to the network’s nodes for validation. No intermediary bank or payment processor sits in between to approve or deny it. MetaMask’s built-in transaction security alerts flag potentially risky interactions before you sign, but the final decision is always yours.

Decentralization exists on a spectrum. Some networks are more decentralized than others, depending on the number and geographic spread of validators, the concentration of staked capital, and the governance process for upgrades. The term isn’t binary. Evaluating any network’s decentralization means examining these variables rather than taking it at face value.

Public vs private blockchains

Not all blockchains work the same way. The distinction between public and private reflects a core design choice about who gets to participate.

A public blockchain is open to everyone. No registration, identity check, or permission is needed to run a node, submit transactions, or read the ledger. Bitcoin and Ethereum are public blockchains. Their transaction histories are fully transparent, and anyone can inspect any transaction using a block explorer like Etherscan. Public blockchains maximize decentralization and censorship resistance, which makes them the right fit when trust between participants can’t be assumed.

A private blockchain restricts participation to a known group. An organization or consortium controls who joins, who validates transactions, and who reads the data. Private blockchains trade decentralization for higher throughput, lower latency, and finer access controls, which makes them useful in enterprise settings where businesses need to share data with partners without exposing it publicly.

A consortium blockchain splits the difference. A group of organizations jointly governs the network, with each member running a node and consensus reached among participants. It balances some of public blockchain’s transparency with the performance and privacy that business applications require.

The right choice depends on the use case. Applications that need censorship resistance and public verifiability, like decentralized finance (DeFi), live on public chains. Applications that need controlled access and regulatory compliance may run on private or consortium networks.

Limitations worth understanding

Blockchain is maturing fast, but it still has constraints worth knowing about. Irreversibility is one example. Irreversibility means that once a transaction is confirmed, it's final. That's what makes blockchain settlement so reliable, but it also means there's no intermediary to reverse a mistake. Self-custodial wallets like MetaMask put you in direct control of your assets, and that control comes with responsibility. Tools like MetaMask's transaction security alerts and address poisoning detection help by flagging suspicious activity before it happens, so that users have advanced protection against scams. That combination of user control and built-in safeguards is how secure self-custody works in practice.