What Is a Blockchain? A Plain-English Explanation

"Blockchain" is one of the most overused and least-understood words of the past decade. It has been variously described as the next internet, a database that cannot be tampered with, the basis for a new world order, and a solution in search of a problem.

What it actually is — at the technical level — is straightforward. This guide explains how blockchains work, why the design has the properties it has, and what the technology can and cannot meaningfully do.

The simplest possible definition

A blockchain is a database that is replicated across many independent computers, records transactions in batches called blocks (where each block cryptographically links to the one before it), and has a rule for those independent computers to agree on what the next block should be without any of them needing to trust each other.

That is it. Everything else — Bitcoin, Ethereum, smart contracts, NFTs, DeFi — is built on this foundation.

Why "chain"?

Each block in the chain contains a cryptographic hash of the previous block. A hash is a fixed-length fingerprint of arbitrary data: any change to the data, however small, produces a completely different fingerprint.

So if someone tries to modify a transaction in an old block, the hash of that block changes. The next block, which contains the previous block's hash, would then have a stale reference. To successfully tamper with one transaction, an attacker would need to redo every block since — and convince the rest of the network to accept their version. This is what makes the database append-only in practice.

Why "decentralized"?

If a blockchain were stored on a single computer, the operator of that computer could manipulate it freely. The whole point of the design is to remove this single point of trust. Instead, the blockchain is replicated across many computers (called nodes) operated by many independent parties around the world. Every node has its own complete copy.

For an attacker to tamper with the blockchain, they would need to control a majority of the nodes — or, in some designs, a majority of the computational power or stake. On large networks, this is enormously expensive.

How nodes agree on what's true

The two dominant approaches are Proof of Work (used by Bitcoin and several others), where validators compete to solve a cryptographic puzzle, and Proof of Stake (used by Ethereum since 2022), where validators stake the chain's native currency as collateral and are pseudo-randomly selected. Both achieve Byzantine fault tolerance by different means. PoW is older and more battle-tested; PoS is more energy-efficient.

What blockchains are good at

Recording transactions where no party can be trusted to maintain the database — most obviously, money. Settlement of value across borders, which can move from one wallet to another anywhere on Earth in roughly an hour with low fees. Programmable money on smart-contract chains. Censorship resistance — transactions broadcast to a sufficiently decentralized blockchain are extremely difficult to prevent or reverse. Auditable records that the state of the database can be independently verified by anyone.

What blockchains are bad at

Storing large data — every node stores the full history, making blockchains a poor fit for video, images, or large datasets. Frequent updates to existing records — the append-only design means "updates" are actually new entries that supersede old ones. Privacy — most blockchains are pseudonymous; every transaction is public forever. Performance, at the base layer — Bitcoin handles ~7 transactions per second; Ethereum's base layer handles ~15-20. Reversing mistakes — sent crypto to the wrong address? It is gone.

The "blockchain not Bitcoin" question

For a stretch in the late 2010s, many large institutions were excited about blockchain technology while skeptical of cryptocurrency. The pitch was that you could use blockchain databases for supply chain tracking, healthcare records, identity, and a hundred other applications, without involving any volatile token.

The results have been mixed. Several useful private blockchain implementations exist in trade finance, settlement, and specific niches. But many "blockchain" projects in this category turned out to be solutions in search of problems — the technology is excellent at one specific thing (decentralized consensus without trust) and is overkill or ill-fitted for problems that do not require that property. The killer applications, so far, have been the ones that genuinely require censorship resistance and trust-minimized settlement: cryptocurrency itself, stablecoins, and decentralized finance.

The bottom line

A blockchain is a way of running a database that does not need a trusted operator. That sounds modest, but it is genuinely novel — there was no good solution to that problem before Bitcoin demonstrated one in 2008. Whether you find blockchains revolutionary or merely useful depends on how often you encounter problems where the absence of a trusted operator is the constraint that matters.