Block

Blocks are how a blockchain turns a firehose of transactions into an agreed-upon shared history. Users broadcast transactions into a mempool where they wait. Validators (or miners on PoW chains) collect a batch of pending transactions, verify they’re all valid, package them into a block, and propose the block to the network. If the network agrees, the block is added to the chain and everyone updates their local copy of the ledger.

Block time varies widely. Bitcoin produces a new block every 10 minutes. Ethereum does 12 seconds. NEAR runs at 600ms. Solana aims for 400ms. Faster block times mean faster user experience (transactions confirm quickly) but also more bandwidth and computational load on validators, and sometimes more forks where multiple valid blocks compete before consensus settles. The right block time for a chain depends on what it’s optimising for: security and decentralisation (slower) or throughput and responsiveness (faster).

Block size and transaction density matter too. A block can contain anywhere from a handful of transactions (Bitcoin, ~2,000-3,000 per block) to tens of thousands (high-throughput L1s like Solana). More transactions per block means more throughput but also more data to propagate across the network, which pushes up hardware requirements for running a node and centralises infrastructure over time. This is one of the core tradeoffs every L1 wrestles with.

The critical thing about blocks is that they’re ordered and immutable. Once a block has enough subsequent blocks stacked on top of it (called “confirmations”), rewriting history requires redoing the work for every block since. On Bitcoin, 6 confirmations is considered final (about an hour). On NEAR, a block is usually final in 1.2 seconds due to BFT consensus. Faster finality is a big usability win: users don’t have to wait for confirmations before trusting that a transaction has landed, which changes how applications can be built on top of the chain.