Throughput
Throughput is the amount of work a blockchain network can process in a given time, usually measured as the number of transactions it can confirm per second. It describes the network's overall capacity rather than how fast any single transaction completes.
Throughput depends on how much data fits in each block and how often new blocks are produced. Block size, block time, the consensus mechanism and the cost of executing each transaction all set the ceiling. When demand exceeds that ceiling, pending transactions wait in the mempool and users compete with higher fees to be included sooner.
For users, throughput affects cost and waiting time. A network with low capacity can become congested during busy periods, pushing fees up. Networks raise throughput in different ways: Layer 1 changes such as larger blocks or sharding, which splits processing across parallel segments, or Layer 2 systems such as rollups, which bundle many transactions off-chain and post them to the main chain as a single batch.
As an example, if a chain handles a limited number of transactions per block and a popular token launch floods it with requests, many users face high fees or delays. The same activity on a rollup can be processed more cheaply because thousands of transfers share the cost of one posting to Layer 1.
A common misconception is treating headline transactions-per-second figures as real-world performance. Advertised numbers are often theoretical peaks measured under ideal conditions. Throughput is also different from latency and finality: a network can process many transactions overall while each one still takes time to become irreversible. Raising throughput often involves trade-offs, such as heavier hardware requirements for nodes, which can reduce decentralization.